I previously asked What measures might an ancient civilization take to protect itself from extreme heat? The answers gave some useful suggestions as to how cities could be adapted. For convenience in the next question I will call these people the “Androy”.

The principle Androy city is now suitably adapted to the heat, but uses a lot of water. The aquifer they rely on is becoming depleted and on average each year the Androy need to delve an additional meter to ensure that their supply of water is maintained.

Assuming the aquifer is as deep as necessary, what methods could the Androy use to maximise the water extraction process, roughly how deep can they get and what ultimately prevents the Androy from obtaining enough water?

Assume a surface layer of sand and mud with a water bearing sublayer of limestone 5 - 10 meters below.

The city population is at least 50,000 but may be adjusted if it helps

The Androy can use any technology that was available to any ancient society in our world prior to 400 CE

Note this is the second in a series of four related questions

Edit A hand dug well 1285 feet deep seems possible! http://www.mybrightonandhove.org.uk/page_id__6948.aspx Point od interest from my researching

  • $\begingroup$ (1)You'd have to keep building new Archimedes screws. (2) One meter per year: how big is the aquifer? (3) How "expansive" is the soil? youtube.com/watch?v=SW-NoiM726U $\endgroup$ – RonJohn Oct 28 '17 at 9:23
  • $\begingroup$ @RonJohn How big is the aquifer? In depth they don’t know, but it can be as deep as you want. Expanse wise it can be hundreds of square miles or much less, but the key point is that the depth needed must increase by 1m every year to keep up. The material they are digging into is limestone so no expansion problems below the surface. Top few meters open to the air so plenty of room to expand. $\endgroup$ – Slarty Oct 28 '17 at 18:52
  • $\begingroup$ This is the kind of situation that prompted the first aqueducts, you could start bringing in water from the mountains. $\endgroup$ – John Oct 30 '17 at 13:29

My grandfather dug a well by hand on his ranch in central Washington. 6 foot square, 160 feet (50 meters) deep. The digging was fairly easy, as it was the Ringgold formation (wind deposited fine sand and clay from a few ice ages ago. Similar wells have been dug in the Middle East. However he didn't pump by hand. He build a float at the bottom, and put in an electric pump.

Let's extend this to 100m.

To lift a kg of water up 100m you are going to expend about 1000J

A person can do useful work at about 100 W for long periods of time if well fed. https://en.wikipedia.org/wiki/Human_power (claims 75 w, but with training it will improve) A watt is a joule per second. So a person using a perfect device could bring 6 kg -- 6 liters or roughly 1.5 gallons of water to the surface per minute.

You would likely do this either with a chain of buckets on a rope so that the weight of the bucket coming up is balanced by the weight going down. This would be powered by a walking cage to use the big muscles in your legs. If you can get the friction down to 50% of the energy one person brings up about 200 liters -- about 1 barrel -- of water per hour. If you have 12 hour shifts (slave labour) each slave brings up 600 gallons a day.

In a hot dry climate a person can sweat out about 4 gallons of water a day. You will need some additional water for cooking, washing. Call it 10 gallons a day. So at this level you need 1 slave on the water wheels for each 60 people total.

Crops. You can do dryland wheat farming on 10 inches of rain a year, if you fallow the land alternate years. Takes 16 inches for annual cropping. In a desert climate it will take more. Tomatoes take 60 inches of irrigation per season in the San Joaquin valley.

Barley can do 100 bushels/acre, or about 3 tons. (60 lbs/bushel for most grains) Romans marched on 2 lbs of grain a day. 700 pounds of grain per person per year. An acre then supports 4 people and a few mice. An acre is forty odd thousand square feet. So a foot of water is 40,000 cubic feet of water or about 300,000 gallons. So it takes 500 man days to pump the water. You are not irrigating with human muscle power. At 1/100 the depth (1 meter deep) your workforce is larger than the productivity.

Ok. Don't flood irrigate. The Israelis are good at using small amounts of water for crops, putting it exactly where it's needed.

Or have it rain some. Soon as you get rain on a reasonable schedule, you have vegetation that isn't farmland. You can graze critters. This is also how the Sahara, much of the middle east, and Greece became barren -- over grazing.

This in turn modifies the climate. If plants aren't transpiring into the air, there is less water to make rain.

Animals don't help much. Straw doesn't have enough nutrition to keep a cow going by itself (has calories but little protein) , so you need some of your grain too for the oxen.

