• Federal Highway Administration >
• Publications >
• Research Publications >
• Hydraulics Publications >
• 05072 >
• 03.Cfm >
• Assessing Stream Channel Stability At Bridges in Physiographic Regions

Assessing Stream Channel Stability At Bridges in Physiographic Regions

3. FIELD OBSERVATIONS

Numerous stream-bridge intersections were observed across the United States to develop and test the stability assessment method. The streams were to reflect a broad range of stream types and physiographic regions; thus, 57 site visits were conducted in 13 physiographic regions and subregions, including Pacific Coastal, Basin and Range, Trans Pecos, southern Rocky Mountains, Great Plains, Central Lowlands, Interior Lowlands, Ozark-Ouachita Plateau, Appalachian Plateau, Valley and Ridge, Piedmont regions, and Atlantic Coastal Plain.

In addition to collecting observations at streams covering a variety of erosion issues, sizes, and physiographic regions, the following criteria also were used in selecting appropriate sites:

• All channels were alluvial or partially alluvial(occasional rock outcrops were acceptable).
• Engineered (straightened or widened) channels were included, although manmade canals were not.
• The streams had to be wade-able or partly wade-able.
• The streams and bridges had to be safely accessible.
• A reasonable level of personal physical safety had to be satisfied.
• The streams were located within a reasonable distance of the travel corridor.

The data for each of the streams are summarized in tables 5-7. Streams that are named N# in table 5 are unnamed on topographic maps. Table 5 provides the locations and global positioning system (GPS) coordinates of the bridges, the physiographic Province, land use, and stream classification. Each of the channels was classified according to the Montgomery-Buffington scheme. The Montgomery-Buffington scheme does not include engineered or altered channels. However, this method is still useful as a basic descriptor of the primary processes in the stream (e.g., transport versus response) and is, therefore, included. To include altered streams, the USACE method (10) and a simple observation of channel pattern (based on both field observation and aerial photos) also were used to classify or categorize the stream types. The resulting stream type is provided as a combination of these methods in table 5. The Rosgen classification method was not considered because it is unnecessarily data intensive for the purposes of assessing channel stability in the vicinity of a bridge. Table 6 provides the bed and bar material, the percent of sand (F_s), and any controls observed in the banks or on the bed. Table 7 provides observations made on the banks, including vegetation, bank material, bank height, and any erosion characteristics. In the next section, observations made in each of the physiographic regions are described.

*C = cascade, S = step pool, P = plane bed, R = pool-riffle, D = dune-ripple, B = braided, MT = mountain torrent, MA = meandering, MO = modified, S.R = State Route, Cr. = Creek, R. = River

*C = cascade, S = step pool, P = plane bed, R = pool-riffle, D = dune-ripple, B = braided, MT = mountain torrent, MA = meandering, MO = modified, S.R = State Route, Cr. = Creek, R. = River

*C = cascade, S = step pool, P = plane bed, R = pool-riffle, D = dune-ripple, B = braided, MT = mountain torrent, MA = meandering, MO = modified, S.R = State Route, Cr. = Creek, R. = River

*C = cascade, S = step pool, P = plane bed, R = pool-riffle, D = dune-ripple, B = braided, MT = mountain torrent, MA = meandering, MO = modified, S.R = State Route, Cr. = Creek, R. = River

*C = cascade, S = step pool, P = plane bed, R = pool-riffle, D = dune-ripple, B = braided, MT = mountain torrent, MA = meandering, MO = modified, S.R = State Route, Cr. = Creek, R. = River

*C = cascade, S = step pool, P = plane bed, R = pool-riffle, D = dune-ripple, B = braided, MT = mountain torrent, MA = meandering, MO = modified, S.R = State Route, Cr. = Creek, R. = River

Fs = portion of sand, u/s = upstream, d/s = downstream, W = width, R. = River, Cr. = Creek

Fs = portion of sand, u/s = upstream, d/s = downstream, W = width, R. = River, Cr. = Creek

Fs = portion of sand, u/s = upstream, d/s = downstream, W = width, R. = River, Cr. = Creek

Fs = portion of sand, u/s = upstream, d/s = downstream, W = width, R. = River, Cr. = Creek

