This US Creek Was Dying — Then One Man Used Dead Invasive Trees to Bring the Water Back.

Pick up a piece of tamarisk. Dry, pale, about as thick as your forearm. It's been baking on a creek bank in northwest Colorado for over a year. A year ago, it was an invasive tree. Right now, it's building material. This dead branch is about to be woven into a structure in a creek called Yellow Creek, a tributary so small most people drive past without noticing. That creek feeds the White River, which joins the Yampa and eventually reaches the Colorado, the river that waters 40 million people. More than half the stream miles that feed the Colorado are creeks like this one. And the cheapest way anyone has found to fix them is with the dead wood of a tree that was never supposed to be here. If you've watched the earlier videos in this series, you know the technique. Posts driven into a stream bed, branches woven between them, mud packed into the gaps, all of it a copy of a beaver dam built by hand. You know the cost, a tenth of what heavy machinery charges per mile. A degraded creek can't support the beavers it needs, so humans have to build the starter dams first. Yellow Creek has all of that. It runs through ranch country in the northwest corner of Colorado, Rio Blanco County, population 6,000. A tributary of the White River, which joins the Yampa and eventually reaches the Colorado. On a dry August afternoon, the creek barely makes a sound from the road. What makes this story different is not the creek, it's the material. Joseph Leonhard is the riparian restoration project manager for the Nature Conservancy in Colorado. He's based in Grand Junction on the western slope. Before TNC, he worked at an organization called the Tamarisk Coalition, a group dedicated to fighting a single invasive tree. That organization later changed its name to River's Edge West, but when Leonhard worked there, the mission was clear, kill Tamarisk. Tamarisk, also called salt cedar, introduced from Eurasia in the 1800s as an ornamental. Americans planted it on purpose, a pretty feathery tree for parks and erosion control. It spread. Today, it is the third most common woody plant in the American southwest. Not the third most common invasive, but the third most common woody plant in general. Colorado lists it as a class B noxious weed. A single tree can produce 600,000 seeds a year, per tree, per year. The seeds are tiny and wind-carried. It grows in thickets so dense you cannot walk through them. Walls of woody stems that choke out everything else, and it fights on fronts. Its roots pull salt from deep in the ground and deposit it on the surface through its leaves. The salt accumulates in the topsoil, poisoning it for cottonwoods, willows, native grasses. Everything the wildlife depends on. And when it dries, it burns. Fast, hot, and often. Fire crews call it ladder fuel. It carries flame from the ground up into the canopy, turning a manageable ground fire into a crown fire. In a state where fire seasons are getting longer, every acre of tamarisk along a creek is another acre of fuel waiting for a spark. There's a number that has followed tamarisk through the press for decades. 200 gallons of water a day, per plant. It has been printed so many times that it became common knowledge. A USGS hydrologist named Pat Shafroth decided to actually measure it. The real figure came in at about 32 gallons, roughly the same as native cottonwoods and willows. Shafroth pointed out that the 200-gallon figure had been repeated until no one bothered to check it anymore. That matters because it changes the argument. The case against tamarisk is not that it drinks the river dry. The case is that it burns, it salts the soil, and it pushes out every native species that holds the ecosystem together. Removing it is not about saving water. It's about saving the plants and animals that depend on the riverbank the tamarisk has taken over. So across the southwest, crews cut it. They've been cutting it for decades, and every time they face the same problem. What do you do with the biomass? They pile it and wait for safe weather to burn it. Or they truck it out, which costs fuel and money. Either way, the tamarisk is waste, a disposal line item. Leonhard spent years working in that world, cutting tamarisk, managing tamarisk, fighting tamarisk. And here is something that, based on how conservation programs tend to be structured, likely goes unnoticed by most people outside the field. Invasive species removal and stream restoration often operate as separate programs. Different agencies, different budgets, different crews. The team that cuts the tamarisk goes home. A different team may come back months later to build restoration structures, hauling in material from somewhere else. Leonhard had worked on both sides. When he moved to TNC and started building restoration structures in creeks, he looked at those piles of dead wood sitting on the bank. Wood his old job had cut, and asked a question nobody had thought to connect. What if the material we're throwing away is the material we need? The Nature Conservancy is now using cured cut piles of tamarisk, the same invasive wood the crews just removed from the banks, to build low-tech restoration structures in the stream. The tree that was wrecking the creek is now the material that fixes it. That sentence is the whole story. The technique is documented in a research paper co-authored by Leonhard himself and John Leary of River's Edge West. It works like this. The crews cut the tamarisk and stack it in piles on the bank. Then they wait. Over a year. 12 months of doing nothing to it. Letting it bake in the high-altitude Colorado sun. On a timeline where funders want quarterly results, that takes a certain kind of patience. The waiting is not laziness, it's biology. Tamarisk is tenacious. Cut