U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590

Skip to content U.S. Department of Transportation/Federal Highway AdministrationU.S. Department of Transportation/Federal Highway Administration

Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations

This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-RD-99-156
Date: August 2004

Enhanced Abutment Scour Studies for Compound Channels



Recent bridge failures caused by local scour around piers and abutments have prompted a need for better technical information on scour prediction and scour-protection measures.(1-3) In 1987, the I-90 bridge over Schoharie Creek near Albany, NY, failed because of local scour around the pier foundations, resulting in the loss of 10 lives and millions of dollars for bridge repair/replacement.(2,4) During the 1993 upper Mississippi River basin flooding, more than 2400 bridge crossings were damaged.(5) In 1994, tropical storm Alberto caused numerous bridge failures as a result of the 100-year flood stages being exceeded at many locations along the Flint and Ocmulgee Rivers in central and southwest Georgia.(6)

There are more than 480,000 bridges over water in the United States.(4) As demonstrated in the past decade, the potential for loss of life and serious disruption of a local economy in the event of a bridge foundation failure caused by an extreme flood is very significant. As a result, a comprehensive effort has been undertaken by the Federal Highway Administration (FHWA) to require all States to evaluate highway bridges for scour potential.(7) Approximately 17,000 bridges have been identified by State departments of transportation (DOTs) as "scour critical" with the potential for scour-related failure of the foundations as a result of a flood disaster. Several technical publications have been developed by FHWA, including Hydraulic Engineering Circular Number 18 (HEC-18),(8) to provide guidance to the engineer on evaluating scour problems at bridges. Unfortunately, the alteration of flow patterns by bridge crossings and the concomitant scour process are quite complex and have defied numerical analysis for the most part. The scouring is the result of flow separation around a bridge pier or abutment and the formation of three-dimensional, periodic horseshoe and wake vortexes that interact with a movable sediment bed. In this situation, the engineer has been forced to rely on laboratory results that are numerous and sometimes conflicting because of idealized laboratory conditions that have been used in the past. In particular, the task of predicting local scour around bridge abutments remains problematic, with many engineers not trusting the current empirical formulas given in HEC-18 that are based on laboratory experiments in rectangular flumes.(9)

Previous laboratory experiments on abutment scour have emphasized the abutment length in a rectangular channel as one of the primary variables affecting scour. In an actual river consisting of a main channel and adjacent floodplains, an abutment terminating in the floodplain is not subject to the idealized, uniform approach velocity distribution obtained in previous laboratory experiments in rectangular flumes.(10) Instead, the scour is a function of the redistribution of flow between the main channel and the floodplain as flow through the bridge opening occurs. In other words, abutment length is certainly important; however, the same abutment length may result in different scour depths depending on the approach flow distribution in the compound channel and its redistribution as it flows through the contracted opening.(11)

Currently, FHWA recommends the prediction of abutment scour with a regression equation developed by Froehlich(12) that is based entirely on results from experimental investigators using rectangular laboratory channels. Laursen(13) has developed an equation for clear-water abutment scour that is based on contraction hydraulics, but that relies directly on abutment length.

Melville(14) has proposed a methodology for predicting maximum abutment scour that also depends on abutment length for short- and intermediate-length abutments, and it does not include the effects of overbank flows or of flow distribution in compound channels. However, Melville(15) has also considered the case of compound channels under the condition that the abutment extends into the main channel rather than terminating on the floodplain. This case of encroachment into the main channel itself would be less common in practice than the abutment terminating on the floodplain or at the edge of the main channel.

The concept of flow distribution in a compound channel depends on the interaction between main-channel flow and floodplain flow at the imaginary interface between the two where vortexes and momentum exchange occur. The net result is that less discharge occurs in the compound channel than would be expected from adding the separate main-channel and floodplain flows that would occur without interaction. The research by Sturm and Sadiq,(16) for example, suggests methods by which predictions of flow distribution between the main channel and the floodplain can be improved in the case of roughened floodplains. Wormleaton and Merrett(17) and Myers and Lyness(18) have also proposed techniques for predicting flow distribution in compound channels. A more detailed review of the literature on compound-channel hydraulics can be found in chapter 2 of this report.

