Advanced Methodology to Assess Riprap Rock Stability At Bridge Piers and Abutments

APPENDIX A. ANNOTATED BIBLIOGRAPHY

This appendix contains an annotated bibliography of literature addressing riprap as a countermeasure for pier scour and related topics. Much of this literature describes failure Modes and layout guidelines.

RIPRAP AT BRIDGE PIERS

Riprap is commonly used to protect bridge piers from scour and has been studied for decades. In 1929, Engels produced one of the earliest reports of the use of riprap at bridge piers that was based on the results of hydraulic model experiments undertaken in 1893.⁽¹⁴⁾ Prior to these experiments, it was the general opinion that the greatest danger to pier foundations occurred at the downstream end of the pier rather than the upstream end and it was common practice to place riprap around the piers up to the low water mark. Engels concluded that the protection was most needed at the upstream end of the pier and that the riprap should be placed flush with the bed.

In 1959, de Sousa Pinto completed his thesis “Riprap Protection Against Scour Around Bridge Piers” under the supervision of Dr. C. J. Posey at the State University of Iowa.⁽¹⁵⁾ His study examined the use of riprap protection against erosion around a circular pier. Riprap layouts around the pier were circular with “filter” grading with respect to the sand bed. Various layout diameters and levels of placement were studied and a correlation was obtained between the needed size of the protection and the dimensions of the unprotected scour hole. An exploratory study of possible modifications on the design of the protective layer was undertaken. Among the conclusions from this study were as follows:

• The Terzaghi-Vicksburg riprap specifications for a horizontal layer around the pier provide an efficient protection against scour. It was observed, moreover, that a certain deviation from the Terzaghi-Vicksburg specifications can be tolerated without appreciable loss of filter properties.
  
  
• Unless the velocity of flow is too high for the size of the riprap, the stability of the layer as a whole depends on the conditions at the edge. When the edge is unexposed or leveled with the bed, no movement of the particles is observed. If the sand is eroded around the layer, then failure of the protection starts to take place.
  
  
• For a circular pier, a criterion was obtained to determine the required areal extension of the riprap protection. The riprap layer should be extended to the dimensions of the unprotected scour hole at the corresponding level of emplacement. All indications are that such a criterion can be extended to other pier geometries; however, further study is recommended before a definite conclusion can be reached.
  
  
• A test indicated the possibility of modifying the riprap layer by sloping down the layer and reducing the thickness toward the edge. Future study should indicate the most convenient solution from the points of view of safety and economy.

The Schoharie Creek bridge failed in 1987 and was attributed to inadequate pier riprap, resulting in a significant increase in the interest in riprap protection of bridge piers.⁽⁵⁾ The failure of the I-90 bridge over Schoharie Creek near Albany, New York on April 5, 1987, cost 10 lives. The foundations of the four bridge piers were large spread footings without piles. The footings were set into the stream bed in very dense ice contact stratified glacial drift, which was not considered erodible by designers at the time. However, subsequent flume studies of samples of the stratified drift showed that some material would erode at velocities that might occur at design flood flows.

Design plans for the Schoharie Creek Bridge called for the footings to be protected with riprap. Over a period from 1953 to 1987, much of the riprap was removed by high flows. The National Transportation Safety Board (NTSB) gave as the probable cause “...the failure of the New York State Thruway Authority (NYSTA) to maintain adequate riprap around the bridge piers, which led to severe erosion in the soil beneath the spread footings. Contributing to the severity of the accident was the lack of structural redundancy in the bridge.”

The NYSTA inspected the bridge annually or biennially with the last inspection before the failure on April 1, 1986. A 1979 inspection by a consultant hired by the New York State Department of Transportation indicated that most of the riprap around the piers was missing; however, the 1986 inspection failed to detect any problems with the condition of the riprap at the piers. Based on the NTSB findings, the conclusions from this failure are that inspectors and their supervisors must recognize that the presence of some riprap does not necessarily make a bridge safe from scour, and inspectors must be trained to recognize when riprap is missing and the significance of this condition.

As reported in 1993, Ruff and Nickelson conducted experiments to examine the effect of riprap size and coverage on reducing local scour at bridge piers.⁽¹⁶⁾ Pier diameter, bed material size, and riprap size were varied during the study. Scour depths with no riprap were compared with scour depths with a single layer riprap mat with 100- and 50-percent areal coverage. They found that scour depths could be reduced 40 to 99 percent by placing a riprap mat around the pier that is 5 to 8 times the diameter of the pier.

