Home Archived January 13, 2017

Potential Drift Accumulation at Bridges

Effects on Bridges

Drift has always been an important cause of bridge failures in the United States, especially in earlier years when quantitative hydrologic information was scarce and the importance of scour was not recognized (Edwards, 1959). The high design-discharge capacity, massive structure, and sturdy foundations typical of modern bridges help them withstand the effects of drift, but drift has remained a significant cause of failure and other damage (Chang, 1973).

Scour is the leading concern related to drift, followed by lateral forces (Pangallo and others, 1992). The presence of drift enhances pier and contraction scour. At most bridges identified as having drift problems, scour was the cause of damage or failure. In many cases, this scour occurred away from the drift accumulation.

Accumulated drift, acted upon by flowing water, exerts significant forces on piers and superstructures. At some bridges where flood waters reached the low chord of the bridge, drift forces were blamed for serious damage independent of scour effects (O'Donnell, 1973; Gannett Fleming Corddry and Carpenter, Engineers, 1974; Brice and others, 1978a; Chang and Shen, 1979). Forces exerted by drift on piers in combination with scour may have contributed to several pier failures.

The impact forces produced by drift striking bridges, and road overflow associated with drift accumulations, are less important effects of drift. Drift impact typically produces minor damage. Loss of conveyance in the bridge opening due to drift accumulation increases flood stages, road overflow, and embankment erosion.

Drift-Related Scour

Based on the cited case studies, scour associated with drift accumulation is the most common cause of bridge failure that involves drift. However, few studies of scour consider drift. Because the size, shape, location, roughness, and porosity of drift accumulations are highly variable, the effect of drift on scour is likewise variable (Laursen and Toch, 1956; Highway Research Board, 1970; Makowski and others, 1989; Richardson and others, 1991; Sherrill and Kelly, 1992; Becker, 1994). The potential for drift accumulation is relatively high for the design discharges (100-year, 500-year, and minimum overtopping discharge) used in the analysis of potential scour in the United States. Limited computer modeling of drift-related scour shows that drift can have a significant effect on scour, as well as on bridge-backwater effects (Prasuhn, 1981).

Scour associated with naturally accumulated drift is difficult to measure and has been measured at only a few sites (Becker, 1994). A few physical model studies of scour related to drift have been performed (Laursen and Toch, 1956; Foster, 1988; Dongol, 1989; Sterling Jones, Federal Highway Administration, written commun., 1996).

In this study, scanning sonar was used at the FM 2004 bridge over the Brazos River near Lake Jackson, Texas to observe scour around a large single-pier drift accumulation near the peak of a major flood. The drift accumulation extended from the bed to the surface, and had a maximum width of about 23 m (75 ft). Flow approaching the pier had a depth of about 8 m (25 ft). The maximum depth just upstream from the drift accumulation was about 11 m (35 ft), and under the spans adjacent to the accumulation the maximum depths were 12 m (40 ft) and 14 m (45 ft). Thus the maximum scour depth was about 6 m (20 ft) relative to the bed of the river upstream from the drift accumulation. The deepest scour occurred next to the widest lateral extensions of the accumulation, and evidently was a combination of local and contraction scour.

Laursen and Toch (1956) studied the effect of drift accumulations on scour in sand around the foot of a model pier 0.06 m (0.2 ft) wide. Their models of drift accumulations included accumulations of floating twigs allowed to form in the flume, bundles of twigs tied with string, and flat pieces of masonite fixed perpendicular to the flow. They found that drift increased scour depths except in the case of buried drift around the base of the pier. They did not attempt to quantify this effect, and concluded that:

"Debris, in effect, enlarges the pier and thus results in increased scour depths and areas. The difficulty in evaluating even qualitatively the effect of debris is that the permeability and the position are as important as the overall size of the debris mass." (Laursen and Toch, 1956)

Hopkins and others (1980) used the concept of effective pier width. Their study used field data from a pier in the Brazos River that characteristically had drift lodged in the piling cluster. They estimated that the effective width of the pier was approximately doubled by the addition of "complete debris buildup" to the footing and piling cluster. However, the authors were not able to measure actual drift accumulations and compare them to the corresponding measured scour depths.

