03/09/2026
In September 1991, a typhoon flooded 30,000 homes across low-lying northern Tokyo.
Japan started digging the following year.
What they built sits fifty metres beneath the outskirts of the city and is now called the Underground Temple. Five concrete silos, each 65 metres tall and 32 metres wide, connected by six and a half kilometres of tunnels. At the end of it, a chamber 177 metres long and 78 metres wide, its ceiling held up by fifty-nine massive pillars. When rivers rise, floodwater pours into the silos, travels the tunnels, and collects in that chamber. Then pumps push it out into the Edo River at up to 200 tonnes of water every second.
It took fourteen years to build and cost around 2.2 billion dollars.
Here is what it has done. The facility activates roughly seven times a year and has been used more than seventy times. Its site chief says it has cut the number of homes suffering water damage in the surrounding area by about 90 percent. Official studies credit that single facility with saving 148 billion yen in disaster cleanup costs.
And it is not even Tokyo's only one. The city has ten other underground reservoirs and three more flood tunnels, with additional projects still being built. Osaka is currently constructing something similar at a cost of 3.5 billion dollars, scheduled for completion in 2044.
Now, before anyone says Tokyo is simply rich, look at what Tokyo actually is.
It is a megacity cut through by more than one hundred rivers, built largely on a floodplain, sitting inside the same typhoon belt that produces the storms we get. Its geography is not easier than ours. In several respects it is harder.
Singapore is the second example, and its geography is also unkind. A small, low-lying, intensely tropical island that receives enormous rainfall in short violent bursts. Its answer was the Marina Barrage, a dam across the mouth of the Marina Channel completed in 2008. When heavy rain coincides with low tide, gates open and the excess drains straight to the sea. When it coincides with high tide, giant pumps push the water out instead. That single structure protects the low-lying heart of the city, and Singapore has reported reducing its flood-prone area from thousands of hectares in the 1970s to a tiny fraction of that today.
And then the Netherlands, which is the most extreme case of all. Roughly a quarter of the country sits below sea level. By any reasonable reading of a map, it should be uninhabitable.
What is instructive about the Dutch is not the size of their walls. It is that they changed their entire philosophy. After decades of building ever-higher barriers, they concluded that fighting water indefinitely was a losing strategy. Their Room for the River programme did something counterintuitive. It deliberately gave rivers more space, lowering floodplains, relocating d***s further back, and creating areas designed to flood safely so that towns would not.
They stopped trying to defeat the water and started designing where it was allowed to go.
Now here is the honest part, and it is the reason this post exists.
None of these were quick. Tokyo's took fourteen years of continuous construction. Osaka's will not finish until 2044, which is eighteen years from now. The Dutch programme spanned decades and multiple governments.
These are not projects. They are generational commitments, and they only work because each successive administration continued what the previous one started rather than beginning again.
That is the real lesson, and it is the least glamorous one. The engineering is genuinely difficult, but it is solved engineering. Other countries have already done it, published how, and in Tokyo's case will happily give you a tour of the result. What is hard is the continuity. Building something for fourteen years that only proves its worth in the years afterward, and that nobody gets to cut a ribbon on quickly.
There is also a brutal accounting truth buried in the Tokyo numbers.
That facility cost 2.2 billion dollars and has saved 148 billion yen in cleanup costs alone, before counting the homes not destroyed, the businesses not shuttered, the school days not lost, and the savings not wiped out.
Prevention is almost always cheaper than recovery. It simply arrives on a different schedule, and it is invisible when it works. Nobody films the flood that did not happen. Nobody thanks a tunnel for a quiet week. The entire reward for building it correctly is that nothing occurs.
Which is exactly why it keeps losing to more visible things, everywhere in the world, until a country decides otherwise.
So the point of this post is not comparison. It is proof.
Three countries with genuinely worse geography than ours looked at recurring flooding, decided it was unacceptable rather than inevitable, and then spent decades building their way out of it. Their citizens do not have to be resilient. They just have ordinary Tuesdays.
That outcome is not a fantasy and it is not a matter of luck. It is a design problem with known solutions, a long timeline, and a heavy price tag.
The engineering exists. The blueprints are public. Tokyo will literally show you around theirs.
What it takes is deciding, and then continuing to decide, for longer than any one term.
