Evaluation of Bridge Scour Research: Pier Scour Processes and Predictions (2011)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

72 CHAPTER 5 PIER SITE COMPLICATIONS 5.1 Introduction Several factors complicate pier sites at bridge waterways, and thereby introduce additional processes and parameters to be considered when estimating pier scour depth. The additional processes of practical importance can be grouped in two categories: 1. Factors affecting pier flow field: i. Bridge structure (pier form, abutment proximity, deck submergence); ii. Debris or ice accumulation at a pier or across the bridge waterway; iii. Channel geometry; and, iv. Tide-affected flows. These factors may substantially change the overall flow field at a pier, relative to the flow field at an isolated pier in a steady approach flow. Some factors (e.g. pier-shape complexity, or abutment proximity) change the flow field so much that the scour-depth estimation methods discussed subsequently in Chapter 6 do not readily apply. Other factors (notably, debris at an individual pier), considered with certain limitations, may be treated through the proposed use of an effective or equivalent pier width. 2. Factors complicating assessment of the erosion resistance of foundation material: i. Multiple strata of alluvium; ii. Cohesive soil (clay); and, iii. Weak rock. The considerations associated with these factors cause foundation material to erode in a less predictable manner (at least given the present state of knowledge) compared to erosion of a continuous homogeneous boundary of non-cohesive foundation material. A third factor of less practical importance, though needing to be addressed here, concerns the possible influence on scour depth of washload sediment. This factor may affect the approach flow and the erosion resistance of the approach bed. When considering the forgoing site complications, Eq. (4.2) must be broadened to include combinations of parameters from the framework indicated in Eq. (5.1);

73 ݕ௦ ൌ ݂ݑ݊ܿݐ݅݋݊ ۏ ێێ ێێ ێێ ێ ۍ݂݈݋ݓ ሺߩ, ߤ, ܸ, ݕ, ݃ሻ, ܾ݁݀ ݉ܽݐ݁ݎ݈݅ܽ ൫ܦ, ߪ௚, ߩ௦, ܿ, ௖ܸ , ൯, ݌݅݁ݎ ሺܽ, Ω, ߠሻ,ݐ݅݉݁ ሺݐሻ, ܾݎ݅݀݃݁ ݏݐݎݑܿݐݑݎ݁, ܾ݀݁ݎ݅ݏ ݋ݎ ݅ܿ݁ ܽܿܿݑ݉ݑ݈ܽݐ݅݋݊, ݄݈ܿܽ݊݊݁ ݃݁݋݉݁ݐݎݕ, ݂݋ݑ݊݀ܽݐ݅݋݊ ݉ܽݐ݁ݎ݈݅ܽ ܿ݋݉݌݈݁ݔ݅ݐ݅݁ݏ, ݓܽݏ݄݈݋ܽ݀ ݅݊ ܽ݌݌ݎ݋݄ܽܿ ݂݈݋ݓ ے ۑۑ ۑۑ ۑۑ ۑ ې (5.1) Design estimation to account for these possible complications can be difficult when the pier is of unusually large size, especially relative to flow depth, or when the pier is of an unusual form. The design estimation of scour depth at wide cylindrical piers (as defined in Table 4-1) is then compounded by the additional number of variables to be considered. This chapter reviews knowledge regarding pier site complications, but does not attempt to identify all the main variables and parameters associated with each complication. Instead, references for further information are provided. Also, this chapter does not elaborate scour in tide-affected flows, except to note that the pier flow field is subject to tidal ebb and flow, which may cyclically reverse the approach flow to a pier. 5.2 Pier Structure Chapter 4 discusses scour at simple cylindrical piers extending to depth. However, pier structures typically comprise multiple components as illustrated in Figure 5-1. The additional components add uncertainty to design estimation of scour depth. Design estimation of scour depth must account for pier structure.

74 Figure 5-1 Common pier structures comprise a column on a slab footing, or column on pile cap with underpinning piles (Melville and Coleman 2000) For piers tapered on the upstream and downstream faces, the slope, in elevation, of the leading edge of the pier affects the local scour depth. Downwards-tapering piers induce deeper scour than a circular pier of the same top width, and vice-versa. Shape factors for local scour at tapered piers have been proposed by Neill (1973), Chiew (1984), and Breusers and Raudkivi (1991). For piers founded on a (wider than the pier) slab footing, caisson or pile cap, placing the footing, cap or caisson with its top below the boundary level can be effective in reducing the local scour depth through interception of the down-flow. However, if the top of the wider foundation element comes to the bed level (in the undisturbed flow region away from the pier), or even above, the scour depth is increased. Unless definite predictions are possible, it is risky to rely on limitation of the scour depth through flow interception by the wider base element. The results of laboratory investigations of local scour at piers founded on slab footings and caissons have been quite extensively reported in studies before 1990 (notably, Chabert and Engeldinger 1956, Tsujimoto et al. 1987, Imamoto and Ohtoshi 1987, Jones 1989), and in publications since 1990 (Jones et al. 1992, Fotherby and Jones 1993, Parola et al. 1996b, and Melville and Raudkivi 1996). Local scour at piers founded on piles

