How wetlands and floodplains blunt floods and storms

The short answer

Wetlands, floodplains, and coastal marshes absorb heavy rain, slow it down, and release it gradually instead of letting it flood homes and towns all at once. For most of the last two centuries, the US has drained, channelized, and paved over that natural infrastructure faster than it has protected it, then built directly in the space that infrastructure used to occupy. The 2025 Central Texas floods and recent hurricane damage are, in part, the bill for that decision coming due.

The numbers, plainly

Since 1700, the US has lost at least 40% of its wetland area to agriculture and development, and losses have accelerated rather than slowed. Wetland loss rates increased by 50% between 2009 and 2019 alone, meaning the pace of loss is speeding up even now, not leveling off.²

A 2026 study in Nature Water put a specific price on part of what that has cost: wetland loss between 1985 and 2023 alone increased US residential flood insurance claims by more than $10 billion. Since flood insurance covers only about 30% of average annual flood losses, the true cost of that historical wetland loss likely exceeds $33 billion once uninsured damage is counted.¹

Coastal wetlands do the same job against storm surge that inland wetlands do against river flooding. Roughly every 2.7 square miles of marsh can absorb about a foot of storm surge. Louisiana, which holds 40% of all coastal wetlands in the lower 48 states, has lost an average of 34 square miles of wetland a year over the past 50 years, and is projected to keep losing coastal land at a similar pace without major intervention.³

Coastal marsh at Aransas National Wildlife Refuge, Texas. Habitat like this absorbs storm surge before it reaches inland communities, the same mechanism that made Gulf Coast wetland loss so costly after Katrina and Rita.

The July 2025 Central Texas floods put a specific, local number on this. Kerr County took the worst of it, but the flooding reached well beyond it: Travis, Williamson, and Burnet counties recorded 18 deaths between them, with the Sandy Creek neighborhood in northwest Travis County and the San Gabriel River corridor in Williamson County among the hardest hit. Statewide, the flooding killed at least 139 people and caused an estimated $1.1 billion in damage, only a fraction of which was covered by flood insurance.⁵

The part of the story that usually gets left out

The history here runs parallel to the wildfire story in a way that's easy to miss. Just as a century of fire suppression let fuel build up in forests that needed periodic burning, a century and a half of "reclaiming" wetlands for farmland and building levees to straighten rivers removed the natural capacity that floodplains had to absorb water, then built homes, roads, and towns directly in the space that capacity used to occupy.

Most of that loss wasn't accidental or incidental, it was policy. Federal programs actively subsidized wetland drainage for agriculture through much of the 20th century, and the US Army Corps of Engineers built thousands of miles of levees specifically to disconnect rivers from their historic floodplains, both to protect farmland and to make rivers like the Mississippi more useful for shipping. That approach worked as intended for the specific, narrow goal it was built for. It also meant that when heavy rain did come, water that used to spread out gradually across a floodplain instead stayed confined to a narrower channel, moving faster and rising higher before it had anywhere to go.

The result, over time, is what hydrologists sometimes call the flashiness of a watershed: rivers that used to rise and fall over days now rise and fall over hours, giving people far less warning before floodwater arrives. Channelization and urban development both make a watershed flashier, and the Hill Country terrain around the Guadalupe, Pedernales, and San Gabriel rivers, with its thin soil over limestone and steep drainage, was already prone to fast flooding before any of that development happened. Removing the floodplain's natural capacity to absorb water made an already flood-prone landscape more dangerous, not less.

Why wetlands and floodplains actually work

Dune vegetation at South Padre Island, Texas. Barrier island dunes work alongside coastal marshes as part of the same natural buffer system against storm surge.

A healthy floodplain does three things a channelized river can't. It gives floodwater somewhere to spread out and slow down instead of staying confined to a single fast-moving channel. It lets water soak into the ground gradually, recharging groundwater instead of rushing downstream all at once. And upstream wetlands specifically reduce the volume and speed of water reaching downstream communities in the first place, which is the exact mechanism behind that $10 billion insurance-claim study: more upstream wetland area measurably means smaller flood claims downstream.

Beavers are one of the most direct examples of this mechanism in action, and not a coincidence, they evolved to do exactly this. A beaver dam slows water, spreads it across a wider area, and raises the local water table, functionally rebuilding in miniature the same kind of floodplain capacity that levees and channelization removed at scale. It's part of why beaver reintroduction keeps showing up as a flood-mitigation strategy, not just a biodiversity one.

What's changing now

The engineering field itself is shifting, unevenly, toward giving rivers room again rather than confining them further. The Netherlands' Room for the River program is the best-known example: 30 projects that lowered floodplains, set levees back from the river, and widened side channels along four major rivers, specifically to reduce flood height by giving the water somewhere to go. In the US, the Army Corps of Engineers now uses the term engineering with nature for a similar shift, and floodplain reconnection projects, removing or setting back levees, removing aging dams, restoring wetland habitat, are becoming a more common flood mitigation tool alongside, not instead of, traditional levees and reservoirs.⁶

This is not a fast or a finished shift. Levee removal and floodplain restoration are expensive, politically difficult in areas where people already live and farm on former floodplain, and slower to show results than building a wall of concrete. But the direction is real, and it's the same direction as prescribed fire in forests that need it: not abandoning engineered flood protection entirely, but recognizing that the natural systems we removed were doing real, measurable work, and that rebuilding some of that capacity is often cheaper and more effective than trying to out-engineer a flood with levees alone.

Frequently asked questions