Fs = portion of sand, u/s = upstream, d/s = downstream, W = width, R. = River, Cr. = Creek

Fs = portion of sand, u/s = upstream, d/s = downstream, W = width, R. = River, Cr. = Creek

Fs = portion of sand, u/s = upstream, d/s = downstream, W = width, R. = River, Cr. = Creek

LB = left bank, RB = right bank, u/s = upstream, d/s = downstream, R. = River, Cr. = Creek

LB = left bank, RB = right bank, u/s = upstream, d/s = downstream, R. = River, Cr. = Creek

LB = left bank, RB = right bank, u/s = upstream, d/s = downstream, R. = River, Cr. = Creek

LB = left bank, RB = right bank, u/s = upstream, d/s = downstream, R. = River, Cr. = Creek

LB = left bank, RB = right bank, u/s = upstream, d/s = downstream, R. = River, Cr. = Creek

LB = left bank, RB = right bank, u/s = upstream, d/s = downstream, R. = River, Cr. = Creek

PHYSIOGRAPHIC REGIONAL OBSERVATIONS

Thirteen physiographic regions and subregions were included in the data set. A wide variety of land uses were observed in the various regions, including natural, agricultural, grazed, rural, and suburban. Width-to-depth ratios also varied widely, from 5 to 24. As expected, the largest ratios were associated with braided or semibraided channels. Very low width-to-depth ratios were associated with incised streams or those that had been engineered. Bed materials varied from very fine materials (silt and very fine sand) in the Midwest and coastal areas to coarser materials mostly associated with higher elevation streams. Bank materials varied widely. For example, the banks of the streams in the Central Plains tend to be made up of fine loess and silt deposits, which erode easily. In the Appalachian Plateau region, in contrast, the bank materials are far more cohesive and tend to be less susceptible to high erosion rates even when bank vegetation was limited.

Although stream types vary widely within any of the physiographic regions, certain common characteristics were observed. A general summary of those observations is given here. The photos of the sites are organized by physiographic region in appendix A. Many, but not all, of them are referred to in the discussion below.

Pacific Coastal

There are two striking features of stream channels within the border and lower Californian subregions of the Pacific Coastal region (see appendix A). The first is the wide diversity in types of channels. Streams range from perennial, cascading channels to arroyos. The second feature common to most streams in these subregions is the frequency of human interference and alteration. The channel bed material in most of the streams (other than first-order streams) was predominantly fine-to-medium sand, while the channel banks were sandy. The average width-to-depth ratio was 12.1. The streams tend to be very high energy; they are typically ephemeral, so they carry water only when there is rainfall. Many of these streams (and arroyos) tend to be naturally unstable, particularly in the lateral direction, and have relatively high width-to-depth ratios. Because of the high degree of channel instability and flash flooding in this region, many, if not most, of the channels in suburban to urban settings were either concrete lined or at least heavily armored with rock. Channels in the outlying areas were unlined.

Intermontane

Observations at bridge-stream intersections were collected in the Basin and Range, Colorado Plateau, and Trans Pecos subregions within the Intermontane physiographic region. Many channels within the Colorado Plateau are bedrock or semialluvial channels in which stability is a function of bedrock erosion. In the Basin and Range, however, where the climate is arid to semiarid over much of the area, the streams are ephemeral with high energy, flashy flows (see appendix A). The energy of these streams combined with the highly erosion-prone sand beds and banks creates unstable channels, particularly at bridges. Due to the high sediment load carried by these streams, the width-to-depth ratios are relatively high, with an average of 25.0. Streams in the Trans Pecos region (see appendix A) tend to have high, steep banks or valley walls, which create valley side failures and subsequent failure material to the stream. The bed materials are sand, and banks are comprised of a mix of noncohesive materials, primarily sand, with minor amounts of cohesive silts and clays. The predominant bank vegetation was desert shrubs, and mass wasting was a common form of erosion. The average width-to-depth ratio was 10.5.

Rocky Mountain System

The southern Rocky Mountains were visited in this region (see appendix A). The channels contain large bed and bank material. The streams tend to be stable, transport streams that are less disturbed by human activities than in other physiographic regions. Stream channel banks are a mix of cohesive silts and clays and noncohesive gravels and larger materials. The average width-to-depth ratio was 16.5.