75 similarly has been extensively studied (early studies by Hannah 1978, Raudkivi and Sutherland 1981, Jones 1989), with more recent publications by Richardson and Davis 1995, Sheppard et al. 1995, and Salim and Jones 1996). A consideration for all of these studies is that pier structure can vary quite extensively. Melville and Raudkivi (1996, also in Melville and Coleman 2000) identify five cases of pier scour for piers comprising a column supported by a slab footing, pile-cap on piles, or caisson (Figure 5-2): Case I, where the top of the footing, cap or caisson remains buried below the base of the scour hole; Case II, where the top of the footing, pile-cap, or caisson is exposed within the scour hole below the general bed level; Case III, where the top of the footing, pile-cap, or caisson is above the general bed level; Case IV, where the top of the footing, pile-cap, or caisson is at or above the water surface level; and, Case V, where the pile-cap is clear of water surface. The submerged portion of the pier comprises a set of piles; e.g., like a pile bent. For Case I, the local scour is unaffected by the presence of the footing, pile-cap, or caisson, while for Case II, the local scour typically is reduced from that at a uniform pier, due to interception of the down-flow. For Case III, the local scour depth can be increased or decreased compared to that at a uniform pier. The Case III scour depth at a pier founded on a larger size pile-cap or caisson is increased, with the maximum local scour occurring for Case IV, where the top of the pile-cap or caisson is at or above the water surface level. Cases III and IV scour depths at piled foundations may be increased or decreased. Figure 5-2 schematically illustrates the trends, in which the level of the top of the footing, pile-cap, or caisson, Y, is measured from the general bed level and is positive downwards. Melville and Raudkivi (1996) made a detailed laboratory study of local scour at a non- uniform cylindrical pier, comprising a circular pier (diameter a) founded on a larger diameter circular caisson (diameter a*). The diameter ratio (a/a*) was varied systematically between 0.12 and 1.0, while the level of the top of the caisson (Y) spanned Cases I, II, III and IV. Their data are plotted in Figure 5-3.

76 Figure 5-2 Scour depth variation for four cases of non-uniform pier shape (Melville and Coleman, 2000). Not shown is the case when the pile cap is fully above the water surface

77 Figure 5-3 Influence of non-uniform shape on local scour depth at piers (Melville and Raudkivi 1996) For scour at piled foundations where the pile cap is clear of the water surface (Case V), the comprehensive study by Hannah (1978) remains useful. He found that the maximum scour depth is closely related to the dimension of the pile group as a whole, as seen from upstream. Accordingly, he recommended that a single line of piles should be used in preference to piers for angles of attack greater than 8o. Similar results were obtained

78 subsequently by Nazariha and Townsend (1997). Based on experiments by Jones (1989), Richardson and Davis (1995) recommend for Case III that the larger of two scour estimates be used: one based on the width of the pier column, and the other on the width of the pile cap, the latter using the flow depth and flow velocity in an assumed flow zone obstructed by the pile cap. This recommendation facilitates a design estimate, but does not reflect the flow field causing the scour. Jones and Sheppard (2000) proposed a fairly involved approach to estimating scour depth at a pile-supported pier. The method is presented also by Richardson and Davis (2001) and FDOT (2010). Essentially, the method entails dismantling the pier’s elements then summing the scour depths attributable to the pier column, pile cap, and pile group, as indicated in Figure 5-2 and Eq. (5-2). Figure 5-4. The disassembly approach proposed by Jones and Sheppard (2000) for estimating scour depth at a pile-supported pier ys = ys-pile stem + ys-pile cap + ys-pile group (5-2) in which total depth of pier scour is the sum of scour-depth component for the pier column in the flow, ys-pile stem; the scour component for the pile cap or footing in the flow, ys pile cap, and the scour component for the piles exposed to the flow, ys-pile group . The variables in Figure 5-4 are defined by Richardson and Davis (2001). The procedure entails calculating a scour depth for each element treated as an equivalent cylinder subject to flow depths and velocities, and with height adjustment for the column and pile group. Richardson and Davis (2001) outline the calculation steps. Though each of these structural elements affects scour depth, the summation of scour attributable to individual parts lacks a physical basis, because it does not relate to the flow field producing scour. The foregoing review indicates how increased complication of pier shape introduces uncertainty in design estimation of scour depth. Scour at some piers of typical form can be estimated using pier-shape factors developed for them, but such factors do not exist for many typical pier forms, and can be unreliable when pier alignment to flow varies. 5.3 Abutment Proximity Many piers are sited near the toe of an abutment, as illustrated in Figure 5-5. Consequently, the pier is within the flow field generated by the abutment, and pier scour

79 occurs within the region of abutment scour. Because an abutment may develop a much larger scour than usually occurs at a pier, scour depth at the pier is dominated by abutment scour. Design estimation of scour depth should check whether abutment proximity will influence scour depth. Figure 5-5 Abutment proximity close to a pier may substantially alter the flow field and scour at a pier Despite the potential severity of scour at piers near abutments, this aspect of pier scour has received little attention. To quite varying extents, three studies investigated how pier and abutment proximity affect scour at an abutment and pier in simple rectangular channels. Hong (2005) examined the influence of pier presence on bridge contraction scour and pier scour, and concluded that pier presence affects the location of deepest contraction scour in the vicinity of an abutment, but gave no relationship for estimating scour depth at a pier near an abutment. An earlier study by Croad (1989) investigated scour at the specific situation of a pier close to the spill-slope of a spill-through abutment. He found that the depth of scour at the pier essentially was determined by abutment scour, and proposed that pier scour be estimated as 0.9 times abutment scour depth. Ettema et al. (NCHRP 24-20, 2009) conducted a set of extensive laboratory experiments on scour depth at piers near abutments. Figure 5-6 shows the abutment proximity layout

