In 2009, a restoration project in eastern Oregon installed 76 simple wooden structures along a degraded stretch of Bridge Creek, a tributary of the John Day River. The goal was modest: provide a foothold for the beavers already living there so they could do the work they were already attempting. The project cost $374,000 and introduced no new animals.

The creek had been in decline for decades. By the time work began, it had carved a deep, narrow trench into the valley floor, dropping the water table well below the reach of the surrounding vegetation. Willows and cottonwoods along the banks were dying, and summer water temperatures often approached the lethal limit for the threatened steelhead that spawned there.

The cause of the collapse was historical. Fur trappers removed most of the beaver population in the 1800s, and cattle grazing stripped the woody plants that held the banks together. Without dams to slow the flow and roots to stabilize the soil, the river began cutting downward. Once that started, the process accelerated on its own: a deeper channel meant faster water, which cut the bed further and narrowed the trench even more by the spring floods.

A small number of beaver colonies had persisted in the watershed since the 1990s, building an average of 40 dams a year. But they were building with thin willow shoots, the only material available, and their structures were being destroyed by spring floods before any sediment could accumulate behind them. Most of their dams did not survive a single season.

The project, led by Nick Bowles and Nick Weber with Michael Pollock of NOAA and Joseph Wheaton of Utah State, focused on changing the physical conditions rather than altering the animal population. The installed structures were not rigid barriers. They were semi-porous frames of wooden posts woven with willow, designed to let water leak through at low flows and safely overtop during floods. The idea was to give the beavers a foundation durable enough to last through the one season that had been destroying their work.

Within a few months, beavers began occupying the structures and weaving their own material into the frames. When the spring flood of 2010 arrived, the new structures held. Water passed over and through them instead of slamming against a solid wall. Sediment began collecting behind the dams, and monitoring wells recorded the water table rising one to three feet. Water was reaching the roots of plants that had been cut off from the river for years, and the revived vegetation provided better building material for the beavers.

Between 2010 and 2012, the team added more structures, eventually functioning over a broader area. By 2013, surveys counted 236 dams across the watershed. Slightly more than half were the human-built structures, and the remaining 115 were entirely beaver-built in areas the project had never touched. The upstream dams had dissipated enough flood energy to protect those downstream, and the beaver population was expanding on its own.

The wetted surface area of the stream increased by up to 203 percent, and vegetation density increased measurably along treated banks. The dams also affected water temperature. Although shallow ponds behind dams are often expected to warm in summer sun, the data showed the opposite. Daily maximum summer temperatures in the dammed reaches dropped by up to 2.5 degrees Celsius. The weight of the pooled water forced warm surface water down through porous gravel into the shallow groundwater layer, where it cooled before seeping back into the stream as cold springs hundreds of meters downstream. This natural underground cooling system appeared to buffer the river against warming trends projected for the region.

Monitoring also tracked the effect on steelhead. In the treated sections, juvenile survival rates rose by 52 percent, fish density and abundance rose by 168 percent, and total biomass rose by 175 percent compared with untreated control reaches. The researchers noted it as the largest documented improvement in threatened steelhead productivity from any habitat restoration project in the Columbia River Basin. Spawning migration success remained stable, suggesting the dams did not block adult fish movement.

The original wooden structures have since rotted, which was expected. The beavers replaced them and have continued building new dams, canals, and ponds on their own. The human intervention cost $374,000. A conventional restoration of the same stretch would have run between $3.4 million and $34 million, and it would likely have required ongoing maintenance after the work crews left.

The approach has since been adopted in several western states, and federal agencies have expedited permitting to allow similar projects. The Bridge Creek experiment showed that a river system locked in a self-reinforcing cycle of erosion can be redirected with a relatively small intervention, provided the animal capable of maintaining the system is already present.