Sturm and Janjua(19-20) have proposed a discharge contraction ratio as a better measure of the effect of abutment length, and the flow redistribution that it causes, on abutment scour. The discharge contraction ratio is a function of abutment length and compound-channel geometry and roughness. It can be obtained from the output of the water-surface profile program, WSPRO.(21) The research reported herein, however, attempts to clarify the influence of bridge backwater on the flow redistribution and to improve the WSPRO methodology for computation of flow redistribution by incorporating more recent research results on compound-channel hydraulics.(22)

The abutment scour experiments by Sturm and Janjua(20) and Sturm and Sadiq(23) used a single, uniform sediment size of 3.3 millimeters (mm). The effect of sediment size on the equilibrium scour depth has been incorporated by including, as an independent variable, the ratio of the approach velocity in the floodplain to the critical velocity for the initiation of motion, which depends on sediment size. Experiments with three different sediment sizes at varying discharges and abutment lengths for a vertical-wall abutment were conducted in the present research to verify this method for quantifying the influence of sediment size on equilibrium scour depth. In all cases, the abutment length was large relative to the sediment size in order to remove any effects of energy dissipation caused by large, uniform sediment sizes in the bottom of the scour hole relative to a short abutment length.(24)

The effect of abutment shape was considered in this research by conducting experiments in a compound channel on vertical-wall, wingwall, and spill-through abutment shapes. A single sediment size of 3.3 mm was used for this series of experiments, and the discharge and abutment lengths were varied. The abutment shape effect has been shown to be insignificant for long abutments in rectangular channels(15); however, this behavior has not been verified for long abutments in compound channels. Currently, FHWA procedures assume a reduction in scour for a spill-through abutment of 55 percent in comparison to the vertical-wall abutment.(8)

Additional experiments were also conducted on abutment lengths that approached the bank of the main channel for both the spill-through and vertical-wall abutments. The purpose of these experiments was to test the methodology developed for abutments that terminated on the floodplain for the more complicated three-dimensional flow field that occurs when the contracted flow joins the main channel near the abutment face.

The live-bed scour case was considered analytically for the condition of sediment transport in the main channel with no sediment movement in the floodplain. This case would be of interest for the abutment located at or near the bank of the main channel with the scour hole occurring at least partially in the main channel rather than on the floodplain alone. Although experiments for this case were attempted, they were not successful because of the limitations of the present compound-channel geometry in the flume.

Finally, a brief implementation procedure for the proposed methodology was developed and applied to a hypothetical example. It was also tested on two field cases of scour in Minnesota using data measured by the U.S. Geological Survey (USGS) in 1997.


The objective of the proposed research is to develop better predictive equations for assessing the vulnerability of existing bridges to abutment scour for cases in which the abutment is located anywhere in the floodplain up to the bank of the main channel in a compound-channel geometry. Of primary interest is the effect of compound-channel hydraulics(16) on the redistribution of the main-channel and floodplain flows as the flow accelerates around the end of an abutment of varying shape. In addition, the effect of sediment size on abutment scour needs to be clarified as well as the effect of live-bed versus clear-water scour.

The experimental research reported herein differs from most of the previous experimental studies of abutment scour that have not had a compound channel as the approach channel and that have not considered the effect of very long abutments in wide, shallow floodplain flow. The results of the experimental research have been used to develop a prediction equation for clear-water abutment scour that was tested on limited field data and was compared to the results of the other scour-prediction methods.

The specific research objectives are:

  • Investigate the effects of flow distribution, as affected by abutment length, on clear-water abutment scour in a compound channel for abutment lengths that terminate on the floodplain as well as encroach on the bank of the main channel.
  • Quantify the effects of floodplain sediment size on abutment scour.
  • Explore the influence of abutment shapes, including wingwall, vertical-wall, and spill-through shapes, on equilibrium scour-hole depth and scour-hole form.
  • Determine the relative importance of the live-bed scour case compared to the clear-water case when there is sediment transport in the main channel and the abutment encroaches on the main channel.
  • Combine the experimental results into a methodology for assessing field abutment scour and test it on available field data.

This report provides a brief review of the literature on compound-channel hydraulics, and on clear-water and live-bed abutment scour in chapter 2. The experimental investigation is described in detail in chapter 3. Chapter 4 contains an analysis of the experimental results and a proposed abutment scour-prediction equation that addresses the effects of the alteration of the flow distribution by both short and long abutments, sediment size, abutment shape, time development, and live-bed conditions. A procedure for implementing the research results for the purpose of identifying scour-susceptible bridges in the field is then suggested in chapter 4 along with an example of a field application. The proposed procedure is tested with limited field data. Conclusions and recommendations are given in chapter 5.

Back to Top

Previous | Table of Contents | Next