80 examined. A significant parameter is Lp/yf, where Lp = pier distance from the abutment toe, and yf = flow depth at the abutment toe. For abutments on floodplains, yf = flow depth on the floodplain. Figure 5-7 presents the findings as ySpier/yS0pier versus Lp/yf, where ySpier is the scour depth at the pier with the abutment present, ySabutment is the scour at the abutment and yS0pier is the scour depth at the pier without the abutment. Figure 5-6 A pier close to the abutment is within the flow field generated by the abutment The main findings ensue: 1. When a pier is located within the immediate vicinity of an abutment (notably, near the toe of an abutment), abutment scour dominates scour. For this situation, the existing equations for estimating pier scour-depth do not apply (the equations discussed in Chapter 6 and Appendix A). Scour depth at the pier is essentially equivalent to scour depth generated by flow around the abutment. 2. Figure 5-7 shows that initially ySpier/yS0pier does not vary with Lp/yf, because the pier was enveloped by the abutment flow field. Pier presence did not influence abutment scour depth substantially; ySpier coincided with the abutment scour depth, ySabutment. In other words, abutment proximity fully dominated pier scour development and its depth. As Lp/yf further increased, the abutment’s influence decreased and so did pier scour depth. Eventually, when Lp/yf exceeded about 11 to 12, pier scour depth became equivalent to the local, pier-scour depth, as if no abutment were present; i.e., ySpier/yS0pier = 1. 3. A curve developed from the study’s data could be used in estimating scour depth at a pier close to an abutment (Figure 5-7). When the pier is at the abutment toe, or near it, one curve limit indicates scour depth equivalent to abutment scour depth. When the pier is distant from an abutment, the other curve limit indicates scour depth is equivalent to scour at the pier in isolation. 4. The study’s data show that pier presence does not substantially increase scour depth at an abutment. An increase in abutment scour depth of about 10% occurred when a pier was close to a wing-wall abutment. However, when a pier

81 was at the toe of an erodible spill-through abutment on a floodplain, pier presence reduced scour depth by about 10 to 20%. 5. Pier presence can influence abutment scour location, though not depth. When a pier is a short distance out from an abutment, a wider scour hole may develop than would have occurred at the abutment alone. This effect occurs because the additional scouring influence of the flow field at the pier acts to widen the abutment scour region. Pier presence, if close to the abutment, moves the location of maximum scour depth closer to the abutment’s centerline axis. 6. Bathymetry measurements as scour progresses reveal that the location of maximum scour depth moves from an initial location at the upstream corner of an abutment to a location near the abutment’s downstream corner. Pier presence, if close to the abutment, moved the location of maximum scour depth closer to the abutment’s centerline axis. 7. For abutments on erodible floodplains pier presence gives a different trend than when the floodplain was fixed. When the pier was at the abutment toe, it was protected by embankment spill-slope soil and riprap, which failed and collected around the pier, and thereby preventing scour at the pier. However, when the pier was sufficiently distant from the abutment toe, so that failed spill-slope soil and riprap did not reach the pier, scour occurred at the pier. The maximum scour depth exceeded scour at an isolated pier (yS0pier ), but was considerably less than for the same pier position but with a fixed floodplain. Because the embankment eroded, and thus the depth and width of abutment scour was reduced, the reach of abutment influence on pier scour was less than when the floodplain was fixed.

82 Figure 5-7 Normalized scour depth at pier relative to scour depth at a spill-through abutment, and wing-wall abutment, on an erosion resistant or fixed flood plain (F) or an erodible flood plain (E). The smallest value of Lp /yf coincides with the toe of the spill- through abutment, at the edge of the fixed floodplain. The error bars indicate relative dune height. 5.4 Bridge-Deck Submergence During major floods through some bridge waterways, the bridge deck may become submerged, as illustrated in Figure 5-8. This situation introduces an additional scour process that can erode the boundary at a pier site such that the net depth of scour can greatly exceed the depth associated to pier scour alone. Design estimation of scour depth should check whether deck submergence is likely to influence scour depth.

83 Figure 5-8 Bridge beams submerged by flood water Scour associated with a submerged bridge superstructure is attributable to flow contraction under the bridge. The misnomer “pressure scour” is often used for this scour process. The water surface elevation upstream of a bridge is above the low chord of the bridge superstructure, causing a vertical contraction of the flow, thereby possibly eroding the boundary across the bridge waterway. Several studies have examined scour at submerged bridges, notably Richardson and Davis (1995), Arneson (1977), Arneson and Abt (1998), Umbrell et al. (1998), Lyn (2008), and Guo et al. (2009). The last reference provides the most comprehensive assessment and formulation of scour. The study by Guo et al. (2009) articulates the current state of knowledge regarding this form of scour, which can be summarized as follows: 1. No study has yet closely examined pier scour for conditions when the bridge deck is submerged. Essentially, scour at a pier develops as if the pier were in a narrow region of contraction scour; 2. Three general cases of submerged-deck scour occur: Upstream beam submerged, no water over deck; all beams submerged, no water over deck; and, deck fully submerged. Figure 5-9 illustrates one of these cases; 3. The equilibrium scour profiles are approximately consistent across the main channel of the bridge waterway; 4. The maximum scour depth is located about 15% of the deck width downstream of the downstream edge of the deck; 5. Scour extends about one deck-width upstream of the deck, and a deposition mound forms about 2.5 deck widths downstream of the bridge;

