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.
52.1 PROJECT DESCRIPTION AND OBJECTIVES Channel migration has important implications for the design, maintenance, and inspection of bridges and other high- way facilities. Bend shifting or cutoffs upstream of a bridge reach can produce poor alignment of the channel approaching the bridge, high angles of attack, and an increase in vegetative debris from bank erosion. Migration downstream of a bridge can also have negative impacts. For example, downstream channel migration can lead to a meander cutoff with com- mensurate degradation that can extend upstream into the bridge reach. In either case, the result at the bridge can be excessive abutment and pier scour, debris loading, loss of conveyance, or erosion of bridge approaches. The Handbook was developed under the National Cooper- ative Highway Research Program (NCHRP). NCHRP Project 24-16 had as its principal objective the development of a prac- tical methodology to predict the rate and extent of channel migration (i.e., the lateral shift and down-valley migration) in proximity to transportation facilities (Lagasse et al., 2003). Predicting channel migration requires consideration of both systemwide and local factors. The morphology and behavior of a given river reach are strongly determined by the water and sediment discharges from upstream. In dynamically adjusted systems, the rate of lateral shifting increases with the supply of water and sediment from upstream. Changes in runoff and sedi- ment yield, as a result of natural processes or human activities, will trigger changes in rates and modes of channel migration. Locally, the distribution of velocity and shear stress and the characteristics of bed and bank materials will control channel behavior. Therefore, local channel morphology such as dimensions (width, depth, meander wavelength, and amplitude), pattern (sinuosity and bend radius of curvature), shape (width/depth ratio), and gradient will not only reflect upstream controls but also provide information on the direc- tion and rate of channel migration. While geomorphologists may view channel stability from the perspective of hundreds or thousands of years, for high- way engineering purposes, a stream channel can be considered unstable if the rate or magnitude of change is such that the planning, location, design, or maintenance considerations for a highway crossing are significantly affected during the life of the facility. In the context of a bridge crossing, meanders may be regarded as stable if they do not migrate appreciably during the design life of a bridge crossing (75 to 100 years). In streams with actively migrating meanders, the kinds of planform changes that are of concern are (1) incremental channel shift from meander migration and (2) episodic chan- nel shift (avulsion) that occurs when a meander bend is cut off. Predicting hazards to highway facilities from incremen- tal planform changes is the principal objective of the Hand- book (as stated in Section 2.4). Consideration of hazards related to episodic or avulsive channel shifts is beyond the scope of the Handbook, but in Section 2.5 they are briefly discussed along with the potential for predicting planform change from the second area of concern listed above. The Handbook deals specifically with incremental channel shift from meander migration and provides a methodology for predicting the rate and extent of lateral channel shifting and down-valley migration of meanders. The methodology is based primarily on the analysis of bend movement using map and aerial photo comparison techniques; however, frequency analysis results are provided (Section 7.5) to supplement the comparative analysis. The methodology enables practicing engineers to evaluate the potential for adverse impacts caused by incremental meander migration over the design life of a bridge or highway river crossing and to ascertain the need for countermeasures to protect the bridge from any associated hazards. The methodology can be applied to locate and design a new bridge or highway facility to accommodate predicted channel migration during its design life. Predictions of channel migra- tion could also be used to alert bridge inspection personnel to the potential for channel migration that could affect the future safety of a bridge. The Handbook provides aerial photograph comparison tech- niques to predict channel migration in proximity to transporta- tion facilities that will enable informed decision making at all levels. The methodologies will be useful in reconnaissance, design, rehabilitation, maintenance, and inspection of highway facilities. The end result will be more efficient use of highway resources and a reduction in impacts associated with channel migration at bridges and/or other highway facilities. As dis- cussed in Chapter 1, the prediction techniques can also be used by practitioners responsible for river channel maintenance, river restoration/rehabilitation, and floodplain planning and management. 2.2 RIVER CHANNEL PATTERNS Rivers may be categorized as straight, meandering, braided, or anastomosing. The great majority of alluvial streams have meandering planforms, and the Handbook is intended for use only on meandering streams. CHAPTER 2 BACKGROUND
