Info

FIGURE 3 .8 Relation of mean current velocity in water at least 1 m deep to the size of mineral grains that can be eroded from a bed of material of similar size. Below the velocity sufficient for erosion of grains of a given size (shown as a band), grains can continue to be transported. Deposition occurs at lower velocities than required for erosion of a particle of a given size. (Reproduced from Morisawa 1968.)

silts and clays, have greater critical erosion velocities because of their cohesiveness.

Once in transport, particles will continue in motion at somewhat slower velocities than was necessary to initiate movement (Figure 3 8). As velocities decrease, grains settle out of suspension, beginning with the largest and heaviest. This occurs when discharge declines following a flood, in reaches of lower gradient, at the inside of bends and behind obstructions.

The shear stress or tractive force (to, force per unit area) exerted by the flow of water on the streambed is estimated as:

where p is fluid density, g is gravitational acceleration, the hydraulic radius R equals channel cross-sectional area divided by its wetted perimeter, and S is the water surface slope. For natural channels with a width much larger than mean flow depth, mean depth is a good approximation of the hydraulic radius.

This equation is important because it relates the resistance of the channel bed and banks to the downstream gravitational tractive force of the water: when the former is exceeded, sediment transport is initiated. Critical shear stress (tc) refers to the shear stress necessary to mobilize a given grain size. For mobile, gravel-bed rivers with bed materials >1 cm diameter, the particle size near the threshold of motion at bankfull flow is approximately equal to the median bed material size (cm). In other words, the D50 is a good indicator of the tractive force on the streambed at bankfull flow.

3.3.3 Sediment load

Sediment load is the amount of sediment passing a point over some time interval. It is estimated by multiplying sediment concentration by water discharge. Matter carried by fluvial systems can be separated into three components (Knighton 1998). These are the dissolved load, which consists of material transported in solution; the wash load, consisting of material between 0.5 |m (the upper limit for dissolved material) and 0.0625 mm (the boundary between silt and sand); and solid load, consisting of material >0.0625 mm. Terms that describe the total sediment load refer either to the source of the material or the mode of transport (Hicks and Gomez 2003) (Figure 3.9).

The dissolved load consists of solutes derived from chemical weathering of bedrock and soils. The dissolved constituents of river water are discussed more fully in Chapter 4. Their contribution is greatest in nonflashy hydrologic regimes where most flow is subsurface, and in regions of limestone geology (Richards 1982). The relative amount of material transported as solute versus solid load depends upon basin

FIGURE 3 9 The components of stream sediment load shown in terms of sediment source and mode of transport. (Reproduced from Hicks and Gomez 2003 )

characteristics, lithology, and hydrologic pathways. In dry regions, sediments make up as much as 90% of the total load, whereas the contribution of solutes is substantially more in areas of very high runoff (Richards 1982). Worldwide, it is estimated that rivers carry approximately 15 billion tons of suspended materials annually to the oceans, which is roughly five times the dissolved load (Holeman 1968, Martin and Meybeck 1979).

By source, the total sediment load is split between wash load and bed material load (Hicks and Gomez 2003). The wash load (so named because this load is "washed" into the stream from banks and upland areas) consists of very fine particles including clay and silt up to very fine sand. It requires only low velocities and minor turbulence to remain in suspension, thus this material may never settle out. The amount of the wash load is determined by its supply from uplands and stream banks rather than by the stream's transport capacity, and is likely to be high where stream banks have a high clay and silt content. The bed material load is derived from the river bed, typically sand or gravel, and its concentration is directly related to the river's transport capacity.

By mode of transport, the sediment load is divided into suspended load and bed load. The flow of water in rivers generally is turbulent, and exerts a shearing force that causes particles to move along the bed by pushing, rolling, and skipping, referred to as the bed load. This same shear causes turbulent eddies that entrain particles into suspension, called the suspended load. The distinction between bed load and suspended load is based on sampling method, and the same material that is transported as bed load at low discharge may become suspended load at higher discharge. Bed load transport is difficult to measure, and often involves a trap or tracer particles (Gordon et al. 2004). Suspended load is fairly easy to sample - a simple grab sample will suffice - but varies with depth and can change rapidly with discharge, and so sampling that integrates across depth and takes place frequently over the rise and fall of the hydrograph is preferred. Because fine sediments tend to be washed into the stream at the beginning of a rain event and entrained by rising water, their concentrations usually are greater during the rise of the hydrograph, and decline during the falling hydrograph due to exhaustion of the sediment supply. As a consequence, sediment concentrations can be different at identical discharges of the rising and falling hydrograph. This is referred to as hysteresis.