So you need a different crop. But you need also to justify some combination of lower water use or higher productivity. (Potatoes AFAIK are the highest calories per acre.)

There is a reason that desert peoples are often nomadic. The exceptions occur when there is a major river in the area. The Nile flooded every spring, soaking the land and allowing long term agriculture. The Tigris-Euprhates rivers also had a flood season, and I think that they ran irrigation canals for miles. One of the reasons to build dams is to get water in a position where you can run it downhill to a field. Aqueducts from the mountains...

  • $\begingroup$ This reference cropsreview.com/transpiration-in-plants.html shows that you have found the key issue. The problem is the enormous amount of water required to grow plants. Although this only considers the loss by transpiration, ignoring percolation and evaporation losses. It might be considered a best case situation in a controlled environment. Even so my rough calculations indicate that each adult needs 1.2 - 1.7 tons of water per day to feed themselves based on corn, half as much for wheat and quarter as much for cowpeas. I need to identify some "drought" crops. $\endgroup$ – Slarty Oct 28 '17 at 19:07

Lets take a look at Roman mine drainage systems.

They used a wheel that had boxes with holes fastened to it. First water would get in the box when it was down, then box would rise and hole would be above water. Then, at the top, box would turn upside down and water would get out through the hole.

enter image description here

At Rio Tinto mines they had a series of eight pairs of wheels like this.

enter image description here

So 16 men elevated 3 liter of water per second from 30 meter depth. In Dakia a remnants of system like this were found. It elevated water from 75 meters! Imagine a hell those men went through, rotating those wheels for years undeground...

  • $\begingroup$ “16 men move 3 litres 30 meters every second”. So 5.6 litre metres per man per second or 20,000 per hour or 160,000 LM/man/8 hour shift shift. Check my comments on Sherwood Botsford’s answer. Let’s say we need 1000 L/person/day at the surface. So if everyone works an 8 hour shift on the machines they can only dig down to 160 m. Need less thirsty plants, better machines or longer shifts to go deeper… perhaps some wriggle room here (growing cowpeas would allow them to dig to 420 m) but the whole thing is very tough. $\endgroup$ – Slarty Oct 28 '17 at 19:39

Your folks are by the ocean, if I recall. Their freshwater may only go down so far. Inland you can go deeper, but a problem with subterranean freshwater near the ocean is that it often exists as a lens atop saltwater or brackish water.

from https://pubs.usgs.gov/circ/2003/circ1262/http://earthsci.org/education/teacher/basicgeol/groundwa/groundwa.html

freshwater lens atop saltwater

This is the Ghyben-Herzberg model explaining freshwater sources atop saltwater. I was interested to see that their studies were done on the island of Malta. If you drill too far you get down into the brackish water. Conversely if you drain fresh beyond its recharge rate (from rain etc), the pressure of the fresh atop the salt decreases and saltwater can intrude on your wells.


A similar problem has already been solved. Suitable solutions are either

a machine activated by water power and used for lifting water into a small aqueduct, either for the purpose of irrigation or for the use in towns and villages. Noria

a machine historically used for transferring water from a low-lying body of water into irrigation ditches. Water is pumped by turning a screw-shaped surface inside a pipe. archimedes'screw

You can adapt them by having animals powering the motion.

  • $\begingroup$ I like the top solution better, but it would probably have to be adapted similar to this you have one cog at the bottom, and another at the top, with a band of some sort (think conveyor belt) attached to them. The buckets are attached to the belt. And since is an ancient civilization, the contraption is likely powered by slaves or convicted criminal. $\endgroup$ – Rissiepit Oct 28 '17 at 8:45
  • $\begingroup$ Noria won't work underground. $\endgroup$ – RonJohn Oct 28 '17 at 9:14
  • $\begingroup$ @RonJohn it would if the water was moving, surely? $\endgroup$ – Joe Bloggs Oct 28 '17 at 13:37
  • $\begingroup$ @JoeBloggs look at the picture in Dutch's answer, and then remember that they're trying to pump water out of a deep well. You'd have to dig a massive pit and use a bunch of them in a staged system: the bottom one feeding the second right above it, which feeds a third, which feeds a fourth, etc. Totally impractical. $\endgroup$ – RonJohn Oct 28 '17 at 17:29
  • $\begingroup$ But then I read the answer given by @Vashu worldbuilding.stackexchange.com/a/96305/8068 although mines are massive pits, not wells. $\endgroup$ – RonJohn Oct 28 '17 at 17:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.