Braided rivers are high-energy rivers with abundant coarse sediment loads that feature multi-thread flow in subchannels (anabranches); mobile braid bars; and rapidly shifting, highly erodible banklines. Braided rivers present particular problems for the designers of bridge crossings (Neil, 1973). Because the analysis of braided rivers is beyond the scope of the Hand- book, the reader is directed to consult Hooke (1997), Thorne (1997), Klingeman et al. (1998), and Knighton (1998) for information on the processes and morphologies of these rivers. Anastomosing rivers feature multiple channels (ana- branches) separated by stable, vegetated islands or floodplain elements and are associated with low-energy fluvial systems (Miall, 1977; Smith and Smith, 1980). The individual ana- branches behave almost independently of each other and are relatively stable compared with multi-thread braided rivers. Highway crossings of anastomosing rivers usually treat each anabranch as a separate alluvial stream. The methodology provided in the Handbook could be used on anastomosing channels displaying a meandering planform to predict the migration rate and risk posed by shifting of individual anabranches. Naturally straight alluvial streams are rarely encountered. Usually, a straight river is constrained by geology or bank protection works, or it has insufficient stream power to erode its boundaries. In such cases, the stream is not alluvial and falls outside the scope of the Handbook. However, some allu- vial streams may possess a straight planform associated with past engineering for flood control or land drainage. Such streams often display a sinuous thalweg with bars or shoals occurring on alternate sides of the channel, spaced at between five and seven times the channel width. In this situation, the hydraulic engineer or bridge inspector should be aware of the potential for deflection of the flow and local bank erosion caused by alternate bars that can initiate meandering of a straight stream with a sinuous thalweg. If evidence of mean- der initiation is present, then the methodology presented here can be used to predict the possible rate of bend growth and migration once bank retreat commences. 2.3 STREAM MEANDERING 2.3.1 Stable Versus Active Meandering This section is a summary of Chapters 2 and 6 of FHWAâs Hydraulic Engineering Circular No. 20 (Lagasse et al., 2001) on the characteristics of meandering streams and the geomorphic factors affecting stream stability and meander migration. Meandering streams can be classified as either stable or actively meandering. An actively meandering stream has suf- ficient energy (stream power) to deform its channel bound- aries through active bed scour, bank erosion, and point bar growth. Active meanders are the result of contemporary flu- vial processes. They evolve and respond to every discharge event with sufficient stream power to mobilize bed and bank sediments. Conversely, a stable meandering stream is one that, under current conditions, has insufficient stream power to 6 erode its banks. Stable meanders do not migrate appreciably during the design life of a bridge and generally pose little or no risk to bridge crossings. Clearly, it is essential to differentiate between stable and active meandering when predicting future rates of lateral migration. In this regard, the first step in screening streams in the methodology presented here is to determine whether a meandering stream is active or stable (see Chapter 3). 2.3.2 Flow Patterns and Cross-Sectional Geometry of Meander Bends The main features of the flow and geometry at meander bends (depicted in Figure 2.1) are the following: ⢠Superelevation of the water surface against the outside bank (see Figure 2.1B). ⢠Helicoidal flow directed toward the outer bank at the sur- face and toward the inner bank at the bed producing a strong secondary circulation superimposed on the main downstream flow in the mid-channel (thalweg) region of the cross-section (see Figure 2.1C). ⢠A small secondary-current cell rotating in the opposite direction of the main helicoidal circulation in the outer bank region, especially where the bank is steep (see Fig- ure 2.1B and C). The strength of this cell increases with the strength of the main, helicoidal circulation. Interac- tion of the main and outer bank cells generates plunging flow that scours the bed and lower bank leading to bank undercutting. ⢠An area of outward-directed transverse flow in the inner bank region caused by convective acceleration above the shoaling point bar (see Figure 2.1C). Outward flow forces the core of maximum velocity rapidly toward the outer bank (Dietrich and Smith, 1983; Dietrich, 1987). ⢠A maximum-velocity thread that moves from near the inner bank at the bend entrance to near the outer bank at the bend exit, crossing the channel at the zone of maximum bend curvature (see Figure 2.1A [ii]). ⢠Progressive shifts in maximum shear stress zone through the bend from close to the inner bank at the bend entrance, to the channel centerline at the