Silted harbours

Why do the oldest quays of many port cities now stand well inland from the working docks?

Walk the medieval waterfront of Bruges and you are nowhere near the sea. Stand on the marble quay of Ephesus and the Aegean is kilometres away across flat ground. Ostia, which handled the grain that fed Rome, sits inland of the modern shoreline. In each case the stones say harbour and the water has gone.

The obvious reading is bad luck — a river that misbehaved, a coast that shifted. But it happened to too many of the best harbours of the ancient and medieval world to be luck. Is it possible that the very thing that made these places good harbours is what filled them in?

Start with what carries sediment. Moving water holds particles in suspension, and how much it can hold depends on how fast it moves. Slow the water and it can no longer carry what it was carrying, so the load comes out of suspension and lands on the bed — the coarse grains first, then the finer, then the finest of all. Deposition is simply what happens wherever flow decelerates. That is the whole of the physics, and everything else follows from it.

Now ask what a harbour is. Not a stretch of coast that happens to have buildings on it: a harbour is a place where water is sheltered — an estuary, a bay behind a spit, a basin behind a mole or breakwater. Shelter means waves are damped and currents are slack, because that is precisely what lets a ship lie safely and be worked alongside.

Put those two observations together and the answer is uncomfortable. A harbour is defined as a place where the water slows down. A place where water slows down is a place where sediment falls out. The shelter is not incidentally accompanied by silting; the shelter is the silting mechanism. Every harbour is, by construction, a sediment trap.

Then ask what is supplied to the trap, because a trap with no supply stays clear. Three feeds, usually together. A river brings its catchment's eroded material downstream — and that supply is not constant, since clearing forests and ploughing hillsides upstream can multiply the load arriving at the mouth. Waves striking the coast at an angle move sand along it in a steady conveyor, longshore drift, which settles the moment it enters the calm. And in an estuary there is a subtler feed: fine river clays meeting salt water clump into larger particles that sink far faster than they would in fresh water, concentrating mud exactly where fresh and salt meet.

Does anything work the other way? Yes, and this is the part that turns the process into a trap in the strong sense. Faced with a shoaling harbour, the natural response is to build: a longer mole, a new breakwater, a training wall. Every one of those calms the water further. The remedy deepens the disease.

Which leaves dredging — lifting the deposit out and carrying it away. That works, and it is why great ports still exist on silting estuaries today. But look at what kind of solution it is. The supply does not stop. Next season brings another load onto the bed you cleared. Dredging is therefore not a repair but a permanent operating cost, budgeted annually forever, and the bill rises whenever ships get deeper. A port is solvent only while the trade it handles is worth more than the mud it must lift.

When the sums stop working, the trade does not vanish; it moves seaward. Bruges lost its access when the Zwin inlet shoaled and was eclipsed by rivals with better water; Ostia's basins filled while the Tiber built its delta out past them; later ports answered the same pressure by building outports downstream. That is the answer to the original question. The old quay did not move, and neither did the sea. What changed is the ground between them.

Two refinements and one correction.

The refinement first: not every silted harbour tells the same story. In some the basin filled where it stood, burying the quay under its own bed. In others the shoreline advanced seaward as a delta or spit prograded, leaving the quay inland by adding land beyond it. Both end with old stones far from the water, and telling them apart needs the local sediment history, not the general principle.

The correction is about blame. It is tempting to read these as engineering failures — if only they had dredged sooner, built the mole differently. Some were. But the process is not a mistake anyone made; it is a consequence of what a harbour has to be. Any sheltered basin on a coast with a sediment supply is filling from the day it is built, and the only question is the rate.

Two honest limits. The rates are highly local — catchment geology, tidal range, wave climate and human land use upstream all move them by orders of magnitude, so nothing here predicts a timescale. And the historical cases carry more causes than sediment: Bruges's decline involved commerce and politics alongside a failing channel, and treating silt as the sole explanation flattens the record.

• Why building a jetty or groyne starves the coast downdrift of it while accumulating sand updrift.
• How dredging spoil disposal became its own constraint on modern port expansion.

Every term the collection defines is gathered in the glossary.