CHAPTER 5: INCEPTION OF SUSPENDED-LOAD MOTION

5.1. Introduction
As discussed in the previous chapters, the inception of sediment motion is related to the shear velocity
(or to the bed shear stress). Considering a given channel and bed material, no sediment motion is
observed at very low bed shear stress until exceeds a critical value. For larger than the critical
value, bedload motion takes place. The grain motion along the bed is not smooth, and some particles
bounce and jump over the others. With increasing shear velocities, the number of particles bouncing
and rebounding increases until the cloud of particles becomes a suspension.

Note: The onset of sediment suspension is not a clearly defined phenomenon.

5.2. Initiation of Suspension and Critical Bed Shear Stress

Considering a particle in suspension, the particle motion in the direction normal to the bed is related to
the balance between the particle fall velocity component and the turbulent velocity
fluctuation in the direction normal to the bed.

With this reasoning, a simple criterion for the initiation of suspension (which does not take into account
the effect of bed slope) is:

[5.1]
Several researchers proposed criterion for the onset of suspension (Table 5.1). In a first approximation,
suspended sediment load occurs for:

[5.2]
Table 5.1: Criterion for suspended-load motion

CVS 549E Page 1

Example 5.1
Considering a stream with a flow depth of 3.2 m and a bed slope of, sin   0.001 indicate whether a 3
mm gravel bed will be subjected to sediment motion. In the affirmative, indicate whether the sediment
motion is bed load, saltation or suspension. Take kinematic viscosity as 10-6 m2s-1.

Solution
Assuming a wide rectangular channel and assuming that the stream flow is nearly uniform
equilibrium, the shear velocity equals:
DH
V*  g sin   gd sin 
4
 9.81*3.2*0.001  0.18 m / s
where d is the flow depth.

For a 3 mm gravel bed, the Shields parameter is:

V* 0.182
*  
 1  s  gd s  2.65  1 *9.81*3*103
 0.67
Assuming s=2.65.
The particle Reynolds number Re* is:
V*d s 0.18*3*103
Re*    540
 1*106

(Re* , * ) the Shields diagram predicts sediment motion.

For these values,

Let us estimate the settling velocity. For a single particle in still fluid, the fall velocity equals:
4 gd s
w0   s  1
 24  
3  1.5 
  w0 d s 
It yields
0  0.20 ms 1
V 0.18
Then *  is about unity (0.9)
w0 0.20
 Sediment transport takes place with bed load and saltation.
 The flow conditions are near the initiation of sediment suspension (equation 5.2).

5.3. Onset of Hyperconcentrated Flow

1. Hyperconcentrated flows are sediment-laden flows with large suspended sediment
concentrations (i.e. typically more than 1% in volume).

2. Hyperconcentrated flows exhibit different flow properties from clear-water flows. For example,
the fluid viscosity can no longer be assumed to be that of water. In practice, numerous
researchers observed that the properties of hyperconcentrated flows differ from those of
‘classical’ sediment-laden flows when the volumetric sediment concentration exceeds 1–3%.

CVS 549E Page 2

 3. Sometimes the term might also include ‘mud flow’ and ‘debris flow’, but the rheology of these
flows differs significantly from that of hyperconcentrated flows. In mud and debris flow, the
volumetric sediment concentration might be larger than 50%!

 Mud is a sticky mixture of fine solid particles and water (i.e. soft wet earth).
 Debris comprise mainly of large boulders, rock fragments, gravel-sized to clay-sized
material, tree and wood material that accumulate in creeks.

4. Hyperconcentrated flows are often associated with severe land degradation in the catchment

5.4. Exercises
1. Considering a stream with a flow depth of 2.3 m and a bed slope of sin Ө = 0.002, the median
grain size of the channel bed is 1.1 mm. Calculate:
(a) Mean flow velocity,
(b) Shear velocity,
(c) Settling velocity,
(d) Occurrence of bed-load motion and
(e) Occurrence of suspension.
(Assume that the equivalent roughness height of the channel bed equals the median grain size.)