apex, and to the outer bank at the bend exit as a result of the bar-pool topography and cross-sectional asymmetry characteristic of meander bends (see Figure 2.1A [i]). ⢠A highly asymmetrical cross section featuring deep scour in the thalweg adjacent to the outer bank, a shallow point bar at the inner bank, and a steeply sloping point bar face separating these two regions (see Figure 2.1 C). The flow pattern at a bend is not constant but varies with discharge. Primary flow effects are dominant at high dis- charges because the main flow follows a straighter path, but secondary currents are relatively strong at intermediate dis- charges (Bathurst et al., 1979). Consequently, the point and severity of maximum erosive attack of the outer bank moves as a function of stage. At intermediate stages, attack is con-
7centrated between the bend apex and exit, leading to simulta- neous lateral growth and down-valley migration of the bend. However, during high in-bank flows, the point of greatest attack moves downstream of the bend exit, leading to rapid bend migration downstream. The degree of superelevation of the flow and the strength of the secondary circulation increase in tighter bends (low ratio of bend radius of curvature [RC] to channel width [W]). In bends where RC/W < 2, flow impinges on the outer bank at an acute angle, causing flow separation and generating a strong back eddy along the outer bank near the bend apex, possibly inducing sedimentation along the outer bank upstream of the bend apex (Hickin and Nanson, 1975, 1984). The channel width/depth ratio (W/D) also exerts a major influence on flow pattern (Markham and Thorne, 1992). Point bar development is more extensive in bends with high width- to-depth ratios, and the shoaling effect over the bar directs the inner bank flow radially outward over a large percentage of the width, concentrating helical flow close to the outer bank. Conversely, in narrow, deep channels, especially where W/D < 10, wide bars are less likely to form, reducing the shoaling effect and allowing an inward movement of near- bed flow. The morphological features typical of meander bends are shown in Figure 2.2. Point bars form at the inner banks of meander bends because of sediment accumulation there. The width of the point bar is often taken to indicate the intensity of meander growth and migration. For example, the presence of a wide and unvegetated point bar may be attributed to a rate of bar growth and lateral advance that is too rapid for vegetation establishment. However, the estab- lishment of vegetation on a point bar also depends on factors other than the rate of growth, such as climate, availability of seeds for pioneer species, and the recent record of major floods. Therefore, the absence of vegetation on a point bar is not conclusive evidence of rapid channel migration. Where point bars become very wide, they often feature chute channels cutting across the inside of the bend. These chute channels are activated during high in-bank flows and can indicate that the channel may have the potential to cut off. Figure 2.1. Flow patterns in meanders: (A) (i) location of maximum boundary shear stress (Ïb) and (ii) flow field in a bend with a well-developed point bar (after Dietrich, 1987); (B) secondary flow at a bend apex showing the outer bank cell and the shoaling-induced outward flow over the point bar (after Markham and Thorne, 1992); and (C) model of the flow structure in meandering channels (after Thompson, 1986). Black lines indicate surface currents and white lines represent near-bed currents. (Figure caption is from Knighton, 1998). SOURCE: Knighton, 1998
Chute channels are associated with relatively high stream powers, abundant bed material loads, erodible banks, and high rates of lateral migration. 2.3.3 Characteristics of Meandering Streams The intensity of meandering is most often characterized by channel sinuosity, which is defined as the ratio of the channel centerline length to the valley length. The valley length is either measured along the valley centerline or along a straight line connecting the ends of channel reaches defined by regional changes in channel direction. Straight stream reaches have a sinuosity of one. The maximum value of sinuosity for natural single-thread meandering streams is about four. Idealized meandering streams are depicted as a sequence of symmetrical curves with similar geometries and dimensions. In fact, symmetrical meander loops are uncommon, and a sequence of two or three identical symmetrical loops is rare in nature. In addition, meander loops are rarely of uniform size. The variability of bank materials and the fact that lateral activ- ity by the stream encounters and produces features, such as clay plugs, distort bend form to produce a wide variety of bend shapes. For example, Fiskâs (1944, 1947) work on the lower Mississippi River indicates that the form of most meanders is influenced by local variations in the erodibility of the mat- erials encountered in the outer bank. More generally, analysis of actual meandering streams indicates that the largest bends are commonly about twice the diameter of the smallest. Meander loops move at unequal rates because of their dif- ferent stages of development and variations in the erodibility of the banks. Consequently, a bend at any moment may be atypical of morphological conditions, and years of observation may be required before planform parameters characteristic of average conditions in the stream can be identified. Bends in meandering streams are connected by crossings (short straight reaches) that are generally shallower and nar- rower than typical cross sections at bends (see Figure 2.2). 