84 6. As expected, the scour depth increases with greater difference in water levels either side of the bridge deck, and decreases with reduced erodibility of the boundary material; 7. Recent research (e.g., that by Guo et al. 2009) is leading to improved formulations of scour depth due to deck submergence, and likely will replace the current method in HEC-. Figure 5-9 Scour at piers beneath a partially submerged bridge deck (adapted from Guo et al. 2009) 5.5. Woody Debris, or Ice, Accumulation During flood and high-flow conditions, many rivers carry appreciable quantities of floating debris, especially tree branches and trunks. Rivers subject to frigid winter conditions also may convey large amounts of ice. Consequently, bridge piers in such rivers are prone to accumulate debris and ice, as illustrated in Figures 5-10a, b. The presence of large accumulations of debris or ice has been a significant factor in the failure of several bridges. An accumulation may amplify scour depth, and it may increase the streamwise load water exerts against a pier. Design estimation of scour depth at a pier site should check whether debris or ice considerations need to be taken into account. Given the numerous variables likely introduced with debris or ice accumulation, and uncertainties associated with accumulation extents, it can be difficult to develop a single method with which to account for the effects of accumulations on scour depth at a pier. An immediate question, for instance, is whether the accumulation will be limited to individual piers, or will it extend across several piers, possibly the entire bridge waterway. It is quite usual for debris to accumulate at individual piers (in the case of longer bridges, Figure 1.1) or across several piers (shorter bridges, Figure 1.2). For ice, however, it is more usual for ice to accumulate across the entire waterway (as in Figure 5-10b). In accordance with these accumulation extents, three design scenarios arise:

85 1. Accumulation occurs at a single pier, modifying the pier’s effective width and form; and, 2. Accumulation occurs across several piers; 3. Accumulation extends across the bridge waterway, modifying the waterway form, backing-up flow, and thus possibly exacerbating contraction scour, as well as pier scour and abutment scour. This scenario, more typical for ice accumulation, is akin to scour at a submerged bridge (Section 5.3), because the channel flow is deflected under the debris in a manner similar to flow deflection under a bridge deck. None of these design scenarios readily lend themselves to design estimation of scour depth. The first scenario has received some study. The second and third scenarios entail numerous uncertainties, and have not yet been treated.

86 (a) (b) Figure 5-10 Accumulation of woody debris (a), and ice rubble (b) at bridge waterways, affects flow locally at a pier and through the entire bridge waterway When debris accumulates at a bridge pier, it may do so in masses normally referred to as debris rafts. The accumulated debris causes a larger obstruction to the flow than the pier without debris (Figure 5-11). Depending on the accumulation thickness and elevation at a pier, the additional flow obstruction may cause pier scour depths in excess of depths under conditions without debris accumulation.

87 Figure 5-11 Woody debris accumulation at a single pier (Lagasse et al. 2010) The likelihood for debris accumulation at bridge foundations depends on a number of factors, including the availability of debris material, the potential for such material to be washed into streams and rivers, and the shape of the bridge foundations. In a study of woody debris transport in a Tennessee River, Diehl and Bryan (1993) found that the predominant large debris type comprised tree trunks with attached root masses. Such trees usually fall into a river because of bank erosion. Hence bank instability is an important catchment characteristic in identifying basins with a high potential for abundant production of debris. McClellan (1994) found, using small-scale laboratory models, that debris accumulations could be formed such that they extend from the water surface to the streambed in all flow conditions. Under low Froude number conditions, the debris rafts tended to be shallow and extensive in plan area, while under high Froude number conditions, the debris rafts tended to be deep and narrow. Substantial further advances were made by Diehl (1997), who developed guidelines for assessing debris yield from watersheds, and the extent of debris accumulation at piers. This effort was extended yet further by NCHRP Project 24-26 (Lagasse et al. 2010), with the intent of providing bridge designers improved guidelines for assessing debris accumulation extents. The two leading studies on debris effects on pier scour are those by Lagasse et al. (2010) and Melville and Dongol (1992). The studies essentially agree in their findings as to how debris can affect pier scour. 1. Debris rafts can lead to deeper scour holes, with the extent of deepening varying with the shape, thickness, and elevation of the raft. 2. The greatest increases in scour depth were found when the debris formed a thick rectangular mass extending about one flow depth upstream of the pier (Lagasse et al. 2010). Triangular accumulations, thickest at the pier centreline, deflected flow