2. A stream carries a discharge of 8m 3/s. The channel cross-section is trapezoidal with a 2.2m
bottom width and 1V:3H sidewall slopes. The longitudinal bed slope is 15 m/km. The channel
bottom and sidewall consist of a mixture of fine sands (d50 = 0.3 mm).

Predict the occurrence of sediment suspension. Assume that the equivalent roughness height
equals the median sediment size. Assume that uniform equilibrium flow conditions are
achieved.

CVS 549E Page 3

CHAPTER 6: BED-LOAD TRANSPORT
6.1. Introduction
When the bed shear stress exceeds a critical value, sediment particles are transported in the form of bed
load and suspended load (formulations for suspended-load transport will not be discussed in this
course). For bed-load transport, the basic modes of particle motion are rolling motion, sliding motion
and saltation motion (Fig. 6.1).

In this chapter, formulations to predict the bed-load transport rate are presented.

Fig. 6.1: Bed-load motion: (a) sketch of saltation motion and (b) definition sketch of the bed-load layer.

 Saltation refers to the transport of sediment particles in a series of irregular jumps and bounces
along the bed (Fig. 6.1).

 In this chapter, predictions of bed-load transport are developed for plane bed. Bed form motion
and bed form effects on bed-load transport are not considered in this chapter.

6.2. Definitions
The sediment transport rate may be measured by weight (units: N/s), by mass (units: kg/s) or by
volume (units: m3/s). In practice the sediment transport rate is often expressed by metre width and is
measured either by mass or by volume. These are related by:


m s
  s qs
[6.1]

6.3. Empirical Correlations of Bed-Load Transport Rate

 *    *  c Bed load transport

[6.2]
0
* 

Where * is the shields parameter
   s  1 gd s  and
  *  c is the critical shields parameter for
initiation of bed-load transport

CVS 549E Page 4

6.4. Empirical bed-load transport predictions
Many researchers attempted to predict the rate of bed-load transport. The first successful development
were proposed by P.F.D. du Boys which will be discussed later in this chapter. Table 6.1 itemizes cal
equations.

Table 6.1: Empirical and semi-empirical correlations of bed-load transport

Notes:
 Subsequently, numerous researchers proposed empirical and semi-empirical correlations. The most
notorious correlations are the Meyer-Peter and Einstein formulae.

 The correlation of Meyer-Peter (1949, 1951) has been popular in Europe. It is considered most
appropriate for wide channels (i.e. large width to depth ratios) and coarse material.
 The Einstein’s (1942, 1950) formulations derived from physical models of grain saltation and they
have been widely used in America.
 Both the Meyer-Peter and Einstein correlations give close results usually within the accuracy of the
data (Fig. 6.2)

CVS 549E Page 5

Fig. 6.2: Bed-load transport rate: comparison between Meyer-Peter formula, Einstein calculation and
laboratory data.

6.4. Bed-Load Calculations

Bed-load transport is closely associated with inter-granular forces. It takes place in a thin region of
fluid close to the bed (sometimes called bed-load layer or saltation layer) (Figs 6.1 and 6.3).

During the bed-load motion, the moving grains are subjected to hydrodynamic forces, gravity force
and inter-granular forces. Conversely the (submerged) weight of the bed load is transferred as a normal

stress to the immobile bed grains. The normal stresses e exerted by the bed load on the immobile bed
particles is called the effective stress and it is proportional to:

 e    s  1 g cos  Cs s
[6.3]

 0    0  c   e tan s
[6.4]

Where
  0  c is the critical bed shear stress for initiation of bed load and s is the angle of repose.

Fig. 6.3: Sketch of bed-load motion at equilibrium.

