You are on page 1of 30

THE MOVEMENT

OF S E D I M E N T

IN R I V E R S

CHIH TED YANG


U.S. Army Corps of Engineers, North Central Division
Chicago, IlL, 60605 U.S.A.

Abstract. Sediment movement in rivers is a complex phenomenon. The rate of sediment transport is
related to many variables such as water discharge, average flow velocity, stream power, energy slope,
shear stress, water depth, particle size, water temperature, and strength of turbulence. Different theories
of sediment transport were developed by assuming different independent variables as the dominant
variables. This survey provides a comprehensive review of the important theories of incipient motion and
sediment transport. It discusses basic concepts and findings upon which knowledge of sediment transport
is based and presents mathematical derivations and equations only in sufficient detail to illustrate some
basic concepts. Data collected from natural rivers and laboratory flumes are used to compare the accuracy
and applicabilityof different sediment transport equations. Finally, procedures are suggestedfor selecting
sediment transport equations under different flow and sediment conditions.
1. Introduction
Sediment m o v e m e n t in rivers has been studied by both hydraulic engineers and
geologists for centuries because of its importance in the understanding of river
hydraulics, river morphology, and related matters. Sediment transport is complex
and often subject to semiempirical or empirical treatment. Most theoretical treatment is based on the idealized and simplified assumption that the rate of sediment
transport can be determined by one or two dominant factors, such as water discharge,
average flow velocity, energy slope, shear stress, etc. Numerous equations have been
published. Each equation is supported by limited laboratory data and, occasionally,
by field data. The calculated results from various equations often differ drastically
from each other and from the measured data. Consequently, none of the published
sediment transport equations has gained universal acceptance in confidently predicting sediment transport rate, especially in rivers.
This p a p e r provides a comprehensive review of the important theories of incipient
motion and sediment transport. Emphases are placed on basic concepts and findings
upon which knowledge of sediment transport is based, rather than on the mathematical derivations. The reliability of each equation discussed in this paper is judged from
the validity of assumptions used in its derivation and the comparisons with measured
data from laboratory flumes and rivers. Emphases are placed on the comparisons
between calculated and measured results from rivers. Because of space limitations
some sediment transport equations are not discussed here. This, however, should not
be interpreted as a sign of disagreement or lessening of their importance.

2. Incipient Motion
In the study of sediment transport, the first thing to consider is the flow condition
under which particles of sediment on the bed start to move, i.e., the condition of
incipient motion. Most sediment transport equations require certain parameters at
Geophysical Surveys 3 (1977) 39-68. All Rights Reserved
Copyright (~) 1977 by D. Reidel Publishing Company, Dordrecht-Holland.

40

CHIH TED YANG

incipient motion. Most criteria for incipient motion can be obtained from a balance
of forces acting on a sediment particle as shown in Figure 1, or from a balance of
moments due to these forces. The magnitudes of these forces can be determined from
either a shear stress or a velocity approach.
V

i/

F[
Vd- /

WS
Fig. 1. Diagram of forces acting on a sediment particle in open channel flow.

2.1. SHEAR STRESS APPROACH


The shear stress or tractive force acting at the bottom of an open channel is the pull of
water on the wetted area. In a uniform flow, the tractive force equals the effective
component of the gravity force acting on the body of water, parallel to the channel
bottom, i.e.,
r = yDS,

(1)

where ~, is the specific weight of water, D is the average water depth, and S is the
energy or water surface slope. The two most significant and widely known analyses of
incipient motion from the shear stress approach are those made by White (1940) and
Shields (1936).
White assumed that the influence of slope S and left force F~ in Figure 1 have on
the incipient motion is of little significance and hence can be neglected when they are
compared with other factors. The drag force Fa is proportional to the shear stress,
r and the square of particle diameter d, i.e.
F,~ = C a r d 2.

(2)

THE MOVEMENT OF SEDIMENT IN RIVERS

41

The overturning moment can then be written as


M o = C1C2"rd 3.

(3)

The resisting moment is the product of submerged weight of the particle C3(ysy r ) d 3 and its moment arm C4d, i.e.
M r = C3C4("ys - "yf)d4,

(4)

where ys and yf are the specific weight of sediment and fluid, respectively. A particle
will start to move when the shear stress is such that M0 = Mr. This critical value of
shear stress can be obtained from Eqs. (3) and (4), i.e.,
9 cr = C ( v s - v i ) d ,

(5)

where C is a coefficient. Thus the critical shear stress is proportional to the particle
diameter.
Shields (1936) believed that it was very difficult to express analytically the forces
acting on a sediment particle. He used dimensional analysis to determine some
dimensionless parameters and established his criterion of incipient motion. Shields
assumed that the factors which are important to the determination of incipient
motion are shear stress r, the difference in density between sediment and fluid
(Ps -Pt), the particle diameter d, the kinematic viscosity v, and gravitational acceleration g. These five parameters can be grouped into two dimensionless quantities

d('T/p[)l/2/p = dV:~/1/

(6)

~-/d(p~ - p / ) g = " r / d y 1 ( p J p / - 1)

(7)

and
where Yr is the specific weight of fluid; and U, is the shear velocity. The relationship
between these two parameters was then determined from experiments as shown in
Figure 2. Although the Shields diagram shown in Figure 2 has been accepted by most

Motion
~ ,

0.01

l_l_k_

O. 1

~ Sands in Turbulen~ Boundary Layer

~
1.

10

100

U*d
v
Fig. 2.

S h i e l d ' s D i a g r a m ; r e l a t i o n s h i p b e t w e e n d i m e n s i o n l e s s critical s h e a r
stress a n d s h e a r velocity R e y n o l d s n u m b e r ,

1000

42

CHIH TED YANG

engineers as a criterion for incipient motion, it still has some basic deficiencies. Some
of these deficiencies were pointed out by Yang (1973).
2.2.

VELOCITY APPROACH

It has been established, at least experimentally, that the drag force acting on a
particle is proportioned to the square of the relative velocity between fluid and
particle. Thus, a criterion for incipient motion can be expressed as a function of
average flow velocity or the local velocity where the particle is situated on the bed.
Neill (1967), Bogardi (1968), and Yang (1973), among others, established criteria
for incipient motion from velocity approach. Yang's analysis is used here to explain
some basic concepts of velocity approach.
The assumptions and relationships used by Yang to develop his critical velocity at
incipient motion are: (1) The drag and lift forces acting on a sediment particle are
proportioned to the square of the relative velocity between the particle and fluid; (2)
Flow velocity follows the logarithmic law distribution; (3) The resistance force acting
on a particle equals the product of a friction coefficient and the difference between
submerged weight of the particle and the lift force acting on the particle; and (4)
Incipient motion occurs when drag force equals resistance force. Yang's criteria for
incipient motion are:
VSo) = {2.5/[log (U,d/~,) - 0,66]} + 0.66,

1.2<U,d/u<70

(8)

and

Vcr/w = 2.05, 7 0 -< U, dl~,

(9)

where Vc~ is the critical average flow velocity at incipient motion; and ~o is the
terminal fall velocity of sediment particle. The comparison between Equations (8)
and (9) and the measured data by eight different investigators are shown in Figure 3.
2.3.