Suspended sediments cause turbidity by restricting the transmission of light through water due to scattering and absorption. By measuring light transmission through a water sample, turbidity meters provide a simple approximation of suspended sediment loads. These usually are reported as nephalometric turbidity units (NTUs), which can be calibrated against measured sediment concentrations (mg L_1). There are additional sources of turbidity, however, including algae and colloidal matter, and so turbidity is not solely a measure of suspended sediments.

The majority of sediment transport is due to the suspended load, which typically exceeds bed load by a factor of 5-50 (Gordon et al. 2004). The bed load will be a lower fraction in bedrock streams than in alluvial streams where channels are composed of easily transported material. However, bed load transport increases substantially during floods, and is particularly important in determining channel shape. For a stream channel in equilibrium, the transport of bed material requires that it be replaced by material derived from upstream banks and channel, in a cycle of scour and fill that accompanies the rise and fall of flood waters; if not, the bed will be downcut. For example, a flood in the Colorado River at Lees Ferry increased bed depth by approximately 1.5 m. Redeposition of sediment as the flood receded reestablished bed elevation at very close to its previous value (Leopold 1962), further evidence of the dynamic equilibrium between erosion and deposition. Since the closing of the Glen Canyon Dam the Colorado River immediately downstream has been downcut by more than 9 m, demonstrating the consequences of the loss of its upstream sediment supply (Postel and Richter 2003).

3.3.4 Factors influencing sediment concentrations and loads

A stream's capacity is the total load of bed material it can carry. This increases with velocity and discharge unless the supply of sediment becomes depleted; the larger the flow, in general, the larger the quantity of sediment transported (Richards 1982). Throughout most of the year discharge usually is too low to scour, shape channels, or move significant quantities of sediment, although sand-bed streams can experi ence change much more frequently. Although one might suppose that extreme events also account for the greatest proportion of total sediment transport, flow events of intermediate frequency actually move more sediment over the years. The discharge at which sediment transport peaks is called the effective or dominant discharge, and it is found from the product of the discharge frequency curve and the curve describing sediment transport rate as a function of discharge (Figure 310). Because the effective discharge accomplishes the most geomorphic work compared to other flows, it follows that fluvial landforms are shaped by frequently occurring moderate floods, rather than by rare, cataclysmic floods (Wolman and Miller 1960).

The effective discharge often is very close to the bankfull discharge estimated from the 1.5 year flow (Q1:5), as Andrews and Nakervis (1995) report for 17 gravel-bed rivers of the western United States, and the D50 closely approximates the particle size that is mobilized at the effective discharge. Given the difficulty of directly determining effective discharge, the usefulness of bankfull discharge and the D50 is clear as both are field measurements that provide estimates of the discharge responsible for transporting most of the annual sediment flux, and thus having the greatest influence on channel shape. An analysis of suspended-sediment transport data from more than 2,900 sites across the United States, sorted into ecoregions, supported the use of the Q15 as a measure of the effective discharge (Simon et al. 2004). Median values of the recurrence interval of the effective discharge for 17 ecoregions ranged from 1.1 to 1.7 years, and the detection of differences among regions argues for the use of regionalized curves. The concept of a channel-forming discharge is widely used in river restoration designs because it suggests ways to estimate, fairly easily, the equilibrium channel dimensions. These efforts have met with both successes and failures; reasons for the latter include application to an area where conditions may be different, inadequate

FIGURE 3.10 The relationship between frequency and magnitude of discharge events responsible for sediment transport: (a) suspended load, (b) bedload. Curve 1 depicts the increase in sediment transport rate with increasing magnitude of discharge, and curve 2 describes the frequency of discharge events of a given magnitude. Their product (dashed line) is the discharge that transports the most sediment, referred to as Qd, the dominant or effective discharge. Qd is approximately Qbkf for suspended sediments, and is in the range Q1:5 —Q10 for bedload. (Reproduced from Richards 1982.)

FIGURE 3.10 The relationship between frequency and magnitude of discharge events responsible for sediment transport: (a) suspended load, (b) bedload. Curve 1 depicts the increase in sediment transport rate with increasing magnitude of discharge, and curve 2 describes the frequency of discharge events of a given magnitude. Their product (dashed line) is the discharge that transports the most sediment, referred to as Qd, the dominant or effective discharge. Qd is approximately Qbkf for suspended sediments, and is in the range Q1:5 —Q10 for bedload. (Reproduced from Richards 1982.)

consideration of past history, and the problems inherent in applying general relationships to specific cases (Smith and Prestegaard 2005, Doyle et al. 2006).

