Understanding River Characteristics
Understanding River Characteristics
1.1.1 Introduction
The primary function of a river channel is the conveyance of water and sediment. It should be
appreciated that this primary function cannot be stopped. Nor can the long-term average be
changed by measures carried out in the riverbed. Thus alterations in space and time can only be
made within the context of ultimate equilibrium.
The most conspicuous aspect of a river channel, apart from its size, is the amount of water it
carries. This is best shown in a hydrograph. A hydrograph is a time series of water level data or
discharge data. Changes in discharge cause changes in water level in the river channel. At very
high discharges a river channel overflows its banks on to the adjacent land. This periodically
flooded land is called the flood plain. Whilst in the upper reaches the flood plains are usually
narrow or even non-existent, in the lower reaches of a river the flood plains could be tens of
kilometers wide.
The imaginary line delimiting various watersheds is known as water dividing line or water-
divide. Its configuration depends on the topography only, and it runs along the highest points
of the surrounding area. Precipitation falling outside the area enclosed by this line will form a
runoff flowing to another stream, over another catchment
catchm area.
The farther downstream along the stream the point of concentration, the more tributaries will
join the stream and the larger the respective watershed. Sudden increase of area indicates the
inflow of a large tributary, the point at which the whole partial watershed of the tributary joins
the watershed of the main watercourse. Gradual increases are derived mainly from overland
flow areas or small gullies.
Figure 1.2: Watershed drainage patterns. Patterns are determined by topography and
geologic structure.
WU_KIOT- WRIE 2
GBU, CoET, WRIE, By Koang G.
River Engineering River Characteristics
Strahler’s stream ordering system is portrayed in Figure 1.3. The uppermost channels in a
drainage network (i.e., headwater channels with no upstream tributaries) are designated as
first-order streams down to their first confluence. A second-order stream is formed below the
order channels. Third-order streams are created when two second
confluence of two first-order second-order
channels join, and so on. Note in the figure that the intersection of a channel with another
channel of lower order does not
not raise the order of the stream below the intersection (e.g., a
fourth-order
order stream intersecting with a second-order
second stream is still a fourth-order
order stream below
the intersection).Within a given drainage basin, stream order correlates well with other basin
parameters, such as drainage area or channel length. Consequently, knowing what order a
stream is can provide clues concerning other characteristics such as which longitudinal zone it
resides in and relative channel size and depth.
Figure 1.3:
3: Stream ordering in a drainage network
Channel and Ground Water Relationships: Interactions between ground water and the
channel vary throughout the watershed. In general, the connection is strongest in streams with
gravel riverbeds in well-developed
developed alluvial floodplains.
Figure 1.4 presents two types of water movement:
Influent or “losing” reaches lose stream water to the aquifer.
Effluent or “gaining” reaches receive discharges from the aquifer.
Figure 1.4: Cross sections of (a) influent and (b) effluent stream reaches.
Practitioners categorize streams based on the balance and timing of the stormflow and
baseflow components. There are three main categories:
Ephemeral streams flow only during or immediately after periods of precipitation. They
generally flow less than 30 days per year.
Intermittent streams flow only during certain times of the year. Seasonal flow in an
intermittent stream usually lasts longer than 30 days per year.
Perennial streams flow continuously during both wet and dry times. Baseflow is
dependably generated from the movement of ground water into the channel.
Discharge Regime: Streamflow is one of the variables that determine the size and shape of the
channel. There are three types of characteristic discharges:
Channel-forming (or dominant) discharge: If the streamflow were held constant at the
channel-forming discharge, it would result in channel morphology close to the existing
channel. However, there is no method for directly calculating channel-forming discharge.
An estimate of channel-forming discharge for a particular stream reach can, with some
qualifications, be related to depth, width, and shape of channel. Although channel-forming
discharges are strictly applicable only to channels in equilibrium, the concept can be used
to select appropriate channel geometry for restoring a disturbed reach.
and in processes such as erosion and deposition. Even among different types of streams, a
common sequence of structural changes is observable from headwaters
headwaters to mouth.
The overall longitudinal profile of most streams can be roughly divided into three zones:
Zone 1, or headwaters (or upper course),
course), often has the steepest gradient. Sediment erodes
from slopes of the watershed and moves downstream. Typically
Typically erosive stream
characteristics.