CVS 549E Page 6

Note:

6.4.1. Bed-load transport rate

The bed-load transport rate per unit width may be defined as:
qs  Cs sVs [6.5]
Where Vs is the average sediment velocity in the bed-load layer

Notes:
 Physically the transport rate is related to the characteristics of the bed-load layer: its mean sediment

concentration Cs, its thickness s which is equivalent to the average saltation height measured
normal to the bed (Figs 6.1 and 6.3) and the average speed Vs of sediment moving along the plane
bed.
 A steady sediment transport in the bed-load layer is sometimes called a (no suspension) sheet-flow.
 Note that the volumetric sediment concentration has a maximum value. For rounded grains, the
maximum sediment concentration is 0.65.

Several researchers proposed formulae to estimate the characteristics of the bed-load layer (Table 6.2).
Figure 6.4 presents a comparison between two formulae. Overall the results are not very consistent. In
practice there is still great uncertainty on the prediction of bed-load transport.

Table 6.2: Bed-load transport rate calculations

CVS 549E Page 7

Fig. 6.4: Dimensionless bed-load transport rate qs/(dsV*) as a function of the dimensionless Shields
parameter

6.5. Examples
Example 6.1
The bed-load transport rate must be estimated for the Mississippi river at a particular cross-section. The
known hydraulic data are flow rate of about 530m 3/s, flow depth of 4.27 m and bed slope being about
0.0011. The channel bed is a sediment mixture with a median grain size of 0.012 m and the channel
width is about 34 m.
Predict the sediment-load rate using the Meyer-Peter correlation, the Einstein formula, and equation
(6.5) using both Nielsen and van Rijn coefficients. Assume kinematic viscosity, v = 1.007x10 -6 m2/s and
  *  c  0.05 ,   998.2 Kg / m3 .
Solution
Assuming a wide channel, the mean shear stress and shear velocity equals, respectively:
 0   gd sin   998.2  9.81 4.27  0.0011
 46.0 Pa
V*  gd sin   9.91 4.27  0.0011
 0.215 ms 1
The Shields parameter equals:
 0   gd sin   998.2  9.81 4.27  0.0011  46.0 Pa
0 46
*    0.237
  s  1 gd s 998.2  1.65  9.81 0.012
Assuming s = 2.65 (quartz particles). And the particle Reynolds number is:
V*d s 0.215  0.012
Re*    2558
 1.007 106
For these flow conditions the Shields diagram predicts sediment motion:
 *  0.237    *  c  0.05
Note that V*/wo is small and the flow conditions are near the initiation of suspension. In first
approximation, the suspended sediment transport will be neglected.

1. Meyer-Peter correlation

CVS 549E Page 8

 3/2
qs  4 0 
   0.188 
 s  1 gd s3    s  1 gd s 
3/ 2 3/ 2
 4 0   4  46 
  0.188     0.188 
   s  1 gd s   998.2  2.65  1  9.81 0.012 
 0.66386
qs
 0.66386
 s  1 gd 3
s

 qs  0.66386  1.65  9.81 0.0123

 0.0035 m 2 / s
2. Einstein function

Work out the answer (given

qs  0.0045 m 2 / s )

3. Bed-load calculation (equation 6.5)

The bed-load transport rate per unit width equals:
qs  Cs sVs
Using Nielsen’s (1992) simplified model, it yields:

With the correlation of van Rijn (1984a), the saltation properties are:

And the sediment transport rate is:

CVS 549E Page 9

Summary
Meyer-Peter Einstein Eqn 6.5(Nielsen) Eqn 6.5 (van Rijn)

Points to Note
Four formulae were applied to predict the sediment transport rate by bed load. Three formulae give
reasonably close results. Let us review the various formulae.
1. Graf (1971) commented that the Meyer-Peter formula ‘should be used carefully at […] high
mass flow rate’, emphasizing that most experiments with large flow rates used by Meyer-Peter
et al. (1934) were performed with light sediment particles (i.e. lignite breeze, s _ 1.25). Graf
stated that one advantage of the Meyer-Peter formula is that ‘it has been tested with large
grains’.
2. The Einstein formula has been established with more varied experimental data than the Meyer-
Peter formula. And the present application is within the range of validity of the data i.e.
qs  s  1 gd s3  0.85 10

