Jakarta, Indonesia experienced tidal flooding in January. Many parts of Southeast Asia have higher sea levels than previously thought, according to a new study. | Claudio Pramana/NurPhoto via Getty ImagesHumans are a coastal species. More than one in ten people in the world live within three miles of the shore, and about 40 percent of us live within an hour’s drive of the ocean. These shoreline regions generate a massive force in the global economy — in the US alone, coastal counties account for one-third of GDP. But the same oceans that draw so many people near them pose a threat when the water rises and pours inland during tidal floods or storm surges. Water is often the deadliest, most destructive, and costliest element of many natural disasters. And with coastal populations growing, the economic toll of disasters in coastal areas is increasing, especially as sea levels rise due to climate change. Already, global average sea levels have increased around 9 inches since 1880, one of the most tangible impacts of a warming world as ice sheets melt and the ocean expands. Sea levels are rising in some places faster than others, particularly as land subsides. So it’s alarming then that in many of the most populated parts of the world, we’ve been significantly underestimating the level of the sea, a basic, consequential fact of life on the coast. That’s the conclusion of a new study published today in the journal Nature. On average, conventional estimates of the height of the ocean are about one foot too low, though in some parts of the world, the error is more than three feet. These apparent errors aren’t just a scientific question — sea level estimates are used to create hazard maps that govern where people live, where protective barriers get built, and whether insurance companies will offer protection for your home. The fact that we have this wrong means some of our infrastructure may already be facing more threats from the ocean than we were expecting. Add to this miscalculation the fact that further warming is now locked in thanks to the greenhouse gases we’ve already emitted, which will inexorably lead to rising seas, and we’re looking at a scenario where tens of millions more people face threats to their lives and livelihoods in the coming decades. By 2100, global average sea levels are projected to rise additionally on average between 9 inches and more than 3 feet, largely depending on how quickly we reduce carbon emissions. That’s on top of the underestimated sea levels outlined in the study. The result is that many people may not know how vulnerable they are today, and how much more danger they face in the future. How so many scientists miscalculated the height of the seaIt may seem bizarre that so many researchers could be wrong for so long about something so fundamental, but measuring global sea levels is surprisingly complicated. You can’t just stick a ruler in the ocean. The water is constantly moving with the waves as well as rising and falling with the tides. Land can also sink over time. Even if you managed to get a decent average measurement in one location, it won’t translate to other places. The planet is round, but not a perfect sphere, which means gravity doesn’t act evenly. It’s also rotating, making it bulge along the equator. The water itself throughout the oceans has different salt concentrations and temperatures, which alter its density. And winds can push the surface of the ocean and cause it to pile up in some areas and thin out in others. The end result is that sea levels vary around the world, and even across countries — in the US, the sea level is higher on the West Coast of the US than on the East Coast. What do scientists do when they’re faced with a system as fiendishly complex as the oceans? They use a simplified model, in this case one of the Earth called a geoid that simulates where the oceans would settle if they were only under the force of gravity and Earth’s rotation, disregarding factors like currents and winds. This geoid model, often constructed from satellite data, is used as a reference level for land and sea elevation. What the geoid model doesn’t do is include actual sea level measurements, explained Katharina Seeger, a coauthor of the new study. The whole process simplifies the math, but it leaves a lot of critical information out, particularly when you zoom into specific regions. Seeger saw this discrepancy at work in her studies of the Ayeyarwady Delta in Myanmar, where there was sparse mapping of lowlands vulnerable to flooding, making it difficult to predict where the waters will rise during regular monsoon rains and storm surges. Philip Minderhoud, a scientist at Wageningen University & Research who coauthored the study, saw something similar when he was researching land subsidence in Vietnam’s Mekong River Delta, an area that routinely floods. “Being in the delta itself, I witnessed there that the water levels were much higher than those maps implied,” Minderhoud said. Their experiences spurred them to investigate whether there were similar discrepancies in other parts of the world, and if there was a bigger underlying problem. What started as a side project ended up becoming