Zone 2, the transfer zone (or Middle course),
course), receives some of the eroded material. It is
usually characterized by wide floodplains and meandering channel patterns. Longitudinal
slope of the stream gradually eases; tributaries
tributaries join the main stream, and therefore often
sudden changes of flow regime. Although stretches of erosion and deposition frequently
exchange, both in space and time, this transitional reach of the stream is on the whole
generally the most stable and
and balanced part. Stream characteristics obtained from the
middle course are frequently used as basis for design of stream training projects.
Zone 3, the depositional zone (or Lower course):
course): Longitudinal slope flattens; discharge
increases in Zone 3, the primary
rimary depositional zone (ref Fig. 1.5 a and 1.5b)
1.5b). gradual
deposition of sediment eroded upstream, hence relatively short-period
short period shifting and
changing of the main stream channel.
Figure 1.5: Three longitudinal profile zones. Channel and floodplain characteristics change as rivers
travel from headwaters to mouth.
WU_KIOT- WRIE 5
GBU, CoET, WRIE, By Koang G.
River Engineering River Characteristics
Though the figure displays headwaters as mountain streams, these general patterns and
changes are also often applicable to watersheds with relatively small topographic relief from
the headwaters to mouth. It is important to note that erosion, transfer, and deposition occur in
all zones, but the zone concept focuses on the most dominant process.
Figure 1.6: Changes in the channel in the three zones. Flow, channel size, and sediment
characteristics change throughout the longitudinal profile.
1.2.1 Introduction
Although the hydraulic behavior of alluvial channels is known to be in many respects different
from the behavior of rigid-boundary channels, there is to this day no theoretical set of
equations applying specifically to the flow in movable-bed channels. In general, the accepted
procedure is to use the kinematic and dynamic equations as developed for rigid-boundary
channels, and to introduce empirical coefficients or relationships supposed to suitably modify
them in order to better fit movable-bed conditions. This is certainly not a satisfactory solution
but it is the only feasible one with the present state of the art.
It can hardly be sufficiently stressed that the limitations and constraints of theoretical
considerations, when applied to alluvial streams, should always be borne in mind by practicing
engineers, if a correct understanding of actual physical processes is to be reached. It has
already been pointed out in the opening sentences of the present chapter that substantial
differences exist between the rigid-boundary and movable-bed hydraulics, and at this point it
may be of interest to mention some of them.
As soon as the flow has started, an alluvial channel begins to continuously adapt and change its
deformable boundaries. Indeed, its characteristic roughness is determined not only by some
mean grain size protruding into the water, but usually by bed forms as well, or often even more
by the latter factor than by the former. In the latter case of form roughness, the ratio between
the roughness height (the effective height of the roughness elements, a measure of linear
dimension designating their effect upon the flow) and depth of flow is generally several orders
of magnitude larger than for fixed-bed channels.
Moving sediment elements and their continuous shifting of position are likely to induce
additional shear stresses; progressive movement of bed forms, on the other hand, may cause
disturbances in flow pattern due to changing water depth. Moreover, suspended sediment
carried by the water, especially when in higher concentrations, often has also an influence upon
the turbulence level of the flow.
Although the quantitative analysis and the extent of the mentioned and other influences due to
the movable-bed conditions of alluvial channels largely elude our present knowledge, they
WU_KIOT- WRIE 7
GBU,CoET, WRIE By Koang G.
River Engineering River Characteristics
should be carefully kept in mind whenever applying flow concepts developed for rigid-
boundary conditions to flows in deformable conduits.
1.2.2 Types of flow and Water Movement in Rivers
Where μ is dynamic viscosity, ρ is density of water and ν is kinematic viscosity ( ν = 10-6 m2/s
at 200C).
Most flows in nature are turbulent. Turbulence is generated by instability in the flow, which
trigger vortices. However, a thin layer exists near the boundary where the fluid motion is still
laminar. A typical phenomenon of turbulent flow is the fluctuation of velocity
U u u ; W w w (1.2)
Where U and W are instantaneous velocity, in x and z directions, respectively
u and w are time-averaged velocity, in x and z directions, respectively
u and w are instantaneous velocity fluctuation, in x and z directions, respectively
Turbulent flow is often given as the mean flow, described by u and w.
In turbulent flow the water particles move in very irregular paths, causing an exchange of
momentum from one portion of fluid to another, and hence, the turbulent shear stress
(Reynolds stress). The turbulent shear stress, given by time-averaging of the Navier - Stokes
equation, is
(1.3)
Note that u w is always negative. In turbulent flow both viscosity and turbulence contribute to
shear stress. The total shear stress is
du
t
dz
u w (1.4)
Steady versus unsteady: A flow is steady when the flow properties (e.g. density, velocity,
pressure etc.) at any point are constant with respect to time. However, these properties may
vary from point to point. In mathematical language,
GBU,CoET,WRIE By Koang G.
