I'm envisioning a city-fortress, existing on a plateau atop of a mountain in an island not very far from the coast. I'm trying to figure out how that fortress could have a constant fresh water supply.

I still haven't decided on the population and size of the city, but let's try for starters something like $15 \text{Km}^2$ and 30.000 inhabitants. It could change, but it should be as densely populated as possible.

The technology level of that civilization would be similar to that of ancient Greece or Rome.

This fortress exists on a temperate climate, with plenty of rain. However I'm not sure if they should or could rely just on collecting those waters, without any other source.

However, I don't think it would be technically possible for them to build aqueducts coming from the continent into the island. Not to mention that the altitude of the mountain would make it almost impossible for an aqueduct to maintain a gradient that would keep the water flowing.

During peace time, that city is highly visited by travelers from overseas. Maybe they could import the fresh water. Would it be feasible? But, what about war time?

I'm more inclined to have them dig wells and build cisterns, but I don't know if that is possible in high altitudes.

What are your thoughts?

  • $\begingroup$ How high an altitude are you looking for? Bear in mind that beyond 11,500 feet (3,500m) and the average human being won't be able to cope $\endgroup$
    – Raisus
    Oct 23, 2016 at 21:36
  • $\begingroup$ Between 1900 and 2300 m. $\endgroup$ Oct 23, 2016 at 22:25
  • $\begingroup$ en.wikipedia.org/wiki/… $\endgroup$
    – John_H
    Oct 23, 2016 at 23:36
  • $\begingroup$ The Romans were masters of water supply with e.g. aqueducts and running water and public baths and sewage systems. The Greeks, not so much. $\endgroup$
    – ohwilleke
    Oct 24, 2016 at 3:29
  • $\begingroup$ Forget water, where do they get food? They can't grow enough on that size area to support that population. If they normally import it, and store sufficient for a siege of whatever duration, they can do the same with the water. $\endgroup$
    – jamesqf
    Oct 24, 2016 at 4:11

4 Answers 4


A lot of things depend on how much freshwater supply do you require for your people. As in, do they need freshwater only for drinking and cooking or for washing and other household work, too?

If the freshwater is required only for food purposes, and assuming that each individual requires 5 glasses of water for daily drinking (1.25 liters) and another 0.5 liters for his share of water in cooking, that would make 1.75 liters of water for every person everyday. This would add up to 52500 liters of water for the whole community for one day.

There are three primary approaches about bringing this much water into the colony everyday.

Handpumps at ground level

Handpumps are an ancient type of machinery, which can be used to collect groundwater easily. In your setting, you would require about a dozen large-caliber pumps of this type at ground level (at the base of the mountain). These pumps would be operated by draft animals (donkeys, oxen, bulls and elephants come to mind) and pump out water at all times. The water would be filled in medium sized drums, which would have ropes attached to them. The other end of the ropes would be tied to a wheel at the top of the mountain (in the colony) where draft animals would rotate the wheel once a water-filled drum is ready to be sent. Of course this would require dedicated, low-friction, smooth pathway on one side of the mountain where the drums would softly roll as they are pulled up into the colony.

You would require nearly 60 hand pumps and equal number of rope-wheels for this method to work. Each drum should would contain 70 liters of water and have a total weight of 80 kgs when full. This would make it a total of 375 trips per day for water supply for the whole colony.


  • Very easy to operate during peacetime. Also, water supply is nearly automatic, once the setup is finished. Hassle-free.

  • Considering that there are 60 rope-wheels at different places in the colony and that it takes 30 minutes for one drum to be filled, pulled up the slope, emptied into containers at the top and then slowly released back to ground level, it would mean 8400 liters of water coming up to the colony every hour. If this system is run everyday for 10 hours, it would supply the colony with 84000 liters of water daily, which should be easily enough for food purposes (drinking and cooking) and still some water would be left for storage.


  • Building the initial setup would take a lot of investment, labor and skill.

  • Easy target for enemies during wartime. The whole colony would be thrown in a drought if the water supply is severed.

  • Difficult to operate during rain. But during rain, water is already available in excess to the people.