8 Sediment eroded from the outside bank is deposited in the crossing to form a prominent bar, termed a riffle, and large sandbars may form if the channel is laterally unconfined. The banklines in crossing reaches are usually less mobile than at bends, although significant erosion of the down-valley bank may occur during large in-bank and overbank floods. 2.3.4 Meander Planform Geometry Meander bends are defined by their shape, bend radius, amplitude, and wavelength (see Figure 2.2); and general mor- phological descriptions and relationships exist between these planform characteristics and channel width. Langbein and Leopold (1966) characterized the shape of meanders as a sine-generated curve, which closely approxi- mates the curve of minimum variance, or least work, in turn- ing around the bend. This concept describes the form of symmetrical meander paths relatively well. However, real meanders are asymmetrical and deviate significantly from idealized, perfectly symmetrical, sine-generated curves. Bend asymmetry occurs because the point of deepest scour and maximum attack on the outer bank in a bend is usually located downstream of the geometric apex of the bend, causing bends to become skewed in the down-valley direction as they migrate downstream (see Section 2.3.2). Leopold and Wolman (1957, 1960) established that the meander wavelength (λ) is generally 10 to 14 times the width. They further noted that the radius of curvature (RC) of a well- developed bend is generally two to three times the width at the crossing. They found these relations to hold over several orders of scale of flow in a variety of natural environments. Schumm (1968) analyzed large empirical data sets for sand bed channels in an attempt to account for the effect of bound- ary materials on meander wavelength explicitly by using a weighted silt-clay index of the bed and bank sediments. He determined that meander wavelength decreases as the pro- portion of fine material in the bed and banks increases. This Figure 2.2. Plan view of a meandering stream. MEANDER WAVELENGTH DOWN VALLEY CROSS VALLEY AM PL IT UD E (A )
9indicates that the greater erosion resistance of silt-clay banks results in tighter, shorter wavelength bends than those chan- nels with less cohesive, easily eroded banks. Schumm has also demonstrated that changes to the weighted siltâclay index (M) produce changes in the channel sinuosity and width/depth ratio. The relationship links the characteristic pattern of a mean- dering channel to its cross-sectional shape and the nature of the boundary materials. By combining the width and bend radius described above, it can be shown that radius of curvature (RC) is approximately equal to about two to three times the channel width (W) in mature bends. For RC/W values of 2 to 3, Bagnold (1960) showed that energy losses caused by the curving of flow in the bend were minimized. Plots of both meander migration rate and bend scour depth as a function of bend tightness also peak sharply at an RC/W of between 2 and 3, indicating that these bends are the most effective at eroding their bed and banks. The fact that many bends in nature develop and retain an RC/W value of 2 to 3 while migrating across the floodplain may be con- sistent with their conformance to the most efficient hydraulic shape, which also maximizes their geomorphic effectiveness (Thorne, 1997). Thorne (1992) examined the distribution of bend scour with bend geometry in a study of the Red River and determined that in very long radius bends (RC/W > 10) mean pool scour depth is about one and a half times the mean riffle (crossing) depth, while the maximum scour depth is between one and seven tenths and two times the mean crossing depth. Scour depths ranged from two to four times the mean crossing depth for bends with RC/W values between 2 and 4, with the deepest scour associated with an RC/W of about 2. Evidence suggests that maximum scour depths decrease with decreasing bend radius for extremely tight bends with RC/W < 2. Although these morphological relationships are useful, cau- tion should be taken when using them to predict channel geometry and scour depth at specific sites. As Knighton (1998) states: âThese various relationships indicate a self-similarity of meander geometry over a wide range of scales and environ- mental conditions. However, the regularity which they imply is not everywhere apparent, and the use of single parameters provides only a partial and often subjective characterization of meander form.