Knowledge LibraryEngineeringCivil EngineeringKL-ENG-HIST-1601

Water and Mud: Irrigation, Flood Control and the First Engineered Cities

The first engineered works were not monuments but levees, canals and drains — because settled agriculture in a river valley is impossible without controlling the river, and because the resource and the hazard are the same thing.

Part 2 of 14 Period c.2500-700 BC Milestones 4 Reading 6 min Updated 2026-08-04

01Executive summary

Engineering begins as a response to water — too much of it, too little of it, and in the wrong place.

The first engineered works were not monuments. They were levees, canals, reservoirs and drains, built because settled agriculture in a river valley is impossible without controlling the river. Four milestones stand for the whole: the canal and levee systems of Mesopotamia, the household drainage of Mohenjo-daro, Sennacherib's aqueduct at Jerwan, and the Siloam tunnel driven from both ends beneath Jerusalem.

c.3000 BCRise of cities, and with them the first engineering as an occupation

~30 mWidth of the Euphrates embankments, running for hundreds of kilometres

~530 mLength of the Siloam conduit, cut through solid rock from both ends

Base 60The Babylonian number system, still used for angles and time

02Why the river valleys and not somewhere else

The Tigris-Euphrates and Nile valleys are agriculturally rich and materially poor. There is clay for brick and very little else — no building stone, no timber, few metals. That combination is what generates engineering. Abundant food supports a population large enough that some people need not produce food; the absence of materials means that anything substantial requires transport, organisation and technique.

Both rivers also flood destructively. Agriculture depends on the water and is destroyed by it, which is the oldest engineering problem there is: the resource and the hazard are the same thing. Every response — levee, diversion, storage basin, irrigation canal — is an attempt to separate them in time and space.

Principle

Store the surplus, release it when needed

A flood is water arriving at the wrong time. Reservoirs and storage depressions convert a destructive peak into a usable supply. Every dam since is an instance of the same move, and so is a grid battery.

Principle

Give the excess somewhere harmless to go

Spillways carried extreme flows into desert depressions covering vast areas. Designing the overflow path rather than assuming it will not be needed is a discipline that failures in every later era have had to relearn.

The mathematics went with the works. Clay tablets from the old Babylonian period deal with practical problems: areas of land, volumes of masonry, the cubic content of a canal excavation. These are not abstractions but quantity take-offs — the same calculation an estimator does today, done four thousand years ago in a base-sixty system that survives in the way we still divide an hour and a circle.

03Mohenjo-daro: drainage as a public system

The Indus Valley sites contain what is, by any reasonable reading, the earliest engineered urban sanitation. By around 2500 BC houses of ordinary means had a well, a bathroom and a latrine. Waste water ran through pottery pipes into covered brick drains laid along the streets, and from those into culverts discharging outside the settlement.

What makes it engineering rather than plumbing

Any household can dig a soakaway. What distinguishes this is that the drains are continuous along streets, covered, and connected into a network discharging clear of the buildings — which means the layout of the drainage and the layout of the town were decided together. A system of this kind cannot be assembled by individual householders acting independently; it requires an authority that plans, specifies and maintains it. The engineering achievement and the institutional achievement are the same achievement.

It is worth being honest about a detail the sources note without dwelling on: the well and the latrine in these houses are often uncomfortably close together. Whatever was understood about carrying waste away, the relationship between sewage and drinking water was not among it. That connection was not established until the nineteenth century, and the first series in this set covers what followed when it was.

04Bringing water in: Jerwan and Siloam

A city can grow only as large as its water supply, and a city under siege can hold out only as long as its supply is inside the walls. Both of the surviving milestones in this part are responses to that constraint.

Jerwan

Around 700 BC the engineers of Sennacherib built a masonry conduit to bring water to Nineveh from sources in the hills. Where it crossed a valley it ran on a bridge over 300 metres long and about 20 metres wide, rising some 9 metres above the valley floor and carried on corbelled arches. The bed of the conduit was paved with a lime-based concrete that retained measurable strength when excavated in the twentieth century. Two things are notable: the scale, and the fact that Sennacherib's own inscription credits the workmen — unusual for a monarch describing his own works.

Siloam

At about the same period, a conduit roughly 530 metres long was cut through rock to bring water from the Gihon spring to a pool inside Jerusalem. It was driven from both ends. An inscription found at the site in 1880 records that the two gangs, working in darkness, heard each other's picks and met after several attempts.

Driving from both ends

Two headings meeting underground is the oldest hard survey problem in engineering, and it recurs in every part of this set: at Samos in the sixth century BC, under the Alps, beneath the English Channel in the 1990s, and in the mid-span closure of any long bridge. The problem is that error accumulates along each drive and there is no way to check against the other heading until they are close enough to hear. That the Siloam gangs needed several attempts is not a failure; it is what a correct account of the difficulty looks like. What changed over the following three thousand years was not the concept but the accuracy of the instruments and the discipline of the control network.

05What we do not know, and why

Honest treatment of this period requires saying what the evidence cannot support. The great Mesopotamian hydraulic works are gone — eroded, buried, or lost as river courses shifted. We know embankments existed and roughly how large they were, because the mounds remain. We do not know how they were made watertight, or how large reservoirs were opened and closed under control. Those are exactly the questions a working engineer would ask, and archaeology has not answered them.

Ancient accounts overstate

Pliny's report that Babylon's walls were 60 miles long and 200 feet high is contradicted by measurement of the site. Kings writing for posterity, and historians repeating legends, are not reliable dimensional sources.

Some famous works are barely documented

The Marib dam in southern Arabia is described in one account as over 3 kilometres long and in another as a quarter of that size and built of earth. Both cannot be right, and no engineer surveyed it.

Absence of evidence is not evidence of absence

Plans on parchment or papyrus would not survive. That only crude sketches remain does not establish that detailed drawings were never made.

06Takeaways for current practice

The resource and the hazard are frequently the same thing. Separating them in time and space is the core of water engineering and of a good deal else.
Design the overflow path. Spillways into desert depressions are four thousand years old; assuming an extreme event will not occur is not.
A network requires an institution. Continuous covered drainage implies a body that plans and maintains it. Infrastructure and governance arrive together or not at all.
Two headings meeting is a survey problem, not an excavation problem. It has been the hard part since Siloam.
Treat contemporary accounts of dimensions with suspicion. This applies to ancient inscriptions and to modern project reporting alike.