They also concluded that the extent of riprap coverage is an important parameter in predicting scour depths. Scour essentially ceased when the riprap slid or rolled into the scour hole and formed a mat of approximately 100 percent of riprap coverage in the scour hole when it started at 50-percent coverage on the bed. This study emphasized the need for periodic inspections of bridge pier riprap protection because time and floods can remove individual rocks in a riprap mat and reduce the coverage, thereby reducing the degree of protection originally intended.

Parola investigated the stability of riprap at bridge piers by considering the influences of the 3D flow on the trajectory of bedload sediment, the seepage gradient within the streambed, and boundary stresses.⁽¹⁷⁾ Maximum mean boundary stresses were inferred from velocity measurements to assess riprap stability. The stability of riprap was considered with respect to several dimensionless parameters. Relative size of the rock compared with the pier and elevation of riprap placement were shown to significantly influence the stability of riprap. Parola also made recommendations regarding the minimum extent of riprap protection required to protect the streambed around piers.

Parola also noted that the local pressure variation on the streambed can be substantial. He built on the work of others who showed that the change in pressure along the streambed can vary significantly from the front corner of a pier to the side of the pier and that such pressure differentials could set up seepage gradients that cause the removal of fine-grained material from beneath the riprap protection.^(18,19) Parola concluded that pressure fluctuations on the streambed near separation zones and under wake vortices could cause streambed material to migrate through riprap protection. To prevent this migration, a properly designed filter should be placed below the riprap protection, especially near corners of rectangular piers and in the region of wake vortices. Riprap should be extended to cover regions of high boundary stresses and regions where sediment is diverted from the streambed. Parola noted that the extent of the riprap layer is sensitive to the angle between the approach flow direction and the longitudinal axis of the pier; therefore, all likely angles of attack should be considered when designing riprap protection. Protection of the region downstream of the pier is required for circumstances in which scour holes downstream of the pier are unacceptable.

Bertoldi et al. completed a study of several scour protection countermeasures as remedial alternatives for scour at bridge piers.⁽²⁰⁾ Countermeasures were evaluated in terms of failure Modes and techniques for analyzing expected stability. Alternatives to riprap vary in size, shape, and mass as well as in their design flexibility. The authors evaluated grout mats and grout bags, extended footings, tetrapods, cable-tied blocks, high-density particles, tile mats, and anchors (used in conjunction with other countermeasures). While riprap is the most common and best documented scour protection at bridge piers, alternatives are used for many situations, such as when riprap is not available, unreasonable riprap sizes would be required for high velocity streams, or riprap placement would be difficult. The study provided insight into the overall behavior and effectiveness of various countermeasures:

• Confirmed previous recommendations to extend the countermeasure at least two pier widths laterally from the pier to provide adequate protection from local scour.
  
  
• Confirmed the need for filter fabric (or other filter material) to seal voids between the pier and countermeasure.
  
  
• Determined that two techniques were appropriate for analyzing stability: (1) particle displacement criteria patterned after Shields and Isbash incipient motion formulas and (2) drag coefficients to characterize overturning forces. Both of techniques involve dimensionless parameters that can potentially be transferred from laboratory to full-scale conditions.
  
  
• Confirmed that loose particle countermeasures such as rock riprap, tetrapods, or other precast concrete particles and high-density particles can be analyzed by particle displacement criteria and can be compared with one another by using an equivalent spherical diameter as a characteristic size.

Chiew published a paper addressing the mechanics of riprap failure at bridge piers.⁽²¹⁾ In this study, experiments conducted in a laboratory flume identified three different Modes of riprap failure at a cylindrical bridge pier. Acting alone or in combination, these Modes of failure are as follows:

• Shear failure: the riprap stones are too small or light to withstand the down flow and horseshoe vortices associated with the pier scour mechanism under the given flow condition.
  
  
• Winnowing failure: winnowing of the underlying finer bed material through the voids or interstices of the coarse riprap stones.
  
  
• Edge failure: the instability at the edge of the coarse riprap layer stones and the finer bed material initiates the formation of a local scour hole that affects the stability of the riprap.

The study proposed a semi-empirical method to size stones for riprap protection. The experimental data showed that a thick riprap layer can prevent winnowing in the absence of a filter layer. The thick riprap layer can also sustain a partial breakup of the layer with the capability of rearmoring the scour hole, preventing a total disintegration of the riprap layer. Finally, the study proposed empirical relationships that describe the effects of riprap thickness and cover on the stability of the riprap layer.

In a subsequent effort, Lim and Chiew examined failure mechanisms of riprap layers around a cylindrical bridge pier under live-bed conditions.⁽²²⁾ They concluded that the most important factors affecting the stability of the riprap layer are the turbulent flow field around the pier and fluctuations of the bed level caused by the migration of bed features past the pier. The latter phenomenon causes the riprap stones to lose their support. Observations showed that the riprap stones will eventually degrade to a level defined by the height of the largest dune and the local pier scour.

