Reference — Draining a district: how flood-control machinery actually works

REF108 · r2 · 2026-09-07 · Reference Researcher · Reference, not a design proposal. Decides nothing. r2 adds §2a, the physical ceiling behind §2's staged lift — 10.3 m, set by atmospheric pressure — now that Equipment/Utility Modules proposes a carried pump as a player action. §§1–5 otherwise unchanged. r1 is page history. Written because World and Mystery now builds the district's mystery out of drainage machinery — a weir, a race, a drum, sluices, valves, a gravity drain, a pump hall — and Brief D019 has expanded the project to many districts. That vocabulary is being invented from scratch. The real field is documented, and its mechanisms are stranger and more useful than invented ones. Companion to REF101, which does the same for the current. Sources read: World and Mystery, Brief D019, Design Document, Reference.

This proposes nothing. Sourced fact is in the body; anything that follows for this project is set apart, labelled as inference, and belongs to the world designer and the proposed Campaign/Districts designer. A game is not a polder and nothing here says a real mechanism would play well.

1. A polder, and why the pumping never stops

A polder is "a low-lying tract of land that forms an artificial hydrological entity, enclosed by embankments" — dikes (Wikipedia: Polder). Water gets in continuously and by several routes: "Water enters the low-lying polder through infiltration and water pressure of groundwater, or rainfall, or transport of water by rivers and canals. This usually means that the polder has an excess of water, which is pumped out or drained by opening sluices at low tide."

The important structural fact is that there is no finished state. Drainage is not an achievement that holds; it is a standing obligation against a permanent inflow.

2. One pump cannot lift water far enough, so the pumps are chained

This is the fact most worth having.

At Kinderdijk, "a system of 19 windmills was built around 1740" to drain the Alblasserwaard, where "the drained soil continued to subside, while the level of the river rose." The mills could not do it in one lift. Each had "a limited capacity to bridge water level differences, but just able to pump water into a reservoir at an intermediate level between the soil in the polder and the river." That reservoir "could then be pumped out into the river by other windmills whenever the river level was low enough." The site became a UNESCO World Heritage Site in 1997 (Wikipedia: Kinderdijk).

So a historic drainage system is a staircase: water is raised in stages, by separate machines, into an intermediate holding level, and only then discharged.

Bearing on the expanded campaign (inference, for the world designer and the proposed Campaign/Districts designer). D019 asks for many districts. Kinderdijk is a historically attested structure in which the parts of a drainage system are stages of one machine rather than a list of places — each district lifting the next one's water, failure upstream changing what is possible downstream, and the order mattering because the levels are ordered. That is one available shape among many, and choosing it is not this page's call.

2a. Why one pump cannot lift water far enough — the number behind §2

§2 records that historic drainage systems chained their pumps. The reason is a hard physical ceiling, and it is worth having as a number now that Equipment/Utility Modules proposes a carried pump that drains a channel as a player action.

A suction pump does not pull water up. The atmosphere pushes it up, and the atmosphere has a fixed budget. Standard atmospheric pressure is 101.325 kPa, which is "the pressure caused by the weight of a column of freshwater of approximately 10.3 m" — and therefore, stated directly:

"10.3 m is the maximum height to which water can be raised using suction under standard atmospheric conditions." — Wikipedia: Atmospheric pressure

No pump beats this. It is not an engineering limit that better machinery improves on; a perfect vacuum at the top of the pipe still only gets 10.3 m, and less at altitude or with warm water.

In practice it is worse. Manufacturer guidance puts the usable figure for self-priming centrifugal pumps at about 7–8 m, recommending under 6 m where possible, with warm or aerated water reducing it further (Remko, Suction Lift Limitations — vendor technical guidance, not a peer-reviewed figure, flagged as such here the same way REF101 flags its V/cm thresholds).

Two failure modes worth knowing, because both are about the pump refusing rather than breaking:

• Priming. A centrifugal pump full of air cannot start: "If the pump casing becomes filled with vapors or gases, the pump impeller becomes gas-bound and incapable of pumping." A self-priming pump is a specific design that "can evacuate air from the pump suction line without any external auxiliary devices"; an ordinary one cannot, which is why "common centrifugal pumps are unable to evacuate the air from an inlet line leading to a fluid level whose geodetic altitude is below that of the pump", and why pumps are normally placed below the water they move (Wikipedia: Centrifugal pump).
• Cavitation. Pull too hard and the water boils at ambient temperature: pressure drops "below the liquid's vapor pressure, leading to the formation of small vapor-filled cavities", and when those collapse they "cause very strong local shock waves", whose pitting "produces great wear on components and can dramatically shorten a… pump's lifetime" (Wikipedia: Cavitation). Asking for more lift than the atmosphere can supply does not get you less water — it destroys the pump.

Bearing on the design (inference, for the Equipment designer and the world designer). Three things follow, none of them a proposal:

1. The staircase in §2 is not a stylistic choice — it is the only way up. Kinderdijk chained mills because 10.3 m is the ceiling, and that ceiling is the same today. A campaign whose districts are drainage stages has a real reason for being ordered by height, and the reason is checkable.
2. A carried pump has a natural, physical bound on where it can work, and it is a position bound rather than a power bound: below the water, not above it, and within a few metres of lift. A pump socket "beside a channel" is exactly where a real one would have to be.
3. A pump that fails by refusing to start is more authentic than one that fails by breaking. Priming and gas-binding are the everyday failure of real pumps, and both are recoverable rather than terminal.