Interior Plains

Three subregions-the Great Plains, Central Lowlands, and Interior Lowlands-were visited within the Interior Plains. In the Great Plains (see appendix A), vegetation in riparian areas and in the flood plains was thick, lush, and dense, except where cattle were permitted to graze. The channel beds were composed of more than 70 percent sand. Bank material was noncohesive silt, loam, and sand, but the thick vegetation helped to keep banks stable. The average width-to-depth ratio was 17.3. Erosion processes within the stream channels are primarily fluvial; observed channel banks were not sufficiently high to create significant mass wasting. Slow to moderate degradation occurred where cattle grazing was permitted.

The Central Lowlands (see appendix A) had silt and loess banks that eroded easily. Many of the streams observed had been straightened in addition to having extensive hoof and/or grazing damage. The channel beds degrade rapidly since the bed material is predominantly silt with some clay and sand. The silt banks then become overheightened, and mass failures result. High, failing banks were common even where a wider riparian buffer existed, but the rate of failure was slower (for example, see figure 7). The average width-to-depth ratio was much lower than that observed in the Great Plains, with an average of 8.3. This may be due in part to channel modifications, such as straightening.

Streams in the Interior Lowlands (see appendix A) seemed less fragile than those in the Central Lowlands due to larger bed material (sand and gravel) and more cohesive materials in their banks (clay and silt). However, where vegetation had been removed, banks failed even when they were not greatly overheightened (see figure 8). A single row of trees in the riparian areas slowed bank failure dramatically. These streams tend to have a low width-to-depth ratio, with an average of about 10.5, and remain stable even when the surrounding land has been disturbed. This may be due to the existence of rock outcrops in the beds and banks.

Interior Highlands

In the combined Ozark-Ouachita Plateau (see appendix A), bed material was larger, containing some gravel. The bank material contains a significant percentage of cohesive clays. Natural erosion occurred at bends with increased mass wasting at bends where vegetation had been removed. Overheightened banks remained stable when more than one row of trees was in place. The average width-to-depth ratio was 8.5.

Appalachian Highlands

Within the Appalachian Highlands, the Appalachian Plateau, Valley and Ridge, and Piedmont regions were visited. In the Appalachian Plateau (see appendix A), bed material was coarser (mainly very coarse gravel to cobbles), with bank material composed of cohesive clay, silt, and minor sand. Critical bank heights appeared to be about 1.5 to 1.8 m, which result in low width-to-depth ratios. For the sites visited, the average ratio was 11.0. Watersheds are heavily forested where vegetation is undisturbed. Stream channel erosion and destabilization occurs through removal of vegetation and/or channel straightening. Overheightened banks may fail, but heal quickly if vegetation is allowed to re-establish; thus, stability tends to be fair at worst.

Stream channels within the Piedmont region (see appendix A) had cohesive banks that could stand at high angles without failure. Bank vegetation, if undisturbed, was dense and provided bank stability with about one river width of woody vegetation. Banks with angles steeper than about 60E tended to have leaning or fallen trees. The potential for debris jams is high. Occasional bedrock outcropping was noted at all streams that were visited. Bed material was sand and gravel with occasional larger material. The average width-to-depth ratio was 15.3.

The Valley and Ridge region of the Appalachian Highlands is comprised of a series of ridges separated by stream valleys. The streams in this region are often very steep, especially coming down from the ridges (see appendix A). Cascade and step-pool morphologies are common. Thus, bed materials are commonly large, such as cobbles and boulders, and often armor the bed. Banks are cohesive clays and silts with some larger materials mixed in, strongly held together by the lush vegetation found in this area. Disturbance to the banks by removal of vegetation may result in ragged, scalloped banks, but erosion of the banks is typically at a relatively slow rate. The average width-to-depth ratio was 15.0.

Figure 7. Failing banks in the Central Lowlands.

Figure 8. Failing banks in the Interior Lowlands.

