Water, water, nowhere—how do we get some?

There are no permanent rivers and lakes in the Sahara that can be tapped for our city's water supply. There's groundwater, but nowhere near enough for our city. So we need to find some, deliver it, and store it, before we can start inviting people to live there.

There are two options: desalination, and icebergs. The icebergs idea is a bit sci-fi, but interesting. The amount of ice that breaks off Antarctica every year, only to melt into the sea unused is so vast, it is mind-boggling. Some of the icebergs are as big as small countries, and they happen to be made of some of the purest water on Earth. People have conceived various creative ways of steering an iceberg (far too big for a mere tugboat) to some coastal destination and then getting it onto land, and they've costed their plans plausibly. However, no-one has actually transported an iceberg yet, so really knowing how feasible and expensive or inexpensive such a project would be is impossible for now. There are some good clues on this site, but for the moment, we'll concentrate on desalination, because desalination idea is here-and-now.

There are numerous desalination plants around the world, and it is not hugely expensive to desalinate water. The standard ways are distillation and reverse osmosis (filtering, basically). Reverse osmosis has become very popular of late, because it is usually more energy efficient. However, as we're in a hot desert, we might let the sun do our distilling for us. Solar stills can distill 5-6 litres per sq m day (less in less sunny climes), using free energy. A square kilometer will provide 5-6,000,000 litres.

Our city will be growing by 2,000,000 people a year (as we aim that it should reach a population of 100 million in 50 years), and we want to get their water to the new residents before they arrive. So, every year we need to desalinate (or melt) enough water for 2,000,000 people, at least.

How much is enough? World bank data offer the following figures for the US in 2000:
According to the World Bank, the US made use of the following quantities of water in the year 2000:

type of use	billion kilolitres	per capita, kilolitres	per capita per day, litres
Industrial	291.0	896	2,448
Domestic	35.8	121	331
Agricultural	120.9	410	1,120
All uses	447.7	1518	1,223

We could use these as a guide. Because the Sahara is drier than the US, we'll choose a slightly higher number -- 1.5 kilolitres per capita per day, which translates into 3,000,000 kilolitres in total per day. Solar desalination would require about a thousand square kilometers, though clever stacking of the solar stills could reduce the amount of land required perhaps a quarter or a fifth of that, so it is a possibility. However, there's the question of what to do with the salt. If the distillation occurs inland (where we already have lots of land), we'll need some way of disposing of the salt (perhaps by burial). If the distillation is to occur on the coast, so that the salt can be returned to the sea, we don't have a lot of land there, and we're relucant to acquire so much.

If we resort to using reverse osmosis, the energy cost of seawater desalination would be about 3 kWh/kL. We'd need the capacity to produce 9 GWh per day. That would require 180,000 solar dishes and 362 sq. km. of land. Just as bad, really.

Whether we desalinate our water or get it from Antarctic icebergs, we need to get it from the coast to our inland city, and this will require a pipeline. The energy cost of doing so will be affected by the distance inland and the height above sea level of our city's reservoirs. Assuming the cost of a water pipeline would be similar to that of an oil pipeline, I looked at some pipelines, such as the Baku-Tbilisi-Ceyhan pipeline so see how much power they used for pumping. From there, I worked out that to pump three million kilolitres a day one thousand kilometres inland would require 43 GWh of power per day, achievable with 2-5 GW of installed capacity (depending on is used). The power would probably be delivered at four pumping stations, 250 km distances apart. Four GW of Stirling engine solar dishes would need 160,000 dishes and 36 sq. km. of land (1GW at each pumping station). It would cost perhaps $12 billion.

Such a project would not be exceptionally large by global standards, but it would be tremendous by regional standards. Indeed, 4 GW is greater than the total installed generating capacity of the entire Sahel region. If, as this city grew, it donated just a small part of its electricity and water to the rest of the region, it could end poverty in that region for ever. In relation to our central point, our city could indeed be supplied with enough water and power for everyone, without having to rely on any petroleum.

After the city was built, we would still have to keep desalinating water, because even with a carefully designed system that recycled as much water as possible, a certain amount would be lost through evaporation. We might even need to increase the flow somewhat, to compensate for losses. There'd be no danger of the sea running dry as a result of this process, even if it continued for millions of years, as the water that escaped from the city would enter the hydrologic cycle and eventually wind up back in the sea.