The Water That Is Not Coming Back

Water is not consumed. It circulates: evaporates, falls, flows, evaporates again, in quantities that have not meaningfully changed in human history. Scarcity is therefore never about the total. It is always about where the water is, when it arrives, and what condition it is in.

For surface water this is a management problem, hard but tractable. Reservoirs move water in time, canals move it in space, and a river refills with the next rain. Groundwater is a different category of problem, and it is where the serious trouble lies.

Aquifers are not underground lakes. They are layers of permeable rock and sediment holding water in pore spaces, and many of them were recharged over thousands of years under climates that no longer exist. This is fossil water, and extracting it is mining rather than harvesting. Large parts of North Africa and the Arabian peninsula irrigate with it. So does a substantial share of American grain production, drawing on an aquifer whose saturated thickness has fallen by a third or more in places. The relevant question for such a resource is not whether it will be exhausted but when, and what the region does afterwards.

Worse than depletion is what depletion does to the container. As water leaves the pore spaces, the clay layers compact under the weight of everything above. Parts of California's Central Valley have dropped by more than eight metres over the past century. The ground does not rise again, and the storage volume does not reopen, so even a return of the rain cannot refill what has closed. Subsidence also cracks canals, buckles roads and shears well casings, adding a large repair bill to the permanent loss of the reservoir.

For a long time this was hard to see, because measuring groundwater meant drilling and trusting whatever well records existed. That changed with satellites that measure gravity. Water has mass, and removing enough of it from beneath a region reduces the local gravitational pull enough to alter the separation between two spacecraft flying in formation. Several of the world's major agricultural aquifers are now visibly declining year on year, independently of anyone's reporting.

Trade complicates the picture in a way worth naming. Growing a tonne of wheat takes on the order of a thousand tonnes of water, so importing grain imports the water that grew it without moving a drop. This virtual water lets arid countries feed populations their rainfall could not support. It is a rational response to geography, and it makes the food supply dependent on rainfall somewhere else and on trade routes staying open.

Which leaves desalination, usually offered as the answer. Costs have fallen a long way and it is viable for coastal cities that can pay. But it produces fresh water at sea level, and agriculture, which uses the overwhelming majority of the world's fresh water, is generally inland and uphill. Pumping bulk water in that direction costs more energy than making it did. Desalination addresses municipal supply, which is a small share of the problem, and does very little for the sector actually consuming the resource.