Doing a watershed assessment is essential before considering active restoration, since poor conditions in the uplands can quickly subvert any work done on the stream.

Streambank vegetation is crucial to a healthy stream system. Roots of riparian plants hold the streambank together and prevent erosion into the stream. Vegetation shades streams and lowers the water temperature. Overhanging branches and roots provide cover for fish. For all these reasons, replanting vegetation is an important part of stream restoration.

Replanting alone, however, will not turn around an eroding, denuded streambank. A heavily used area will need protection for a period of time-from a year to a decade-to allow new plants to take hold and grow. That's why replanting works best when it is part of an overall watershed approach that coordinates management changes in uplands and riparian corridors. Restoration starting in the headwaters is more likely to last.

Many native plants are cultural resources for Columbia Basin tribes. They are gathered for food and ceremonial uses and are treated with respect. In the culture of the Columbia Basin peoples, the plants and animals that provide food for people are not just utilitarian things, like products bought and sold in the market; they have intrinsic identity and value.

The species chosen should root easily from cuttings, be adaptable to the site conditions, and be available from local sources. Willow, alder, cottonwood, salmonberry, dogwood, elderberry, and hawthorn are some of the species commonly used for bioengineering in the Northwest. It is best to use a variety of species to minimize failure caused by insects or disease. Cuttings should be harvested during the dormant season, roughly September or October to March. They should be planted immediately after harvesting.

The NRCS is a good source of information on bioengineering. To match the technique and species to site conditions, you will need to inventory the site's topography, exposure, geology, soils, and hydrology. NRCS can provide technical assistance.

The floodplain slows flood waters and acts as a natural reservoir; its vegetation holds the soil in place; its soil holds water for later release. The ponds and marshes that form after a flood are rich in nutrients and support a variety of organisms. Flooding adds nutrients to the stream, too. It deposits sediment on the banks, encouraging the growth of riparian species. Cottonwood trees, for example, need wet, recently deposited sediment for germination of seeds. Without it, no young cottonwoods grow, and eventually the riparian forest dies off.

The European-American settlers in the Columbia Basin often straightened the stream channels, drained the adjacent marshes, cleared off the trees, and dug ditches to bring water to the new crops

growing in the floodplain. Straightening the channel and disconnecting it from the floodplain actually increases the force of floods, because it increases the slope of the channel and the velocity of the water. Sediment is not dispersed on the floodplain but stays in the water, further increasing its erosive force and damaging fish habitat.

The periodic cycling of nutrients from floodplain vegetation to stream channel is lost. The productive backwaters that are refuge and nursery to young fish and other aquatic life are gone. The connections between groundwater and surface water are altered or severed locally.

Restoring the stream to its natural channel and reconnecting channel and floodplain can produce benefits on several fronts.

• Reduces flooding downstream.
• Reduces sediment load.
• Raises water table.
• Lowers water temperature.
• Enhances vegetation.
• Restores fish and wildlife habitat.
• Increases forage for livestock.

At the least, reestablishing a well-vegetated riparian buffer zone is essential-the wider and the more trees the better.

It is highly desirable to keep backwaters and side channels available for high flow events. Sections of the old channel or ditch that fill up only in flood stage, then gradually subside, serve many purposes:

• Bank stabilization by reducing erosive force

• Groundwater recharge

• Refuge for fish during floods

• Habitat for other wildlife

If a long reach of river is diked or leveed to protect property on the floodplain, more water goes directly downstream, putting pressure on the system to cause worse flooding somewhere else. The 1993 Mississippi River floods stimulated development of the concept of "cell" flooding-the idea that certain areas or cells, such as agricultural fields, parking lots, or other areas of lower value, can be flooded while other cells containing houses, barns, or other structures are protected with well-maintained dikes. It's a matter of setting priorities. Some dikes may be removed or breached, culverts may be installed, or streambanks lowered to relieve instream pressure and reduce erosion.

The McCoy Meadows project in the Upper Grande Ronde watershed of northeastern Oregon is an example of successful reconnection of creek to floodplain and restoration of a meandering channel.