E N E R G Y A N D STATISTICAL A P P R O A C H E S

Yalin (1963) presented photographic evidence that sediment particles can leave their
position in a vertical upward direction. The energy required for this vertical
movement is in the form of energy due to turbulent vortices, which is proportional to
the square of shear velocity. At incipient motion

pK3U~cr/(Ps -P)g K4 = ~'S(% - T)K = constant,

(10)

where K is a coefficient; p is the density of water; U , cr is the shear velocity at


incipient motion; and ~'cr is the shear stress at incipient motion. Basically, Equation
(10) is another form of Shields parameter.
The magnitudes and directions of the lift, drag, and resistance forces shown in
Figure 1 are all subject to change in turbulent flow. They vary with respect to time
and location. Gessler (1965, 1970) made an attempt to measure the probability that
grains of a specific size will stay. He found that the probability depends not only on
the statistical variation in turbulence but also on the location and orientation of }he

THE MOVEMENT

28 "' 9 I

OF SEDIMENT

I I 1 ]IIII

43

IN R I V E R S

II I I I ' I I I

I l'I I I I

EXPLANATION

26

0 Casey
~z Grand Laboratory
A Gilbert
o Kramer
9 Thljsse
V Tison
9 Vanonl
9 U.S. Waterways Experiment
Station

24
3

22

20

18

>

16
14
-SMOOTH

TRANSITIOR

>

COMPLETELY ROUGH

12
i0
8

i.
~W~ o

Vcr

]'~O

i~.., J

2.5
"

U.d

'

+ 0.66

`~~

L
1

L L Lll{tl
I0

I I IllllI

I Ill

I00

SHEAR VELOCITY REYNOLDS NUMBER~

ILl.
i000

Res

U~d
~j

Fig. 3. Yang'sDiagram; relationshipbetween dimensionless critical velocity and shearvelocity Reynoldsnumber.


individual grain. Gessler also found a close relationship between this probability and
Shields' critical shear parameter. From a practical point of view, one might say that
for a stable channel of uniform material, when the tractive force reached Shields'
parameter, the probability for any individual grains to stay is 0.5.

44

C H I H TED Y A N G

3. Theories of Sediment Transport


The rate of sediment transport in rivers is related to many variables, such as water
discharge, average flow velocity, water depth, energy slope, shear stress, stream
power, bed configuration, channel pattern, intensity of turbulence, particle size, and
gradation, as well as water temperature. Some of these variables are interrelated and
dependent on each other. It is virtually impossible for us to consider all these
variables simultaneously and to develop an equation which contains all of them
explicitly. In the process of developing a sediment transport equation, it is important
to consider the relative importance of these variables and select one or two as the
dominant factors governing the rate of sediment transport. Different theories and
equations have been developed based on assuming different dominant factors as the
independent variable. Most sediment transport theories and equations were
developed from one of the following approaches.
3.1. REGIME APPROACH
Fundamentally, a river tends to adjust its width, depth, slope, roughness, and channel
pattern to transport the given water discharge and sediment load. A river is 'in
regime' when there is no long term scour or deposition. 'Regime theory' is an
inductive theory of channel self-formation that has grown systematically since about
1890. Regime equations were developed as a result of long term field measurements
of channel dimension, water discharges, sediment loads, and hydraulic parameters.
Most regime equations express channel dimension as well as sediment load as power
functions of water discharge. Blench (1969) made a systematic study of the regime
theory and equations. Because all the regime equations were derived from measured
field data, their applications are limited to those rivers with similar conditions from
where the regime equations were derived.
3.2. S H E A R S T R E S S A P P R O A C H
DuBoys (1879), among others, believes that the rate of bed-load transportation is
dominated by excess shear stress. Bed load is defined as the part of sediment load that
moves by rolling or sliding along the bed. DuBoys assumed that the sediment moves
in layers. These layers move because of tractive force (shear stress) applied to them.
The DuBoys type equation can be expressed by the general form

qs = a l ( ' r - ' r c r ) B1,

(11)

where qs is sediment load per unit width; ~- is shear stress; ~rcris critical shear stress;
and A1, B1 are coefficients.
The original DuBoys equation can be expressed by
qs = K l ~ ( r - ~rcr),

(12)

where K~ is a function of sediment size. According to Straub (1954) experiments

Kt = 0.39i/d 3/4,
where d is the particle size.

(13)

THE MOVEMENT OF SEDIMENT IN RIVERS

45

In his process of developing a criterion for incipient motion, Shields measured


various values of shear stress and bed load at least twice as large as the critical value
and then extrapolated to the point of zero bed load. In so doing, a sediment transport
equation was established. Shields' equation can be expressed by

(q,/yq)(y, - y/Sy) = 10(z - "rcr)]](')/s "y)d,


-

(14)

where q is water discharge per unit width; Ys and y are specific weight of sediment
and water, respectively, and S is energy slope.
Laursen (1958) developed an equation which can be used to calculate either total
load, with wash load excluded, or bed load. Wash load is defined as the part of
sediment load that consists of particles finer than those found in the stream bed.
Suspended load is the part of sediment load transported mainly in suspension. Thus,
total load is the sum of bed load alad suspended load. Laursen's equation can be
expressed in a dimensionally homogeneous form

Ct

E{[pi(d/D)V/6/('l'/7"c,
i

- -

1)]f(U,/o))},

(15)

where Ct is total sediment concentration; Pi is the fraction of a given grain size d; and

f ( U , / w ) is a function of (U,/w). Two differentf(U,/o2) were developed by Laursen


for the calculation of total load and bed load, respectively.
Engelund and Hansen's (1967) equation is a more recent one obtained from a
shear stress approach. They divided the total shear stress into two parts, one caused
by skin friction and the one caused by bed forms. Their equation is

qs = l t.6U, Caa[Ii In (30D/k )+ I2]

(16)

where U, is the shear velocity; Ca is the sediment concentration at reference level a ;


a is often selected as the distance equal to the surface roughness which is equivalent
to the average bed material size; k is the grain roughness; and Ii and I 2 c a n be read
from the work charts published by Einstein (1950).

3.3.

VELOCITY APPROACH

Donat (1929) and Colby (1964) expressed the rate of sediment discharge as functions
of average flow velocity. The basic form of this type of equation can be expressed by:

qs = A 2 ( V - Vc,)~2,

(17)

where A2 and B2 are coefficients; and V and Vc, are average velocity and critical
velocity, respectively.
Colby (1964) investigated the effect of flow velocity, water depth, particle size,
water temperature, and concentration of fine material on the bed-material discharge. Bed-material discharge is defined as the part of sediment load that consists of
particles large enough to be found in appreciable quantities in the stream bed. Colby
developed four graphical relations from measured data to express the bed-material

46

CHIH TED YANG

load mainly as a function of average flow velocity with correction factors to reflect the
influences of particle size, water depth, water temperature, and concentration of fine
material on bed-material load. In spite of many inaccuracies and uncertainties in the
data used by Colby, some agreements between calculated and observed results can
be found in the literature.
3.4. DISCHARGE APPROACH
Based on laboratory experiments, Schoklitsch (1934) suggested an equation which
has the basic form
qs = A3(O - O , ) ~3,

(18)

where A3, B3 are coefficients; and O, O~ are water discharge and critical discharge,
respectively. Most sediment rating curves obtained at gaging stations have the basic
form of Equation (18). The Schoklitsch equation can be written as
qs = 7 0 0 0 ( $ 3 / 2 / d 1/2)(q _ q~,),

(19)

where q and q , are water discharge and critical discharge per unit width, respectively. Another well-known sediment formula in this category is the Meyer-Peter et
al. (1934) equation which can be expressed by
0.4 q2/3/d = qZ/3 s/ d - 17.

