A city can lose several metres of elevation without any mountain moving and without a chasm opening up. Sometimes all it takes is pumping the water beneath its foundations. The phenomenon is less spectacular than an earthquake: it advances slowly, neighbourhood by neighbourhood, until it changes the relationship between the city, its subsurface and, on the coasts, sea level.

The simplest image is that of a squeezed sponge. Beneath a city, aquifers are not gigantic caverns full of water. Water occupies the spaces within a collection of underground materials. When large quantities are removed, the structure can compact. At the surface, a few millimetres or centimetres lost each year eventually become a considerable problem.

It is not only the water that disappears

Excessive pumping lowers groundwater levels and can cause the land to settle. This mechanism is sometimes compared with a water-filled balloon deflating. The comparison helps explain the general idea, but it has an important limitation: the real subsurface consists of different geological layers and does not react in the same way everywhere.

That is precisely what makes the phenomenon deceptive. A city can continue to function normally while its elevation slowly changes. The problem becomes particularly serious on coastlines: land subsidence then adds to sea-level rise. For a coastal city, losing elevation therefore amounts to locally accelerating relative sea-level rise.

Jakarta provides an extreme illustration. Some parts of the city are sinking by more than 10 centimetres a year. Other estimates put the rate at 5 to 6 centimetres a year in the worst-affected areas, with local rates of 20 to 25 centimetres. Only 64% of its population has access to piped water, pushing millions of residents to rely on unregulated private wells. Pumping thus becomes less an individual choice than a collective infrastructure problem.

In Mexico City, InSAR measurements taken between 2014 and 2021 showed subsidence exceeding 35 centimetres per year in some parts of the city. Curiously, neighbouring areas are rising by up to 2 centimetres per year. Researchers attribute this second phenomenon to elastic unloading of the crust linked to the loss of water mass. Even around a single metropolis, therefore, the ground does not behave like a uniform surface.

Osaka shows that the trajectory can change

Osaka is particularly instructive because it allows the problem to be viewed over nearly a century. Between the 1920s and 1960s, intensive pumping lowered groundwater levels by as much as 30 metres. In some places, the ground subsided by more than two metres.

In the early 1960s, Japanese authorities imposed strict groundwater regulations. Groundwater levels then began to recover gradually. Recent work shows that sustained recharge of an aquifer can also cause the land surface to rise.

It would be misleading, however, to imagine that simply putting water back will restore the ground exactly as it was. Some compaction may be irreversible, and the observed uplift is not uniform. Faults notably play a role by guiding underground flows. Two nearby areas can therefore react differently to the same recovery of the water table.

That may be Osaka's most interesting lesson: a sinking city is not necessarily doomed to keep sinking indefinitely. Restoring an aquifer, however, is not like inflating an air mattress. Geology retains a memory of decades of withdrawals.

Successes rarely rely on a miracle measure

A study published in Science examined 67 cases of aquifer recovery around the world. Three broad families of action stand out: using other water sources, introducing suitable public policies or environmental mechanisms, and artificially recharging aquifers.

The most revealing figure is probably this one: more than 80% of successful initiatives involved finding an alternative water source. In other words, simply asking residents, farmers or businesses to pump less is generally not enough. They need to be given other water to use.

Most of the successes studied actually combine several types of intervention. That makes sense. Regulation without a replacement solution may merely shift the problem. Artificial recharge without control over withdrawals is like filling a reservoir while leaving its tap open.

Scott Jasechko, a professor at the Bren School at the University of California, Santa Barbara, sums up the lesson from these experiences: groundwater depletion is not inevitable, and different communities have already found ways to curb it.

This perspective matters all the more because more than a third of the world's aquifers are declining and groundwater supplies 40% of global irrigation. The issue therefore extends far beyond a few megacities built in the wrong place.

A global risk, but very unevenly documented

The best-documented situations give an idea of the scale of the problem. In Tianjin, a Chinese city of 15 million people, continued subsidence could put 15% of the population below water by 2120 if the trend is not corrected. Pumping-related phenomena have also been reported in San Diego and major Iranian cities.

In New Zealand, about 80% of the urban coastline is subsiding. Groundwater contributes to the phenomenon in some areas, particularly around Christchurch and Wellington. This is a reminder that the issue is not limited to overcrowded tropical megacities.

And France? We need to resist the temptation to fill in the blanks. The research considered here does not make it possible to identify with certainty a French city experiencing urban subsidence caused by groundwater pumping. Claiming otherwise without local measurements would turn a legitimate question into artificial certainty.

Europe is not inherently protected either: the physical mechanism depends on pumping and geology, not on the continent. But identifying a city that is genuinely affected requires local observations capable of distinguishing groundwater effects from other possible causes of ground movement.

The real solution starts at the tap

I find there is an interesting reversal in this story of the way we usually think about cities. We look at their buildings, levees and roads as though the city stopped at pavement level. Yet part of their stability depends on an invisible resource sometimes located far beneath our feet.

Experiences of aquifer recovery suggest above all a method: reduce dependence before imposing bans, diversify water resources, control withdrawals and, where conditions allow, encourage aquifer recharge. It is not a single recipe, because local geology can profoundly alter the outcome.

But that is precisely what makes Osaka more useful than another story of a doomed city. The ground can respond when pressure on the aquifer is reduced over the long term. Not everywhere in the same way, not necessarily back to its original elevation, and certainly not instantly. Yet the direction can change.

A city sinking by a few centimetres a year can easily give the impression of a geological phenomenon beyond human reach. In many cases, however, the cause is linked to a very everyday decision: where does the water we consume come from? It is also there, far less spectacular than a giant levee, that part of the solution begins.