MANNING River Hydraulics and The Channel
MANNING River Hydraulics and The Channel
MANNING River Hydraulics and The Channel
You will recall from the previous chapter that the continuity relationship can be expressed by
equation (3.1) as:
wd = A = Q/v (3.1)
where the cross-sectional area of the channel is a direct function of the discharge and the flow
velocity. Our task in this chapter is to learn what controls the mean velocity, v. Although most
of this discussion is concerned with the mean velocity and mean flow conditions, we know
already that velocity varies with position in the channel. It can also vary with time (that is, it can
fluctuate). Even though velocity does vary in space and time it is useful at a broad level of
generalization to assume that flow in a river channel can be characterized simply by the mean
flow. Nevertheless, it will be useful before we begin our journey here to acknowledge a few
definitions that will come in handy now but will be even more useful as we continue.
Hickin: River Geomorphology: Chapter 3
A Few Definitions
• Open-channel flow is the name given in fluid mechanics to flow through a channel that
has a free water-surface. It contrasts with closed-conduit flow (flow in a pipe) in which
there is no free water-surface. The difference is important because, unlike flow in closed
conduits, pressure variation in open-channel flow is expressed as a change in flow depth
(the water surface can rise or fall) and forces at work at the water/air interface can give
rise to surface waves. Such waves can exert significant control on the mean flow. The
definitions to follow all relate to open-channel flow although most can also be adapted to
pipe-flow as well.
• Steady flow is flow in which flow velocity (v) and flow depth (d) do not change with
time. To use the language of the calculus we say that flow velocity and flow depth are
time invariant (dv/dt = 0 and dd/dt = 0). Of course, the continuity relationship dictates
that, for a given discharge through an open channel with a rigid boundary, if dv/dt = 0
then it must also be true that dd/dt = 0.
• Unsteady flow, on the other hand, is flow in which flow velocity (v) and flow depth (d)
do change with time. Again, to use the language of the calculus we say that flow velocity
and flow depth vary with time (dv/dt ≠ 0 and dd/dt ≠ 0). Once again, the continuity
relationship dictates that, for a given discharge through an open channel with a rigid
boundary, if dv/dt ≠ 0 then it must also be true that dd/dt ≠ 0.
• Uniform flow is flow in which velocity and depth of flow do not change in space (ie,
with distance downstream, s). Again, to use the language of the calculus we say that flow
depth (and therefore flow velocity) does not vary spatially (dv/ds = 0 and dd/ds = 0).
Note that this does not mean that the channel bed and the water surface have to be planar
(flat). Uniform flow can also be achieved over a wavy bed if the water-surface has the
same parallel waveform so that flow depth does not vary downstream. It turns out,
however, that this kind of in-phase condition of the bed and water surface is possible but
an unlikely outcome in most river channels.
• Non-uniform flow is the converse of uniform flow: flow depth varies downstream along
the channel so that the depth of flow varies spatially along with the mean-flow velocity
(or dv/ds ≠ 0 and dd/ds ≠ 0).
-45-
Hickin: River Geomorphology: Chapter 3
Most flows in rivers are unsteady and non-uniform although it is sometimes convenient in fluid
mechanics to assume that they are in fact steady uniform flows. Flow can be steady and
uniform, unsteady and uniform, unsteady and non-uniform but never unsteady and uniform or
steady and non-uniform. Think about a rectangular channel carrying a constant discharge; can
you see why these last two conditions (unsteady uniform flow and steady non-uniform flow) are
not possible?