WU_KIOT- WRIE 8
River Engineering River Characteristics
(1.5)
In the case of turbulent flow, steady flow means that the statistical parameters (mean and
standard deviation) of the flow do not change with respect to time. If the flow is not steady, it
is unsteady.
Uniform versus non-uniform: A flow is uniform when the flow velocity does not change
along the flow direction.
Boundary layer flow: Prandtl developed the concept of the boundary layer. It provides an
important link between ideal-fluid flow and real-fluid flow. Here is the original description.
For fluids having small viscosity, the effect of internal friction in the flow is appreciable only
in a thin layer surrounding the flow boundaries. However, we will demonstrate that the
boundary layer fills the whole flow in open channels.
The boundary layer thickness (δ) is defined as the distance from the boundary surface to the
point where u = 0.995 U. The boundary layer development can be expressed as
GBU
WU_KIOT- WRIE 9
GBU,CoET,WRIE By Koang G.
River Engineering River Characteristics
(1.6)
Assuming the vertical instantaneous velocity fluctuation having the same magnitude
(1.7)
Where negative sign is due to the downward movement of the fluid parcel; the turbulent shear
stress now becomes
(1.8)
The turbulent shear stress can be expressed in a way similar to viscous shear stress as follows
(1.9)
WU_KIOT- WRIE 10
GBU, CoET,WRIE By Koang G.
River Engineering River Characteristics
b P x g A x sin
Where P is the wetted perimeter and A the area of the cross section. By applying the hydraulic
radius (R = A/P) we get
(1.11)
In the case of wide and shallow channel, R is approximately equal to h; eq (2.11) is identical to
eq (1.10).
Friction velocity: The bottom shear stress is often represented by friction velocity, defined by
(1.12)
WU_KIOT- WRIE 11
GBU, CoET, WRIE By Koang G.
River Engineering River Characteristics
b
The term friction velocity comes from the fact that has the same unit as velocity and it
Viscous shear stress versus turbulent shear stress: Eq (1.10) states that the shear stress in
flow increases linearly with water depth. As the shear stress is consisted of viscosity and
turbulence, we have
(1.14)
On the bottom surface, there is no turbulence (u = w = 0, u´ = w´ = 0), the turbulent shear stress
Therefore, in a very thin layer above the bottom, viscous shear stress is dominant, and hence
the flow is laminar. This thin layer is called viscous sub-layer. Above the viscous sub-layer,
i.e. in the major part of flow, the turbulent shear stress dominates.
Measurements show the shear stress in the viscous sub-layer is constant and equal to the
bottom shear stress, not increasing linearly with depth.
Classification of flow layer: Figure 1.9 shows the classification of flow layers. Starting from
the bottom we have
WU_KIOT- WRIE 12
GBU,CoET, WRIE By Koang G.
River Engineering River Characteristics
1. Viscous sublayer: a thin layer just above the bottom. In this layer there is almost no
turbulence. Measurement shows that the viscous shear stress in this layer is constant. The
flow is laminar. Above this layer the flow is turbulent.
2. Transition layer: also called buffer layer. Viscosity and turbulence are equally important.
3. Turbulent logarithmic layer: viscous shear stress can be neglected in this layer. Based on
measurement, it is assumed that the turbulent shear stress is constant and equal to bottom
shear stress. It is in this layer where Prandtl introduced the mixing length concept and
derived the logarithmic velocity profile.
4. Turbulent outer layer: velocities are almost constant because of the presence of large
eddies which produce strong mixing of the flow.
In the turbulent logarithmic layer the measurements show that the turbulent shear stress is
constant and equal to the bottom shear stress. By assuming that the mixing length is
proportional to the distance to the bottom (l = kz), Prandtl obtained the logarithmic velocity
profile.
Various expressions have been proposed for the velocity distribution in the transitional layer
and the turbulent outer layer. None of them are widely accepted. However, by the modification
of the mixing length assumption, the logarithmic velocity profile applies also to the transitional
layer and the turbulent outer layer. Measurement and computed velocities show reasonable
agreement.
Therefore, from engineering point of view, a turbulent layer with the logarithmic velocity
profile covers the transitional layer, the turbulent logarithmic layer and the turbulent outer
layer.