  • Working with 120 large-bore hand-pumps around all directions of the mountain and then also with 60 rope-wheels above them (in the colony) requires a lot of running costs for replacing old barrels, weakened rope, draft animals etc.

Boiling seawater

The principle is simple: when seawater is boiled, pure water turns to water vapor and salt is left behind as residue.

This method would require a large water supply reservoir, filled with seawater, available to everyone in the colony. Water is supplied to the reservoir everyday through large drums (each drum carrying 200 liters of water, and brought up the mountain on carts), all times of the day.

People would collect seawater from the reservoir, take it to their homes, where they would boil it in a close apparatus so that the water vapors forming at one end of the apparatus condense at the other end to form liquid water again.


  • Having a large reservoir means that the city can hold on its own for sometime even during a siege.

  • Building another such reservoir for rainwater would make things yet easier for the people.

  • Possibly the only option in case draft animals are not easily available.


  • Water gathered through this method would require to be boiled, adding a lot to fuel requirements.

  • Due to the high fuel requirements, time and hassle in this method, it would mean that people would have to do with the minimum water intake for healthy lifestyle. Freshwater would only be available for drinking and cooking. Even for brushing, seawater would have to be used.

Supplying through a river/stream

This works the same way as the handpumps idea. Except that this one is more practical and somewhat easier to manage. When you have a permanent source of freshwater such as a river or stream, you can set up a type of gigantic pump near it, driven by draft animals, which would pump the water from the river/stream directly into the colony through large-bore pipes.

The reason why I did not mention this method (channeling water directly into the colony with pipes) for hand pumps method is that it takes a lot of effort to pull up water from underground reservoirs to ground level. Pulling that water further up onto a mountaintop would be extremely difficult and exhausting for the draft animals. While it can be done with some high power electrical motors in our times, it would be completely impractical idea to try doing that with draft animals. However, it takes very little effort to pull up water from a few feet below ground level (from river or stream), so with some added effort, the draft animals should be able to pump it all the way to the top of the mountain, into the colony.


  • Probably the best method for constant, automatic supply of water into the colony.

  • Initial investment would not be as much as compared to drilling out hand pumps all over around the mountain.

  • With one huge pump, it would be much easier to manage the system as compared to 60 hand pumps!


  • Works only if there is a river or stream very near the mountain. What if there is no such river or stream close to the mountain?

  • Also, rivers and streams are highly seasonal in most tropical areas. The river might flood and overflow in rainy season and dry up completely in dry season. This would make it very difficult to use this method and the primary method of water supply for the colony.

  • $\begingroup$ I think your numbers are a bit off. 1.75 liters times 30.000 equals 52,500 liters per day.... $\endgroup$ Oct 24, 2016 at 6:16
  • $\begingroup$ @AndreasHeese: Ah yes. Demons be cursed! I did miss a little "0" and it adds a 0 to all the subsequent figures, meaning 120 handpumps in total. This would also mean a labor force of nearly 180 people (3 per pump) on the ground and 60 rope-wheels in the colony. And yes, that is a huge investment in infrastructure. $\endgroup$ Oct 24, 2016 at 12:22
  • $\begingroup$ This was a very interesting answer, regarding handpumps, but I didn't understand some details (sorry, english is not my native language): 1) From where would the water be removed? From subterranean wells or rivers / springs? 2) You said that "this would require dedicated, low-friction, smooth pathway on one side of the mountain where the drums would softly roll as they are pulled up into the colony"; But couldn't the drums be pulled vertically, along a pit or a ravine? $\endgroup$ Oct 24, 2016 at 22:12
  • $\begingroup$ @PedroGabriel: The freshwater would be dug up from underground reservoirs (sort of like wells, but hand pumps work much faster than wells). The water drums would be pulled up on a side of mountain because that appears (to me) to be the safest method. Also, remember that when we have 120 wells and 60 water drums to be transported between the colony and ground level at all times, we cannot expect to find vertical slopes on the mountain for all those 60 places. Basically, these people are going to have to invest a lot of infrastructure on freshwater if they adopt this method. $\endgroup$ Oct 24, 2016 at 23:08
  • $\begingroup$ +1 for thoroughness and information - Could I suggest changing the first con for rivers/streams into a pro? If there is a mountain, runoff will be present, it's less of a "what if there isn't" and more of a "there will be" $\endgroup$
    – Zxyrra
    Oct 25, 2016 at 5:19