Military journals from the early 1800s indicate that McCoy Meadows, a large meadow complex on McCoy Creek, was previously a hub of tribal use in spring, summer, and fall. It was a campsite holding as many as 200 lodges, a base from which men and women would go out to hunt, gather, and fish. There are several archeological sites in the area.

The meadow has been in private ownership since white settlement in the mid-1800s. It was originally part of a large sheep ranch; subsequently, various parcels were sold off and the meadow was grazed, farmed, and irrigated.

By the mid-twentieth century McCoy Creek had been straightened and confined to a ditch running along one side of the meadow. In the summer the creek ran shallow and warm, supporting small fish tolerant of warm water, such as shiners, dace, suckers, and northern pikeminnow. No salmon had been seen here since the 1970s, although summer steelhead spawn in the subbasin.

Before the restoration project began, the creek's former meandering channel was visible as a complex of willows and other wetland vegetation. A small trickle called McIntyre Creek entered the meadow from the opposite side and ran through part of the old channel. Beaver activity on McIntyre Creek helped to maintain a narrow band of willows and other vegetation along it, but McIntyre Creek often dried up in the fall.

In 1992 the Confederated Tribes of the Umatilla Indian Reservation, with the assistance of the USEPA, obtained a grant of $180,000 from the Oregon Department of Environmental Quality for a large meadow restoration project. The tribes believed that reconnecting McCoy Creek to its old channel and natural floodplain would benefit fish habitat. They expected riparian vegetation to rebound, water temperature to drop, beaver to return, and the water table to rise. The owners of McCoy Meadows were willing to fence their cattle out of the meadow and contribute in-kind assistance to the project. Negotiations and planning took several years.

Before starting the reconstructive work, tribal biologists surveyed stream conditions. McCoy Creek was in poor condition; it ran straight, with few or no pools, little streamside vegetation, and high water temperature. It did not provide safety or adequate food for juvenile fish, and its high temperature and lack of pools could be lethal to spawners.

Project hydrologists used an elevation survey of the entire meadow and a 1937 aerial photograph to identify the former channel. Putting the creek back in its channel required the use of heavy equipment. A large meander was created by blocking off the ditch with a long gravel berm and sending the creek to the left. Parts of the ditch were left connected to the creek for use as refuge for juvenile fish during times of high water. Salmon Corps workers planted 10,000 willow, cottonwood, and other native plant shoots. When the work was done the areas that had been compacted by the bulldozers and trucks were churned, raked, and replanted.

McCoy Creek now ran in its old channel. Where the beavers had managed to survive on a trickle from McIntyre Creek, there was now ample water. New beaver dams sprang up, in some places literally overnight, creating new pools, and raising the water table to nourish riparian plants.

It was inferred that the nearly immediate temperature reduction occurred for two reasons:

The underground flow of water, or subinfiltration, popularly called "subbing," lowers water temperature quickly. Subinfiltration began immediately when the dam across the old ditch was complete. Water from the creek began to filter underground through the dam and come out a short distance downstream.

And the landowners are pleased. They know that in the long run the meadow will produce four or five times the forage it produced in the past, and if they want to try short grazing rotations they may do so.

"We never would've been able to get this restoration underway without backing from the tribes, NRCS, and others," says landowner Mark Tipperman. "We're happy with the results. It seems to be working."

Beavers modify streams by building dams out of brushy debris and mud, slowing the water and creating a stair-step profile in the stream channel. The dams trap sediment, raise the streambed and the water table, and change the stream flow regime. The floodplain and associated wetlands expand. The plant community diversifies. Trapping sediments and slowing the water allows nutrients to build up. The number and diversity of aquatic insects increase. (Aquatic insects are an important part of the food chain that supports fish.)

Beavers alter the riparian area by felling deciduous trees along the stream. Clearing these trees allows

Why bring them back? In addition to improving fish habitat and stream function, beavers can help landowners. In arid areas, beaver dams help to maintain moisture even during the worst droughts. The meadows that form at the edges of beaver ponds produce more forage for livestock. With proper management of both beavers and livestock, ranchers can increase the feed available for cattle.

Although historically many landowners have despised beavers, opinions are changing as we realize that beavers are an important part of the ecosystem. They can add value to property and return streams to healthy conditions hospitable to fish.