At the viscous sublayer the effect of the bottom (or wall) roughness on the velocity distribution
was first investigated for pipe flow by Nikurase. He introduced the concept of equivalent grain
roughness ks (Nikurase roughness, bed roughness). Based on experimental data, it was found
u* ks
1. Hydraulically smooth flow: 5; bed roughness is much smaller than the thickness
of viscous sublayer. Therefore, the bed roughness will not affect the velocity distribution.
u* ks
2. Hydraulically rough flow: 70 ; bed roughness is so large that it produces eddies
close to the bottom. A viscous sublayer does not exist and the flow velocity is not
dependent on viscosity.
WU_KIOT- WRIE 13
GBU,CoET,WRIE By Koang G
River Engineering River Characteristics
u* ks
3. Hydraulically transitional flow: 5 70 ; the velocity distribution is affected by bed
in which u denotes average point velocity at a distance z from the boundary (the averaging is
related to turbulent fluctuations in time). C is a constant of integration, which must be
determined from the boundary conditions, requiring that close to the boundary the turbulent
and the laminar velocity distributions must join each other. On the other hand, since within the
laminar sublayer turbulent-flow conditions are no longer valid, the constant of integration may
be adjusted to give a zero velocity at some distance zo within the sublayer (see Fig. 1.11). The
distance zo is presumed chosen in such a way as to ensure a smooth blending to the profiles
somewhere in the transition zone.
Hence, when z = zo, u = 0, and the constant of integration is obtained C = -(u*/k (ln z0)). Eq
(1.16) can, therefore, be written in the form
u*
u (ln z ln z 0 ) (1.17)
k
Order of magnitude of the laminar sublayer is about υ/u*, υ denoting kinematic viscosity.
Accordingly, z0 = mυ/u*, in which m is a dimensionless constant. Eq. (2.17) can now be
written in the dimensionless form
u 1 z u*
(ln ln m) (1.18)
u* k
Finally, denoting A = l/k and B = -1/k(ln m), the resulting logarithmic equation is
u zu
A ln * B (1.19)
u*
Eq. (1.19) is generally known as the Prandtl – von Karman universal velocity distribution law,
valid for all types of conduits.
Laboratory studies have shown that velocity distribution is to a great extent influenced by the
roughness of the boundary. Consequently, distinct expressions are obtained for different ranges
of roughness conditions.
Z0
In order to derive a suitable expression of Eq. (1.17) for hydraulically rough boundaries, we
have to introduce the roughness height, Ks, having the dimension of length, Fig. 1.12.
WU_KIOT- WRIE 16
GBU, CoET,WRIE By Koang G.
River Engineering River Characteristics
K s u*
B ' 8.5 2.5 ln (1.21)
v
Hence, using again Eq. (2.19)
u zu K u
2.5 ln * 8.5 2.5 ln s * (1.22)
u* v v
Or finally
u z z
5.75 log 8.5 5.75 log 30 (1.23)
u* Ks Ks
For the transition zone between the smooth and completely rough rigid boundaries (~
5 K s u* / v ~70), velocity-distribution equation is found also to be of the general form
u z
5.75 log B' (1.24)
u* Ks
Here B' has not a constant value, but is according to Nikuradse a function of the dimensionless
parameter K s u* / v . Subsequent investigations have shown, however, that the transition curve
obtained by Nikuradse for uniform-grain distribution does not hold for non-uniform roughness
conditions generally encountered in engineering practice. Indeed, transition curve giving B'
values for non-uniform roughness seems to be much more gradual.
A more simplified approach regarding velocity distribution for flow over rigid boundaries was
later proposed by H.A. Einstein. It is supposedly valid for all three boundary conditions
(hydraulically smooth, rough or in transition region), and is given by
u z
5.75 log ( 30.2 x) (1.25)
u* KS
Roughness height Ks denotes here d65, i.e. grain diameter determined so that 65% are equal to
it or smaller, and x is a correction factor, which depends on the ratio K s / , where stands
for the thickness of the laminar sublayer, Fig 1.13.
Figure 1.13: Correction factor x in the logarithmic velocity- distribution Eq. (1.25)
By integrating Eq. (2.25) over the vertical section, Einstein, following previous Keulegan’s
deductions, has proposed expressions for the average velocity in turbulent flow,
Smooth boundary:
V Ru *
5.75 log 3.67
u* v
Rough boundary and transition
V R
5.75 log 12.27 x (1.26)
u* Ks
K su*
Rough boundary ~ 70
v
WU_KIOT- WRIE 18
GBU,CoET,WRIE By Koang G.
