FEMA flood maps may undercount at-risk homes by about 6 million, one estimate says

One estimate puts the number of properties at risk at 14.6 million, about 6 million more than FEMA’s 100-year flood maps indicate.

The maps are commonly outdated, some by two decades, and, even when updated, typically rely on historical flood data rather than future risk projections. FEMA maps answer only whether water is likely to reach a neighbourhood, not how deeply it might affect an individual home or how well that home can withstand flooding.

To examine these differences, researchers Yilei Yu, Dylan Connor, James R. Elliott, Aaron Flores and Sara Meerow built a detailed flood risk score for nearly 1 million homes in Houston. They combined four types of information: how deep floodwaters would likely get at the home in a 100-year flood event, the age and construction quality of the building, nearby road traffic and its potential effects on mobility during floods, and financial factors including repair costs and the share of nearby homes with flood insurance.

They found that vulnerability scores differed more from one house to the next within a single neighbourhood than they did between different neighbourhoods. This means that a neighbourhood that appears safe on average can contain some of the county's riskiest homes, while a risky neighbourhood may hold many well-protected homes.

The researchers note that much of the problem lies at the scale of individual buildings rather than neighbourhoods or ZIP codes. Their past work in Houston found that more than half of 100,000 documented flood-prone buildings in the city currently lie outside official flood boundaries, and that these overlooked buildings are more likely to be in poor condition and uninsured.

Homeowners can check their own risk by asking about past flooding, looking up the building’s age in property records, having an inspector assess its condition, and reviewing flood insurance coverage, the researchers advise. They say communities can apply the same approach with local data to locate particularly vulnerable housing and invest limited flood-protection money wisely.