the focus of years of research. Seeger and Minderhoud looked at coastal impact assessments around the world in 385 peer-reviewed articles between 2009 and 2025. They found that calculations of land elevation from the geoid sea level didn’t line up with the direct measurements of sea level. In fact, more than 90 percent of coastal hazard assessments Minderhoud and Seeger looked at underestimated coastal sea levels because they didn’t correct their geoid calculations with direct local sea level measurements. The scale of the errors wasn’t evenly spread around the world. Regions like North America and Europe had more local measurements and thus had smaller gaps between their assumed sea level and the actual sea level. The largest discrepancies were in less well-studied regions like Southeast Asia and the Indo-Pacific, including densely populated island countries like Indonesia and the Philippines that have proportionately more coastline, and thus more people and infrastructure in the path of danger. “The findings are pretty stunning,” said Torbjörn E. Törnqvist, a professor of earth and environmental science at Tulane University, who was not involved in the study. “We’ve dropped the ball a little bit. It’s such a basic thing that a lot of us haven’t really paid much attention to.”We are building on imperfect informationStill, it’s not that scientists didn’t know that geoids were just an approximation of sea level. Oceanographers have known this for ages. So how was this missed for so long? It’s hard to pin it on one definitive answer. Part of the reason is that connecting the physics of sea level rise to the dangers for people on the coast is an inherently interdisciplinary problem, and the experts aren’t always great at talking to each other. “A lot of the issues that this paper points to is in the translation between these two communities,” Törnqvist said. “I’m actually really curious to see what this [new study] is going to do.”One immediate possible solution jumps out: could the authors of the studies that underestimated sea levels just swap out the old number for the correct one in their hazard calculations? “It could be almost as simple as that,” Minderhoud said. “It’s not that these studies are methodologically wrong, but they have relied on an assumption that wasn’t acceptable.” It will take still more local measurements to fill in the gaps in many parts of the world to get the correct baseline sea levels. These more accurate measurements can then help communities better plan how to adapt to the warming we can’t avoid. We’ve already seen that while many types of disasters have become more costly, they are generally killing fewer people as societies become wealthier. Better forecasts and early warning systems have helped people get out of the way of looming typhoons and tsunamis. According to the World Meteorological Organization, the world averaged 170 disaster-related deaths per day in the 1970s, dropping to 40 deaths per day in the 2010s.But while people can escape rising seas, the built environment can’t. Protecting it requires preparation, but even the simple act of picking a sea level rise scenario to plan around can be politically contentious – nobody wants to find out their home is in a flood plain. In the end, other factors may prove more important to what happens along the coast than scientists’ sea level rise estimates, explained Robert Kopp, a professor at Rutgers University who studies sea level rise and was not involved in the study. People who live along the coast are well aware that the ocean doesn’t sit at one level and many areas that do have the resources to plan for the future are already considering what to do when, say, a hurricane pushes several feet of water inland. Additionally, people are not going to be sitting still. As the economy shifts, as more residents become aware of the dangers, and as insurers stop protecting properties near the shore, communities will likely retreat or build seawalls and floodgates. That may end up playing the dominant role in how much danger we face from the sea if we take the challenge head-on. “It is easy for physical scientists to focus on our part of the system, but the truth is that future coastal risk is determined as much if not more by the evolution of the human system as by current and future sea level,” Kopp wrote. But those human systems depend on good information, and that information is still lacking in some of the poorest and most populated parts of the world, as the study showed. These regions also have the fewest resources to adapt to what lies ahead. “Many areas, for example in Southeast Asia, may actually be lower lying relative to local sea level than what we thought,” Törnqvist said. “That means there’s millions more people who are in harm’s way than we thought previously.”It’s a stark reminder that we’re making billion-dollar decisions based on imperfect information, and we have to adapt to a world whose outcomes are still being shaped by our actions today. Models of risk do promise greater precision, but they are only as good as their initial data, and that’s always worth double-checking. Otherwise, we may be in more danger than we realize.