Further work by Lim and Chiew examined the failure behavior of a riprap layer around a cylindrical bridge pier.⁽²³⁾ The study confirmed the inherent flexibility of a riprap layer to offer a self-healing process. This flexibility helps to reduce further erosion under a steady flow condition allowing an equilibrium state to be attained when the erosion ceases. When the flow velocity was increased steadily, the riprap layers eventually failed in two Modes: total disintegration and embedment. The failure occurred when the erosion power is higher than the self-healing ability. The study showed that the eventual failure Mode can be determined by comparing the relative magnitude of the critical shear velocity of the bed sediment and the adjusted threshold velocity of the riprap layer.

Melville et al. performed a laboratory study of the stability of riprap protective layers at piers under mobile-bed conditions.^(24,25) The authors emphasized the role of the placement level of the riprap layer in maintaining stability. The results indicated that the deeper the initial level of the riprap layer, the lesser the depth of local scour at the pier, i.e., the better the protection afforded. They also noted decreasing riprap stability with increasing flow velocity and discussed mechanisms of riprap layer disintegration.

Melville et al. reported four failure mechanisms during the live bed study:

• Shear failure: the riprap stones are entrained by the flow because they are too small to resist movement.
  
  
• Winnowing failure: the finer underlying bed material is eroded through voids between riprap stones.
  
  
• Edge failure: scouring at the periphery of the riprap layer undermines stones.
  
  
• Bedform-induced failure: passing bed forms cause riprap settlement.

The first three Modes were previously observed under clear water conditions.⁽²¹⁾ The additional mechanism relates to the presence of bed forms in a mobile bed. When a deep trough passes the pier, settlement of the riprap layer is initiated. Further settlement can occur if there are subsequent deep troughs that pass. The fluctuating bed level caused by the movement of bed features causes the riprap stones to lose support and stability. If the trough of the bed form is deeper than the riprap layer, stones are undercut and move down into the lower trough region. Ahead of the bed forms is an area of high turbulence and correspondingly high shear stresses are induced for short periods. Riprap stones can be plucked from the layer and be transported to the lee of the pier and beyond. Removal of stones allows increased winnowing.

Hoffmans and Verheij created a scour manual to provide civil engineers with practical methods to calculate the dimensions of scour holes and to furnish an introduction to the most relevant literature.⁽²⁶⁾ The manual reflects the cumulative findings of Dutch scour research projects. Among other topics, the manual provides specific guidance on the schematically simplified shape and extent of a pier scour hole as a function of the side slopes, the radius of an upstream half circle, and the length of the axis of a downstream half ellipse. From these schemata, the volume and minimum extent of bed protection can be determined. Citing other literature, the manual recommends estimating a minimum length of bed protection around circular bridge piers based on the projected width of the pier.^(19,27) However, if the width of the pier influences the flow pattern strongly (e.g., if the pier width is half of the flow width), the cited relations are not applicable.

In another study to determine countermeasure effectiveness to protect bridge piers, Parker et al. determined that placing a geotextile under a riprap layer with the same areal coverage as the riprap layer, resulted in a relatively poor riprap performance.⁽²⁸⁾ With live-bed conditions, the rocks at the edges tended to slide or be plucked off, exposing the underlying geotextile and ultimately resulting in failure of the riprap layer as successive bed forms passed and plucked more rocks from the riprap layer. Additional test results from the study confirmed that riprap performance was best when a geotextile filter extended two-thirds the distance to the periphery of the riprap.

Fotherby and Ruff evaluated large scale riprap and concrete armor units to prevent pier scour.⁽²⁹⁾ They combined the data generated from their study with that of previous studies to demonstrate that countermeasure equations are adequate when a ratio of the characteristic axis length of the countermeasure to the pier width is less than 0.15, but are conservative for greater values of the ratio. In addition, they proposed a design procedure that enveloped the points of incipient motion of riprap and several shapes of concrete armor units. The procedure incorporates the relative pier size and provides the recommended dimensions for the countermeasure.

Chiew and Lim performed live-bed experiments to examine the failure behavior of riprap at a cylindrical bridge pier.⁽³⁰⁾ They observed that riprap fails in one of two Modes: total disintegration or embedment. The former refers to the break-up of the entire riprap layer where the stones are washed away by the flow field generated at the pier. The latter refers to the burying (embedment) of the riprap in the sediment. The authors proposed a criterion to demarcate the limiting condition between the two types of failure. In addition, they observed that embedment failure is the more common riprap failure Mode under live-bed conditions. The causes of embedment failure are twofold: (1) bed feature destabilization and (2) differential mobility. Bed level fluctuations caused by the propagating bed features resulted in bed feature destabilization, whereas differential mobility is the result of the different response of the riprap stones and bed sediments to the flow field. Experimental results also showed that the riprap layer can degrade to an equilibrium level for a given flow condition. Finally, the authors proposed a semi empirical equation to compute the maximum depth of riprap degradation, which occurs at the upper end of dune regime.