We also need to recognize two general kinds of flow that control the way in which velocity
varies within the cross section of a channel in two quite different ways. The first is laminar flow
and the second is turbulent flow (Figure 3.1):
Figure 3.1: Some definitions in open-channel flow (from Knighton, 1998, Figure 4.1)
-46-
Hickin: River Geomorphology: Chapter 3
• Laminar flow is flow in which the rate of fluid deformation (for example, the change in
velocity with height above the channel bed) is controlled only by the viscosity of the
fluid. Viscosity is that internal molecular property that determines how easily a fluid will
flow. We say that water is less viscous (more runny) than higher viscosity honey or oil,
for example. Laminar flow is envisioned as occurring as stacked laminae or layers of
flow in which the water particles do not move vertically between enveloping layers of
water. Any force that tends to create these vertical motions is dampened and overcome
by the opposing viscous forces. Laminar flow is not common in nature and probably
never occurs in natural channels. We consider it here because it provides a basis of
comparison for examining the alternative more realistic “turbulent flow.”
• Turbulent flow is flow in which the rate of fluid deformation is controlled, not by fluid
viscosity (because it is too weak a force), but overwhelmingly by the internal chaotic
fluid motions we call turbulence. In turbulent flow the orderly arrangement of flow
layers found in laminar flow are disrupted because, as vigour of the flow increases,
viscosity is no longer strong enough to overcome the large forces tending to move water
particles vertically, thus allowing mixing of the flow.
In steady uniform flow the velocity is constant and we therefore know that the impelling and
resisting forces must be in balance (zero net force). Because it is the case that, over short reaches
along a channel (say a channel width or so), flow velocity in a natural channel is sensibly
constant, we can safely assume that it approximates steady uniform flow. This assumption
allows us to derive some relationships that lead us to some very useful tools for analyzing flows
in rivers. This task will be aided by reference to Figure 3.2.
-47-
Hickin: River Geomorphology: Chapter 3
w
d
FI
A = Wd
W FR
α
W = volume.density.gravity
W = ALρg
P = 2d + w
Figure 3.2: A control volume of fluid in a rectangular channel of length, L, width w, and depth, d. The
impelling gravitational force F1 is opposed by the force of boundary resistance, FR.
Figure 3.2 shows a control volume of fluid flowing in a rectangular channel (we want to keep
things simple for now but the principles we will derive can be applied to a cross-section of any
shape).
The impelling force, FI , is the force of gravity acting over unit area on the bed. This force is
simply the weight of the water (the product of the water volume (wdL), fluid density (ρ) and the
acceleration of gravity (g = 98.1 m/s2) resolved in the downstream direction. In other words we
can say that:
FI = sinαALρg
where α is the angle that the bed slopes downstream. FI, like any force in mechanics, is
expressed in units of mass x acceleration or kg x m/s2; these force dimensions are known as
Newtons (N). So we might say, for example that the water exerts a force of 5.0 N on the channel
boundary.
-48-
Hickin: River Geomorphology: Chapter 3
The resisting force, FR, is taken to be the force of friction being exerted by the bed and vertical
banks on the flow moving in our rectangular channel. This is the friction force operating in the
opposite direction to the impelling force that the flow has to overcome to accelerate or at least
equal if the velocity is to remain constant. This frictional force acts in the plane of the channel
boundary so it is a shear stress (τo), a force per unit area, acting over the total area of the channel
boundary supporting our control volume of water. Stated mathematically:
FR = τo(2d+w)L
Since FI = FR in our model we can say that, for our control volume of flow:
FI = FR
sinαALρg = τo(2d+w)L
A
or τ o = sin αρg (3.2)
2d + w
3. Boundary€Shear Stress
Equation (3.2) provides us with a means of measuring the shear stress exerted on the channel
boundary by the flowing water. Shear stress is a force per unit area and is expressed as so many
N/m2.
-49-
Hickin: River Geomorphology: Chapter 3
4. Flow Resistance
As shear stress increases in a river, so does the flow velocity. The degree to which the mean
velocity responds to increasing shear stress depends on the resistance to flow in the channel. Put
another way, any process in a river that bleeds off energy from driving the mean flow, represents
resistance to flow. The principal source of flow resistance for most rivers is the frictional
resistance that the water encounters as it moves across the channel boundary. But this is not the
only source. Indeed, under certain conditions, boundary friction can be overwhelmed by other
sources of flow resistance.