(20)

The Meyer-Peter et al. equation can be applied only to coarse material with particle
size great than 5 ram. For mixture of non-uniform material, d should be replaced by
d35, that is, 35% of the mixture is finer than d35.
3.5. SLOPE APPROACH
After 14 years of research and analysis, Meyer-Peter and Miiller (1948) transformed
the Meyer-Peter equation to an energy or slope type equation. This type of equation
has the basic form of

qs =m4(s-gcr) B4,

(21)

w h e r e A4, B4 are coefficients; and S, Scr are energy slope and critical slope,

respectively.
Meyer-Peter and Miiller believed that only a portion of the total energy loss, the
one due to grain resistance St, is responsible for the bedload motion. Their equation
can be written as
yRS~ = O.047(ys-y)d + 0 . 2 5 p

1/32/3
q~ ,

(22)

where R is the hydraulic radius; y and Ys are specific weight of water and sediment,
respectively; and p is the density of water. The Meyer-Peter and M/iller formula
enjoys great popularity in central Europe.

THE MOVEMENT OF SEDIMENT IN RIVERS


3.6.

47

STREAM POWER OR UNIT STREAM POWER APPROACH

Bagnold (1966), Ackers and White (1973), Maddock (1973, 1976), and Yang (1972,
1973) believed that there should be a balance between the rate of energy expenditure
and the rate of sediment transport. Bagnold, and Ackers and White defined the
stream power as the product of shear stress per unit bed area and average velocity.
Yang defined the unit stream power as the rate of potential energy expenditure per
unit weight of water. Bagnold's equation for total load (wash load excluded) can be
written as

qs = "rV[eb/tan a + 0.01( V/o))],

(23)

where r V is the stream power, eb is the bedload efficiency factor, and tan a is the
friction coefficient. The values of eb and tan a can be determined from charts
developed b~ Bagnold.
The basic approach of Ackers and White is very similar to that of the Bagnold's.
Ackers and White emphasized the use of dimensional analysis and expressed the rate
of sediment transport as a function of several dimensionless parameters. A total of 25
equations are required to determine these parameters and the rate of sediment
transport. These equations are not presented here. The application of Ackers and
White's equation is limited to subcritical flow with Froude number less than 0.8.
Yang (1973) used the concept of unit stream power and dimensional analysis to
develop his dimensionless unit stream power equation.
log Ct = 5.435- 0.286 log w d / u - 0.457 log U . / w
+(1.799 -0.409 log wd/u -0.314 log U , / w )
9log (VS/to - VcrS/w)

(24)

where C, is the total sediment concentration in parts per million by weight, o) is the
fall velocity of sediment with particle diameter d, u is the kinematic viscosity of
water, U. is the shear velocity, the velocity slope product VS is the unit stream
power, and VcrS is the critical unit stream power required at incipient motion. The
value of VcrS/o can be obtained from the product of slope S and Vc,/w calculated
from Equations (8) and (9). Yang made one of the most elaborate verifications of the
validity of his equation. He used 1093 sets of laboratory data (Yang, 1973) and 156
sets of field data (Yang and Stall, 1976) to test the accuracy of Equation (24).
Maddock's (1973, 1976) basic approach is similar to that of Yang's (1972). Both
Yang and Maddock emphasized the importance of the velocity-slope product, i.e.,
unit stream power9 Maddock's equation combines the concept of unit stream power
and regime theory. A graphical solution is used for Maddock's equation.
3.7. STOCHASTIC APPROACH

There are two basic differences between this approach and previously mentioned
approaches; (1) The critical value at incipient motion is not considered in this
approach because it fluctuates and is difficult to determine. (2) It is suggested that the

48

CHIH TED YANG

bedload transport is related to the fluctuation of the independent variable, say


velocity, rather than the average value.
Einstein (1950) made one of the most significant contributions to the study of
sediment transport from a stochastic approach. He expressed the beginning and the
end of the particle motion with the concept of probability, which relates the
instantaneous hydrodynamic lift force to the particle's weight. He considered most of
the variables which might influence the rate of sediment transport in his bedload
function and computation procedure. Einstein's procedure is extremely complex and
difficult for engineers to use in the field. However, because of the theoretical
approach used and because calculated results often agree with the measurements
within reasonable accuracy, his procedure has been widely used in the United States
despite its complexity. His calculation procedure was modified or simplified by Colby
and Hembree (1955), Colby and Hubbell (1961), Colby (1964), Bishop et aI. (1965),
and Toffaleti (1969) for engineering purposes.
3.8. REGRESSION APPROACH
Shen and Hung (1972) assumed that sediment transport is such a complex phenomenon that no single Reynolds number, Froude number, or combination of them can be
found to describe sediment motion under all conditions. Instead of trying to find a
dominant variable which dominates the rate of sediment transport, they recommended a regression equation based on 587 sets of laboratory data. They expressed
the total sediment concentration by the regression equation
log C, = - 107 404.459 381 64+324 214.747 340 85 X
- 3 2 6 309.589 087 39 X2+ 109 503.872 325 39 X 3

(25)

where
X = [ VS~

0.32]0.007 501 89

(26)

Favorable comparisons between computed results and measurements, especially


from laboratory flumes and small rivers, were reported by Shen and Hung.
4. Validities of Different Approaches
Most of the equations discussed here were derived under the assumption that the rate
of sediment transport can be determined from one dominant variable. These
equations were supported by limited data collected under carefully designed laboratory conditions. When such an equation is applied to other flow conditions, the
agreement is often very poor. Laboratory data collected by Guy et al. (1966) from a
laboratory flume with 0.93 mm sand are used here as an example to examine the
validity of these assumptions.
Figure 4 shows the relationship between total sediment discharge and water
discharge. For any given value of O two different values of qt can be obtained. Field

THE

MOVEMENT

I0

OF

SEDIMENT

IN RIVERS

I I I Illl

I I I Il~

49

w
J,l
torY
UA

1.0
)

r~
z
0

n~
,,i

0. i
mF
0
_J

i 84

4J
cr
0.0[

13s
<
i
o

E X p LAN.AT I ON

co
7,

F-"

LIA 0 . 0 0 1

co

Plane bed

Dune

Transition

&

Standing wave

-A
<
O
1"-

0.1

WATER DISCHARGE,
Fig. 4.

I I I fill

I I If Ill

0. 0001

iO

Q, IN CUBIC METERS PER SECOND

Relationship between total sediment discharge and water discharge


for 0.93 mm sand.

data obtained by Leopold and Maddock (1953) also indicate similar results. Some of
Gilbert's (1914) data indicate that there is no correlation at all between water
discharge and sediment discharge. Apparently, different sediment discharges can be
transported by the same water discharge, and a given sediment discharge can be
transported by different water discharges. The same sets of data shown in Figure 4

50

CHIH

TED

YANG

are plotted on Figure 5 to show the relationship between total sediment discharge
and average velocity. Although qt increases steadily with increasing V, it is apparent
that for approximately the same value of V the value of qt can differ considerably due
to the steepness of the curve. Some of Gilbert's data also indicate that the correlations between q~ and V are very weak. Figure 6 indicates that there is no single
relationship between total sediment discharge and slope. Different amounts of total

1 l llll

I 1 I1111

1.0
cIc
b'--

rF
(ZL
Z
O
(J
V')

0, I

(-w
L~
Q-

F
O
....2

C,

Z
0.01

i o

Ll.I
(..9
QC
-'r"
O

CO
I-Z

0.001

--O

I
EXPLANATION

,H

Plane bed

Dune

Transition
Standing wave

0.0001

I I Illl

1.0

I I t Ill

I0

AVERAGE VELOCITY~ V, IN METERS PER SECOND


Fig.