Sometimes all it takes to bring beavers back to a stream is to quit shooting them. Beaver families migrate and colonize new areas when they reach the limits of the food sources near home. As beavers help to return a stream to its natural functioning state, salmon habitat is improved.

Reintroducing a species, however, always requires caution. A stream must have enough riparian vegetation to keep the banks stable after beavers have felled a number of trees. Otherwise, beaver activity could cause rapid erosion. A stream with relatively low gradient is a good choice.

For successful reintroduction, landowners need to be willing to change grazing regimes, or already have cattle at a distance from the stream. Transplanting beavers into a site where grazing is not properly managed may result in failure.

While beavers may help to create the conditions for re-establishing willows and other riparian species, they also use these same species for building materials and food. If they are being reintroduced into degraded riparian areas in the shrub-steppe zone, it may be helpful to provide a pickup truckload of aspen or other trees near the site to encourage the beavers to stay and to allow cuttings time to get established.

Consider whether moving beavers to a new location will cause damage to an irrigation system or clog culverts. Avoid areas with high road densities, where the woody debris beavers add to the stream is likely to plug culverts and cause sedimentation problems.

Instream structures, from engineered structures such as concrete revetments to natural structures such as root wads, are often used with the intention of stabilizing the bank, diffusing the current, encouraging sediment deposition, creating pools, or otherwise improving fish habitat. To be successful, however, instream structures should be combined with improvements in land management throughout the watershed. It is also essential to figure out beforehand how any instream addition will affect conditions both upstream and downstream.

Tribal staff prefer using natural objects such as root wads or woody debris when instream modifications are necessary. These can add habitat complexity to a stream in the short term while wider changes in the watershed get underway.

Sometimes instream structures can be successfully incorporated into a larger restoration strategy, if carefully planned by an experienced fluvial geomorphologist (a river specialist) or hydrologist. Unless you are a trained restoration scientist, however, do not attempt to place instream structures on your own. Although it may appear simple, avoiding unintended consequences downstream is quite tricky.

In healthy streams, woody debris and root wads can create habitat diversity by forming pools, waterfalls, and places for sediment to collect, while also adding organic nutrients to the water. Often the best management for improving stream habitat is to allow trees to remain where they fall. It was previously considered proper stream management, however, to remove woody debris, in the belief that it impeded fish passage.

The water runs more swiftly and gains more power to scour streambanks downstream or incise channels. More sediment is carried downstream. Where the channel widens, the stream loses its power to transport sediment and deposits it. In addition, as flood waters rise and the natural channel is forced through a section of levees, it may back up and flood upstream.

Riprap can be an even more difficult problem than levees, because the cumulative effect of individual uses can freeze entire sections of the river into place.

Barriers to fish passage created by human action range from Grand Coulee Dam to minor culverts. Many thousands of miles of habitat can be opened again to anadromous fish if we are willing to remove the barriers or provide fish passage around them. In the case of Grand Coulee, that would require a technologically challenging redesign and a lot of money. But in many other cases a relatively small expenditure can accomplish a great deal.

Large dams are discussed earlier in this section, as are small dams such as gravel berms or pushup dams. Many small, permanent irrigation dams also can or do block fish passage if they were not designed with fish in mind. It is often possible to modify these at relatively low cost to allow fish passage. Technical assistance and some funding is available from NRCS, Bonneville Power Administration, and the Bureau of Reclamation.

Standard hatchery operation is "concrete-to-concrete" management. Fish are released from the hatchery and return to the hatchery to be artificially spawned. In this case, the purpose of hatchery production is to raise fish strictly for harvest.

Supplementation relies on the natural environment for most of the life cycle, and is less intrusive than expensive captive rearing programs. It works because young salmon "imprint" on the stream they are released into. It is believed that young salmon learn cues such as the smell and taste of the water in the particular tributary they rear in. When, as adults, they leave the ocean and head back into fresh water to spawn, they search for their own unique stream.

Supplementation was used to bring salmon back to the Umatilla River basin, where the last chinook and coho spawned more than 70 years ago. (See Good Partnerships section.) Fall chinook released from acclimation ponds on the lower Yakima River near Prosser, Washington, are also returning in numbers large enough for a modest harvest season in 1998.