A study published in Nature on March 4, 2026, found that more than 99% of coastal hazard assessments conducted over the past 16 years used flawed sea-level data, meaning actual ocean levels are significantly higher than scientists and planners assumed. The error puts millions of additional people in the path of extreme flooding, with the gap between modeled and real-world water levels reaching 30 centimeters on average globally and up to 150 centimeters in parts of southeast Asia and the Indo-Pacific. The findings call into question the reliability of flood risk maps, insurance models, and infrastructure plans that governments worldwide have built on faulty baselines.
Over 99% of Coastal Studies Got the Data Wrong
The research team behind the Nature paper systematically evaluated 385 peer-reviewed coastal hazard and sea-level rise impact assessments published between 2009 and 2025. Their central finding is stark: more than 99% of those studies handled sea-level and land elevation data inadequately, typically by failing to convert between different vertical reference systems before comparing ocean heights to terrain. Roughly 90.6% of the assessments likely omitted vertical datum conversion entirely, a technical step that aligns satellite-derived sea-level measurements with local elevation benchmarks so that flood projections reflect real conditions on the ground.
This is not a minor bookkeeping detail. Vertical datums are the invisible scaffolding that determines whether a given stretch of coastline sits one meter or two meters above the tide line. When researchers skip the conversion, they effectively lower the ocean in their models, making coastal land appear safer than it actually is. The result is that flood hazards have been systematically underestimated across nearly the entire body of published impact science for more than a decade. Every downstream product that relied on those studies, from municipal zoning decisions to billion-dollar infrastructure investments, inherited the same blind spot, and the authors warn that adaptation plans built on such foundations may already be out of date.
Southeast Asia and the Indo-Pacific Face the Widest Gap
The global average discrepancy of 30 centimeters is already enough to redraw flood maps in many low-lying areas, but the regional picture is far worse. In southeast Asia and the Indo-Pacific, the gap between assumed and actual sea levels reaches up to 150 centimeters, according to the analysis by Tara Russell and colleagues. That disparity reflects the fact that data-sparse regions tend to rely more heavily on simplified global ocean models that ignore local dynamics such as mean dynamic topography, the persistent differences in sea surface height caused by currents, temperature gradients, and salinity patterns that can raise or lower the water line by tens of centimeters.
For densely populated coastal megacities across Indonesia, the Philippines, Vietnam, and Bangladesh, a 150-centimeter underestimate does not simply shift risk categories on a chart. It means that communities believed to have decades before regular tidal flooding could already be experiencing conditions that planners thought lay far in the future. Climate finance decisions, which channel adaptation funding based on modeled risk, have also been shaped by these same flawed baselines. If the true water line is already a meter and a half higher than assumed, the cost of protective infrastructure rises sharply, and the window for effective intervention shrinks as routine high tides begin to overlap with what were once considered rare storm surge levels.
Elevation Errors Compound the Problem
The datum conversion gap identified in the 2026 study compounds a problem that earlier research had already flagged from the land-elevation side of the equation. A 2019 study published in Nature Communications demonstrated that switching from older shuttle-radar elevation data to an improved model called CoastalDEM tripled some estimates of populations on land threatened by sea-level rise and high-tide flooding. That work showed that conventional elevation datasets systematically overstated how high coastal terrain sits, making it appear that fewer people lived in flood-prone zones than was actually the case, particularly in low-lying river deltas and subsiding urban areas where even small vertical errors translate into large differences in exposed population.
When both errors run in the same direction, the combined effect is severe. Overestimating land height while simultaneously underestimating sea level creates a double miscalculation that dramatically narrows the true margin of safety for coastal residents. Most published risk assessments have not corrected for either error, let alone both, which means that widely cited global exposure figures are likely biased low. The implication is that global estimates of flood-exposed populations, which already run into the hundreds of millions, could be significantly larger than any official tally currently reflects. Correcting the land-side data alone was enough to triple exposure counts in some regions; layering in the newly identified sea-level baseline errors would push those numbers higher still, potentially reordering which countries and cities are judged most in need of adaptation support.
U.S. Planning Tools Built on the Same Assumptions
American coastal planning is not immune to these problems. Federal agencies rely on scenario frameworks such as the NOAA technical report on sea-level rise scenarios for the United States, which synthesizes projections and extreme water level context used by state and local planners. The Department of Defense maintains its own regionalized sea-level change database that tracks exceedance probabilities and combined water level trends at military installations and critical infrastructure sites. Both tools depend on the same class of baseline references that the Nature study found to be unreliable in the broader scientific literature, including the careful alignment of satellite altimetry, tide gauge records, and land elevation models via consistent vertical datums.
If the vertical datum misalignment identified in the global review also affects U.S. scenario inputs, the practical consequences are significant. Flood return periods, the statistical intervals used to set building codes, insurance rates, and evacuation triggers, would need to be recalculated. A storm surge event currently classified as a 100-year flood could in reality occur far more frequently if the baseline sea level feeding the model is 30 centimeters too low. That, in turn, would alter cost-benefit analyses for seawalls, levees, and buyout programs, potentially shifting federal investment away from short-lived stopgaps toward more transformative adaptation strategies such as managed retreat. The authors of the Nature study note that robust solutions will require not only better models, but also more transparent access to the underlying sea-level datasets, which are currently gated behind portals such as the Nature article authentication system that controls full-text scientific content for many practitioners outside academia.
Rebuilding Coastal Risk Science on Solid Ground
The emerging consensus from these findings is that coastal risk science needs a foundational overhaul rather than a minor technical patch. At a minimum, future assessments will have to document their choice of vertical datums, apply explicit conversions between satellite and local reference frames, and pair improved sea-level baselines with higher-fidelity elevation data such as CoastalDEM. The authors argue that journals, funders, and professional societies should treat datum handling as a core element of methodological quality, subject to peer-review scrutiny in the same way that emissions scenarios or climate models are evaluated today. Without such standards, they warn, new studies risk replicating the same structural bias that has already propagated through more than a decade of published work.
For policymakers, the message is equally blunt: existing coastal hazard maps and adaptation plans should be treated as provisional until they are rechecked against corrected baselines. In practice, that means revisiting zoning rules for low-lying neighborhoods, reassessing the design heights of critical infrastructure such as ports, wastewater plants, and coastal highways, and revising national adaptation priorities to reflect the likelihood that many communities are closer to chronic flooding than previously believed. Because sea-level rise will continue regardless of these corrections, the study does not change the direction of travel so much as reveal that the world has already traveled farther along the path than planners assumed, leaving less time and less room for error in the race to protect people and assets along the coast.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