5.

Relationship between total s e d i m e n t discharge and average water


velocity for 0.93 m m sand.

THE

MOVEMENT

PY
W

OF SEDIMENT

I I II111

51

IN R I V E R S

I I IIIII

W
>or"

1.0

EXPLANATION

r-~
Z
0
c_)

LIJ
0s
IJJ
O_

0
_J

Plane bed

9
[]

Dune

Transition

Sta

/~
/
f
I

0.1

Cr

<
0.01

7
UJ

UJ

<
p0
b-

0.001

I I Iltll

O. 0001

WATER
Fig. 6.

I I I11tl

0.001

SURFACE

SLOPE,

0.01

S,

IN METERS

PER METER

Relationship between total sediment discharge and water surface


slope for 0.93 mm sand.

sediment discharges can be obtained at the same slope, and different slope can also
produce the same sediment discharge. Figure 7 shows that there is a one to one
correlation between total sediment discharge and shear stress when total sediment
discharge is in the middle range of the curve. For either higher or lower sediment
discharge, the curve b e c o m e s vertical, which means for the same shear stress
numerous values of sediment discharge can be obtained.
It is apparent from Figures~4, 5, 6 and 7 that more than one value of total sediment
discharge can be obtained for the same value of water discharge, velocity, slope, or
shear stress. The validity of the assumption that total sediment discharge of a given

52

CHIH

10

T-q~l

TED

YANG

lqll

I F-

I.kl
i11
~E
1.0
LU

0
1:3
Z
(J
LU
Q3
m.."
LU
r.,
(I)
:Z:

0 [3

0 , 1 ~

0
.d
w

-s

0.01
tO
n,EXPLANAT 1ON

"xL)
03

Plane bed

Dune

F--

Transition

IaA
~2

0.0Ol

G'I
.d

Standing

w a v e

b-O
b-

o.oootJ

I
0.t

SHEAR
Fig. 7.

I I ,

1.0

S T R E S S ~ r,

Relationship

I I IlillL
IN K I L O G R A M S

PER

SQUARE

METER

between total sediment discharge and shear stress for


0.93 mm sand.

particle size can be determined from water discharge, velocity, slope, or shear stress
is now open to question. Because of the basic weakness of these assumptions, the
generality of an equation which was derived from one of these assumptions is also
questionable. When the same sets of data are plotted on Figure 8, with unit stream

THE

MOVEMENT

20,000

OF

SEDIMENT

IN RIVERS

I I IIIII

53

I I l lIll]

i0,000

5
J
d

E X P L A N A T I ON

HI
t~

P]ane bed

Dune

Transition

Standing

wave

OZ

tO00

(J
]

13

I-[I
u

I-z
c~
ej
I'-,Z

IO0

,,g
_.1
}-

10

I
UNIT

8.

I I l llll

O.O01

STREAM
PER

Fig.

I I I lil

0.0001
POWERj

VS,

KILOGRAM

0.01

IN M E T E R - K I L O G R A M S
PER SECOND

Relationship between total sediment concentration and unit stream


power for 0.93 mm sand.

power as the independent variable, the correlation between total sediment concentration and unit stream power is significantly improved. This close correlation exists
in spite of the presence of different bed forms such as plane bed, dune, transition, and
standing wave. However, data shown in Figure 8 also indicate some deviation from
the straight line when the total sediment concentration is below 30 ppm. This is due
to the fact that at low sediment discharge, or concentration, total sediment concentration depends not only on the values of unit stream power and critical unit stream
power but also on water depth. This phenomenon was explained in detail by Yang
(1976). The close relationship between total sediment concentration and unit stream
power exists not only in straight channels but also in those channels which are in the
process of changing their patterns from straight to meandering, and to braided
channels as shown in Figure 9. These data were collected by Schumm and Khan

(1972).

54

CHIH

TED

YANG

Z
o

10,000

I Ill

I I Ill

SCHUMM AND KHAN'S DATA


m

9 STRAIGHT CHANNEL
9 MEANDERING TIIALWEG
A BRAIDED CHANNEL

Z
m

J
Z
O

J
I000
z

_-

,J

u
z
o

7-,

O
I00~

I I Ill

I I I Ill
0.01

0.001

UNIT

STREAM

POWER,

VS,

KILOGRAM

IN M E T E R - K I L O G R A M S

PER

PER

SECOND

Fig. 9. Relationship between total sediment concentration and unit stream


power during process of channel pattern development from straight to
meandering, and to braided.

The stream power approach emphasizes the rate of energy expenditure along the
bed. Efficiency coefficients and factors were used by Bagnold (1966) and Ackers and
White (1973) to determine the amount of power used in transporting bedload and
suspended load. The calibration of these coefficients and factors were based on
limited laboratory data. The pronounced scattering along some of their best fit
calibration curves indicates that it is extremely difficult to determine these coefficients and factors accurately.
The assumptions used in stochastic approach are more difficult to verify. However,
even in a stochastic approach, certain deterministic assumptions are generally made.
For instance, in his stochastic approach, Einstein (1950) assumed that when a
sediment particle makes a step to move forward, the step length equals 100 times its
particle diameter. Experiments made by Yang and Sayre (1971) indicate that step
length of a given particle size is not a constant; it follows a gamma distribution.

T H E M O V E M E N T O F S E D I M E N T IN R I V E R S

55

A major drawback of the regression approach is the lack of physical meaning of


those parameters in the regression equation. The choice of variables, the form of
equation, and the determination of coefficients depend entirely on the data used in
the regression analysis.
In summary, none of the assumptions used in deriving existing sediment transport
equations are perfect. The non-generality of these assumptions is the basic reason
why it is extremely difficult to recommend a universal equation for sediment
transport. A detailed comparison of these assumptions (Yang, 1972) by using more
than 1000 sets of laboratory and field data suggests that total sediment discharge or
concentration depends on unit stream power more than on any other variables.

5. Sediment Discharge in Rivers


Measuring sediment discharge from natural rivers is an expensive and time consuming process. Suspended load can be measured fairly accurately by a depth integrated
sampler. However, there is no reliable bedload sampler which can be used in rivers.
Total sediment load in a river has to be measured from a contracted section or from a
man-made turbulent section where all the sediment particles are kept in suspension.
In order to compare the calculated results from a sediment transport equation and
the measurements from a river, sediment and hydraulic data have to be measured
simultaneously, or within a short time span, to insure that these data are collected
under approximately the same flow condition. All the above mentioned factors
significantly reduce the number and source of field data available to compare the
accuracy and reliability of different sediment transport equations. Data used in this
article are those with total loads and their corresponding hydraulic data measured
within a 24-hour period. Wash load is excluded from total load when the former is
significant, and total load is replaced by bed material load. Field data used herein are
considered by most engineers as reliable, and quite often used by them to compare
the accuracy of different sediment transport equations.
The ASCE Task Committee on Sedimentation (1971) made an analysis of the data
collected from Niobrara River near Cody, Nebraska (Colby and Hembree, 1955).
The calculated results from 13 different equations by the ASCE Task Committee are
shown in Figure 10. Figure 10 also shows the calculated results from the unit stream
power equation, i.e. Equation (24). Among these 14 equations, the unit stream
power equation has the highest accuracy rating. Colby's, Laursen's and Toffaleti's
equation and Einstein's bedload function can all provide reasonable estimation of
the total sediment discharge in the Niobrara River.
Einstein (1944) measured the total sediment discharge from Mountain Creek at
Greenville, South Carolina. Figure 11 shows the comparisons between measurements and computed results from 7 different equations by Vanoni et al. (1960), and
the computed results from the unit stream power equation by the author. Figure 11
shows that the unit stream power equation is the only one that can provide a close
estimation of the total sediment discharge in Mountain Creek. The Schoklitsch
equation ranks second in accuracy.