88 laterally as well as downwards, producing a shallower but wider scour hole than if no debris were present. 3. The main process whereby debris deepens scour is the deflection of flow downwards to the pier base. This process is akin to flow deflection at a submerged, or partially submerged, bridge deck (Section 5.3). The approach used by Lagasse et al. (2010) and Melville and Dongol (1992) to quantify the influence of debris on scour depth has been to convert the accumulation into an equivalent pier width. The conversion uses the ratio of scour at a pier with debris to that a pier without debris. This approximation is useful when the debris accumulation is at a single pier. When debris extends across several piers, the conversion becomes more approximate. If the extent of debris accumulation, even at a series of single piers, causes a substantial backwater effect, the conversion may become overly approximate. Lagasse et al. (2010) give an equation for estimating an equivalent pier width for use with the Richardson and Davis (2001) equation for estimating pier scour depth; ( )( ) ( ) ( ) ( ) 1.0 /y for L 1.0 /y for L / d 11 d 11 2 ≤ −+ = > −+ = y aHKyWHK a y aHKyyLWHK a ddddd e dd K dddd e d (5-3) where a = pier width (without debris) normal to the approach flow, Hd = thickness of debris, Wd = width of debris normal to the flow, y = depth of approach flow, Ld = length of debris upstream from the pier face, and Kd1 and Kd2 = coefficients depending on the shape of the debris accumulation. Values are given for triangular and rectangular debris masses. Figure 5-12 illustrates a rectangular debris accumulation, and indicates the variables involved in Eq. (5-3). In developing the equation, nine geometrically unique rectangular (in planform and profile) debris shapes and seven geometrically unique conical debris shapes (triangular in planform) were tested.

89 Figure 5-12 Rectangular accumulation of debris at a pier (Lagasse et al. 2010) The equation was verified, using laboratory data, for use with the equation proposed by Richardson and Davis (2001). It has not been verified for use with any other scour depth equation. Such verification is the topic of further research (Chapter 9). Melville and Dongol (1992) investigated local scour depths at circular bridge piers with debris rafts. The debris was modelled as an impervious circular cylinder, concentric to the pier and having its upper surface at the water surface level. They proposed the following expression for the equivalent size, ae , of the uniform circular pier that induces about the same scour depth as the actual pier with accumulated debris: y aHyWTa ddde )52.0(52.0 −+ = (5-4) where Hd and Wd = thickness (vertical dimension) and width of the floating debris raft; and a = pier width, as shown in Figure 5-13. The equivalent width can be used to estimate local scour depth where debris is present, if the dimensions of the likely debris accumulation can be estimated. Figure 5-13 also shows trends in the data. For the study, Td/b varied from 0.52 to 1.64, while Wd/a varied from 2.1 to 6.9. The maximum local scour depth recorded was 3.6a, representing a 50% increase over that at a uniform circular pier (ys = 2.4a). The maximum scour depths occurred when the debris raft extended to about the undisturbed bed level, that is, Td ≈ y, as also reported by Lagasse et al. (2008). Derived from data for circular piers and debris masses, it is unclear how Eqs (5-3) and (5-4) apply to other pier shapes.

90 Figure 5-13 Local scour-depth variation with quantity of floating debris (Melville and Coleman 2000) The literature on pier scour at bridge piers subject to ice accumulation is quite sparse. Only a handful of studies have been completed (e.g., Zabilansky 1996, Hains 2004, Hains et al. 2004, Zabilansky and White 2005). Commonly ice accumulates across the entire bridge water waterway, such that the approach flow depth rises, and pier scour is not significantly affected. If a thick accumulation of ice jam occurs at a bridge waterway, the flow and scour condition becomes similar to scour at a submerged bridge deck, as elaborated in sub-section 5.4. 5-6 Channel Morphology Channel morphology may influence pier scour by locally affecting flow behaviour, sediment movement, and bathymetry at a pier site. NCHRP 24-27(03), Evaluation of Bridge-Scour Research: Geomorphic Processes and Predictions, a companion to the present project, reviews existing knowledge regarding morphologic processes related to channel stability at bridge sites. The evaluation outlines that, though the processes

91 themselves are well understood, there can be considerable uncertainty as to how they impact pier scour. Some influences can be difficult to take into account, especially when they trigger unplanned changes in channel alignment and bathymetry. The following aspects of channel morphology complicate reliable estimation of pier scour depth: 1. Long-term degradation and aggradation of the channel; 2. Lateral migration 3. Non-uniform cross-sectional shape of the approach channel; 4. Pier site proximity to a channel bend, a large bar, confluences of channels, or a local irregularity in channel bed or bank alignment; 5. Pier closeness to a channel bank; 6. The lateral distribution of flow velocity to a pier; 7. The effect of the cross-sectional shape of a compound channel on the flow intensity parameter V/Vc 8. The interaction of pier scour and contraction scour at a pier site. , at a pier; and, Some aspects often are foreseeable, such as the lateral migration of the approach channel (or its thalweg), or the bathymetric features associated with a channel confluence. Other aspects may not be foreseeable at the time of pier design, especially inadvertent consequences of a subsequent engineering activity (e.g., a dam) that alters the stage- discharge relationship at a bridge site. Common major concerns are channel bed degradation or aggradation, and bed lateral migration. These concerns can be addressed during pier design as long as the velocity and depth of approach flow used to estimate pier scour depth represent foreseen flow conditions during the design life of the pier. In this regard, hydraulic laboratory modeling or numerical modeling may be needed to assess flow conditions at a bridge site. Then, the following methods may be used to address the channel-morphology concerns: 1. Guide the approach flow so as to conform to required design conditions. This is achieved by means of channel-training works (e.g., hard points, spur dikes, guide- banks); and, 2. Monitor bridge waterway conditions to check for potentially adverse channel morphology effects. A wise old saying (Neil 1980) holds that “person who overlooks water under bridge will find bridge under water.” The need for monitoring increases when erosion processes within bridge waterways (at abutments, piers, channel banks, shift in thalweg) are complicated by increasing interactions with each other, and may cause scour depth estimations to be of uncertain accuracy. Design relationships for scour at piers and abutments typically are derived from laboratory tests of piers and abutments simulated in simplified conditions that do not replicate the complexities of actual bridge sites. Therefore, as indeed recognized by most agencies responsible for bridges, there is on-going need to monitor bridges to ensure that their foundations and approach embankments are not imperilled by the various erosion that may occur.