Mexico City is at 2300m, has what could be described as a temperate climate (between 45 F and 80 F all year) and plenty of rain (more than London, less than New York). Its water supply is extensively covered. It was built in a mountain valley where there was natural lake (which was more like a swamp). The residents of Teotihuacan and Tenochtitlan maintained high population densities by buildign cities around hte lakes, and used floating platforms for agriculture. The Spanish drained the lake. Most of the water in the city still comes from groundwater but some is pumped (uphill) from other areas.

Going down a list of other 1 million+ high altitude cities here:

  • La Pax/El Alto: Built on the Choqueyapu river

  • Quito: Used to be well fed; does not have sufficient water, now gets via aqueduct from dam on Rio Quijos

  • Bogota: Built on Bogota river

  • Addis Ababa: Used to be spring fed from springs on slopes of the Entoto Mountains (to the north of the city). Now gets via aqueduct from Gafsara Dam.

  • Sana'a: Used qanats from the wetter mountains to the south. The qanats broke down over the last few centuries to they turned to well water. The water table then dropped over 1000m since the 1970s, so now the city basically has no water. Also there is a civil war there, so no one really knows how many people still live in Sana'a.

  • Xining: Built on Hangshui river.

So there are a variety of ways to get water to a fortress city: Build it surrounded by lakes and swamps, build it on a river, build aqueducts/qanats from nearby mountains, or rely on well water and try to avoid overpopulation and civil wars.

  • 1
    $\begingroup$ But the groundwater for Mexico City comes from a very large catchment basin, which wouldn't exist on a small island. $\endgroup$
    – jamesqf
    Oct 24, 2016 at 4:14
  • $\begingroup$ A geologically active small island could have geysers and hot springs like Iceland does. Most large islands that are isolated in the middle of nowhere far from a continental shelf are geologically active and built on 'hot spots" that shift somewhat over time as geologic plates slowly move. $\endgroup$
    – ohwilleke
    Oct 24, 2016 at 6:04
  • $\begingroup$ The Valley of Mexico is a closed basin, so no water from the outside. I measured it on a map at around 50km by 100 km, so 5000 km^2. Not the smallest island, but not huge. That is about the same size as volcanic islands with big mountains, like Bali or Hawaii Island. OP doesn't really say the island size. Also, Mexico has 20 million people, so a city of 30,000 could be in a much smaller basin. $\endgroup$
    – kingledion
    Oct 24, 2016 at 12:47

I would suggest that they use a mix of methods.

1) Collect as much water as possible from the sky. This includes large outdoor water collecting buckets and also making the water in drains be collected rather than letting it run off. Also, collect and store storm water for times when water is less plentiful.

2) Collect water from any natural springs/streams on the mountain.

3) Dig wells and boreholes. I see no reason this wouldn't work at high altitude.

Between these three methods there should be enough water as long as the population is careful about water usage. Also, they may need to melt ice/snow for water over winter.


You might want to look at Hawaii; and capture incoming rain.

"Annual means range from 204 mm (8 inches) near the summit of Mauna Kea to 10,271 mm (404 inches) near Big Bog on the windward slope of Haleakalā, Maui. In general, high mean rainfall is found on the windward mountain slope..."

Which means you'll want to make sure that you're ready, and able to catch the incoming rain - and designed specifically to exploit that. Ie: not build your city on the side of the island facing the continent, or aqueduct from the windward side (ie: your fortress is not on the top of the mountain).

  • $\begingroup$ That's a little tough, since I wanted the city to be atop of the mountain and to it to be visible from the continent. However, the city would span the entire plateau of the top of the mountain, which would include the windward side, so they could collect the water from there, I suppose. $\endgroup$ Oct 24, 2016 at 22:04

You must log in to answer this question.

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