Turbulent resistance refers to the dampening of the mean flow by the presence of the small-
scale chaotic motions of water particles. These motions are three-dimensional and consist of
small parcels of water than spin and twist and tumble as the flow moves along. Some have
proposed that we might think of these small eddies in the flow as constituting an additional type
of viscosity: eddy viscosity. Eddies are generated at the bed of a river as the water flows over the
boundary but they are convected into the flow where they are carried along and slowly diffuse as
they lose their rotational energy to viscous resistance.
-50-
Hickin: River Geomorphology: Chapter 3
The distinction between small-scale turbulence and the scale of larger disturbances to the flow
(such as large eddies shed from bank protuberances or eddy shedding from boulders on the bed,
for example) is entirely arbitrary. Some scientists have suggested that this larger scale of
discrete flow structure should be termed macroturbulence and this term has become quite widely
adopted in England and North America. For our purposes any flow structure scaled up to the
depth of flow we can call turbulence and is thus a source of turbulent resistance.
Internal distortion resistance refers to the energy lost to still larger flow structures such as the
energy lost to the mean flow that is bled off to drive the secondary circulation in bends (helical
flow). Other things being equal, there is greater resistance to flow through a meandering channel
than through a straight channel. Whenever there are sudden changes in bank alignment causing
abrupt changes in channel width or depth, the flow encounters internal distortion resistance.
Spill resistance is encountered where there are such severe changes in channel morphology
(from narrow to wide) that flow literally runs into the slack water ahead of it. The most extreme
example of spill resistance is a waterfall where the flow plunges into a pool. Spill resistance can
far exceed all other kinds of flow resistance in certain special circumstances. In mountain
streams, for example, it is not uncommon to encounter a high-energy kind of flow (called
supercritical flow) that is associated with severe water-surface deformation including breaking
surface waves and chutes that “collide” with downstream pools. Here spill resistance can be
extreme.
Boundary or skin resistance, however, is the principal source of flow resistance in most natural
channels. Boundary resistance depends on boundary roughness: smooth boundaries exert low
boundary resistance and rough boundaries give rise to high boundary resistance. It is difficult in
practice to decouple boundary resistance from turbulent resistance because both increase as
boundary roughness increases. Boundary roughness, as we shall soon see, is itself not an
uncomplicated concept. It is often thought of as consisting of two components: grain roughness
and form roughness. Grain roughness is the roughness that relates to the individual particles or
grains making up the sediments forming the channel boundary. Form roughness is the roughness
that relates to aggregates of particles forming bedforms and bars. These features are more
-51-
Hickin: River Geomorphology: Chapter 3
difficult to analyze because, unlike the fixed grain size, they are transient and change as the flow
changes.
It turns out that, by countless repeated observations in open-channel flow, shear stress has been
shown to be proportional to the velocity squared (also shown to be the case by dimensional
analysis):
τ ∝ v2 or
v∝ τ
Since τ = γds, we can say that
v = C ds and
€
V
C= (3.4)
ds
€
The coefficient of proportionality, C, is called Chezy C and v = C ds is known as the Chezy
€
equation after the 18th-century French hydraulician who introduced it (http://chezy.sdsu.edu/).
Chezy C is a measure of flow conductance or efficiency. Other things being equal, flow velocity
€
increases as C increases (that is, it is the inverse of flow resistance).
Chezy C by itself was found difficult to assess and several attempts were made to evaluate it in
terms of variables that could be easily determined in the field. In 1891 Robert Manning, an Irish
engineer determined Chezy C empirically in terms of parameters that could be related to a river
channel in the field:
1
d6
C=
n
where n is the Manning roughness factor and ranges over an order of magnitude from 0.01-0.10
and d is flow depth.
