Dam Removal

Have we been too one-dimensional when it comes to reconnecting river systems by focusing on longitudinal connectivity? When it comes to restoring river systems, are more blue-line miles enough? Or do we also need green-field acres? Are there ways we can better integrate lateral and vertical connectivity into our efforts to restore free flowing rivers?

While longitudinal reconnection is critical to the regeneration of a fully functioning river ecosystem—and of course to fish passage—so too is the lateral connection to the floodplain and the vertical connection through the hyporheic zone to the shallow ground water.

One possible solution is the integration of concepts learned through successful “Stage 0” ecological restoration projects into future dam removal projects.  Stage 0 restoration is based on an ecologically led geomorphic concept, developed in 2013 by fluvial geomorphologists Brian Cluer and Colin Thorne, that looks to slow down and spread flows out by restoring complexity to an aquatic system.  To explore this further we sat down with Brian Cluer, one of the developers of the Stage 0 concept.

Cluer first started envisioning the Stage 0 concept during visits to Elk Meadow in Idaho’s Sawtooth National Recreation Area.  “Slogging across that wetland in all kinds of seasons,” said Cluer, “and realizing how rich and anomalous it was and yet how simple and probably ubiquitous that condition used to be before we drained all of our valleys is what got me thinking about Stage 0 and how to link that condition to the channel evolution model.”

Cluer was referring to the Schumm’s Channel Evolution Model (Schumm et al. 1984) which, for the past three decades, has assisted river restoration practitioners in predicting channel response to disturbances and guiding restoration efforts. The Schumm model presents the evolution of a channel through various stages of response to disturbance, the pre-modified condition, labeled “Stage 1,” is depicted as a sinuous, single thread channel.

Cluer, having spent much of the last 35 years looking at remote valleys in the western US from the vantage of his small airplane, and Thorne, who works on rivers all over the globe, compared notes and agreed that many undisturbed stream valleys in various parts of the world looked nothing like Stage 1—or any stage, for that matter—of the Channel Evolution Model. Some undisturbed valleys featured a river-wetland corridor. Others, a network of small, anastomosing channels. Some had no channel at all and were best described as rivers of grass. All, however, were geomorphically complex, rich in diverse habitat, and hydrologically connected to the floodplain for long periods. These observations led Cluer and Thorne, in 2013, to propose a new model: the Stream Evolution Model (Cluer, B. and Thorne, C. R. 2013 (online), 2014 (in print)).

While the Channel Evolution Model focused solely on the channel itself, the Stream Evolution Model expands to include floodplain processes and connectivity. “It is a model that links channels to the valleys in which they sit, and to the processes that stored sediment in the valley” said Cluer.  This expansion of the model is also a predictor of when a Stage 0 approach might be best suited to improve 3D connectivity for a dam removal project.  For dams that submerged a river in a steep, V-shaped valley, the Stage 0 approach to dam removal may not directly apply.  However, for dams that submerged rivers with broad valley bottoms, where meandering channels or forested floodplains with multi-threaded channels once were, a Stage 0 approach to dam removal could ideally restore longitudinal, lateral, and vertical connectivity. And downstream from dams where alluvial valleys exist, it is possible to reconnect floodplains that were disconnected by land use and dam impacts.

In addition to valley configuration, factors such as a site’s geology, hydrology, and land ownership must be considered in order to determine the feasibility of applying a Stage 0 approach to a dam removal/river restoration project.  The fundamental requirements, according to Cluer, are alluvium, a downstream base-level or grade-control (preferably a natural one), and space for deposition processes to function. There is a spectrum of approaches that have emerged for returning alluvial valleys to Stage 0, ranging from regrading anthropogenic sediment features to fill over-sized channels to adding substantial complexity to channels with natural materials in the form of engineered log jams or beaver dam analogs, which mimic the complexity once added by beavers. Then, you let processes do the work and the floodplain wetland evolves on its own. The result is lateral, longitudinal, and vertical connectivity, and a whole lot of ecological uplift.

Cluer revisited Elk Meadow this summer, and its conditions have not changed, despite the region’s drought. “Half of the valley had ankle-deep water that was so cold it was giving me an ice cream headache on a 90-degree August day,” he said.  According to Cluer, Elk Meadow contains a network of subtle beaver dams, even though there had been no beaver activity on the valley for more than 100 years.  “This valley has been resilient to a whole century of change in climate, some past heavy livestock grazing and ongoing locally harmful ATV use,” he said.