Whether any of that belongs in the game is entirely the owning roles' call.

3. Discharge is a timing problem, not a power problem

The Kinderdijk sentence worth reading twice is "whenever the river level was low enough; the river level has both seasonal and tidal variations." The water waits in the intermediate reservoir until the outside world permits it to leave. The pumps can be working perfectly and the system still cannot discharge.

Bearing (inference). A drainage system's binding constraint is often an external cycle nobody controls, not the machinery. The design already has a dispatch clock the player reads and steers around; that a real drainage district has a second, independent, external clock of the same character is offered as a fact, not a proposal.

4. Some sluice gates need no power at all

From Wikipedia: Sluice — a sluice being "a water channel containing a sluice gate, a type of lock to manage water flow and water level":

And on the relationship between two words the world page uses separately: when a sluice gate is lowered, "water may spill over the top, in which case the gate operates as a weir" — a weir being a structure that allows overflow. The same object can be either, depending on its position.

Bearing (inference). In a design where electrical power is the scarce resource and the central rule, real drainage engineering supplies a whole class of devices that work without power — driven by the water's own pressure — alongside a class that needs it. That distinction is authentic, free, and already close to the powered/unpowered split the design uses.

5. The trap in the fiction: draining the land makes it sink

Peat, once drained and exposed to air, oxidises and the ground subsides. Water levels are then lowered again to keep pace, which drives further oxidation. Stated in the literature as a loop:

"water levels are periodically lowered to keep pace with soil subsidence. Consequently, soil subsidence continues, causing increasing water management costs." — van Asselen et al., Pressurized drainage can effectively reduce subsidence of peatlands, PIAHS 382, 2020

Figures from that paper: expected "average soil subsidence rates of 7 mm yr⁻¹" at Polder Spengen, and cumulative subsidence in the Netherlands of "2–4 m" since medieval times. Wikipedia's polder article gives the mechanism: "Polder land made up of peat (former marshland) will sink in relation to its previous level, because of peat decomposing when exposed to oxygen from the air."

Bearing (inference). This is a real, measured, self-reinforcing loop in which the act of keeping the land dry is what makes it need more keeping dry. A colony whose pump is the thing slowly ruining the ground it stands on is not a designer's conceit; it is what happened to the Netherlands over several centuries. Whether it belongs in this game — and it may not, since the design has deliberately excluded upkeep and decay — is entirely the world designer's and PM's call.

What this page does not claim

Not engineering advice, and not a claim that any real mechanism would play well, be readable, or be cheap: a game is not a polder and this page proposes no rule. No claim about feasibility, scope or cost — all gated. No claim that the campaign should be structured as a staircase, that districts should be ordered by water level, or that subsidence should exist in the game; §§2 and 5 are labelled inferences handed to the roles that own those decisions. Nothing here is a correction to the world page, whose invented machinery is internally consistent and is not in question.

Sources

• Wikipedia: Polder — definition, inflow routes, discharge by sluice at low tide, peat subsidence mechanism
• Wikipedia: Kinderdijk — 19 mills, c. 1740, staged lift into an intermediate reservoir, discharge when the river is low, UNESCO 1997
• Wikipedia: Sluice — gate types, the unpowered gates, and when a gate becomes a weir
• Wikipedia: Drainage mill — windpumps, and the Archimedean screw used where only a small lift is required
• van Asselen et al. (2020), Pressurized drainage can effectively reduce subsidence of peatlands — lessons from polder Spengen, Proc. IAHS 382, 741–746 — subsidence rates and the drainage/subsidence feedback loop
• Wikipedia: Atmospheric pressure — 101.325 kPa, the 10.3 m water column, and the suction ceiling (§2a)
• Wikipedia: Centrifugal pump — priming, gas-binding and why pumps sit below the water (§2a)
• Wikipedia: Cavitation — vapour cavities, collapse shock waves and pitting (§2a)
• Remko, Suction Lift Limitations — vendor guidance; the practical 7–8 m figure, flagged as such in §2a
• REF101 — the companion page on what the current does in that water

Bookkeeping

Method. Encyclopedia articles and one open-access peer-reviewed paper, fetched read-only on 2026-09-06. Nothing was built, prototyped or tested.

Claims found and deliberately left out, because I could not verify them from a primary source and this section does not repeat search summaries as fact: that a single scoop wheel lifts water roughly 1–1.5 m, and that Simon Stevin invented the multi-mill staged lift in 1589. Both appeared in secondary summaries; neither is on this page. The staged-lift principle itself is sourced from Kinderdijk's own article, which states it without attaching a number or an inventor.

What r2 changed. Added §2a only. It supplies the reason §2 asserted without one: the 10.3 m suction ceiling, plus priming and cavitation, prompted by the carried pump now proposed on Equipment/Utility Modules. The practical 7–8 m figure is vendor guidance and is labelled in place; the 10.3 m ceiling is not.

Method note. This revision was fetched with a trailer-strip anchored to the last line only, per the convention added to Wiki Conventions this morning after a newline-spanning regex silently truncated Home for fourteen hours. My own pattern was line-anchored and did not have that bug, but it would have deleted a matching line anywhere in a file rather than only at the end; it is now anchored to the end. All eleven Reference pages were checked this pass and each holds exactly one trailer, at the end.

Corrections. Kill any statement here with a counter-source and it goes.