€
Combining the Chezy equation and the Manning’s estimate of Chezy C yields
-52-
Hickin: River Geomorphology: Chapter 3
v
C=
ds
1
d 6
v
= 1 1
n
d 2s2
1 1 1 2 1
d 6d 2s2 d 3 s2
v= =
n n
2 1
d s
3 2
or v= (3.5)
€ n
Equation (3.5) was once known as the Chezy-Manning equation but now is known simply as the
Manning equation. The equation is in metric units but is not dimensionally balanced.
€
The Manning equation is very important because, in spite of its antiquity, it is widely used in
river engineering and applied work in fluvial geomorphology today. Manning n is thought of as
an index of channel roughness when it is based on velocity, depth and slope measurements in the
field, but as such it is really a coefficient of proportionality that reflects all sources of flow
resistance and not just simply channel roughness.
-53-
Hickin: River Geomorphology: Chapter 3
Throughout our discussion of flow resistance and channel roughness to this point we have
avoided a very difficult problem: what precisely is channel roughness? The conventional
wisdom is that it consists of two parts: grain roughness and form roughness.
Grain roughness is taken to be that component of roughness that relates directly to the size (grain
diameter) of particles constituting the boundary materials. For example, at a qualitative level we
can say that a boundary made of sand is less rough than one made of fine gravel because the
grains of sand are smaller than the particles of gravel. Making a more precise quantitative
statement about grain roughness requires us to digress for a bit because grain diameter is itself
not a straightforward concept.
Grain diameter in natural sediment varies considerably and in order to characterize the
size of grains in a sample of sediment we must have some basis for generalizing such as using an
average or some other measure of central tendency. It turns out that many geomorphologists
think that the most important grain-size influencing channel roughness for the purpose discussing
flow resistance is not the average size but rather the size of larger particles in any given sample.
In order to bring some objectivity and consistency to measures of these coarser particles the
grain-size distribution has become the basis of defining effective grain size. In general
geomorphologists use the Wentworth grade scale (borrowed from sedimentary geologists) when
discussing grain size (Figure 3.3):
Figure 3.3: The Wentworth grade scale used to describe grain size.
-54-
Hickin: River Geomorphology: Chapter 3
The Wentworth grade scale is based on the median size of a sample of sediment grains. We
designate the median size as D50 because it is the grain size in the cumulative distribution of
grain size, half of which (50%) is coarser and half therefore is finer. An example of how we
construct a cumulative grain-size distribution graph is shown in Figure 3.4.
120
100
80
Percent
finer
60
40
20
0
0 0.5 1 1.5 2 2.5 3 3.5
Diameter, cm
Figure 3.4: A cumulative grain-size analysis and graph for determining grain-size percentiles
-55-
Hickin: River Geomorphology: Chapter 3
Here the median grain size, D50, is 1.25 cm and D84 is 2.50 cm.
In research on flow resistance a number of studies have shown that resistance is more sensitive to
the higher grain-size percentiles than D50. For example, it is common for D84 to be used as a
measure of effective grain size. Remember, D84 is simply mathematical shorthand for saying
“that grain size in a sediment sample in which 84 per cent of the grains are finer (16 per cent are
coarser). In the case of the example in Figure 3.4 we say that “84 per cent of the grains in the
sample are finer than 2.50 cm”. For the statistically-inclined reader, D84 represents two standard
deviations above the mean if the grain-size distribution is normal.
Form Roughness is that component of roughness that relates to aggregates of grains in the form
of bedforms and bars. Form roughness can be considerably greater than grain roughness,
especially in steep rugged mountain channels. Examples of bedforms in sandbed rivers are
shown in Figures 3.5 – 3.7:
Figure 3.5: Gravel and sand have accumulated around young willows to create rib-like sedimentary forms that
have form roughness larger than the sediment grains that form them.
-56-
Hickin: River Geomorphology: Chapter 3
Figure 3.7: Dunes and ripples on dunes formed on a point-bar surface of a sand-bed river.