56

CHIH

TED

I IIIIII

YANG

I
i00
Z
C)
CJ
,,i
GO

uJ
(~_
GO
,=C
OC
0
..J
v
Z

I0

eY

I-LtJ

r,
_J
1-0
1--

EXPLANATION

1.0

9 Measured
o Unit S t ~

*I* I I I l l l l
1.0

_ _
Power Equation

I I I I llll

I0

I00

WATER DISCHARGE, Q, IN CUBIC METERS


PER SECOND

Fig. 10. Comparison between measured total sediment discharge of the


Niobrara River near Cody, Nebraska and computed results of various equations.

The total sediment discharge in the Middle Loup River at Dunning, Nebraska, was
measured by Hubbell and Matejka (1959). Their computed total sediment discharge
from 5 different equations as well as the computed results from unit stream power
equation by the author are shown in Figure 12. Among the 6 equations used in Figure
12, only the unit stream power equation and the modified Einstein method can
provide good estimates of the total sediment discharge in the Middle Loup River (as
shown between the two dashed lines). Other equations cannot even indicate the
general trend of variation of total sediment discharge at different water discharges.

THE MOVEMENT OF SEDIMENT IN RIVERS

l Ill

57

/i I~'I

E
9
i.O

__

Measured
Stream

Unit
o

Power

Equation

r~
z
0
(]3

rw

r~
~D
0
-A

O.1l-uJ
<.9
r~

~o~ I
" 6'

-r
L)

7,

:. ~ , . " .

z
IJJ
8 "
8.'

..J
I--(2:1
FO"01 t-------~

0.1

WATER DISCHARGE,

I l]lll

1.0

Q, IN CUBIC METERS PER SECOND

Fig. 11. Comparisonbetween measured total sediment discharge of the


Mountain Creek at Greenville, South Carolina and computed results of
various equations.

The bed material discharge in the Mississippi River at St. Louis, Missouri, was
measured by Jordan (1965). Figure 13 shows the comparisons made by Jordan
between measured and computed results from four different equations. The computed total sediment discharges from the unit stream power equation are also shown
in Figure 13. It is considered that the measured bed material discharge can best

58

CHIH

z
0

<

w
(i.)

TED

YANG

[] I

rY
w

Q-

(D

W
z

~d

~/

oo

7~

I--

tJ
im

t~
03
_1

EXPLANATION

kO
k1.1.1
O-

(D

o
&
o
9
9
9

I
----

U n i t S t r e a m Power
Modified
Einstein
Straub
Kalinske
Schoklitsch
Meyer Peter-Muller

ZL_LLL10
MEASURED
AT

SEDIMENT
IN

100

DISCHARGE,

KILOGRAMS

---

IILLL

TOTAL

SECTION

Equation

PER

qt'

SECOND

Fig. 12. Comparison between measured total sediment discharge of the


Middle Loup River at Dunning, Nebraska and computed results of various
equations.

represent the total sediment discharge in a river with wash load excluded. The wash
load is independent from the hydraulic factors of a river; it depends only on the rate
of supply from the upstream reach and watershed. Figure 13 indicates that the unit
stream power equation is the most accurate one. Calculated results l~rom modified
Einstein and Colby's method are two to three times as large as the measurements.
The bed material discharge in the Rio Grande River near Bernaliilo, New Mexico,
was measured by Nordin (1964). Figure 14 shows that the computed results from the

THE MOVEMENT

OF SEDIMENT

59

IN R I V E R S

Z
0
10,000

I I IIII 11]

i I 1 11r11

i000

/e

I00

,sXS,-I/,/

i
u

ee

//,4

"

to

EXPLANATION

Unit Stream Power


Equation

1
0

MEASURED

I IIIIII
I0
BED

MATERIAL

I I llili

I
I00

DISCHARGE,

I I IIIII

I i ilil

i000
q B M ) IN K I L O G R A M S

i0,000
PER

SECOND

13. Comparison between measured bed material discharge of the


Mississippi River at St. L o u i s , Missouri and computed results of various
equations.

Fig.

unit stream equation and Laursen's equation compare favorably with the measurements. On the average the computed results from the modified Einstein method are
three times as large as the measured bed material discharge.
A total of 156 sets of data collected from six river stations were used to compare
the applicability of different sediment transport equations. These data cover a wide
range of variations of water and sediment discharges, velocity, slope, water temperature, and river size. Based on these data, the unit stream power equation is rated
superior to other equations. The comparisons between computed results from the
unit stream power equation and measurements from six river stations are summarized in Figure 15. The agreements are very good.
White et al. (1975) recently reviewed sediment transport theories. With the
exception of Yang's (1973), and Shen and Hung's (1972) equations, most equations
discussed in this paper were reviewed and compared by them. Their comparison was
based on over 1000 flume experiments and 260 field measurements. Data with

60

CHIH

TED

YANG

z
0

i0,000

I IIIItl

I I IIIII

I l illtt

po

I000

F_

k.
Z~ 9

z~

.B

J 9

B Q 9149

'~, o~

Lx
i00

17

9.o7 /
~'~

EXPLANATION
o

: Section ~2} Unit St.....


Section
Power Equation
Section FA2} Modified
Section
Einstein

/
/

d
0
I--

~
0
U

-/
1

/I

9 Section
Sectio~?}
I IIII1

I I IIIII

10

Laursen

I I IIIII

100

MEASURED BED MATERIAL DISCHARGE, qBM'

--

1000

I I [III
10,000

IN KILOGRAMS PER SECOND

Fig. 14. Comparisonbetweenmeasuredbed materialdischargeof the Rio


Grande River near Bernalillo,New Mexicoand computedresultsof various
equations.
Froude numbers greater than 0.8 were excluded by them. They used two dimensionless parameters for comparison purpose, the dimensionless particle size Dgr and the
discrepancy ratio. The discrepancy ratio is defined as the ratio between calculated
and measured sediment loads. Dgr is defined as
Dgr = [ g ( p / p - 1)/u211/3d,

(27)

where g is the gravitational acceleration, Ps and P are the density of sediment and
water, respectively, u is the kinematic viscosity of water; and d is the particle
diameter. Comparisons made by White et al. (1975) indicate that the Ackers and
White's (1973) equation is the most accurate one, followed by the Engelund and
Hansen (1967), Rottner (1959), Einstein (1950), Bishop et al. (1965), Toffaleti
(1969), Bagnold (1966), and Meyer-Peter and Miiller's (1948) equations. A comprehensive test of the accuracy of Yang's (1973) equation is shown in Table I. A total
of 1093 sets of flume data and 154 sets of river data were available to the author. All
these data have particle size in the sand range with flow in both subcritical and

