Billions of ‘Uncounted’ People? The Real Mystery Is Where the Maps Think Everyone Actually Lives

Josias Láng-Ritter found a way to check population maps by looking at places that had disappeared underwater.

A postdoctoral researcher in water and environmental engineering at Finland’s Aalto University, Láng-Ritter collected records of people across the globe who’d been relocated due to dam construction. He wanted to know whether population maps showing those places before they flooded accounted for the residents who had to move.

In a 2025 study in Nature Communications, Láng-Ritter and his colleagues compared records from 307 dam projects across 35 countries. They looked at population estimates for 1975 through 2010 and found far fewer people on the maps than in the resettlement records. Depending on the dataset, the maps underestimated the populations anywhere from 53 to 84 percent.

The finding inspired plenty of hand-wringing headlines, suggesting there might be billions more humans on the planet than anyone realized. But Láng-Ritter tells Popular Mechanics that his research “cannot support any revision to the global population total.”

The more unsettling possibility is that we could have a reasonably accurate count of a country’s people and still put them in the wrong places. And that could have disastrous consequences.

“Global population maps are widely used for planning and risk assessment,” Láng-Ritter says. They help estimate how many people live in flood zones and guide decisions about warning systems, shelters, and relief supplies. If rural communities appear smaller than they are, he says, they would likely receive “fewer resources, or lower priority, than towns and cities.” An inaccurate map could affect whether a village got a shelter at all. If the map shows fewer people at risk there than elsewhere, officials might give other communities priority when deciding where to put warning systems, shelters, and emergency supplies.

The problem, Láng-Ritter notes, is that a census and a population map answer very different questions. A census records how many people live in a particular province, while a map divides that total among smaller areas, estimating where all of those bodies actually reside. Too many might be assigned to a town, or too few to nearby villages, without changing the province’s total count. “While rural populations are underrepresented in these maps, urban populations might be overrepresented due to erroneous allocation of population counts to rural and urban areas,” Láng-Ritter says.

Some researchers disagree about whether rural populations are widely underestimated. In a critique posted to arXiv in February 2026, Till Koebe, a postdoctoral researcher at Germany’s Saarland University, and five colleagues argued that the large shortfalls in rural population numbers reflected problems with the comparisons and where historical maps placed people, rather than rural residents going uncounted per se.

Researchers from WorldPop and the European Commission’s Joint Research Centre raised a more specific objection in a separate response. Many population maps estimate where people live by detecting buildings in satellite imagery—but the imagery available for a reservoir site was often captured after the valley had already flooded. Where a village once stood, those images show only open water, so the map registers no settlement there. Even when such a map is meant to represent the population in the years before the dam was built, it may not show a village that was, at the time, still standing and full of residents.

The critics accuse Láng-Ritter’s study of confusing missing data with misplaced data. The people, they argue, are still in the maps, “just in slightly different locations.”

But a slightly different location can be the difference between the dry side of a flood line and the wet side. Put a village on the wrong side, and the officials tallying who needs help won’t count anyone in it.

Láng-Ritter doesn’t oversell his case. He calls the critique “a valuable contribution to the ongoing discussion,” cautions against reading his percentages “too literally,” and treats them as a signal of a problem worth investigating rather than a figure to plan around. He also distinguishes a plausible failure from a documented one: “I’m not aware of a documented case where a population map error was shown to cause an inadequate response.”

Such a case would be hard to prove. Too much else goes into an emergency response to pin a bad outcome on a bad map. And the error often never gets that far, because the people deciding what to send know the village is there; local officials, Láng-Ritter says, usually have “better information than global maps.” His worry is what happens when that knowledge runs out—when the map is the only thing in the room.

What he can’t say is whether the problem persists. His records stop in 2010. Closing that gap would take a fresh comparison: a survey of rural households, or resettlement records from a recent dam project, measured against what the maps claim for the same place and period. Until someone does that work, his advice is to trust local figures over global ones. And when a global map is all there is, assume it undercounts the countryside. Don’t just ask whether a shelter has enough room—plan for the night more people arrive than it has beds for.