-57-
Hickin: River Geomorphology: Chapter 3
The case of bedforms that occur in sand-bed rivers has been much studied although we still have
a great deal to learn about the role that bedforms play in accounting for the resistance to flow in
channels. In Figure 3.8 is reproduced David Knighton’s Figure 4.3 from his fluvial
Figure 3.8: Form roughness elements (from Knighton, 1998: Figure 4.3).
The flow-resistance effects of changing bedforms in sandbed rivers are sometimes evident in
discharge rating curves where measured discharge is related to the flow stage. In Figure 3.9
-58-
Hickin: River Geomorphology: Chapter 3
Discharge
Gauge height
(stage)
Figure 3.9: This synthetic discharge rating curve shows typical responses to changing bedform regime.
the discontinuities or “kinks” in the graph relate to the shift in the bedform domains as discharge
increases. For example, in Figure 3.9A, discharge increases as an orderly direct function of
gauge height while ripples dominate the bedforms but as they are replaced by dunes in the
transition zone, discharge increases with very little change in gauge height because the lower
resistance allows the velocity to increase and accommodate most of the change in discharge
(refer to Figure 3.8). Once ripples begin to form again on the surface of newly developed dunes
on the bed, however, flow resistance increases yet again and discharge and gauge-height increase
together once more. Figure 3.9B shows the same set of transitions with gauge height as the
dependent variable. The rate of change in the stage flattens off in the transition zone as changes
in velocity accommodate the increasing discharge and then increases once again as flow
resistance increases in the zone of ripples on dunes.
-59-
Hickin: River Geomorphology: Chapter 3
Direct measurement of flow resistance is usually done when the flow resistance is the
variable of interest (rather than the prediction of velocity). For example, a geomorphologist
might be interested in comparing the flow resistance in a river as she moves along the channel
from the headwaters to the mouth. Measured data required to compute Darcy-Weisbach ff and
Manning n are flow depth, water-surface-slope, and flow velocity at the downstream sequence of
measurement stations; substitution in equations (3.5) and (3.7) yields the desired result.
Grain roughness is mainly a function of relative roughness (d/D or Rh/D). A widely used version
applicable to open-channels is the Wolman equation. The Wolman equation is a semi-empirical
correlation between ff and the relative roughness, y/D84 (y = flow depth, d, or hydraulic radius,
Rh):
1 y
= 2log + 1.0 (3.8)
ff D84
I say semi-empirical because the structure of equation (3.8) does have some theoretical basis in
early studies of pipe flow although the details need not concern us here. Figure 3.10 shows a
€
graph of a very similar updated function derived by David Knighton and presented in his 1998
-60-
Hickin: River Geomorphology: Chapter 3
textbook (the published equation below actually is incorrect and here the correct version appears
above the original in bold).
1
= 0.82ln(4.35 R D )
ff 84
Figure 3.10: Relationship between friction factor (ff) and relative roughness (from Knighton, 1998).
For Rh/D84>1.0 the Wolman and Knighton equations yield sensibly identical results.
Estimating Manning n from grain size is also based on an empirical correlation, in this
case known as the Strickler equation:
The Strickler equation was developed in the European Alps and applies ideally to gravel-bed
rivers where grain roughness is the primary source of channel roughness. In any case, it yields a
minimum estimate of Manning n.
-61-
Hickin: River Geomorphology: Chapter 3
So, for D50 = 2.5 cm (or 0.025 m) equation (3.9B) yields: n = 0.0478(0.025)1/6 = 0.026.
Combining equation (3.9B) with the Manning equation yields:
d 2 / 3 s1/ 2
v= (3.10)
0.0478D501/ 6
Equation (3.10) is commonly employed today by river engineers in need of flow estimates in
ungauged gravel-bed rivers.