92 One set of questions still to be addressed adequately concerns the combination of pier scour and some other scour forms, notably contraction scour, or scour at some other channel feature such as a confluence of approach flows (at a confluence of two channels, or behind an island or large bar). Figure 5-14 illustrates a case where pier scour occurred within a narrow region of contraction scour on a grassy flood plain. In some respects, this case is akin to Case III of pier scour in layered sediment (sub-section 5.7). There seem to be very few studies examining this aspect of pier scour. There are, however, several field-case photos of pier scour in situations where both pier scour and contraction scour evidently occurred (e.g., Figure 5-14). Figure 5-14 Pier scour and abutment/contraction scour on flood plain. Channel geometry and vegetation substantially affect the approach flow to the pier 5.7. Layered Sediments Many sedimentary deposits are heterogeneous and often distinct layers are present. If a more resistant layer underlies a readily erodible layer, scouring is inhibited and lesser scour depths may develop than predicted using the surface material characteristics. In some instances, coarse sediments cover deposits of finer material. Ettema (1980) (also, Breusers and Raudkivi, 1991, Melville and Coleman 2000) identified the following four cases of coarser material overlying finer sediment, as shown in Figure 5-15: • Case 1, where the covering layer is thicker than the local scour depth, is the usual scour problem (in the coarser sediment). • Case 2, the local scour penetrates the covering layer and induces a disintegration of the layer in both the upstream and the downstream directions for a considerable distance. The end conditions are those of local scour in the underlying sediment due to the new hydraulic conditions. The total (local) scour depth is ys + (yf - yc), where yc and yf = uniform flow depth over a flat bed of particle sizes equivalent

93 to the upstream coarse surface particles and the underlying surface fine particles, respectively. Ettema (1980) determined that the total local scour depth for Case 2 was always less than that for Cases 3 and 4. • Case 3, the covering layer disintegrates in the downstream direction only, leaving a step just upstream of the pier. The local pier scour develops at the bottom of this step. This case was found to give the deepest scour (Ettema 1980). Breusers and Raudkivi (1991) give the following expression for H (Figure 5-15): H = 2.6 (yf – yc ) (5-5) An expression for the relation between yc and yf is given also (Breusers and Raudkivi 1991). The total local scour is the sum of H and the local scour depth, ys • Case 4, the covering layer disintegrates only over a small area downstream of the pier but remains intact at both sides of the pier and upstream of it. For this case, the scour depth was about 3a, the increase from that at a pier in uniform sediment being mainly due to reduced downstream support of the scour hole. , in the finer sediment, as for Case 2. Bed-material layers vary locally in extent and details, such that scour-depth estimation should take into account site extents of sediment layering, and consider the merits of placing a pier in a location least at risk from the consequences of layering. This complication remains a topic of active research (Raikar and Dey 2009), and relates closely to the design of riprap protection of piers (e.g., Mashahir et al. 2010, Lagasse et al. 2007).

94 Figure 5-15 Local scour in layered sediments (Breusers and Raudkivi, 1991); yc = depth to top of coarse layer, ys = depth to top of fine layer 5.8 Scour of Cohesive Soil Since 1990, insight regarding pier scour in cohesive soils or clay has advanced substantially. NCHRP 24-15 (Briaud et al. 2004a; also, Briaud et al. 2004b), in particular, provides comprehensive insights. Other notable studies are reported by Hosni (1995), Molinas and Abdeldayem (1998), Ansari et al. (2003), Debnath and Chauhuri (2010). The principal findings of these studies can be summarized as follows: 1. Increased clay content in a predominantly sand bed reduces scour depth, for constant approach flow conditions, because the addition of cohesive strength increases sediment resistance to erosion. The extents of scour reduction vary with clay mineralogy (Molinas and Abdeldayem (1998), Ansari et al. (2003), Debnath