Transit New Zealand and its predecessor the National Roads Board had long recognized the national importance of bridge scour research in New Zealand. Since about 1972, these bodies have supported many bridge scour research projects, the majority of which were completed at the University of Auckland with the involvement of Dr. Bruce Melville. Recognizing the need for a summary of the state of knowledge of bridge scour, Transfund facilitated the preparation of the Bridge Scour Manual.⁽⁶⁾ This manual provides comprehensive coverage of bridge scour primarily based on procedures, guidance, and experience developed in New Zealand.

Thirty-one detailed case studies of scour-induced bridge failure (primarily in New Zealand) are presented in the Manual. These provide designers with an understanding of processes involved and also cases against which design methodologies can be tested. One chapter presents principles of scour-resistant design along with a comprehensive summary of scour protection methods and remedial methods for preventing bridge scour.

Lim and Chiew performed a parametric study of riprap protection around a cylindrical bridge pier with uniform bed sediments.⁽³¹⁾ The authors examined the role of riprap layer thickness, cover width, and placement level, as well as the median grain size and density of the riprap stone on the riprap stability. Maintaining an undisturbed approach flow depth, the tests were conducted with a sequence formation of ripples, dunes, and transition flat bed. Observations showed that a riprap layer will eventually degrade to a maximum embedment level. The embedment velocity was defined as the velocity at which the riprap layer has embedded to its maximum level. The experimental results showed that variations of the characteristic parameters have no influence on the embedment failure at the upper end of the dune regime. The authors also proposed a maximum embedment velocity that defines a critical flow velocity at which all riprap layers will fail, irrespective of the characteristic parameters.

Lauchlan and Melville evaluated failure mechanisms, stability, and placement level effects for riprap at bridge piers.⁽³²⁾ The authors assessed the ability of riprap stones to protect bridge piers under a wide range of flow conditions. The effects of placing the riprap layer at depth within the sediment bed, rather than level with the bed surface, were also investigated. The study showed that, as the flow velocity increases, the ability of riprap stones to protect a pier decreases asymptotically until the scour depth in the riprap layer reaches that of an equivalent unprotected pier. In addition, it was found that the deeper the placement level the less exposed the riprap was to destabilizing bed forms and the better the protection against local scour. Lowering the placement level also meant that the riprap performed better than for surface-placed layers as the flow velocity increased. The Mode of riprap failure also changed as the placement level below the bed surface is lowered. A pier riprap size prediction equation was proposed, including a parameter to account for placement level. The authors concluded the following:

• Destabilization of riprap layers protecting bridge piers by the progression of bed forms past the pier is the dominant failure Mode under live-bed conditions. Edge, shear, and winnowing mechanisms play secondary roles.
  
  
• Riprap stones subside into the bed sediment with the passage of bed forms past the pier and eventually settle at the level of the passing bed-form troughs. The stones subside as a layer and therefore remain capable of providing protection.
  
  
• Under live-bed conditions, over a range of riprap stone sizes, the relative local scour depth increased asymptotically with relative flow velocity up to a point. This result is in agreement with the findings of others.⁽³⁰⁾
  
  
• Riprap installed below the sediment bed surface is less vulnerable to the destabilizing influence of passing bed forms and is, therefore, more effective in protecting the pier.
  
  
• Under live-bed conditions, a larger stone size provides a greater scour protection for a given flow intensity. It is possible to predict the stone size necessary to provide a certain degree of local scour protection.
  
  
• Increasing layer thickness increases riprap stability, with an approximate 70 percent reduction in local scour when increasing the layer thickness from 1 to 3 times the median rock size.
  
  
• Synthetic filter layers placed beneath the riprap improve riprap stability. The lateral extent of the synthetic filter should be limited to about 75 percent of the lateral extent of the riprap. The reduced coverage of the synthetic filter ensures that edge stones in the riprap layer are able to protect the synthetic filter from being rolled-up by the flow.
  
  
• The pier size, relative to the riprap size, has a relatively minor effect on riprap stability under live-bed conditions. This finding is contrary to that of Parola for clear-water conditions.^(17,33)
  
  
• Neither pier shape nor pier alignment is a significant factor in terms of riprap stability.

