River Work

An unnaturally broad, gravel riverbed defined Anderson Creek along the author’s property, prior to his river work. Author’s photographs

An unnaturally broad, gravel riverbed defined Anderson Creek along the author’s property, prior to his river work. Author’s photographs

Rivers and sunlight, mountains and fish: they are always there, rising up out of exhaustion, a sudden rush of sound and motion, a Wagnerian assault of light and shadow, hissing water, pounding rapids, chilly mountain winds easing inexorably into a requiem of distant rapids, a fish’s silent rise, the splash of blue green water over the back of wet black stones.

Harry Middleton, Rivers of Memory

Mid-February winter storm. It has been raining for days and the soil is supersaturated. The river will be up to its mischief, running bank to bank, dark brown with sediment, violently carrying away all impediments to its will. The memory of last year’s configuration will be all but gone. Every year, we are apprehensive to learn the scale of the damage left in the wake of flood stage. Recently, however, even the dangerously high flows of el niño years have left, literally, truckloads of silt. New depositions of soil are laid down with each ensuing storm, since we placed in-stream structures that are now working effectively to slow the river’s energy behind the constructed deflectors. Trees are quickly planted in every nook and cranny of the “new ground,” and within a year or two, this migrating soil is captured, layer upon layer, to eventually build up and replace what was once lost. For, in one winter’s mayhem eighteen years ago, a neighbor lost two acres of rich alluvial pasture, to a depth of six feet (about 20,000 cubic yards!) in one night. Such is the power of our north coast rivers.

We inherited a river with no riparian canopy, naked, wide—hugely wide—choked full of gravel, running shallow and braided, open to the extreme oscillations of high winter flows and summer’s aridity. The river was “lost,” its summer flow buried under ten feet of gravel. It was hard to imagine steelhead and salmon ever having lived here. Indeed, Anderson Creek, this tributary of the once great Navarro River, is symptomatic of similarly eroded and degraded waterways almost anywhere in the Pacific Northwest, where extinction of the native salmonids is looming.

The same stretch of Anderson Creek now beginning to return to its natural width and equilibrium

The same stretch of Anderson Creek now beginning to return to its natural width and equilibrium

Plants Are the Key

In regard to rivers, I am actually optimistic, for if we “cut them enough slack,” they will rebound in good health and teem with fish again. River restoration is a poorly developed art, and by no means a science; the native ingredients found in our once mighty riparian forests are the miracle workers. Imagine a tree such as alder (red alder for the coast and north, white alder for inland and south) that can put on twenty feet of growth in four years in pure river-run gravel without any soil. Likewise successful in such circumstances are six willow species and three cottonwoods, two maples (box elder and big-leaf), Oregon ash (adapted to handle six months of submersion), the slow but tenacious California bay, coast redwood (at its best in the riparian zone), the majestic valley oak, and, in the south, native sycamore and both of the now rare black walnuts. With a constant water supply and access to newly entrapped alluvial soils, these trees demonstrate the most rapid growth. Along with the re-establishment of the river’s canopy comes a rich wildlife corridor. Sixty-five kinds of birds are known to use valley riparian forests for nesting and foraging, while cottonwood/willow forests support the highest density and diversity of nesting birds of any plant community in California (see Gaines, The Valley Riparian Forests of California: Their Importance to Bird Populations, University of California, Davis). For gardeners and landowners, the strategy seems straight forward: we learn which species can play “offense” in the gravel bar thickets and which species can play “defense” on the flood plain terraces. Even if we must reluctantly fence and water with drip irrigation, these tasks are second nature to a gardener’s cultivating instinct.

Planting trees during the first spring after installation of the “bio-cells,” which are made of geotextile fabric and filled with on-site river-run gravel

Planting trees during the first spring after installation of the “bio-cells,” which are made of geotextile fabric and filled with on-site river-run gravel

The plants and their establishment pose the least of our problems, but provide the most important of our solutions. The real art of river restoration, however, lies in dealing, correctly, with the river’s sinuosity—in our case, its outside curve. Inside curves deposit, outside curves cut. When in flood, our section of the river is like a huge freight train, barreling down upon cut-banks at almost ninety-degree turns. It is a formidable energy to harness, this relentless undercutting and scouring of outside curves, with the eventual bank failures and loss of precious topsoil. The suspended sediment becomes a “non-point source pollution;” topsoil is still the largest export, by weight, of the United States! These problems are exacerbated in the agricultural landscape by the lack of “hard-points”—immovable objects in the riverbed such as large stumps, massive logs, or bedrock outcrops (now buried under migrating gravels)—on the very lands cleared by early settlers for “the best water and the deepest soils.” Until recently, the degradation of our watersheds has continued unabated, under the pressure of ongoing resource extraction from every quarter.