Coastal Plain

The Coastal Plain covers a very large area of the Atlantic and Gulf coastal areas (see figure 6). Sites were visited along both the Atlantic and Gulf coasts (see appendix A). In both of these areas, researchers observed moderate rates of bed degradation and bank failure. A buffer of at least one river width appeared to be sufficient in most locations to keep banks stable. Where undisturbed, lush vegetation on the banks held the banks in place, resulting in excellent stability, even when banks were nearly vertical. Bed material is typically sand with minor amounts of small gravel, and banks are cohesive with clay, silt, and minor amounts of sand. Because of the cohesive banks, strong vegetative resistance, and degradation, width-to-depth ratios tended to be rather low. Streams in this region are often sluggish due to low slopes and backwater from the bays or estuaries into which they flow. Where banks or the flood plain are disturbed, debris jams are frequent. The average width-to-depth ratios were much lower in the Gulf area (9.0) than in the Atlantic area (13.5).

New England

All of the streams visited in the New England region were located in Connecticut. At all streams, the banks were heavily vegetated with large woody vegetation, providing tremendous stability to the streambanks. The bank materials typically were comprised of some cohesive materials combined with silt, sand and, in some places, gravel and larger particles. The bed materials in the New England region are considerably larger than in the Atlantic Coastal Plain to the south. The sand, gravel, and cobble beds were often armored; the width-to-depth ratios reflected this armored condition with an average value of 24. The channels were all meandering, but with beds transitional between plane beds and pool-riffle beds.

General Observations of Streams at Bridges

Channel stability is a function of levels of disturbance to the water and sediment discharges, and susceptibility of the channels to change. In every physiographic region, the disturbance that caused the greatest damage to the streams was the combination of cattle activity, vegetation removal, and channel straightening. The combined impact of these activities was worst where cattle had direct access to streams. Also, susceptibility of the channel banks to erosion significantly impacted the level of damage. Figures 9 and 10 provide examples of this combination of disturbances. All vegetation has been removed either through farming practices or by cattle grazing. The channel apparently had been straightened to provide better drainage and to maximize land for farming. Not only are cattle grazing in this area, but also they have direct access to the stream. Hoof damage is extensive. The combined disturbances have resulted in stream channel destabilization; the channel bed elevation has degraded and the banks have become overheightened and steepened. Figure 10 shows the eroding channel beneath the single-span bridge.

In many cases, maintaining a riparian buffer of an appropriate width is all that is needed to preserve channel stability. As discussed in the descriptions of the streams channels across the physiographic regions, some regions require only a single row of trees to help maintain stability, while others require a much greater width. This is due to bank materials and the susceptibility of the banks to failure. In the cases where channels are degrading because of channel straightening, cattle grazing, and urbanization effects, a vegetation buffer may not be enough to maintain stability. When the channel degrades, banks can become overheightened and fail through mass wasting. In this case, vegetation may help to slow the rate of failure, but usually cannot prevent collapse of the banks.

Figure 9. Stream impacts due to disturbances, including hoof damage,
vegetation removal, and channel straightening.

Figure 10. Impacts of disturbances at bridge (from figure 9).

Another observation that was frequently made at sites in all physiographic regions was that there was often a distinct change in channel stability upstream and downstream of the bridges. This was caused in every case by a change in property management, as it is common for a road (and, thus, a bridge) to divide property ownership. As an example, unnamed stream N 28 is wooded upstream, with a healthy wide band of upright trees keeping the banks stable (see figure 11). Immediately downstream of the bridge, all trees and other vegetation have been removed, resulting in destabilization of the banks (see figure 12).

Aerial photos were examined for each of the sites using http://terraserver-usa.com (these photos are not included in the report because they are readily available online). The photos were examined to check a larger view of the river, specifically looking at land use in the watershed and flood plain, construction areas, the extent of the riparian buffer, channel straightening, and channel pattern. In most cases, the aerial photos reinforced observations that were made on the ground. In a number of cases, the photos helped put the bridge reach into the perspective of the meander pattern, particularly where the bridge was located between meanders or just downstream of a tight meander. Old abandoned meanders also could be detected sometimes, giving an indication of previous lateral movement. Changes in channel pattern, for example from meandering to braided, can be detected on aerial photos. Examining the photos before or after visiting a site helped provide a rating, especially for the watershed condition factor.

Figure 11. Wooded land upstream of bridge.