Chiew conducted an experimental study examining local scour and riprap stability at bridge piers in rivers subject to bed degradation.⁽³⁴⁾ The data showed that the equilibrium bed profile associated with the “with” or “without” pier condition is essentially the same, except for the section around the pier. Total scour depth is shown to be the sum of bed degradation and local pier scour depth. The latter can be computed from the time-averaged live-bed scour depth associated with the undisturbed velocity ratio before bed degradation.

Chiew notes that when an alluvial channel bed degrades, a riprap layers around bridge piers is subject to destabilization. Experimental observations showed that the initial instability of the riprap stones occurred at the riprap layer boundary with the finer bed sediment. As bed degradation starts, edge stones experience undercutting, causing them to slip into the scour region. With time, dune forms move toward the riprap layer causing more stones to be displaced especially when the dune trough arrives at the pier. As general scour took place over a long duration, a mound eventually formed as more edge stones slid onto the degraded bed. An auxiliary test showed that the mound is very vulnerable to another flood flow accompanied by large dunes. This type of riprap instability is defined as bed-degradation induced failure.

The objectives of the NCHRP Report 593 study were to synthesize information and studies evaluating countermeasures to protect bridge piers from scour and to provide recommendations for guidelines and specifications for their design, construction, inspection, maintenance, and performance evaluation.⁽⁵⁾ The report included several pier scour countermeasures including riprap, partially grouted riprap, articulating concrete blocks, gabion mattresses, grout mattresses, and geotextile sand containers. The objectives were accomplished with the support of extensive testing in the hydraulics laboratory at Colorado State University.

The report notes that when properly designed and installed, riprap has an advantage over rigid countermeasures for pier scour protection because it is flexible when under attack by river currents. It can remain functional even if individual stones may be lost and it can be repaired relatively easily. Properly designed and installed riprap can provide long-term protection if it is inspected and maintained on a periodic basis as well as after flood events. For State transportation departments, riprap has been the most common countermeasure installed at bridge piers.

Gaps in the current state of the practice were summarized in NCHRP Report 593 and a specific test, or series of tests, was designed to address each deficiency. Each test series was designed to permit one configuration to be carried forward to the next series. This served to quantify the repeatability of the test program as well as to identify inconsistencies that could arise in the experimental set up. In addition, the testing program addressed stability and performance issues associated with the extent of the countermeasure placement around the pier, and the termination details at the pier and around the periphery of the installation. In addition, various filter types and extents were investigated. The design intent for the riprap coverage tests included examination of the following:

• Areal coverage and edge treatment with recommended geotextiles.
  
  
• Areal coverage variation from HEC-23 with recommended geotextiles.
  
  
• Areal coverage and thickness variation from HEC-23 with recommended geotextiles.
  
  
• Scour hole extent with recommended geotextiles.
  
  
• Scour hole extent without filters.
  
  
• HEC-18 guidelines.⁽³⁵⁾
  
  
• Thickness and filter variation from HEC-23 guidelines.
  
  
• Mounded riprap without filters.

Riprap used for pier protection is often placed on the surface of the channel bed because of the ease and lower cost of placement and because it is more easily inspected, but it may also be placed in a pre-existing scour hole, or in a hole excavated around the pier. Test results indicated that when the stable baseline riprap configuration was mounded on the surface without a filter, the performance was poor. Given the same areal extent of riprap coverage, none of the riprap in the mounded riprap test configurations performed as well as riprap in test configurations with the top of the riprap level flush with the bed. This study concluded that mounding riprap around a pier is not acceptable for design in most cases, because it obstructs flow, captures debris, and increases scour at the periphery of the installation.

Numerous riprap studies suggest that thickness of the riprap layer placed around the bridge piers should be between 2 to 3 times median stone size of the riprap.⁽²⁾ Testing for NCHRP Report 593 indicated that 3 times the median stone size is appropriate for specifying minimum thickness and that performance improved with increasing riprap layer thickness.

Results from riprap testing indicated that riprap areal coverage should be a minimum of two pier widths in all directions. Riprap should be placed in a pre-excavated hole around the pier so that the top of the riprap layer is level with the surrounding channel bed elevation. Placing the top of the riprap flush with the bed is ideal for inspection purposes, and does not obstruct the flow. The riprap layer should have a minimum thickness of 3 times the median rock size.

Additional tests confirmed that the lateral extent of riprap protection at rectangular piers must be increased when the longitudinal axis of the pier is skewed to the flow direction. Tests also confirmed that a filter should not be extended fully beneath the riprap; instead, the filter should be terminated two-thirds the distance from the pier to the edge of the riprap. Other recommendations regarding the riprap extent and layer thickness were included in the report.