THE

10,000

I IIIIIII

MOVEMENT

i l llllll

OF

SEDIMENT

I l llllll

61

IN RIVERS

I IIIIIII

c~
z
o
(o

1000

cY
klJ
I1
~-

<

100

-_

-..I

1.0

-5

I-'-

Z
UA

u-I

?+!<:

lO-

tao
(._9
rY

2-

~<~b~,~,
~

"

v
,
9
9
9

O.

o~O

~/%~176
~

o')

o.o

0.001'

0.001

EXPLANATION
Niobrara River
Middle Loup River
Mountaih Creek
Rio Grande Sec A2
Rio Grande Sec F
Mississippi River

=_

.,/!
IIIII

0.01

t I illlll

I I Illtll

}11111

O. i

COMPUTED SEDIMENT DISCHARGE

I I llllll
i0

I I IIIlll
I00

IN KILOGRAMS

I000

1 ] I ttlH
10,000

PER SECOND

Fig. 15+ Comparison between measured sediment discharge at six river


stations and computed total sediment discharge of Yang's equation.

super-critical conditions. Most of these data were also used by White et al. (1975) in
their comparison. Results in Table I are summarized in Table II and plotted on
Figure 16 to show the average deviation of calculated results by Yang's equation
from the measurements for both laboratory and river data. For the 1247 sets of data
used, the mean discrepancy ratio is 1.03. On the average, computed results from
Yang's equation are only 3% higher than the measurements. Table II also indicates
that 91% of the observations have a discrepancy ratio between 0.5 and 2.0, and 94%
between 0.25 and 1.75. The average discrepancies of Shen and Hung's equation are
also shown in Table II. Shen and Hung's equation and Yang's equation can both
provide accurate estimation of the total sediment concentration in laboratory flumes.
However, Yang's equation is much more accurate when it is applied to natural rivers.
A detailed analysis of the accuracy of Shen and Hung's equation indicates that their

62

CH1H TED YANG


TABLE I
Comparisons between measured and computed results from Yang's Equation
Discrepancy ratio

Particle
size (ram)
(1)

Channel
width (m)
(2)

Dimensionless
grain size
Max
(3)
(4)

0.305
0.305
0,375
0,375
0,375
0.375
0.506
0.506
0.506
0.506
0.506
0.786
0.786
0.786
1.71
1.71

0.40
0.60
0.20
0.30
0.40
0.60
0.13
0.20
0.30
0.40
0.60
0.20
0.30
0.40
0.20
0.30

7.63
7.63
9,38
9.38
9.38
9.38
12.66
12.66
12.66
12.66
12.66
19.66
19.66
19.66
42.78
42.78

0.152

0.27

4.16

0.137

0.85

3.62

0.233
0,549
0.233

0.27
0,27
0,85

6.44
14.97
6.32

2.21
1.53
3.53

0.40

1.22

10.60

1.22

Mean Min
(5)
(6)

0.751.25
(7)

0.51.50
(8)

0.251.75
(9)

0.52.0
(10

No. of
data
(11)

100%
100
88
100
82
89
93
94
98
98
100
81
81
85
100
64

100%
100
98
100
94
96
100
100
100
98
100
100
100
100
100
96

100%
100
100
100
98
96
93
98
100
100
100
81
81
85
100
71

21
33
50
42
51
44
15
63
61
46
49
36
53
26
12
28

83%

100%

92%

12

57%

71%

72%

14

86%
93
85

86%
100
92

93%
100
92

14
14
13

100%

100%

100%

42

86%
89
85
74
88
88
69
50

86%
89
88
97
94
94
86
75

86%
89
91
80
98
97
93
83

29
18
33
34
50
32
29
22

(a) Gilbert'sdata(1914)
1.40
1.49
1.76
1.48
1.91
1.96
0.98
1.68
1.62
1.81
1.14
1.24
0.99
1.10
1.45
1.81

1,01
1.02
1.22
1.07
1.19
1.08
0.76
0.98
0.85
0.81
0.81
0.69
0.64
0.68
0.87
0.77

0.63
0,61
0.64
0.60
0.43
0.49
0.45
0.43
0.55
0.52
0.52
0.38
0.41
0.41
0.51
0.26

8t%
79
56
67
61
59
60
68
67
59
63
36
26
35
33
43

(b) Nomicos' data (1956)


1.52

0.74

0.42

33%

(c) Vanoni and Brooks' data (I957)


6.13

1 . 6 6 0.50

43%

(d) Kennedy's data (196 l)


1 . 2 5 0.68
1.02 0.74
1.29 0.73

57%
86
62

(e) Stein's data (1965)

(f)
0.19
0.27
0.28
0.45
0.47
0.93
0.32
0.33
(Uniform)
0.33
(Graded)
0.54

0.88

0.67

86%

Guyetal.'sdata(1966)

2.44
2.44
2.44
2.44
2.44
2.44
0.61
0.61

4.59
6.22
6.36
11.23
12.27
29.39
8.03
8.26

3.25 1.10
3.99 1.29
4.36 1.19
1.68 0.89
1.96 1.02
2.00 1.14
2.56 1,32
2.09 1.45

0.51
0.66
0.61
0.21
0.41
0.65
1/.61
1.05

52%
61
70
41
50
56
31
42

0.61

8.55

1.91 0.90

0.39

43

64

93

79

14

0.61

15.19

2.24

0.44

51

63

91

91

35

1.22

THE MOVEMENT OF SEDIMENT 1N RIVERS

63

Table I (continued)
Discrepancy ratio
Particle
size (ram)
(1)

Channel
width (m)
(2)

Dimensionless
grain size
Max
(3)
(4)

Mean Min
(5)
(6)

0,75 ~
1,25
(7)

0.51.50
(8)

0.251.75
(9)

0.52.0
(10

No. of
data
(11)

(g) Williams' data (1967)


1.35

0.30

33.54

0.25

2.44

6.40

2.38

1 . 1 5 0.19

30%

49%

84%

70%

37

81%

94%

84%

31

95%

97%

61

100%

25

80%

87%

15

100%

100%

15

87%

100%

15

96%

65%

23

(h) Schneider's data (1971)


2.41

0.87

0.25

45%

(i) Einstein's data from Mountain Creek (1944)


1.00

3.38-5.18 26.35

2.05

1.25

0.34

44%

77%

(j) Colby and Hembree's data from Niobrara River (1955)


0.283

21.0-21.9

0.16-0.24

37.5-46.6

6.42

1.51

0.94

0.56

80%

96%

100%

(k) Hubbell and Matejka'sdata from Middle Loup River (1959)


4.50

2.19

1.16

0,58

47%

80%

(1) Nordin's data from Rio Grande River Sec A2 (1964)


0.23-0.39

81.1-83.2

0.22-0.45

106.7196.6

7.43

1.65

1.09

0.50

60%

87%

(m) Nordin's data from Rio Grande River Sec F (1964)


7.61

1.78

1 . 2 5 0.52

40%

80%

(n) Jordan's data from Mississippi River (1965)


0.21-0.78

464.5532.2

8.71

2.33

0.81

0.31

44%

61%

TABLE I1
Summary of average discrepancies of Yang's Equation and Shen and Hung's Equation
Discrepancy ratio