â 2.3.5 Bank Erosion in Meander Bends Bank retreat in actively meandering streams is responsible for lateral channel migration. There are two mechanisms by which stream banks retreat: (1) fluvial entrainment (detach- ment of grains or aggregates by the flow) and (2) mass fail- ure (slumping or sliding caused by gravity). The specific retreat mechanisms at a given location are related to the char- acteristics of the bank material. Severe bank retreat often results when mass failure and fluvial entrainment act in con- cert. Fluvial erosion scours the lower bank and bank toe, leading to oversteepening and mass failure of the upper bank. Removal of failed bank material from the base of the bank follows through fluvial erosion and the cycle of retreat is repeated. It is important to note that fluvial erosion of previously failed bank material plays a significant role in determining the rates of bank retreat. Fluvial activity controls the state of basal endpoint control, or the amount of basal accumulation of bank material, which ultimately controls the stability of the bank. Removal of the failed material results in the formation of steeper banks and may induce toe erosion by removing the material along the toe that buttresses the bank slope (Thorne, 1981). These factors renew the process of bank erosion by mass failure. Without basal erosion, mass failure of the bank material would lead to bank slope reduction and stabilization within a relatively short period of time (Lohnes and Handy, 1968; Thorne, 1981). Curved flow around a bend causes erosion at the toe of the outer bank and subsequent bank failure because of high near bank velocities, plunging flow, and increased shear stresses around and downstream of the bend apex. As a bend tightens, flow may approach the outer bank at a very acute angle, generating extremely severe toe scour, ero- sion, and bank retreat through impinging flow. However, if the bend tightens further, flow at the outer bank sepa- rates, leading to reduced erosion in the zone of separation and possible bar accumulation at the outer bank. Separated flow at the outer bank may deflect the filament of maxi- mum velocity so that it attacks the inner bank in the bend or the down-valley bank in the crossing downstream. In either case, the distribution of bank retreat and meander migration will be radically altered in a manner that is very difficult to predict. 2.3.6 Modes and Rates of Meander Migration Although no two meanders migrate in exactly the same way, the movement of bends in a particular stream reach tends to conform to one of the several modes of behavior illustrated in Figure 2.3, which is based on a study of about 200 sinuous or meandering stream reaches (Brice, 1977). Rates of lateral movement generally scale to the size of the channel with annual migration rates generally being 10 percent of the channel width but sometimes reaching as high as 20 percent (Hooke, 1997). In large rivers, these dimensionless rates convert into considerable absolute rates, and meanders have been observed to move at 2,500 ft/year (750 m/year) in large alluvial rivers (Lagasse et al., 2001). Mode a in Figure 2.3 represents the typical development of a loop of low amplitude, which decreases in radius as it migrates downstream. Mode b occurs only where meanders are confined by artificial levees or by valley sides on a narrow floodplain. Well-developed meanders on streams that have moderately erodible banks are likely to follow Mode c. Mode d applies mainly to larger loops on highly sinuous or tortuous streams. The meander has become too long in relation to stream width, and secondary meanders have developed, making a compound loop. Mode e also applies to highly sinuous mean- dering streams, usually with relatively narrow point bars and no chutes. An elongated loop has formed without a chute cut- off, but the neck of the loop is gradually being closed and a neck cutoff will eventually occur. Modes f and g occur mainly
10 Figure 2.3. Modes of meander loop development: (a) extension, (b) translation, (c) rotation, (d) conversion to a compound loop, (e) neck cutoff by closure, ( f ) diagonal cutoff by chute, and (g) neck cutoff by chute. SOURCE: Modified from Brice, 1977 Figure 2.4. Relationship between erosion (migration) rate and bend curvature. SOURCE: Hooke, 1991 in less sinuous meandering streams with easily eroded banks and wide point bars. Low-amplitude loops are cut off by chutes that break diagonally or directly across the point bar. The rate of migration is largely controlled by bend geome- try, especially by channel curvature as represented by RC/W. Hickin and Nanson (1975, 1984) and Nanson and Hickin (1986) demonstrated through detailed studies of more than 125 bends on 19 river reaches in Canada that the rate of migration reaches a maximum when 2 < RC/W <3. This rela- tionship between form and rate of change for meanders (see Figure 2.4) has been further substantiated by Biedenharn et al. (1989) and Hooke (1987). The rate decreases rapidly on either