€
where each term is defined in Figure 3.14. The term n0 assigns a roughness based on boundary
materials alone. The term n1 is an estimate of the effect of within cross-sectional irregularity,
negligible for a very regular shape (trapezoidal or semi-circular, for example) and up to an
additional 0.20 roughness units for channels that have severe boundary irregularities such as
those caused by bank slumping. The term n2 also refers to channel irregularity but this time for
changes along the channel. If the changes along the channel are minor or gradual no adjustment
-62-
Hickin: River Geomorphology: Chapter 3
-63-
Hickin: River Geomorphology: Chapter 3
Manning n = 0.026
Manning n = 0.059
Figure 3.12: Photographs from a field manual designed to assist a field technician in estimating
Manning n (from Barnes, 1967)
is necessary but severe alternations of narrow and wide sections, for example, might add up to
0.015 roughness units to the total estimate of Manning n. Term n3 recognizes the effects of
individual obstructions (a large boulder, a fallen tree, grounded logs, etc). Obstructions assessed
here must not have been included in n1 or n2 (that is, no double counting of effects). Severe
obstructions can add as much as 0.06 roughness units to the total roughness. The term n4
recognizes the effects of vegetation in the channel and as Figure 3.14 illustrates, vegetation type
-64-
Hickin: River Geomorphology: Chapter 3
Negligible........... 0.00
n3 Determination of n3 is based on the presence and
characteristics of obstructions such as debris, slumps, 0.010
stumps, exposed roots, boulders and fallen and lodged Minor ............to 0.015
Relative logs. Conditions considered in other steps must not be
effect of reevaluated (double counted) in this determination. In 0.020
obstructions judging the relative effect of obstructions, consider the Appreciable….to 0.030
extent to which the obstructions occupy or reduce
average water area; the shape (sharp or smooth) and 0.040
position and spacing of the obstructions. Severe..............to 0.060
Low: Dense but flexible grasses where flow depth is 2-3 x the
height of vegetation or supple tree seedlings (willow, poplar) 0.005
where flow depth is 3-4 x vegetation height............................................................to 0.010
Medium: Turf grasses in flow 1-2 times vegetation height; stemmy grasses
n4 where flow is 2-3 x vegetation height; moderately dense brush on 0.010
banks where Rh>0.7 m.............................................................................................to 0.025
High: Turf grasses in flow of same height; foliage-free willow or poplar,
Vegetation 8-10 years old and intergrown with brush on channel banks 0.025
where Rh>0.7m; bushy willows, 1 year old, Rh>0.7m............................................to 0.050
Very High: Turf grasses in flow half as deep; bushy willows (1 year old)
with weeds on banks; some vegetation on the bed; trees with 0.050
weeds and brush in full foliage where Rh>5m.........................................................to 0.100
m
Minor: Sinuosity index = 1.0 to1.2............................................................................... 1.00
Degree of Appreciable: Sinuosity index = 1.2 to 1.5.................................................................... 1.15
Meandering Severe: Sinuosity index >1.5........................................................................................ 1.30
Figure 3.13: The Cowan procedure for estimating Manning n [(n = n0+n1+n2+n3+n4)m].
-65-
Hickin: River Geomorphology: Chapter 3
can greatly influence the overall roughness. Grasses and supple seedlings (which likely lay
down in streamlined fashion in the flow) can add up to about 0.02 roughness units to the total
roughness but bushes and brush (that more likely will stand up in the flow) can add as much as
1.00 roughness units to the total roughness.
An adjustment is also made in the Cowan procedure for the degree of channel meandering,
increasing the total bracketed terms in equation (3.11) by as much as 30% where the sinuosity
index exceeds 1.5. The sinuosity index is simply the ratio of the actual channel length between
two points along the channel and the straight-line distance between them.
Like learning to play the piano or the flute, practice makes perfect here as well! River scientists
and engineers who routinely assess Manning n on rivers quickly achieve a remarkable degree of
accuracy in assigning roughness factors in their work. Experience clearly counts for a lot in this
endeavour and you too will find that your estimates of Manning n will greatly improve with
repetition.