Concentrations of suspended sediments vary greatly depending on the factors described above that influence sediment supply, and with discharge and velocity, which determine how much sediment is in transport at any time. Based on some 400-600 stations sampled from 1970-1983 across the United States, site-specific measurements of total suspended sediments (TSS) had a median value of 63mgL—1, but varied by more than three orders of magnitude (Dodds and Whiles 2004). Sediment concentrations and yields vary greatly with region, with human activities leading to erosion, and whether the channel is in a stable or unstable state. Using a model of channel evolution that recognized equilibrium conditions both for predisturbance and disturbance-accommodated channel forms, Simon et al. (2004) estimated that the median values (for suspended-sediment yields at Q1:5) at stable sites within a given ecoregion are generally an order of magnitude lower than for nonstable sites.

Nearly 90% of the variation in TSS in the large data set analyzed by Dodds and Whiles (2004) was explained by turbidity, indicating that the latter is a reasonable surrogate measure, at least to within an order of magnitude. TSS was negatively correlated with catchment land area in forest, and highest values were found at forest covers <20%. Relationships with urban land were less clear, presumably because impervious surfaces result in less erodible soil, so high TSS values were rare in urban catchments. TSS also exhibited pronounced differences associated with ecoregion. Lower sediment concentrations were seen in Eastern Deciduous Forests, and higher values in the Great Plains and North American Desert ecoregions.

Sediment yields from individual rivers are calculated as loads divided by catchment area, and provide a useful comparison of variation in sediment export among rivers and over time. Water discharge alone is a poor predictor of sediment load except within a region. Rivers in just 10% of the world's drainage basins account for over 60% of sediment discharge (Milliman 1990). The Hwang Ho (Yellow River) of northern China is believed to carry the highest suspended load of any river, as much as 40% sand, silt, and clay by weight, during high discharge (Cressey 1963). The great rivers of South America make a significant but nonetheless much smaller contribution to the world sediment flux, and large northern rivers account for considerably less.

Human activities can increase or reduce sediment yields. Deforestation and poor agricultural practices greatly increase erosion, perhaps as much as fivefold in Asia and Oceania. On the other hand, sediment flux is greatly reduced in rivers by thousands of large and millions of small impoundments. Although prior work in the United States has emphasized sediment storage in stream channels and floodplains (Trimble 1983), Renwick et al. (2005) estimate that, at least for the later part of the 20th century, much of the sedimentation in the United States actually is occurring in impoundments. Both the Nile and the Colorado have experienced a complete cessation of sediment export, and the Rhone is estimated to export approximately 5% of its load of a century ago. Thus in a number of large rivers we have the apparent paradox of increased erosion within the drainage basin coupled with reduced export to the oceans.

The global consequences of these trends can be seen in Table 3 2, which summarizes dis charge and sediment fluxes for prehuman and modern times by continent. Combining data and models, Syvitski et al. (2005) estimated the global total prior to human influence to be 14 billion tons annually (15.5 billion tons year1 when bed load is included). Asia produces the greatest quantity of fluvial sediment, whereas Oceania and Indonesia have the highest sediment yields as well as the highest runoff (discharge divided by area). By latitude, warm areas produce the highest sediment yields, accounting for nearly two thirds of global delivery. Modern sediment loads are a moving target because land use has generally accelerated erosion (although reforestation and other improvements have brought about declines in some regions), and impoundments retain sediments. The modern sediment flux is estimated to be 12.6 billion tons year1, about 10% less than the prehuman value. Using the additional information that large impoundments trap 20% of the sediment load of rivers and small impoundments an additional 6%, the current flux would be 16.2 billion tons year1 of suspended sediment in the absence of dams. Thus sediment flux into global rivers due to erosion has increased, while sediment yields to the world's coasts have declined. Coastal retreat that may affect inhabited areas, subsidence of

TABLE 3.2 Landmass area, discharge, predicted sediment flux to the world's coastal zones from world rivers under prehuman and modern conditions, and percentage of sediment load retained in reservoirs. Uncertainty estimates for sediment fluxes and sediment retained in reservoirs have been omitted for simplicity, but range from 15% to 30% of stated values. See Syvitski et al. (2005) for details.

Landmass Area Discharge Prehuman suspended Modern suspended Load retained

(M km2) (km3 year—1) load Qs(MT year—1) load Qs(MT year—1) in reservoirs

TABLE 3.2 Landmass area, discharge, predicted sediment flux to the world's coastal zones from world rivers under prehuman and modern conditions, and percentage of sediment load retained in reservoirs. Uncertainty estimates for sediment fluxes and sediment retained in reservoirs have been omitted for simplicity, but range from 15% to 30% of stated values. See Syvitski et al. (2005) for details.

Landmass Area Discharge Prehuman suspended Modern suspended Load retained

(M km2) (km3 year—1) load Qs(MT year—1) load Qs(MT year—1) in reservoirs

Africa

0 0

Post a comment