11 side of this range. At the lower end of the range, the decrease may be attributable to the large increase in resistance or a decrease in outer-bank radial force as RC/W falls below 2. Observations indicate that river meanders tend to an RC/W value of between 2 and 3. Because of the loss of energy asso- ciated with flow through a bend, a maximum bend sharpness exists beyond which further significant lateral erosion is unlikely to occur. It has been shown that the maximum lat- eral erosion rate for a meander bend occurs when the ratio of radius of curvature to channel width is in the range of about 2 to 4 (see Figure 2.4). For RC/W values less than about 2, the erosion rate reduces sharply because of energy loss in the bend; and, in very tight bends (RC/W < 2), deposition may actually occur along the outer bank of the bend. Under this condition, the rate of lateral migration significantly decreases or migration stops, and the bend either cuts off or avulses (i.e., undergoes an abrupt shift in channel course). Oxbow lakes are formed by the cutoff of meander loops, which occur either by gradual closure of the neck (see Figure 2.3e) or by a chute that cuts across the neck (see Figure 2.3g). Commonly, a new meander loop soon forms at the point of cutoff and grows in the same direction as the previous mean- der. Although recently formed oxbow lakes along a channel are evidence of recent lateral migration, the presence of abun- dant oxbow lakes on a floodplain does not in itself indicate rapid channel migration because an oxbow lake may persist for hundreds of years. Usually, the upstream end of the oxbow lake fills quickly to bank height. Overflow during floods and overland flow entering the oxbow lake carry fine materials into the lake area. The lower end of the oxbow remains open, and drainage enter- ing the system can flow out from the lower end. The oxbow gradually fills with fine silts and clays, which are plastic and cohesive. As the stream channel meanders, old meander bends filled with cohesive materials (referred to as clay plugs) are sufficiently resistant to erosion to serve as semi-permanent geologic controls that can drastically affect planform geometry. The local increase in channel slope caused by cutoff usually results in an increase in the growth rate of adjoining meanders and an increase in channel width at the point of cutoff. Typi- cally, the effects of a cutoff do not extend very far upstream or downstream. The consequences of cutoffs are an abruptly steeper stream gradient at the point of the cutoff, scour at the cutoff, and a propagation of the scour in an upstream direction. Downstream of a cutoff, the gradient of the channel is not changed, and, therefore, the increased sediment load caused by upstream scour will usually be deposited at the site of the cutoff or below it, forming a large bar. 2.3.7 Measuring Meander Migration Before predictive tools for channel migration can be developed, one must be able to measure and describe chan- nel migration. A standard approach to analyzing data sets must be developed, and this approach should be adhered to for all subsequent measurements. The initial or existing meander bend should be represented by a starting point (upstream end), an ending point (down- stream end), a location of the center of bend radius (bend cen- troid), an orientation with respect to a baseline (e.g., down- valley direction), and an outside bank radius of curvature (RC). As shown in Figure 2.5, it can be assumed that the bend starts and ends at the flow crossing (shown as âriffleâ in Figure 2.2). Bend migration can be reasonably described by four modes of movement. Extension is across-valley migration and is easily measured at the bend centroid. Similarly, translation is down-valley migration and is also measured at the bend cen- troid. Expansion increases bend radius; contraction decreases bend radius. Rotation is a change in the orientation of the meander bend with respect to the valley alignment. Predicting four modes of movement is a significant task for every bend of interest (see Figure 2.5). However, actual bend migration is even more complex. For example, one part of the bend may be expanding faster or translating down valley faster than another, resulting in changes in bend symmetry. As a concession to practicality, one must limit the number of modes of movement to the fewest possible. In the methodol- ogy developed in Chapter 7, extension and translation are considered directly (as a vector sum). Expansion (a change in RC) is included because it could have a major impact on the location of the outer bank and because rates of migration appear to be correlated to RC/W (bend radius of curvature/ width). If movement in these three modes can be predicted, the primary threats to a bridge, highway, or other facility will be established. Rotation is considered only indirectly as a component of the combined movement in the other three modes relative to adjacent bends. 