How Engineers Build Underwater Without Anyone Getting Wet

The Impossible Problem: Building Something Solid Where Nothing Will Hold Still

Bridges need piers standing in the middle of rivers. Dams need to be built directly across the water they're meant to hold back. Ports need foundations sitting on a seabed nobody can stand on. In every one of these cases, engineers face the same absurd-sounding requirement: pour solid concrete, lay brick, or dig a stable foundation, in a location that is, at that exact moment, underwater.

There are exactly two honest ways to solve this. You can push the water out of the way and work in the dry hole that's left behind — that's a cofferdam. Or you can build a sealed chamber, sink it into the riverbed, and have people work inside it while it's still underwater, using compressed air to keep the water out — that's a caisson. Both ideas sound almost too simple to work. Both have been responsible for some of the most quietly heroic, and quietly dangerous, engineering in history.

That's the elegant option. The other one is stranger, and far more dangerous — because instead of removing the water, it puts people directly underneath it.

The Cofferdam: A Temporary Dry Island in the Middle of a River

Cofferdams are the more common solution today, and the logic is almost embarrassingly simple once you see it drawn out: build a wall, pump out what's inside, work in the dry. The engineering difficulty isn't the concept — it's holding back the full, relentless force of moving water and unstable riverbed soil for months at a time, sometimes at depths where the water pressure alone could crush an improperly built enclosure.

The largest cofferdam project in American history didn't enclose a small work site — it rerouted the entire Colorado River. To build Hoover Dam in the dry riverbed, engineers first drove four gigantic diversion tunnels, each 50 feet across, through the solid rock canyon walls on either side of the river. Upstream and downstream cofferdams then sealed off a dry construction zone in the middle, forcing the entire river to detour through the tunnels while the dam rose in the space where a river used to be. When construction finished, the tunnels were partially sealed and the Colorado was allowed back into its original bed — flowing, for the first time in years, against a wall of concrete over 700 feet tall.

Smaller cofferdams do the same trick at the scale of a single bridge pier, and they fail in exactly the way you'd expect if the maths is wrong: catastrophically and all at once. Water pressure doesn't announce itself gradually the way a cracking wall does — a breached cofferdam floods in seconds, which is why modern designs are engineered with substantial safety margins and constant monitoring for seepage long before a leak becomes a collapse.

The Caisson: Sinking a Sealed Room Into the Riverbed, With People Still Inside

Cofferdams work well close to the surface. But for very deep water, or a riverbed made of soft, shifting mud instead of solid rock, pumping out an entire enclosure isn't practical. Engineers in the 19th century solved this with something that sounds, on first hearing, almost unbelievable: the pneumatic caisson — an airtight chamber, open at the bottom, sunk into the riverbed with men working inside it, kept dry not by pumping water out from above, but by pumping air in, at a pressure high enough to physically push the water back out through the open bottom.

The Brooklyn Bridge, begun in 1870, needed foundations reaching far deeper than any American engineer had previously attempted — through decades of river silt down to bedrock, 78 feet below the water's surface on the Brooklyn side alone. Chief engineer Washington Roebling adopted the caisson method his own father had studied in Europe, and had two of the largest timber caissons ever built floated into the East River, sunk, and filled with workers digging in candlelit compressed-air chambers, sometimes for hours at a stretch, six days a week.

What happened next is the part of the story engineering textbooks rarely have room for, but should: unable to visit the site himself, Roebling continued directing construction from his apartment window overlooking the bridge, using a telescope — while his wife, Emily Warren Roebling, who had no formal engineering training when the project began, taught herself structural theory, cable analysis, and construction management well enough to carry his instructions to the site daily, resolve disputes with contractors and the board of trustees, and effectively manage the day-to-day engineering of the bridge for over a decade. When the Brooklyn Bridge opened in 1883, Emily Roebling was the first person to cross it.

The caisson disease casualties on the Brooklyn Bridge weren't a freak accident of one project — they were common enough across 19th-century underwater construction that the medical understanding of decompression sickness itself grew directly out of these worksites, before it was ever properly applied to diving. It's a genuinely uncomfortable fact sitting underneath a lot of the world's oldest bridges and tunnels: their foundations are, quite literally, built on the bodies of the people who first figured out how to work safely below the waterline — safety knowledge modern codes now take entirely for granted.

Cofferdams vs. Caissons: Two Different Answers to the Same Impossible Question

Modern practice has moved away from pressurized pneumatic caissons specifically because of the human cost the Brooklyn Bridge made so visible — today's deep-water foundations more often use open caissons sunk without pressurized air, drilled shafts, or massive precast concrete caissons floated into place and sunk by controlled flooding, like the box-shaped piers used on modern crossings such as the Confederation Bridge in Canada or the Øresund Bridge between Denmark and Sweden. The core idea — sink a controlled, self-contained structure to bypass the water entirely — never went away. Only the method of keeping people safe inside it did.

Why This Matters Even If You'll Never Build a Dam

Almost nobody using Estima's estimator is diverting a river or sinking a caisson. But the underlying problem — how do you build something reliable in ground you can't fully see, trust, or control — shows up at every scale, all the way down to an ordinary house. A high water table under a residential foundation, a basement excavation next to a stream, or a footing poured too close to unstable, waterlogged soil all carry a smaller version of the exact same risk that cofferdams and caissons exist to manage: water and unstable ground don't announce their danger gradually, and underestimating them is one of the most expensive mistakes a project can make, because the fix usually happens after the damage, not before it.

The deeper lesson from the Brooklyn Bridge isn't really about caissons at all — it's that the safest, most reliable structures are usually built by people who took the ground beneath them more seriously than anyone thought was strictly necessary at the time. That instinct is worth carrying into a project of any size, at any budget: start with an honest, complete estimate of what the site actually demands, before the first foundation is ever poured.