a fresh stem, lay it in wet soil, and it can send out new roots and start growing again. You don't want your restoration structure sprouting into a new invasion. A full year of UV and dry air kills the root tissue completely. The stem that comes out is dead wood. Structurally rigid, rough enough to trap sediment, but biologically inert. It won't re-sprout. Native willow would seem like the natural building choice, but cutting healthy willows to build restoration structures defeats the purpose. The whole point is to grow more willows, not fewer. Dead tamarisk gives you building stock without sacrificing a single native tree, and it's already sitting at the job site. Free material. Once cured, the dead tamarisk gets woven into structures alongside fresh willow, rock, and sod. The fire hazard comes off the banks, a restoration structure goes into the stream, and the fuel load drops. All in one operation, with one crew, on one budget. Leonhard's team calls it stacked benefits. The stranger part is happening near Moab. On West Coyote Creek in Utah, the team is going further. They're weaving live tamarisk, trees still standing but weakened by beetle infestation, directly into the structures. The beetles deserve more than a footnote. In 2001, after more than a decade of testing, the USDA released the tamarisk leaf beetle into the wild as a biological control agent. It feeds only on tamarisk. It defoliates the trees over and over. It strips the leaves again and again until the tree is barely alive but still standing, a skeleton of its former self. Leonhard's crew takes those beetle-weakened trees and weaves them into the dam. Then they monitor what happens when the water pools, specifically whether the sustained flooding suppresses the tamarisk's ability to re-sprout. The monitoring is ongoing, not conclusive, but early signs suggest it works. The pooled water appears to drown the root zone of the tamarisk woven into the dam. There's something almost circular about it. An invasive tree weakened by a beetle imported to fight it, then woven into a structure designed to heal the creek it was destroying, and drowned by the water that pools behind it. The whole sequence folds in on itself. So now the structure is in the Dead tamarisk packed against fresh willow, sod, and rocks in the gaps. The current slows, pools, and spreads into ground that hasn't been wet in decades. The underground water level starts to rise. A 2024 study in Restoration Ecology by Orr and colleagues measured what happens next. Willow cuttings planted near these structures grew 2.8 to 9.6 times more than those planted far from them. Up to nearly 10 times. The benefit compounds. Wetter soil feeds roots, which stabilize banks, which trap more sediment, which holds more water. In the Gunnison Basin, where crews have been building these structures since 2012, Colorado State University tracked nearly 900 of them through 2020 and found a 40% increase in wetland plant cover, despite 6 of the monitored years being drought years. In the same basin, real beavers returned to Cold Harbor Ranch after a crew installed just 10 structures and moose followed. In some monitored sites, BDA structures reconnected up to 1,000 feet of side channel. Small braids of slow water that had been cut off from the main flow when the creek incised. Those side channels are nursery habitat. Fish fry, tadpoles, aquatic insects. Everything that breeds in water needs slow, shallow, protected pools to survive its first weeks. When a creek cuts down and straightens out, those pools vanish. Reconnecting them doesn't just help adult fish, it brings back the generation behind them. The caveat. BDAs don't create more water, they redistribute it, and they require repair after floods. This is not install and forget restoration. But what they do with the water they have is worth understanding. Hydrologists use a term, base flow. It means the slow, steady trickle of groundwater that seeps out of the soil into a creek long after the last rain. It's what keeps a river alive in August. When headwater creeks are healthy, they absorb spring runoff and release it slowly all summer. When they're incised and bare, the runoff races through in a week and the creek goes dry. What Leonhard's structures appear to do in Yellow Creek, based on what's been measured at comparable sites, is function as base flow generators. They don't add water to the system, they slow what's already there, sink it into the ground, and let the ground release it later. Through July, through August, through the months when the White River downstream is running too warm for its own fish. That urgency is not abstract, it has a date. In 2018, the Yampa River experienced the first call in its recorded history. A call is a legal order under western water law. It means there isn't enough water in the river for everyone who has a legal right to it, and upstream users are forced to stop withdrawing. For a rancher, it means you shut off your irrigation ditch and wait. Before 2018, the Yampa had never run short enough to trigger that order. The math had always worked. That year, it didn't. That was the Yampa. Then in 2026, the White River, Yellow Creek's parent, saw voluntary fish enclosures because flows were too low and temperatures too high for native mountain whitefish. Mountain whitefish are cold-water fish. They've lived in rivers like the White for thousands of years. When the water gets too warm for them, it likely signals a broader shift. The insects the fish eat, the birds that eat the fish, the vegetation that shades the water. Whitefish are widely considered an indicator species for cold-water health, and what they seem to be indicating on the White River is a system with shrinking margin. Nearly 29,000 acres in Rio Blanco County are irrigated, 90% in pasture. Water here is not a