2.4 HAZARDS TO HIGHWAYS CAUSED BY INCREMENTAL MEANDER MIGRATION A number of hydraulic problems may stem from meander migration in the vicinity of a highway crossing. These include the incremental shift of flow direction (angle of attack) approaching bridge piers or abutments; development of an extensive point bar in the bridge reach with the commensu- rate loss of conveyance; lateral channel erosion at piers, abut- ments, or approaches; and possible flanking of the bridge approaches. Meander migration can also exacerbate scour and pressure loading because of debris accumulation. For example, failure of a major bridge on the Hatchie River near Covington, Tennessee, in 1989, has been attributed in part to lateral migration of the channel in the bridge reach (Bryan, 1989; NTSB, 1990). Significant migration of the right bank of the channel between 1975 and 1989 had undermined the foundation of a support bent previously located in the floodplain and contributed in part to the collapse of three spans of the roadway and the deaths of eight people. The maximum bank migration rate of 4.5 ft (1.37 m) per year occurred between 1975 and 1981. The chapters of the Handbook that follow this one provide screening and classification procedures and methodologies to
predict the rate and extent of incremental channel shift as a result of meander migration. Aerial photograph comparison and frequency analysis techniques are provided to predict the potential impact of meander migration on highway facilities for specific classes of meandering channels. 2.5 HAZARDS TO HIGHWAYS CAUSED BY AVULSIONS AND CUTOFFS The scope of NCHRP Project 24-16 specifically limited the investigation of meander migration to incremental channel shift (i.e., lateral channel shift and down-valley migration). 12 It was beyond the scope of the project to develop a method- ology to predict avulsive or catastrophic channel shift; how- ever, the process of avulsion, or cutoff of a meander, can also have a significant impact on the stability of a highway facility. These impacts include accelerated erosion of adjacent bends, especially the bend immediately downstream; degra- dation in the upstream channel reach; and aggradation in the downstream channel reach. An avulsion is a sudden change in the river channel course that usually occurs when a stream breaks through its banks. Avulsions are often associated with a major flood or cata- strophic event. A cutoff is an avulsion related to a single mean- der loop and can be defined as a natural or artificial channel Figure 2.5. Measuring meander migration. Extension (Across-Valley Migration) Translation (Down-Valley Migration) Expansion (Increasing Radius) Rotation All Four Modes of Movement Flow Upstream End (crossing) DownstreamEnd (crossing) Rc Channel Centerline Down Valley Across Valley Initial Bendway
13 that develops across the neck of a meander loop (neck cutoff) or across a point bar (chute cutoff). The following discussion from a study of the cutoff of mean- der bends on the Sacramento River is provided to indicate the potential for developing a predictive methodology for avulsions and cutoffs (WET, 1988; Harvey, 1989). Meander bends eventually cut off when the radius of cur- vature (RC) becomes too short, primarily as a result of the arrest of one limb of the bend by more erosion-resistant material. This is a manifestation of reduced hydraulic slope, which causes a reduction in sediment-transport capacity of the flows within the upstream portion of the bend and leads to sediment deposition within the channel and reduced hydraulic capacity in the channel (Harvey, 1989). Reduced hydraulic capacity in the upstream portion of the bend increases the fre- quency of flows over the point bar, which leads to chute development and eventually bend cutoff. Cutoffs can occur as a result of chute development or neck closure. Recent and historic cutoffs on the floodplain of the Sacra- mento River were investigated to evaluate whether cutoffs could be predicted (WET, 1988; Harvey, 1989). A dimension- less cutoff index, which was defined as the ratio of the center- line radius of curvature to the migration distance (RC/MD), was developed to predict cutoff occurrence. For the coarse- grained meanderbelt section of the Sacramento River, the dimensionless cutoff index was For the fine-grained meanderbelt section, where the flood- plain sediments are more cohesive, the cutoff index was Associated with these values are two other characteristics that were identified on aerial photographs: (1) the presence of a mid-channel bar in the upstream limb of the bend and (2) the presence of chute channels across the point bar. The cutoff index was tested independently on bends in the Butte Basin reach of the Sacramento River, and it was evident that there is a relationship between the cutoff index and the presence of the two ancillary features. It should be noted that the cutoff indices developed for the Sacramento River are site specific and have not been general- ized for application to other river systems. They are presented here only as evidence that it should be possible to develop pre- dictive relationships for channel avulsions or cutoffs. Devel- oping these predictive relationships was beyond the scope of NCHRP Project 24-16, and the Handbook does not provide a methodology to predict episodic channel shift. However, the possibility of a neck cutoff can be identified using the techniques presented in the Handbook when the predicted incremental migration leads to meander neck closure. 2 5 4 3 2. . ( . )< <R MDc 2 1 7 3 7 1. . ( . )< <R MDc 2
