Home Archived January 13, 2017

Potential Drift Accumulation at Bridges


Most drift is wood from trees growing close to the channel (Lyell, 1856; McFadden and Stallion, 1976; Harmon and others, 1986; Perham, 1987; Murphy and Koski, 1989; Robison and Beschta, 1990a; Lagasse and others, 1991; Diehl and Bryan, 1993). Most such trees fall into the stream as a result of bank erosion, while others are felled by wind throw, ice, disease, or old age. Flood plains can provide drift to the channel under some circumstances. Landslides and debris flows also transport large woody debris in areas of steep topography (Everitt, 1968; Keller and Swanson, 1979; Harmon and others, 1986).

Channel characteristics that determine the rate of bank erosion, such as bank height, bank angles, bank materials, erosion rates, and drift concentrations, vary from point to point within a given basin. These variations may be abrupt, and the bridge site may not represent average conditions (Diehl and Bryan, 1993). In order to assess the potential supply of drift at a bridge site, channel characteristics should be evaluated in all upstream reaches of the channel system that can transport drift, not just at the bridge site (Pangallo and others, 1992).

In the present study, most logs from bank-grown trees were recognized by curved trunks and asymmetric root masses (figure 7) (Simon and Hupp, 1992; Diehl and Bryan, 1993). Trees rooted in alluvium stabilize banks by binding sediment together with their roots. When stream banks erode, tree bases typically remain attached and project into the channel beyond the retreating bank. Growth of roots continues in the bank, while the exposed parts of the root mass become burly, develop rougher bark, and are often scarred by the impact of drift. The trunk leans toward the stream as it is deprived of support on that side, while the top of the tree grows straight upward. Logs introduced into the channel through bank erosion could sometimes be recognized by the development of strongly asymmetric root masses with fine roots only on one side. Rapid bank retreat, however, produces symmetrical logs from trees that grew on the flood plain.

120K GIF version of this photograph

Figure 7. Fallen tree with asymmetric root mass and slightly curved trunk.

Bank erosion and the resulting introduction of large woody debris are concentrated on the outside of bends where high shear stress promotes erosion (figure 8) (Elliot and Sellin, 1990; Knight and Shiono, 1990). Because the highest flow velocities are near the outside bank, trees that fall from it are likely to be mobilized and transported.

120K GIF version of this photograph

Figure 8. Bank erosion along outside of curve.

Widespread bank erosion producing abundant drift typically results from channel instability (Brice and others, 1978a; Lagasse and others, 1991). Channel instability is a natural property of some channel types, but may also result from climate change, fire, or human modifications to the channel, flood plain, or river basin. The presence of abundant drift in the channel may aggravate instability (Murgatroyd and Ternan, 1983; Gippel and others, 1992). Although channel instability may be recognized on the basis of geomorphic features and the history of the basin, the extent and rate of channel change is difficult to assess (Mueller and Dardeau, 1990). Large floods and infrequent prolonged periods of high flow may cause abrupt, extensive bank erosion even in stable channels where bank erosion is otherwise localized and slow.

A history of human alteration of the channel system and accompanying channel instability indicates that continued instability and drift production is likely. Where channelization has increased the channel slope of the main stem or major tributaries, the channel system will likely be unstable, and will continue to adjust to its new slope and alignment except where stabilization structures and non-erodible channel boundaries prevent it from doing so (Brookes, 1988; Simon and Hupp, 1992). Such adjustment may occur only in infrequent floods, producing no chronic drift problem, yet still creating high potential for abundant drift delivery during these channel-altering floods.

Extensive ditching and wetland drainage upstream from the site typically produce changes in the flow-duration relation of the stream and thus promote accelerated channel evolution. Urbanization and conversion to agriculture typically involve extensive development of ditches. Drift removal from channels can increase flow velocity, leading to channel instability and further drift production (Gippel, 1989; Gippel and others, 1992; Smith and others, 1992). Clearing of the flood plain can also promote channel instability (Brice and others, 1978a).

Logging has been cited as one of the major sources of drift (Chang and Shen, 1979; Lagasse and others, 1991). Much of the research devoted to drift and untransported large woody debris in the Pacific Northwest has been motivated by concern about timber harvesting practices that leave too much or too little large woody debris in stream channels (Ice and Lawrence, 1985; Bisson and others, 1987; Bilby and Wasserman, 1989). However, mountain and foothill streams of the Pacific Northwest historically contained large amounts of drift before logging began in the region (Orme, 1990). Drift accumulations observed in Washington contained little saw-cut material; most of the large logs in the accumulations included root masses. Of the studies reviewed, none cited forestry in gently sloping basins as a source of drift in streams.