policy debate. It's a livelihood. The Nature Conservancy saw it coming. In 2019, they founded the Yampa River Fund with an endowment of nearly $4 million, later handed to friends of the Yampa. Its mission is water security for communities, agriculture, and the environment. Three interests that usually compete for the same drop. It has since directed almost $1 million to 23 projects across the basin, roughly $200,000 a year. The money goes to exactly the kind of work Leonhard does. Small-scale, community-led restoration on creeks that most maps don't show. Not just construction, planning, monitoring, capacity building, and the logistics nobody sees. And the work itself pulls together groups that don't usually share a truck. TNC, the Bureau of Land Management, Rivers Edge West, the White River Alliance, and local ranchers who have the most to lose if the water runs out. It buys posts and pays for crews and covers the gas to drive out to a tributary nobody will photograph for a magazine. For ranchers, the payoff is concrete. The water stays in the ground longer during the growing season, meaning more forage in July and August when unrestored meadows have dried out. The native grasses that return are better forage than the weeds they replace. And the land stops losing topsoil. Restored banks hold their ground instead of bleeding dirt downstream. A rancher in Idaho named Chris Black, whose land was restored with the same techniques, put it this way. These meadows are like a sponge. They take that water and they hold it and release it slowly into the system. And that benefits downstream users too. He described habitat appearing for spotted frogs, sage-grouse, and beaver, a sequence that suggests the cascade was already underway. The sagebrush sea, the vast shrubland ecosystem that defines, has lost half its historical range. The wet meadows threaded through it are not scenery. Based on what ecologists understand about sage-grouse biology, those meadows are where hens bring their chicks in the first weeks. The only habitat with enough insects and soft wet vegetation to keep a brood alive. Lose the meadows and the adults may persist, but chick survival drops. A species can hold on for years that way before anyone notices the numbers declining. Those same meadows are breeding and feeding ground for mule deer, elk, migratory songbirds, pollinators, and northern leopard frogs. Lose the creeks and you don't just lose water, you lose the landscape. Leonhardt's career is a single line drawn from the invasive tree, is the problem to the invasive tree, is the material. At River's Edge West, he led the Desert Rivers Collaborative, a multi-agency effort to manage invasive vegetation across public lands in western Colorado, and worked with the BLM in the McInnes Canyons and Dominguez Escalante National Conservation Areas, landscapes where Tamarisk had colonized miles of canyon riverbank, choking out the cottonwood galleries that once shaded the water and cooled it for native fish. He spent years learning exactly how the tree behaves, how it re-sprouts, how it responds to cutting and to fire and to flood. Then he landed at TNC and carried everything he knew about the enemy into a new job building with it. In 2025, three restoration events improved 3.7 miles of stream across Meadow, Yellow, and Willow Creeks. 200 more structures are planned for the Yampa and White River Basins. TNC is running programs across seven states and the workshops are training new crews every year. The training model matters as much as the technique. TNC sponsors an annual LTPBR workshop. A few days where land managers, ranchers, agency staff, and volunteers build structures together in a real creek. They learn site assessment, structure placement, how to read a channel, and how to maintain what they build after the first high flow. The point is not to fix that one creek. It's to send people home knowing how to fix their own. On any given build event, the crew is a mix that shouldn't work and does. 19-year-old AmeriCorps members standing knee-deep in the water next to PhD scientists from the BLM and USGS. The photographer Robert Marcos documented the dynamic as a kind of group cohesion. Everyone covered in the same dirt with the same sore backs. Leonhard said his team felt activated as opposed to sitting at a desk writing reports about problems they'd only seen in satellite images. That word matters. The bottleneck in restoration has never been the science. The science has been published and free for years. The bottleneck is people willing to stand in cold water at dawn. Every workshop produces more of them. Leonhard put the logic simply. It's easy to overlook small tributaries, yet work from each community across our seven states adds up to a cumulative impact that benefits waterways as large and important as the Colorado River. But Yellow Creek is still early. No monitoring data has been published from the site itself. The evidence comes from analogues in Gunnison, Idaho and Utah. The physics and the biology suggest Yellow Creek will respond the same way, but nobody has measured the result yet. What exists right now is a man with dead wood, a creek with a chance and the patience to wait for the answer. Saying that is the most honest thing this story can do. There is a creek in northwest Colorado where a man walks in each morning carrying dead wood from a tree that was never supposed to grow here and he builds with it a structure shaped like something a beaver would have built 10,000 years ago. The water slows, the willows root, the meadow starts to hold. No one is going to name a law after Yellow Creek or build a monument beside it, but more than half of the river miles in this country begin in a creek exactly like it. Small, seasonal, easy to overlook and holding the whole system together from underneath.