Figure 12. Downstream of figure 11, vegetation removed.

EFFECT OF CHANNEL INSTABILITY ON BRIDGES

Unstable channels can cause a variety of problems at bridges; however, this is not necessarily the case. For example, the Mojave River in California (see figure 13) can be considered to be a naturally unstable channel, primarily in the lateral direction, in that there is considerable lateral movement of the channel. The channel bed and banks are noncohesive fine sand that adjust readily to sudden changes in hydrology from a dry condition to flash flooding. However, the bridge at the site that was visited spans a wide section of the flood plain, thus providing room for some lateral migration. In many other sites visited, lateral migration of meanders was a potential threat to bridge abutments. In figure 14, lateral migration of a gentle meander bend has forced the channel against the left abutment. This has, in turn, caused additional local scour at the abutment and undermining of the abutments, and could result in an unstable bridge foundation. Lateral and downstream migration of this meander would have a significant impact on the left abutment.

Figure 13. Mojave River, CA.

Figure 14. Meander migrationaffecting right abutment, Hammond Branch, MD.

One of the biggest problems created by channel instability at bridges exists at single-span bridges that are only as wide as the channel. This allows for no or limited lateral or vertical adjustments of the channel. As an example, figure 15 shows a single-span bridge across a channel that is both degrading and widening. Significant widening will result in undermining of the abutment walls.

Even for channels that are unstable, the bridge may not be in danger if adequate structural redundancy is in place. Thus, an observation of channel instability is not a sufficient condition for impending structural failure. The bridge inspector must consider what impact, if any, a channel that is deemed unstable will have during the time period between inspections, especially in the event of a large hydrologic event.

Channel stabilization measures at bridges are quite common. Given the small right-of-way at most bridges, the measures typically are placed directly at the bridge and perhaps a short distance upstream or downstream. By far, the most common type of stabilization measure observed at these sites was riprap. In some cases, the riprap appeared to be effective in holding the bank in place at the bridge. In other cases, however, riprap did not appear to be effective without significant maintenance. For example, at S.R. 445 over Roaring Run in Pennsylvania, there is a high riprap wall composed of graded riprap with a median size of about 152-229 mm (see figure 16). The purpose of the wall is to prevent lateral migration of the tight meander bend just upstream of the bridge. The wall has a bank angle of about 70E. This configuration of loose, undersized riprap in such a steep arrangement has little chance of withstanding the high shear stresses imposed on it at high flows as the high gradient stream makes this tight bend. There is already evidence of riprap wall failure, as much of the stone is deposited in the stream channel just upstream of the bridge. In other cases, stabilization efforts seem to work quite well. As an example, a cross vane has been installed just downstream of the S.R. 144 bridge over Potter Run in Pennsylvania (see figure 17). The cross vane causes the flow to pool just upstream and under the bridge, slowing the high velocity and minimizing scour under the bridge and along the banks.

RELATIONSHIP BETWEEN CHANNEL STABILITY AND SCOUR AT BRIDGES

In HEC-18, scour is defined as having three vertical components: local, contraction, and bed degradation. Local and contraction scours are caused by the bridge and occur within close vicinity of the bridge. Bed degradation, on the other hand, is not caused by the bridge and may be reach-wide or even systemwide. Channel instability includes bed degradation, but also comprises other components, based on the definition given previously, such as channel widening, lateral migration, and bed aggradation. At bridges, channel instabilities can cause:

• Channel bed degradation, which may undermine the bridge foundations.
• Channel widening, which can undermine and outflank bridge abutments and piles in the flood plain.
• Lateral migration, which can undermine abutments and permit local scour to be far more productive as the channel thalweg nears an abutment.

Channel aggradation in itself is not usually detrimental to the bridge structure, but it can lead to increased flooding and channel widening. At many of the bridges observed during this project, narrow, single-span bridges often were impacted more because small lateral movements of the channel could press the stream thalweg up against one abutment, increasing the local scour at that abutment.

Figure 15. Single-span bridge over unstable channel.

Figure 16. Riprap stabilization wall along Roaring Run, PA.

Figure 17. Cross vane downstream of bridge over Potter Run, PA.

Previous | Contents | Next