Logging practices that directly disturb the stream corridor are responsible for most forestry-related drift (Bryant, 1980; Bryant, 1983; Bryant, 1985; Phillip D. Martin, Quinault Tribe, oral commun., 1995). Logging of the stream corridor increased the amount of woody debris in areas of steep topography with channels bounded by bedrock and gravel deposits. Current forestry practices typically include leaving a strip of trees along the stream and avoiding disturbance to the banks and bed. Where such practices are successful, logging may now be a less important source of drift than it has been in the past (Dykstra and Froelich, 1976). However, clear-cutting may increase stream discharge, leading to channel adjustment through erosion (Harr, 1976).

Wind throw of trees growing on the bank produces logs that include root masses. Erosion around wind-thrown logs may cause additional trees to be introduced into the channel (Bryant, 1980; Bryant, 1985). In basins near the Atlantic coast, hurricanes have caused the delivery of large amounts of drift, coincident with high discharge, in streams that are otherwise nearly free of drift (O'Donnell, 1973; Gannett Fleming Corddry and Carpenter, Engineers, 1974; Brice and others, 1978a).

Bank erosion, wind throw, and ice storms involve the same population of trees ---- those that grow on banks, bank tops, or flood plains immediately adjacent to bank tops. Wind and ice also promote the introduction of trees into the channel, but are most effective where erosion has already reduced the strength of the root system. Trees rooted in erosion-resistant materials like bedrock are much less easily detached. The prevalence of erosion over wind throw and ice damage is supported by indications of channel instability at most drift-study sites.

Flood plains may act as sources of drift where flood-plain flow is deep and few trees are present to intercept drift (Benke and Wallace, 1990; Pangallo and others, 1992). These conditions occur in the western United States along rivers where cottonwood trees grow sparsely, and in cleared flood plains where fallen or cut logs do not lie upstream of an effective barrier to drift transport. However, deep rapid flow over flood plains did not topple or break living trees at sites observed in this study.

Some processes producing abundant drift are specific to steep, forested areas (Everitt, 1968; Calver, 1969; Swanson and Lienkaemper, 1978; Keller and Swanson, 1979; Harmon and others, 1986; Hogan, 1987; Orme, 1990). These processes include landslides, debris flows, and debris torrents; some flow events combine features of all three (Pierson and Costa, 1987). These processes occur locally and infrequently in steep slopes or channels, but may occur simultaneously at several locations in a basin in response to torrential rains. In debris torrents, drift and other debris accumulating against the upstream side of standing trees exerts sufficient force to break or uproot these trees. These trees, in turn, become part of the debris in the torrent. Debris torrents are limited to valleys with at least a 3-percent slope (Swanson and Lienkaemper, 1978; Keller and Swanson, 1979; Hogan, 1987).

Human artifacts are often included in drift accumulations. Two small accumulations appeared to be about half trash and half woody debris. Other accumulations contained objects of human origin such as lumber, fragments of buildings, toys, trash, and appliances, but these objects were not large enough, strong enough, or abundant enough to play a significant structural role.

Trees felled by beavers and humans are rarely important in drift accumulations. Beaver-cut logs were rare or absent at drift-study sites. Two reported drift accumulations in Manitoba were composed partly of drift from washed-out beaver dams (James Lukashenko, Penner and Keeler Partners, written commun., 1994). Three other reported accumulations consisted largely of sawlogs from storage areas (Brice and others, 1978b; I. Nagai, California Department of Transportation, written commun., 1992; James Lukashenko, Penner and Keeler Partners, written commun., 1994; Martin Fisher, Washington State Department of Transportation, written commun., 1995). Several accumulations contained a few saw-cut logs or lopped branches, but logging debris was never more than a small fraction of the total drift accumulation.

previous || table of contents || next

U.S. Department of the Interior, U.S. Geological Survey
Last update: Tuesday, 08-Jan-2013 14:46:20 EST
URL: http://tn.water.usgs.gov /publications/FHWA-RD-97-028/generati.htm
USGS Privacy Policy || Disclaimer