Max,
(1)

Mean
(2)

Min.
(3)

0.751.25
(4)

0.51.5
(5)

0.251.75
(6)

0.52.0
(7)

No. of
Data
(8)

Sand in laboratory Yang


flumes
Shen and
Hung

2.05
1.79

1.02
0.93

0.57
0.44

54%
45

84%
81

94%
95

91%
86

1093

Sand in rivers

Yang
Shen and
Hung

1.92
2.17

1.08
1.30

0.47
0.61

53%
38

80%
60

93%
72

92%
77

154

All data

Yang
Shen and
Hung

2.03
1.85

1.03
0.98

0.56
0.46

54%
44

83%
78

94%
91

91%
85

1247

64

CHIH TED Y A N ~
34

--I

-7

32
3O

28

FLUME DATA
------ RIVER DATA
NO. OF F L U M E D A T A = 1093

26

NO. O F R I V E R D A T A =

154

24
22
~

20
18

g
-

10
8
6
4
2

-_

3
\l

o
0.5

5a. A,
I

1.0

1.5

DISCREPANCY

Fig. 16.

__

.
2,0

2.5

3.0

RATIO

D i s t r i b u t i o n of d i s c r e p a n c y ratio of Y a n g ' s e q u a t i o n .

equation can provide good estimation of the bed-material discharge in very small
rivers. It provides poor estimations of the bed-material discharge in large rivers, such
as the Rio Grande River and the Mississippi River. Since both the amount and source
of data used in Table II of this paper and Table 5 of White et al. (1975) are
comparable, these two tables are combined together in Table III to show the
accuracies of different equations. It is apparent that Yang's unit stream power
equation is the most accurate one.

T H E M O V E M E N T O F S E D I M E N T IN R I V E R S

65

TABLE III
Summaryof accuraciesof different equations
Equations

Date

Data with discrepancy


ratio between i and 2

Yang
Shen and Hunga
Ackers and White
Engelund and Hansen
Rottner
Einstein
Bishop et al.
Toffaleti
Bagnold
Meyer-Peter and Miiller

1973
1972
1973
1967
1959
1950
1965
1969
1966
1948

91%
85
68
63
56
46
39
37
22
10

aShould not be applied to large rivers.

6. Selection of Equations
Although there is no perfect assumption which can be used to derive a sediment
transport equation, there are differences in the generalities of these assumptions.
Based on the majority of published data, it appears that the rate of sediment
transport or total sediment concentration is dominated by unit stream power more
than any other variable. Even if perfect assumptions could be found and used in the
derivation of an equation, the coefficients in the equation have to be determined by
comparing the mathematical model with measured data. Thus, the applicability of an
equation depends not only on the assumptions and theories used in its derivation, but
also on the range of data used in the determination of the coefficients in the equation.
Sediment discharge in natural rivers depends not only on those independent
variables mentioned in previous sections, but also on the gradation and shape factor
of sediment, percentage of bed surface covered by coarse material, variation of
hydrological circle, rate of supply of fine material or wash load, water temperature,
channel pattern and bed configuration, strength of turbulence, etc. Because of the
tremendous uncertainties involved in estimating sediment discharge at different flow
and sediment conditions under different hydrologic, geologic, and climatic constraints, it is extremely difficult, if not impossible, to r e c o m m e n d one equation for
engineers and geologists to use in the field under all circumstances. The following
procedures are r e c o m m e n d e d based on the author's experience and his understanding of the assumptions and limits of data used in obtaining different sediment
transport equations.
1. Determine the kind of field data available or measurable within the time,
money, and m a n - p o w e r limits.

66

CHIH TED YANG

2. Examine all the formulas and select those ones with measured values of
independent variables determined from step 1.
3. Compare the field situation and the limitations of formulas selected in step 2. If
more than one formula can be used, calculate the rate of sediment transport by these
formulas, and compare the results.
4. Decide which formulas can best agree with the measured sediment load and use
these formulas to estimate the rate of sediment transport at those flow conditions
when actual measurements are not possible.
5. In the absence of measured sediment load for comparison, the following
formulas are recommended for consideration:
a. Use Meyer-Peter's (1934) formula when the bed material is coarser than 5 mm.
b. Use Einstein's (1950) procedure when bedload is a significant portion of the
total load.
c. Use Toffaleti's (1960) formula for large sand-bed rivers.
d. Use Colby's (1964) formula for rivers with depth less than 10 feet.
e. Use Shen and Hung's (1971) formula for laboratory flumes and very small
rivers.
f. Use Yang's (1973) formula for sand bed laboratory flumes and natural rivers
with total sediment concentration (wash load excluded) greater than 20 ppm by
weight.
g. Use Ackers and White's (1973), or Engelund and Hansen's (1967) equation for
subcritical flow condition in the lower flow regime.
h. A regime equation can be applied to a river only if the flow and sediment
conditions are similar to that of the river from where the regime equation was
derived.
i. Select an equation according to its degree of accuracy shown in Table III.
6. I n c a s e none of the existing sediment transport equation can give satisfactory
results, use the existing data collected from a river station and plot sediment load or
concentration against water discharge, velocity, slope, depth, shear stress, stream
power, and unit stream power. The least scattered curve without systematic deviation from a one-to-one correlation between dependent and independent variables
should be selected as the sediment rating curve for the particular station.

7. Summary and Conclusions


Basic concepts and approaches used in the study of incipient motion and sediment
transport were reviewed and summarized. Although some of the assumptions used in
the derivation of sediment transport equations are more realistic and general than
others, none of these assumptions is perfect in the sense that there should always be a
one-to-one correlation between dependent and independent variables. Due to this
basic weakness and the complex nature of natural rivers, our understanding of
sediment movement in rivers is still very primitive compared to our understanding of
open channel flow with rigid boundary. The validity and applicability of an equation