95 and Chaudhuri (2010). As the clay content increased, the location of deepest scour shifted from the pier face to the pier flanks (Anasari et al. 2003); 2. Observations (Ansari et al. 2003, Briaud et al. 2004a) indicate that clay mostly erodes in clusters or lumps of material. The size of the lumps varies with clay properties; 3. The geometry, location and extent of scour at piers differ substantially in clay from those in cohesionless material. The scour holes reported by Ansari et al. (2003) show deepest scour at the pier flanks, and are similar in appearance as the scour hole depicted in Figure 3-2b (from Briaud et al 2004). Most laboratory scour holes had minimal scour at the pier face; 4. The scour holes observed during laboratory tests were no deeper than developed by scour in a sand bed; 5. As clay erodes from the scour hole it does not deposit as a scour mound immediately downstream of the scour hole; and, 6. Scour typically develops much slower in clay materials than in non-cohesive materials. Briaud et al. (2004) indicate that scour rates can be of the order of 1,000 times slower in clay. The relatively slow rate of scour development may mean that flood durations may be too short to generate significant local scour depths. Briaud et al. (2004a) and Ansari et al. (2003) usefully discuss the clay property variables affecting clay erodibility. They include soil unit weight, mineralogy, moisture content, and a set of physio-chemical factors affecting the levels of electromagnetic and electrostatic inter-particle forces within clay. For differing combinations of these properties, clay may erode at different rates, and in somewhat different manners. Additionally, non-uniformities and cracks in clay can affect clay erodibility. Turbulence structures in the pier flow field evidently exert oscillatory stresses on the clay material, and through a process of fatigue dislodge clay lumps from the scour region. The details whereby the pier flow field interacts with the clay boundary, and scours it, have not been studied. A difficulty with estimating scour depth in clay, and mixtures of clay and non-cohesive material, is determining the erosion resistance or critical shear stress (or velocity) for clay. To address this difficulty, Briaud et al. (2004a, b), and others, use a device termed an Erosion Function Apparatus to ascertain the erodibility of a 75mm-diameter surface of clay and other material. They use this device to estimate rates of scour at model piers. 5.9 Scour of Weak Rock It has been assumed for design estimation of pier scour (e.g., Richardson and Davis 2001) that a rock foundation beneath a pier is not readily subject to scour. This assumption was exposed to question by the failure in 1996 of a pier at Schoharie Creek Bridge (Lagasse et al. 1988, Richardson and Davis 1988). The rock beneath a pier footing was sufficiently weak that flow eroded rock fragments, leading to the pier’s collapse. NCHRP 24-29 Scour at Bridge Foundations on Rock presently is underway to assess scour of weak rock foundations (Keaton et al. 2009). A few individual studies of pier scour in rock have

96 been conducted. Hopkins and Beckham (1999) for example, investigated scour of rock beneath piers at several hundred bridges in Kentucky. In general, observations regarding pier scour of rock can be summarized as follow: 1. Scour of weak rock is a possible concern for piers on slab footings; 2. The scour depths are considerably less than observed at pier scour of a sand foundation material. Hopkins and Beckham (1999) found that, when pier scour occurred, it was less than about 0.25m deep; 3. Rock material commonly fails by virtue of fatigue associated with oscillatory loads imposed by turbulence structures in the pier flow field. Physical and chemical and weathering of joints within rock helps facilitate scour of rock; 4. The concern for scour of weak rock may increase if the rock weakens when exposed to freeze-thaw cycles, such as can occur for footings exposed to frigid air during low flow periods during winter conditions; and, 5. Scour holes in rock are irregular in form, in accordance with patterns of failure of the rock. 5.10 Suspended Sediment (Silt and Clay) in Flow The presence of suspended clay and silt in the flow toward a pier in a planar bed is reported to influence scour depth (Sheppard et al. 2002, Clunie 2002). However, the mechanism for this influence is unclear. Two possible mechanisms need to be examined: 1. The presence of suspended sediment might alter the flow field; and/or, 2. Deposition of suspended sediment might increase bed resistance to erosion. To date, insufficient study has confirmed the veracity of these mechanisms. Studies imply that both mechanisms could be at play under certain conditions of flow and suspended-sediment concentration. Most studies examining these mechanisms involve flow and bed sediment movement on planar beds. Fortier and Scobey (1926) conducted an early empirical study, and found that the velocities for bed erosion inception were significantly higher in turbid water compared to clear water, which points toward a reduction of the bed shear stress with increased turbidity. They suggested that drag reduction should occur because a certain part of the available flow energy will be spent to transport the suspended sediment particles. As suspended sediment dissipates energy from the flow turbulence, more energy is needed to erode bed sediment. To be noted is that the influence reduces scour depth (Sheppard et al. 2002, Clunie 2002). Therefore, this influence is of little practical interest for the purpose of scour depth estimation, though may aid interpretation of field data on scour depth. 5.10.1 Flow Field Modification Suspended sediment may alter the flow field by stratifying flow density within the approach flow that modifies the vertical profile of approach-flow velocity and flow-field

97 turbulence. Larger concentrations of sediment near the bed may cause flow density stratification near the bed, if the concentration of sediment is sufficiently great. Modification of approach-flow distribution and turbulence structure by stratification would affect scour, though suitably controlled studies have yet to be done to yield a quantitative relationship of general use. The experiments reported by Sheppard et al. (2002) using a large-scale, field flume suggest that the presence of suspended sediment in the near bed region reduces scour depth. Other experiments, in a different context, conducted by Best and Leeder (2006) with a seawater/clay suspension over a mud bed show reduced near bed velocities with the increase in the clay concentration. The authors attributed this to the modification of the turbulence structure (e.g., a reduction in the rate of turbulence burst events which reduces the momentum exchange within the boundary layer) in the near bed region due to the increased concentration of sediment. The presence of suspended sediment particles in the near bed region and in the scour hole may globally damp the turbulence and decrease shear velocity at the bed. This mechanism was suggested by Li and Gust (2000). Reduced vertical gradient of approach velocity, and thus magnitude of shear velocity, could reduce scour at a pier. The magnitude of the downflow velocity at the pier face (Figure 3-4) varies in accordance with approach-flow gradient as well as overall magnitude. The complexity of the experimental setup and procedures needed to investigate suspended sediment influence on pier flow field, and much more simply for a straight channel flow, has resulted in the lack of research and data to investigate this influence. Some studies do exist, though. For these few studies, the range of suspended sediment concentration was quite limited, and more data are needed to confirm the trends obtained, as well as identify the range over which these effects can be significant and could affect scour depth. The recent Direct Numerical Simulation study by Cantero et al. (2008) indicates that the presence of self-stratified suspended sediment in a channel flow can substantially decrease the bed shear stress, and thereby diminish the erosion capacity of the flow. Though their study was performed with a no-slip boundary condition at the top surface (closed channel) their findings are directly applicable to the open channel flow. Their simulations show that the mean velocity profile loses its symmetry for sediment having dimensionless settling velocities larger than 1.5x10-2 , and the position of the maximum moves toward the bottom wall. Meanwhile the velocity profiles near the bottom wall were found to deviate from the sharper turbulence profile to a rounder laminar-like profile. This result was attributed to thickening of the bottom wall layer including the viscous sub-layer. The finding is consistent with that of Li and Gust (2000), who related decrease in the bed shear velocity with increase in suspended sediment concentration. The study by Cantero et al. (2008) also showed that, for dimensionless settling velocities larger than 2x10-2 (the ratio between the concentrations at the bottom and top boundaries was close to 10), the bed shear stress decreased more than 50% compared to when no