Periodic and post-flood inspection practices were also evaluated. Pier riprap is typically inspected during biennial bridge inspection programs. However, more frequent inspection might be required for a particular bridge or group of bridges that are designated as vulnerable or after floods exceeding a critical threshold. The following guidance is presented in a National Highway Institute (NHI) course for bridge inspectors:⁽³⁶⁾

• Riprap should be angular and interlocking.
  
  
• Riprap should have a granular or synthetic geotextile filter between the riprap and the subgrade material.
  
  
• Riprap should be well graded (a wide range of rock sizes). The maximum rock size should be no greater than about twice the median size.
  
  
• For bridge piers, riprap should generally extend up to the bed elevation so that the top of the riprap is visible to the inspector during and after floods.
  
  
• Riprap inspections should determine the following, which are strong indicators of problems:
  
  
  • Has riprap been displaced downstream?
    
    
  • Has angular riprap blanket slumped down slope?
    
    
  • Has angular riprap material been replaced over time by smoother river run material?
    
    
  • Has riprap material physically deteriorated, disintegrated, or been abraded over time?
    
    
  • Are there holes in the riprap blanket where the filter has been exposed or breached?

RIPRAP AT BRIDGE ABUTMENTS OR IN STREAM CHANNELS

Riprap is frequently used as a countermeasure to protect bridge abutments from scour, as well as to stabilize stream channels. In a United States Geological Survey (USGS) study for FHWA, Brice and Blodgett developed guidelines supporting the design, maintenance, and construction of countermeasures for reducing bridge losses attributable to scour and bank erosion.⁽³⁷⁾ These guidelines are based on case histories of 224 bridge sites in the U.S. and Canada, interviews with bridge engineers in 34 states, and a survey of published work on countermeasures. Each case history (in volume 2) includes data on bridge, geomorphic, and flow characteristics; a chronological account of relevant events at the site; and an evaluation of hydraulic problems and countermeasures. Problems at piers and abutments were identified at 100 and 80 sites, respectively. While the study does not focus on riprap for pier scour protection, several relevant conclusions were included:

• Hydraulic problems are attributed to local scour at about 50 sites, to general scour at about 55 sites, and to lateral erosion by stream action at about 105 sites.
  
  
• Countermeasure performance is rated primarily on how well it served its intended function, but also on whether it sustained damage or produced any unwanted effects. According to the "principle of expendability," a countermeasure that serves its purpose most economically is likely to be damaged.
  
  
• Flexible revetment has a better record of performance at study sites than rigid revetment such as concrete pavement. Failures of riprap, a flexible revetment, are attributed to factors already taken into account by existing design guidelines.
  
  
• Riprap keyed in place by tamping is more stable than loose riprap, and a lesser volume of rock per unit area is required.
  
  
• As protection for piers, dumped rock riprap proved to be effective at most sites where it could be evaluated. There is some evidence that the riprap should be placed so that the top is below, rather than level with, the streambed.

Blodgett and McConaughy developed guidance for riprap design for stream channels near highway structures.⁽³⁸⁾ Design procedures used for design of rock riprap revetment installations were evaluated using data from 26 field sites. Four riprap failure Modes were identified: (1) particle erosion, (2) translational slide, (3) modified slump, and (4) slump. Particle erosion occurs when individual particles are dislodged by the hydraulic forces generated by flowing water. Particle erosion can be initiated by abrasion, impingement of flowing water, eddy action/reverse flow, local flow acceleration, freeze/thaw action, ice, or toe erosion. Factors associated with riprap failure included stone size, bank side slope, size gradation, layer thickness, insufficient toe or end wall, failure of the bank material, overtopping during floods, and geomorphic changes in the channel. One conclusion from a review of field data and the design procedures suggested that estimates of hydraulic forces acting on the boundary based on flow velocity rather than shear stress were more reliable. Several adjustments for local conditions, such as channel curvature, superelevation, or boundary roughness, may be unwarranted in view of the difficulty in estimating critical hydraulic forces for which the riprap is to be designed. Success of riprap is related not only to the appropriate procedure for selecting stone size, but also to reliability of estimated hydraulic and channel factors applicable to the site.

While not specific to pier riprap, the authors discuss general characteristics of riprap failure and note that inadequate recognition of the type of erosion process that is occurring or improper riprap design may lead to failure of the riprap. Types of erosion that can be successfully controlled by riprap include channel degradation, bank erosion, scour, and changes in alignment associated with meandering, branching, and braiding of streams. The rate of channel erosion varies with time, but is primarily a function of the magnitude of streamflow. The major characteristics of the riprap layer include: (1) thickness, (2) placement method, (3) toe construction, (4) stone gradation, and (5) filter blankets.