Problem 2: A sensibly straight and rectangular channel has a smooth bed of fine to
medium gravel. Calculate the discharge if the bed slope is 0.0005, flow depth is 2.00 m
and the channel width is 150 m.
Solution 2: From continuity we know that Q = Av = (150)(2.00)v = 300v
-66-
Hickin: River Geomorphology: Chapter 3
d 2 / 3 s1/ 2 2 2 / 3 (0.0005)1/ 2
From the Manning equation, v = =
n n
From the table of Manning n we note that, for fine to medium gravel,
300(1.59)(0.0224)
n = 0.03. So Q = € = 356 m 3 s−1
0.03
Note that we could also have used the Strickler equation, n = 0.0151D501/6 (mm)
to estimate Manning n: n = 0.0151(25)1/6 = 0.03
€
Problem 3: A river channel with a rectangular cross-section has 3.00 m-high banks, is
200.00 m wide, and has a slope of 0.0001. Will a discharge of 1000.0 m3s-1 cause
flooding if Mannings n = 0.03?
1000 5
Solution 3: From continuity (Q = wdv) we know that 1000.0 = 200dv and v = =
200d d
d 2 / 3 s1/ 2 d 2 / 3 (0.0001)1/ 2
From the Manning equation, v= = = 0.33d 2 / 3
n 0.03
€
Substituting for v we can say that
5
= 0.33d 2 / 3
€ d
5 = 0.33d 5 / 3
5 3 / 5
d =
0.33
d = 5.11m
It follows that the 1000 m3s-1 discharge cannot be contained by a channel of this size and
that overbank flooding will occur.
€
Although we don’t have quite enough theory yet to deal with this question analytically here in
Chapter 3, our discussion of equilibrium channels in Chapter 2 provides a hint about how we
might go about this kind of analysis. If we could define the critical conditions for moving the
material forming the bed of a river channel we could combine this condition with the hydraulics
we have learned in this chapter to resolve the problem. We will return to this issue in Chapter 4.
-67-
Hickin: River Geomorphology: Chapter 3
Meanwhile, there are some empirical studies that point clearly to the role of bank strength in
determining the form ratio of an alluvial channel, the most famous of which is the work of
Professor Stanley Schumm at the State University of Colorado at Fort Collins. He reasoned that,
because very sandy channels have very low banks (because sandy sediment is not strong enough
to form high vertical bank sections), sandbed channels are always very wide relative to their
small depths. Conversely, streams with mud-rich banks have strong cohesive sediments and can
be much deeper (and therefore narrower) than their sandy counterparts carrying the same
discharge. Schumm formalized these general relationships in an empirical model based on
sampling channel sediments in rivers in the American Midwest (Figure 3.14).
Figure 3.14 shows the strong inverse relationship between the shape of a river channel expressed
as the form ratio (w/d) and the strength of the boundary materials expressed as the percentage of
silt/clay that they contain. Channels consisting of sediments with about 50% or more silt-clay
have form ratios less than 10 (deep and narrow) while those formed in sediments with less than
about 5% silt clay have form ratios greater than 50 (wide and shallow).
-68-
Hickin: River Geomorphology: Chapter 3
Knighton, D. 1998: Fluvial Forms and Processes: A New Perspective, Hodder Arnold Publication: pages 96-107.
Barnes,H.H., 1967: Roughness characteristics of natural channels. Water Supply Paper 1894, US Geological
Survey, Washington, DC, 213 pp.
Schumm, S.A., 1960: The shape of alluvial channels in relation to sediment type. United States Geological Survey
Professional Paper 352B, 17-30.
Simons, D.B. and Richardson, E.V., 1966: Resistance to flow in alluvial channels. US Geological Survey Paper
422J.
-69-