It is not surprising that beavers are often mentioned in tandem with the Stage 0 concept, considering the beaver’s use of wood and vegetation to slow down and spread out moving water and the ultimate impact of their work on the ecosystem.  Beaver naturally achieve many of the same goals of river restoration: raised water tables, reconnected and expansive floodplains, increased base flows and hyporheic exchange, improved water quality, and greater habitat complexity and diversity. Beaver are believed to have played an important role in stream valley processes of temperate North America, Europe, and Asia, until they were virtually extirpated by the fur trade in the 1800s.

Cluer attributes Elk Meadow’s resilience to the vegetation, which controls shear stress via resistance, and valley width, which limits shear stress accumulation. The Elk Meadow site is reflective of the resiliency that is a primary goal of a restored Stage 0 river valley.

Take, for example, the case of Whychus Creek, a tributary to the Deschutes River near Bend, Oregon. Following the record flood in 1964, the channel was dredged and a cobble berm (i.e., lateral dam) had been pushed up along the creek to confine it. This was common practice throughout the world. In 2015, the Upper Deschutes Watershed Council partnered with the Deschutes National Forest and private landowners to restore the creek’s 170-acre floodplain. Applying a Stage 0 approach, they pushed the berm back into the channel and graded additional high ground nearby to completely fill the channel. Whole trees with root wads were placed on the floodplain. Within a week after the project was completed, the hyporheic zone was fully charged and the water table had risen four feet, and that was during the dry fall season.

“Restoration designing for stable, two-year, ‘bankfull’ channels has been an ecological disaster, but it has been the standard of river restoration practice since the 1960s,” said Cluer. “The surface water-groundwater connection via the hyporheic zone is probably the key ecosystem function that was most impacted by all of our channelization and floodplain disconnection, and ‘bankfull’ channels don’t address that,” said Cluer. “That connection is the difference between sage-brush upland and saturated wetland, between life-support and thriving habitat, and that connection is not possible when a valley has a high capacity channel continually draining it.”

As part of his analysis on the Whychus Creek project for NOAA, Cluer used 2-dimensional hydraulic modeling to estimate the project’s habitat delivery. Modeling a variety of flow levels ranging from summer base flow to the 100-year event, he determined that salmon rearing habitat increased by 11-fold and was larger during flood events. rather than smaller as before. He has found this pattern and magnitude of improvement to be true of two other, similar sized Stage 0 floodplain restoration projects he has evaluated.

“It’s obvious how much larger these habitats are,” said Cluer. “If you think about the suitable rearing habit in a single thread channel, it’s really just a small wedge on each bank. But if you take that same flow of water and you slow it down and spread it out through a number of channels or a wetland complex, the entire wetted cross section becomes suitable habitat.”

Yet rarely has the Stage 0 restoration approach been applied to river restoration projects associated with the removal of dams. Ironically, one of the hurdles involves commonly accepted fish passage guidelines meant to promote adult fish abundance.

“We’ve had long-standing fish passage guidelines that were developed primarily around concrete fish ladders,” Cluer said. “For lack of any other guidelines, restoration practitioners and regulators started applying those guidelines to dam removal and river restoration projects.” The same was true when fish passage guidelines for culverts were developed, said Cluer, who recalls once having to issue memo to the restoration community in his jurisdiction clarifying that the guidelines were only intended for culverts, not for channels or for habitat restoration projects.

The beautiful, complex mayhem resulting from the application of Stage 0 designs does not lend itself well to meeting specified minimum depth and velocity criteria published in fish passage guidelines. The commonly applied approach of adding wood or beaver dam analogs into channels often runs into fish passage concerns.

Cluer believes such guideline-based pushback is waning, at least on the West Coast of the U.S., where his work has focused. He remarks that fisheries scientists have only come to recognize the importance of floodplains to salmonid populations in the last 15 years, because of the need to support large numbers of juveniles. Although this knowledge has come at a time when there are few functioning floodplains left in the salmonid world, Cluer is hopeful, and points to the restoration of the South Fork McKenzie River valley in the Willamette National Forest as one example of the potential of Stage 0 restoration to create productive fish habitat despite having conditions that do not align with those required in certain fish passage guidelines.