TIlE MOVEMENT OF SEDIMENT IN RIVERS

67

is judged from the generality of assumptions used in the derivation and the
agreements between the calculated and measured results. This study has reached the
following conclusions:
1. In spite of some basic theoretical weakness in Shields' analysis, Shields'
diagram is still the most widely used criterion for incipient motion. Due to its sound
theoretical approach and agreement with measured data, Yang's (1973) criteria for
incipient motion warrant further consideration and verification.
2. Among all the assumptions used in the derivation of different sediment
transport equations, the assumption that the rate of sediment transport, or total
sediment concentration with wash load excluded, can be determined by unit stream
power is the most general and realistic one.
3. Based on more than 1000 sets of laboratory data and some field data collected
from six river stations, Yang's unit stream power equation is more accurate than
other sediment transport, equations.
4. Because of the basic theoretical weakness and the complex nature of natural
rivers, it is extremely difficult to recommend a universal equation which can be used
to estimate the rate of sediment transport in rivers under different flow and sediment
conditions. Recommendations are made to help engineers and geologists to make a
reasonable choice among published sediment transport equations.
References
Ackers, P. and White, W. R.: 1973, 'Sediment Transport: New Approach and Analysis,' J. Hydraulics
Div., A S C E 99, 2041.
ASCE Task Committee on Sedimentation: 1971, 'Sediment Transportation Mechanics: H. Sediment
Discharge Formulas,' J. Hydraulics Div, A S C E 97, 523.
Bagnold, R. A.: 1966, 'An Approach to the Sediment Transport Problem from General Physics,' U.S.
Geological Survey Profession Paper 422-J.
Bishop, A. A., Simons, D. B., and Richardson, E. V.: 1965, 'Total Bed-Material Transport,' J. Hydraulics
Div., A S C E 91, 175.
Blench, T.: 1969, 'Mobile-Bed Fluviology-A Regime Theory Treatment of Canals and Rivers for
Engineers and Hydrologist,' The Universily of Alberta Press, Edmonton, Alberta, Canada.
Bogardi, J.: 1968, 'Incipient Sediment Motion in Terms of Critical Mean Velocity,' Acta Tech. (Budapest)
62, 1.
Colby, B. R.: 1964, 'Practical Computations of Bed Material Discharge,' J. Hydraulics Div., A S C E 90,
217.
Colby, B. R. and Hembree, C. H.: 1955, 'Computation of Total Sediment Discharge, Niobrara River near
Cody, Nebraska,' U.S. Geological Survey Water Supply Paper 1357.
Colby, B. R. and Hubbell, D. W.: 1961, 'Simplified Method for Computing Total Sediment Discharge
with the Modified Einstein Procedure,' U.S. Geological Survey Water Supply Paper 1593.
Donat, J.: 1929, @ber Sohlangriff und Geschiebetrieb,' Wasserwirtchafi, No. 26, 27.
DuBoys, P.: 1879, 'Le Rhone et les Rivieres a Lit affouillable,' Mem. Dec., Ann. Pont et Chausses, Ser. 5
17.
Einstein, H. A.: 1944, 'Bed-Load Transportation in Mountain Creek,' U.S. Department of Agriculture
Soil Conservation Service SCS-TP-55.
Einstein, H. A.: 1950, 'The Bedload Function for Sediment Transport in Open Channel Flows,' U.S.
Dept. of Agric. Soil Conservation Service, Technical Report No. 1026.
EngeIund, F. and Hansen, E.: 1967, A Monograph on Sediment Transport in Alluvial Streams, Danish
Technical Press, Copenhagen, Denmark, Revised edition 1972.

68

CHIH TED YANG

Gessler, J.: 1965, 'The Beginning of Bedload Movement of Mixtures Investigated as Natural Armoring in
Channels,' W. N. Keck Hydraulics and Water Resources Laboratory, California Institute of Technology, Pasadena.
Gessler, J.: 1970, 'Self-Stabilizing Tendencies of Alluvial Channels,' J. Waterways Harbors Div,, A S C E
96, 235.
Gilbert, K. G.: 1914, 'The Transportation of Debris by Running Waters,' U.S. Geological Survey
Professional Paper 86.
Guy, H. P., Simons, D. B., and Richardson, E. V.: 1966, 'Summary of Alluvial Channel Data from Flume
Experiment, 1956-1961,' U.S. Geological Survey Professional Paper 462-I.
Hubble, D. W. and Matejka, D. Q.: 1959, 'Investigations of Sediment Transport, Middle Loup River at
Dunning, Nebraska,' U.S. Geological Survey Water Supply Paper 1476.
Jordan, P. R.: 1965, 'Fluvial Sediment of the Mississippi River at St. Louis, Missouri,' U.S. Geological
Survey Water Supply Paper 1802.
Laursen, E. M.: 1958, 'The Total Sediment Load of Streams,' J. Hydraulics Div., ASCE 84, 1530.
Leopold, L. B. and Maddock, T. Jr.: 1953, 'The Hydraulic Geometry of Stream Channels and some
Physiographic Implications,' U.S. Geological Survey Professional Paper 252.
Maddock, T. Jr.: 1973, 'A Role of Sediment Transport in Alluvial Channels,' Journal of the Hydraulics
Division, A S C E 99, 1915.
Maddock, T. Jr.: 1976, 'Equations for Resistance to Flow and Sediment Transport in Alluvial Channels',
American Geophysical Union, Water Resources Res. 12, 11.
Meyer-Peter, E., Favre, H., and Einstein, A.: 1934, 'Neuere Versuchsresultate fiber den Geschiebetrieb,'
Schweizerische Bauzeitung 103.
Meyer-Peter, E. and Mfiller, R.: 1948, 'Formula for Bedload Transport', International Association of
Hydraulic Research 2nd Meeting, Stockholm, p. 39.
Neill, C. R.: 1967, 'Stability of Coarse Bed-Material in Open Channel Flow,' Research Council of
Alberta, Edmonton, Canada.
Nordin, C. F., Jr.: 1964, 'Aspects of Flow Resistance and Sediment Transport, Rio Grande near
Bernalillo, New Mexico,' U.S. Geological Survey Water Supply Paper 1498-H.
Rottner, J.: 1959, 'A Formula for Bed Material Transport,' Houille Blanche 4.
Schoklitsch, A.: 1934, 'Geschiebetrieb und die Geschiebefracht', Wasserkraft Wasserwirtsch. 39.
Schumm, S. A. and Khan, H. R.: 1972, 'Experimental Study of Channel Patterns,' Geol. Soc. Amer. Bull.
83, 407.
Shen, H. W. and Hung, C. S.: 1972, 'An Engineering Approach to Total Bed Material Load by Regression
Analysis,' in H. W. Shen (ed.), Proceedings, Sedimentation Symposium, Berkeley, California, p. 14.
Shields, A.: 1936, 'Application of Similarity Principles and Turbulence Research to Bedload Movement,'
transl, into English by W. P. Ott and J. C. Van Uchelen at California Institute of Technology, Pasadena,
California.
Straub, L. G.: 1954, 'Terminal Report on Transportation Characteristics-Missouri River Sediment,'
University of Minnesota, St. Anthony Falls Hydraulics Lab., Sediment Series No. 4.
Toffaleti, F. B.: 1969, 'Definitive Computations of Sand Discharge in Rivers,' J. Hydraulics Div., A S C E
95, 225.
Vanoni, V. A., Brooks, N. H., and Kennedy, J. F.: 1960, 'Lecture Notes of Sediment Transport and
Channel Stability,' Report KH-RI, California Institute of Technology, Pasadena, California.
White, C. M.: 1940, 'The Equilibrium of Grains on the Bed of an Alluvial Channel,' Proc. Royal Soc.,
London, England, Series A 174, 322.
White, W. R., Milli, H., and Crabbe, A. D.: 1975, 'Sediment Transport Theories: A Review,' Proc. Inst.
Civil Engineers, Part 2, No. 59, June, London, England.
Yalin, M. S.: 1963, 'An Expression for Bedload Transportation,' J. Hydraulics Div., A S C E 89, 221.
Yang, C. T.: 1972, 'Unit Stream Power and Sediment Transport,' J. Hydraulics Div., A S C E 98, 1805.
Yang, C. T.: 1973, 'Incipient Motion and Sediment Transport,' J. Hydraulics Div., A S C E 99, 1679.
Yang, C. T.: 1976, 'Minimum Unit Stream and Fluvial Hydraulics,' J. Hydraulics Div., A S C E 102, 919.
Yang, C. T. and Sayre, W. W.: 1971, 'Stochastic Model for Sand Dispersion,' J. Hydraulics Div., A S C E
97, 265.
Yang, C. T. and Stall, J. B.: 1976, 'Applicability of Unit Stream Power Equation,' J. Hydraulics Div.,
A S C E 102, 559.

You might also like