A comprehensive design manual on the use of rock in hydraulic engineering was prepared by the Netherlands Center for Civil Engineering Research and Codes (CUR) and the Rijkswaterstaat, Dutch Public Works Department (RWS) between April 1988 and December 1993 under a collaborative project.⁽⁷⁾ In many countries, rock is a commonly used construction material in hydraulic engineering. This Manual provides a standard reference on this subject, addressing the entire lifecycle of rock structures. The Manual sets out to answer the question "Why use rock in hydraulic engineering?" with the following observation:

Rock is basically used as a material to protect all kinds of hydraulic structures against erosion and as such it has a number of major qualifications. Protective and other layers when made of rock are flexible and follow slight settlements. Besides, local damage or loss is easily repaired, construction is generally not complicated and rock is usually durable and recoverable. Last but not least, depending on transport distance and means of transport, for many locations rock is the most cost-effective material for the protection of erosion-prone slopes of structures or beds in a marine or inland environment.

Section 2.2.2 of the Manual describes various failure mechanisms (or failure Modes) of rock structures, with emphasis on the functions of rock in these structures. Failure can be simply defined as the exceedance of a predefined limit state, which occurs when the loading exceeds the strength. When this exceedance occurs, a failure response of the structure (or parts of it) can be defined. Failures can occur both during construction and operation. Typical loadings and responses for rock are wave height and displacement relative to the as-built position. Both loading and response are functions of time. The response is also determined by the characteristics of the rock system such as weight and shape. The loading may, to a certain extent, be affected by the system, for example through permeability.

Failure corresponds to unacceptable displacements and/or deformations. The associated loading is defined as a failure loading. In general, failure mechanisms (or failure Modes) are named after their consequent displacements or movements and the common characteristic is a relatively large increase of response (e.g., stone movements) resulting from a minor increase in loading (e.g., wave height).

The Manual emphasizes the importance of inspection and maintenance for any riprap installation and notes that a proper balance of costs for inspection and maintenance and the capital costs of the scheme is desirable. To achieve this scheme, an inspection and maintenance plan is necessary. This maintenance plan should be available at the design stage so that the design can be adjusted to suit the inspection and maintenance procedure or vice-versa. For instance, if after construction inspection and maintenance is not possible then the design must have a low probability of failure.

In order to develop an inspection and maintenance plan, it is necessary to consider the following ways riprap armor may fail:

• Sudden collapse of embankment of which the rock structure forms a part:
  
  
  • Sliding of embankment along failure planes in the subsoil, for instance caused by rapid decrease of water level in rivers or reservoirs or caused by scour at the toe of a slope.
    
    
  • Liquefaction caused by earthquakes.
    
    
  • Failure of transitions.
    
    
• Sudden collapse of bed protection layer:
  
  
  • Damage by ships (negative keel clearance, dragging anchors, propeller jets).
    
    
  • Attack by high waves or current.
    
    
  • Scour at the edge, followed by liquefaction.
    
    
  • Vandalism (removal of elements).
    
    
• Sudden local failure of bed protection:
  
  
  • Transition failure.
    
    
  • Vandalism (removal of individual elements of protection layer).
    
    
• Rapid degradation of bed protection from a grossly undersized protection layer.
  
  
• Gradual degradation of embankment and bed protection:
  
  
  • Deterioration of individual elements of protection layer from climate and solar effects.
    
    
  • Deterioration of exposed geotextile caused by ultraviolet radiation.
    
    
  • Clogging up of underlying filter.

The Manual treats the following as special structures regarding bed/scour/bank protection works:

• Pipeline and cable crossings.
  
  
• Fish sluices.
  
  
• Anchoring structures.
  
  
• Bridge piers.
  
  
• Jetties.

As such, the design of these and similar structures is outside the scope of the Manual. However, the Manual does provide example calculations for bridge pier scour (Section 8.4). For designing protection against pier scour, the Manual notes that it is more important to know the areal extent vulnerable to scour than the depth of scour.

In October 1998, the FHWA, the American Association of State Highway and Transportation Officials (AASHTO), and the Transportation Research Board (TRB) sponsored a scanning review of European practice for bridge scour and stream instability countermeasures.⁽³⁹⁾ This review involved a Panel of representatives from FHWA, state departments of transportation (California, Illinois, Maryland, Minnesota, Oregon, and South Carolina), universities, and the private sector. The review included visits to highway research institutes, hydraulic research laboratories, and field sites in four countries: Switzerland, Germany, the Netherlands, and the United Kingdom. Scanning review objectives were as follows:

• Review and document innovative techniques used to mitigate the effects of scour and stream instability at bridges.
  