98 suspended sediment was present. The mixing induced by wall turbulence is inhibited by the flow stratification in the lower part of the channel where the concentration levels are relatively large. A Rouse concentration profile develops. When the settling velocity increases, the gradient of concentration at the bed also increases. Also relevant for dimensionless settling velocities larger than 2x10-2 , the logarithmic region disappeared at the bottom wall while it was still present at the top wall. One should mention the channel Reynolds number was small so for larger channel Reynolds numbers it is expected that a modified log-law region will still be present for a relatively large stratification and associated large sediment concentration gradients in the lower part of the channel. For smaller settling velocities and stratification, a log-law region can still be identified at the bottom wall but the value of the von Karman constant has to be decreased. Thus, a kinematic effect due to the presence of suspended sediment in the flow is present. This effect modifies the mean velocity profile over the depth and decreases the bed shear stress. For a given parameter value, the ratio between the bed friction velocity and its critical value decreases, which means that the local scour depth should decrease. This effect should be relatively easy to account for in classical local scour methods if the quantitative relationship between the mean concentration of suspended sediment and the decrease in the non-dimensional bed friction velocity in the incoming channel is known. The clear-water scour experiments conducted by Sheppard et al. (2004) in a flume in the Conte USGS-BRD laboratory in Turner Falls, Massachusetts resulted in significant differences in the local scour depths as the concentration of the suspended sediment varied in the water supply. Pier diameter varied between 0.11m and 0.91m, and the mean bed sediment size varied between 0.22mm and 2.9mm. As the suspended sediment concentration of the water supply (reservoir of a power plant off the Connecticut River) could not be controlled, scour-depth reduction as a function of the suspended sediment concentration was not quantified. When the water turbidity was higher, lower equilibrium scour depths resulted. The differences were significant; high turbidity diminished the equilibrium scour depth by more than 50% compared to low turbidity conditions. Video recordings of these two experiments conducted with low and high turbidity showed no apparent major difference in flow conditions. The only exception was a reduction of the sediment movement in the bed in the high turbidity experiment. The tests in which the effect of the suspended sediment was the largest were those with a mean sediment size of 0.22mm. Only one test with a bed sediment diameter of 0.8mm was significantly affected by the presence of suspended fine sediment. These tests were conducted over a large range of flow and pier sizes. As the measured concentration of fine sediment in the water column were less than 0.1g/l, the experiments show that relatively small concentrations of sediment are needed to significantly affect the equilibrium scour depth. 5.10.2 Bed Erosion Resistance Deposition of suspended sediment amidst the bed particles can increase bed resistance to erosion (Garcia, 2008). This effect would occur in the scour hole, as well as on the approach bed.

99 As the increase in water turbidity was gradual, Sheppard et al. (2002) and Smyre (2002) concluded the suspended sediment must have a gradual influence on the scour hole development. The influence should be larger in the later stages of the scour process, when bed particles in the scour hole are closer to an incipient motion condition. Also, the presence of deposited suspended sediment amidst bed particles reduces the angle of repose of the bed particles in the scour hole. This effect causes bed particles to slide down the slope of the scour hole more readily than normal, and results in a wider scour hole. In the initial stages of the scour, the development of the scour hole is expected to be similar to the case when suspended sediment is not present, because the erosive forces greatly exceed the erosion resistance of the bed at the pier. As scour progresses, the bed particles become more resistant to entrainment, because the interstices between them infill with fine sediment. The net effect is a shallower, wider scour hole. The observations by Sheppard et al. (2002) confirm that these mechanisms occurred for their experiments. Experiments conducted with live-bed conditions at the University of Auckland offer mixed results regarding the influence of the suspended sediment concentration on the equilibrium scour depth. Two live-bed tests indicate an increase in the scour depth as a result of the presence of suspended sediment in the incoming flow. Meanwhile, clear- water tests confirm that the equilibrium scour depth decreases with the increase in the suspended sediment concentration. 5.10.3 Conclusion The flow and bed-resistance mechanisms stated at the beginning of this sub-section indeed can be at play and affect scour depth. The encouraging conclusion, though, is that washload does not deepen scour compared to scour development in clear-water conditions.