Using the facilities of the Hydraulics Laboratory of the U.S. Army Engineer Waterway Experiment Station, Foster (1988) performed a model study of scour around a large drift accumulation on a large construction trestle in the Mississippi River. The drift accumulation had an area of about 4 hectares (10 acres), a width perpendicular to the flow of about 240 m (800 ft) (the full length of the work trestle), and an unknown thickness, assumed in the model study to be either 2.7 or 5.5 m (9 or 18 ft). The drift was simulated in the model by a layer of rubberized hair material. Based on the model study, the drift caused as much as 6 m (20 ft) of bed scour under the trestle and doubled the depth of local scour around cofferdams.

Dongol (1989) conducted a flume study of scour in sand around a cylindrical pier, using solid wooden shapes fixed to the pier. The models used were various cylindrical disks, an elliptical disk, and an inverted cone. Dongol also modeled drift accumulation around the pilings below the footings of the Wairoa Bridge in New Zealand, using foam rubber held in place by wire netting. Presence of drift at the surface increased scour depth and downward velocity at the pier nose. Dongol gives an equation for calculation of effective pier width and suggests that maximum scour is 2.3 times the effective pier width (Melville and Sutherland, 1988; Melville and Dongol, 1992). Based on field reports and case studies, he notes that drift trapped on piers is forced downward and that ultimately the effective pier width may be equal to the width of drift. This implies that the maximum scour depth "could be anticipated to be" 2.3 times the width of the drift accumulation. However, some of Dongol's models of large drift accumulations caused deposition around the pier, rather than scour. Thus, this maximum anticipated depth is not invariably the result of a large accumulation.

Scour may occur beneath or upstream from a drift accumulation, or away from the drift where increased flow velocity results from constriction of the bridge opening or deflection of the flow (Laursen and Toch, 1956; Klingeman, 1971; Klingeman, 1973; Rowe, 1974; Shen and others, 1981; Foster, 1988; Richardson and others, 1991; Becker, 1994). Downward vertical velocity is important in determining scour depth (Tison, 1961; Dongol, 1989). The presence of disks or plates near the base of a vertical pier reduces scour depth by interrupting downward flow (Tanaka and Yano, 1967; Thomas, 1967). The presence of drift at the base of a pier can also reduce scour depths near pile footings, probably in a similar manner (Laursen and Toch, 1956; Dongol, 1989; Sterling Jones, Federal Highway Administration, written commun., 1996). General scour and deflected flow are common where piers constrict the channel, and vortices at adjacent piers interact to produce additional scour (Blodgett, 1984; Elliot and Baker, 1985). Drift accumulations may increase these effects. Deflected flow may increase the velocity and skew at adjacent piers. Scour associated with drift can include removal of riprap as well as movement of natural bed material (Linder, 1967; Klingeman, 1973).

Drift accumulations at bridges promote several types of channel change. Sedimentation may occur among the logs of a drift accumulation and in the eddy just downstream, producing a bar surrounding a pier or in the eddy just downstream. Steep-faced lingulate bars were observed downstream from span blockages, representing the deposition of material transported under these accumulations or scoured from beneath them. Bars were also observed in the divergent, decelerating flow upstream from the area of local scour at the upstream edge of drift accumulations. Channel widening may expose piling clusters or skewed pile lines formerly on or in the bank. Channel migration promoted by drift may increase skew and thereby worsen the potential for drift accumulation. It has been observed that blocking more than 5 percent of the channel cross section with piers may cause scour (Blodgett, 1984). Even a small drift accumulation can exceed this threshold size; such accumulations are common at selected scour-potential-study bridges (table 4).

Table 4. Percentage of channel blocked by drift at selected scour-potential-study bridges.

previous || table of contents || next

U.S. Department of the Interior, U.S. Geological Survey
Last update: Tuesday, 08-Jan-2013 14:46:18 EST
URL: http://tn.water.usgs.gov /publications/FHWA-RD-97-028/effects.htm
USGS Privacy Policy || Disclaimer