A deep pool created by the “bio-cells,” with native plants revegetating the newly deposited silt

A deep pool created by the “bio-cells,” with native plants revegetating the newly deposited silt

In-stream Structures Begin the Process

We approach our ninety-degree curve with caution, and no small amount of humility—and the full permission of the Department of Fish and Game, which has jurisdiction over such river restoration. A series of in-stream deflectors (vegetative hard-points), which start some distance ahead of the curve, ease the river out gradually, structure by structure, around the bend. The distance between the deflectors is dependent upon their length. Each deflector has a large void on its downstream side—the key to rebuilding lost topsoil and to the quick establishment of riparian trees. The creation of secure, in-stream structures that, over a period of years, enhance the meander of the river, is a most gratifying experience. Under natural flow regimes, pools deepen, complexity returns to the stream with alternating riffle and pool sequences, and the cross-sectional shape of the natural river terraces becomes relatively stable and defined. The closure of the canopy overhead becomes the goal. Our stretch of the river, once more than 250 feet wide in places, is now down to about seventy-five feet, which is natural for this particular point in the watershed. With shaded pools come rebounding aquatic life. Our first summering-over of adult steelhead has taken fifteen years. Not long, really…

Traditional structures for controlling rivers are built out of rip-rap. Imagine two- to three-ton boulders placed in deflective jetties, ten to twelve feet high and up to one hundred feet long. In rowdy rivers like ours, rip-rap has proved unreliable at best. Undermined by constant scour in eighty-inch rainfall winters, even three-ton boulders can be dislodged and half the jetty blown out, lost in a sea of moving gravel. Through trial and plenty of error, we have evolved beyond the need to import thousands of dollars’ worth of rock from the quarry to the river bed—a foolish task when the creek itself is choked in gravel (akin to bringing coals to Newcastle).

A typical cut bank on Anderson Creek before stabilization

A typical cut bank on Anderson Creek before stabilization

A hard-point structure should do four things in a riverbed: provide protection for stabilizing banks, create pools and flush out “fines,” capture migrating topsoil, and establish a secure footing for intensive revegetation. We now use structures that I call “bio-cells” to create a vegetative hard-point. Geotextile fabric is constructed into cells eight to ten feet in diameter, filled with pure, on-site, river-run gravel, and planted in willows, cottonwoods, and alders to create a living jetty that gains strength as the trees grow. These bio-cells are stacked and joined together to form a monolithic unit, far more stable than large boulders that could be undermined and rolled. In two to three years, these structures become green walls, with their fabric frame invisible, buried in foliage and new soil. Our longest structure is 120 feet long, twenty-two feet wide, and nine feet high; the equivalent rip-rap with imported rock would have been cost prohibitive. Ours is now a mass of eighteen-foot saplings, three-year-olds growing at a tremendous pace, oblivious to the fact that they are placed right out in the strongest current—the thalweg of the river.

The same bank after stabilization

The same bank after stabilization

We could have simply armored the existing cut banks with rock, leaving the river to its own devices. However, by necking the river down to its “correct” width (determined by measuring the nearest intact stretch) with these now forested hard-points, the river begins to deepen its channel and find its bedrock again. For where there is a structure, there is a pool and, in between the pools, a shallow riffle. The rhythm between riffle and pool returns. By scouring out the gravel, we are restoring the creek to year-round flows and, thereby, restoring perennial aquatic life as well. The structures have captured wayward soil, now up to eight feet deep in places, and rebuilt the floodplain, while respecting the river’s natural sinuosity. Some of the riparian trees are now approaching sixty feet in height, having been in the ground for just eighteen years. There is now enough real soil and shelter for us to begin planting redwoods—the climax riparian tree for this part of Mendocino County.

The goal for this half-mile stretch of river is to return the riparian zone to the polar opposite condition from which we began: from overheated gravel bars to a closed canopy of trees. For is it not a total surprise to enter a riverbed on a hot summer day and experience the atmosphere of moisture and “coolth,” where fish can survive under a wad of roots in the deep dark shade?


Roots of red alder (Alnus oregona) begin to take over the geotextile fabric of a “bio-cell”

Roots of red alder (Alnus oregona) begin to take over the geotextile fabric of a “bio-cell”

Selected Riparian Trees for River Work

(ideally propagated from local sources)

Box elder (Acer negundo subsp. californicum)

Big-leaf maple (A. macrophyllum)

Red alder (Alnus oregona)

White alder (A. rhombifolia)

Black walnuts (Juglans californica var. californica and J. californica var. hindsii)

Western sycamore (Platanus racemosa)

Narrow-leafed cottonwood (Populus angustifolia)

Fremont cottonwood (P. fremontii subsp. fremontii)

Black cottonwood (P. balsamifera subsp. trichocarpa)

Valley oak (Quercus lobata)

Narrow-leaved willow (Salix lucida)

Coastal willow (S. hookeriana)

Red willow (S. laevigata)

Pacific Willow (S. lucida subsp. lasiandra)

Arroyo willow (S. lasiolepsis)

Scouler’s willow (S. scouleriana)

Coast redwood (Sequoia sempervirens)