The South Fork McKenzie River was once a complex, anastomosing system that provided important habitat for spring Chinook salmon and bull trout. Historic alterations to the river, including diking and draining the floodplain to facilitate logging of old-growth forest, and the construction of the Cougar Dam in 1963, had forced the river into a single thread, incised channel. In 2018, the McKenzie Watershed Council and the U.S. Forest Service applied a valley-resetting Stage 0 approach to begin restoring four miles of the river and 600 acres of valley below Cougar dam.

The design pushed berms into the incised channel and added roughly 30 pieces of wood per acre of floodplain. By 2019, with just the first two of three project phases complete, preliminary monitoring showed a fourfold increase in baseflow wetted area, and biological monitoring revealed exciting results.

“Many people have taken an interest in monitoring the South Fork McKenzie projects,” said Cluer. That includes Kate Meyer, a fisheries biologist with the Willamette National Forest. Cluer said that Meyer’s work showed that “this subbasin of the McKenzie River watershed had seen no more than four Chinook in the last 30-40 years, but months after the first phase of the Stage 0 restoration project, there were 200 reds there.”  “It attracted all of these adults to a terrain that traditional fish passage concepts would consider not optimal spawning habitat,” said Cluer, “but they went to it like gangbusters.”

According to Cluer, another barrier to applying the Stage 0 restoration concept in dam removal projects relates to regulations that favor “stability” over dynamic function and deter any form of fill and sediment transport. This includes a general hesitancy to allow for passive sediment transport post dam removal, even when the impoundment sediment is clean in relation to the river’s background levels, and the downstream river system has been starved of sediment for decades by the dam. “In addition,” explains Cluer, “it is difficult to restore the complexity needed for a Stage 0 restoration design when both the relocation of impounded sediment within the boundaries of the former impoundment and the placement of woody material are characterized as ‘fill’ in the regulations and highly discouraged.”

But Cluer is encouraged by the progress being made on this front in the West Coast. “I think we’ve turned the corner on the traditional view that sediment is bad, and we’re beginning to gain traction on the idea that floodplain flooding is not only good for the ecosystem but can alleviate flood risks downstream,” he said. “We’re just on the cusp of creating new, more holistic guidelines for California that will put these old fish passage criteria and regulatory concepts into proper context and give regulators and practitioners tools to adapt to site conditions and biological needs rather than applying the same criteria everywhere.”

According to Cluer, two factors have made such progress possible: education and partnership. In West Coast jurisdictions, Cluer has worked very closely with NOAA’s science centers and regulatory biologists to help them understand the benefits of reducing the capacity of oversized channels and that sediment is not categorically a problem, but rather the foundation of a functioning ecosystem.

Cluer said that collaborating with regulatory partner agencies has enabled the pursuit of more innovative restoration approaches that reuse sediment to reduce channel capacity. He cited the 2020 removal of a dam on York Creek, a tributary to the Napa River, as an example.  “We placed over 30 large wood structures in 1400 feet of channel downstream from the dam to trap and retain some of the sediment released from the reservoir.” “The idea,” explained Cluer, “is to restore the incised channel bed and improve lateral connectivity to the narrow but important floodplain.” Cluer said that this same concept is planned for the removal of the Searsville Dam on San Francisquito Creek, a project NOAA is working on with Stanford University.

Cluer also attributes this shift, in part, to the development of modeling tools (such as the Dam Removal Express Assessment Model – DREAM, developed by Yantao Cui at Stillwater Sciences), which helped people see when rivers had the capacity to transport sediment post dam removal. Knowing this enabled them to allow restored sediment transport processes while preparing downstream communities and infrastructure to deal with a temporary impact post dam removal.

Cluer encourages the passive release of sediments during the removal of a dam and believes that it is better for the ecosystem as a whole. He also says that releasing the impounded sediment at once should be encouraged because it shortens the duration of the impact. “We try to get our dam removals done as quickly as possible, so recovery can begin as quickly as possible,” He said. “And if we can retain some of that sediment downstream and better connect to lateral habitat, that is an inspired restoration approach.”

This is welcome advice for restoration practitioners who are eager to restore three-dimensional connectivity to streams and rivers newly liberated by dam removal.