Example 6.2
The North Fork Toutle river flows on the north-west slopes of Mount St. Helens (USA), which was
devastated in May 1990 by a volcanic eruption. Since 1980 the river has carried a large volume of
sediments.
Measurements were performed on 28 March 1999 at the Hoffstadt Creek bridge. At that location the
river is 18 m wide. Hydraulic measurements indicated that the flow depth was 0.83 m, the depth-
averaged velocity was 3.06 m/s and the bed slope was sinӨ = 0.0077. The channel bed is a sediment
mixture with a median grain size of 15 mm and d84 = 55mm.
Predict the sediment-load rate using the Meyer-Peter correlation, the Einstein formula and equation
(6.5) using both Nielsen and van Rijn coefficients. Assume kinematic viscosity, v = 1.007x10 -6 m2/s and
  *  c  0.05 ,   998.2 Kg / m3 .

Solution
First calculations
Assuming a wide channel (d = 0.83m << 18 m), the mean shear stress and shear velocity equals,
respectively:

The Shields parameter equals:

0 62.6
*  
  s  1 gd s 998.2 1.65  9.81 0.015
 0.258
Assuming s = 2.65 (quartz particles) and using ds =d50. And the particle Reynolds number is:

CVS 549E Page 10

Note that V*/wo is small (V*/wo ≈ 0.5) and the flow conditions are near the initiation of suspension.
In first approximation, the suspended sediment load will be neglected.

1. Meyer-Peter correlation
3/2
qs  4 0 
   0.188 
 s  1 gd s3    s  1 gd s 
3/ 2 3/ 2
 4 0   4  62.6 
  0.188     0.188 
   s  1 gd s   998.2  2.65  1  9.81 0.015 
 0.777

qs
 0.777
 s  1 gd s3
 qs  0.777  1.65  9.81 0.0153
 0.0057 m 2 / s

2. Einstein function
Work out the answer

3: bed-load calculation (equation (6.5))

The bed-load transport rate per unit width equals:
qs  Cs sVs
Using Nielsen’s (1992) simplified model, it yields:

CVS 549E Page 11

With the correlation of van Rijn (1984a), the saltation properties are:

and the sediment transport rate is:

qs  0.0013 m 2 / s

Exercises
1. A wide stream has a depth of 0.6 m and the bed slope is 0.0008. The bed consists of a mixture of
  2980 kg / m 3
heavy particles ( s ) with a median particle size d50 = 950µm. Compute the bed-
load transport rate using the formulae of Meyer-Peter and the bed-load transport equation
using both Nielsen and van Rijn coefficients for uniform equilibrium flow conditions.
Necessary assumptions can be made.

2. In a section of the river Fujigawa, the bed slope is roughly 0.01. The bed consists of a mixture of
gravels with median particle size d50 = 18 cm. For a flow rate of 24m 2/s, calculate the bedload
transport rate at uniform equilibrium flow conditions using the formula of Nielsen. (Assume
that the equivalent roughness height of the channel bed equals the median grain size.)

3. Considering a 20 m wide channel, the bed slope is 0.00075 and the observed flow depth is 3.2
m. The channel bed is sandy (ds = 50.008 m). Calculate (a) mean velocity, (b) mean boundary
shear stress, (c) shear velocity, (d) Shields parameter and (e) occurrence of bed-load motion. If
bed-load motion occurs, calculate: (f) bed-load layer sediment concentration, (g) bed-load layer
thickness, (h) average sediment velocity in bed-load layer and (i) bed-load transport rate.
(Assume that the equivalent roughness height of the channel bed equals the median grain size.
Use Nielsen simplified model.).

CVS 549E Page 12

CVS 549E Page 13