  
• Evaluate these techniques for potential application in the United States.
  
  
• Share information on U.S. practice.

The scan found that the use of riprap, in combination with a geotextile or granular filter, is by far the most common scour and stream instability countermeasure in all four countries visited. The availability, economy, ease of installation, and flexibility of riprap were considered highly desirable characteristics. As a result, considerable effort has been devoted to techniques for determining size, gradation, layer thickness and horizontal extent, filters, and placement techniques and equipment for revetment and coastal applications. In Europe, riprap is considered an effective and permanent countermeasure against channel instability and scour, including local scour at bridge piers.

The scan also noted that great care is taken in placing riprap at critical locations, and in many cases stones are placed individually in the riprap matrix. Highly specialized equipment has been developed by construction contractors in Europe for placing riprap, particularly for coastal installations. The use of bottom dump or side dump pontoons (barges) is common in both Germany and the Netherlands. Some of the smaller pontoon systems, particularly the bottom dump pontoons developed in Germany could be used to place riprap in water at larger bridges.

European hydraulic engineers consider placing an effective filter beneath riprap in flowing or deep water critical to the success of the installation. The use of large geotextile sand containers at the Eidersperrwerk in Germany, the use of a geotextile mattress filled with granular filter material at the Eastern Scheldt Barrier in the Netherlands, and the use of fascine sinker mats at both locations are examples of these innovative placement techniques. The availability of testing apparatus to ensure that geotextiles perform as required and the development of specific codes to guide the design and installation of geotextiles contribute to the success of these installations. Engineers in all four countries referred to a Dutch design manual for rock applications in hydraulic engineering as a primary reference guide.⁽⁷⁾

NCHRP Report 568 represented a major synthesis effort to develop riprap design criteria, recommended specifications, and quality control guidelines for riprap for a range of applications.⁽²⁾ The applications included revetment on streams and riverbanks, bridge piers and abutments, and bridge scour countermeasures such as guide banks and spurs. This synthesis study did not involve any original laboratory or field work. A fundamental premise of the study was that riprap is an integrated system and that successful performance of a riprap installation depends on the response of each component of the system to hydraulic and environmental stresses throughout its service life.

For pier scour riprap protection, the report includes laboratory results and design recommendations from a concurrent study for countermeasures to protect bridge piers from scour.⁽⁵⁾ The recommendations included filter requirements, riprap extent, and other construction and installation guidelines for riprap at piers. Guidelines for the use of geotextile containers as a means of placing a filter for pier riprap derived from European practice are presented. Construction and installation guidelines and constructability issues are investigated, including: dumping vs. controlled placement, underwater versus dry installation, and buried versus mounded placement.

HEC-23 identifies and provides design guidelines for bridge scour and stream instability countermeasures that have been implemented by various State departments of transportation (DOTs) in the United States.⁽¹⁾ Countermeasure experience, selection, and design guidance are consolidated from other FHWA publications and NCHRP research studies in this document to support a comprehensive analysis of scour and stream instability problems and provide a range of solutions to those problems. Selected innovative countermeasure concepts and guidance derived from practice outside the United States are introduced.

Volume 1 contains a chapter on riprap design, specifications, and quality control as well as an expanded chapter on biotechnical countermeasures. Volume 2 contains nineteen detailed countermeasure design guidelines grouped into six categories: (1) stream instability, (2) streambank and roadway embankment protection, (3) bridge pier protection, (4) abutment protection, (5) filter design, and (6) special applications.

Volume II DG 11 provides specific guidance for rock riprap at bridge piers. DG 11 notes that design of a pier scour countermeasure system using riprap requires knowledge of the river bed and foundation material; flow conditions including velocity, depth and orientation; pier size, shape, and skew with respect to flow direction; riprap characteristics of size, density, durability, and availability; and the type of interface material between the riprap and underlying foundation. The system typically includes a filter layer, either a geotextile fabric or a filter of sand and/or gravel, specifically selected for compatibility with the subsoil. The filter allows infiltration and exfiltration to occur while providing particle retention.

Bridge pier riprap design is based, primarily, on research conducted under laboratory conditions with limited field verification. Flow turbulence and velocities around a pier are of sufficient magnitude that large rocks move over time. Bridges have been lost as a result of the removal of riprap at piers resulting from turbulence and high velocity flow. Usually the loss of riprap does not happen during one storm, but is the result of the cumulative effect of a sequence of high flows (e.g., Schoharie Creek). Therefore, if rock riprap is placed as scour protection around a pier, the bridge should be monitored and inspected during and after each high flow event to ensure that the riprap is stable.