River Engineering River Characteristics
K su*
Transition 3.6 ~5 ~70
v
In order to obtain the thickness of the laminar sublayer, it is assumed that at the boundary
between the laminar and turbulent layer, the shear stress of the laminar flow, (du / dz), is
equal to that of the turbulent flow. Hence
du
u* 2 (1.27)
dz
Let us next denote the velocity at the boundary between the two layers as uo; then du/dz may
be approximately written as u o / , where is the thickness to the sublayer. If the velocity-
distribution curve within the thin laminar sublayer is represented by a straight line, then it is
possible to write
u z
(1.28)
uo
Experiments have shown that the thickness is related to a constant value of the
dimensionless parameter u o / v const., and the numerical value of the constant has been
found to be about 135. Since within the sub-layer (du / dz) (u o / ), it follows that
u o ( / ); substituting this into the former expression for the constant parameter, it can be
11.6 (1.29)
u*
In order to compare the thickness of the sublayer with the height of a roughness element, let us
write Eq. (1.29) in a different form,
Ks 1 K s u*
(2.30)
11.6 v
If now for smooth boundary k s u* / v 3.6~5, it follows that
Ks 1 1 1
(3.6 ~5) ~ (1.31)
11.6 3 .2 2 . 3
For completely rough boundary, the parameter K s u* / v ~70. Hence similarly,
Ks 1
70 6 (1.32)
11.6
WU_KIOT- WRIE 19
GBU,CoET,WRIE By Koang G.
River Engineering River Characteristics
In the transition region the ratio K s / must be between the limits given by Eqs. (1.31) and
(1.32).
From the above approximate comparisons it can be concluded that for smooth boundaries the
thickness of the laminar sublayer is not less than about three times the roughness height Ks. For
rough boundaries, the thickness is not more than about one sixth of the roughness height.
A logical analysis of the dynamic equilibrium conditions in the turbulent flow leads to the
conclusion that the energy spent on the entrainment of sediment particles by the water should
somehow damp the turbulent momentum-transfer mechanism, and hence the random velocity
fluctuations, as compared to the clear water. Following this reasoning, one is likely to conclude
that sediment particles should lag behind the water particles, and that their turbulent mixing
length should be shorter on the average. It implies as well that the von Karman universal
constant for turbulent flow k, so far assumed to be 0.4 would be reduced.
Laboratory experiments reported by the ASCE Task Committee show a reduction of the
constant k as large as about 50% for a suspended sediment concentration of 15,800 ppm, see
Fig. 1.14. The diagram clearly shows that for the same depth, slope and bed surface, velocity
for sediment-laden stream is greater than for the clear water. Taking into account Eq. (1.16),
this implies that the constant k must be smaller.
However, these and more recent experiments have not been conclusive, since they also have
shown that this reduction of flow resistance can be more than offset by the increase in
resistance caused by bed formations. This fact may well explain many field observations which
WU_KIOT- WRIE 20
GBU,CoET, WRIE By Koang G.
River Engineering River Characteristics
tend to indicate that sediment-laden water brings about more resistance to flow, and not less as
suggested by carefully executed experiments.
While experiments by Vanoni et al., suggest that for clear water the value of the constant k
always tends to be close to 0.4, whatever the bed formation, some later experiments, cited by
the Task Committee, seem to indicate that the bed formations may reduce k-values even for
clear water flows. Some field observations have also hinted that the sediment-transfer
coefficient may in some cases be greater than the momentum-transfer coefficient for water
particles.
Figure 1.14: Comparison of the velocity profiles for clear water and suspended sediment, after
ASCE Task Committee; d = 9 cm, ds = 0.1 mm. S = 0.0025.
Einstein et al. have tried several times to experimentally determine the variation of the von
Karman coefficient k in dependence of characteristic sediment parameters, and subsequently to
derive an equation for the velocity distribution in alluvial streams. These equations have been
unduly complicated for engineering applications. Adopting the view that there should exist an
average value of k for both the velocity and the suspended-load distributions down to a point
close to the bed, Einstein and Abdel-Aal finally proposed an experimental correlation curve
between the coefficient k and sediment parameters (see Fig.1.15). Here w35 denotes the settling
velocity for ds35 –size of sediment (35% by mass smaller than the given size); ds65 – grain size
for which 65% by mass of all grains are smaller; q _ specific discharge (discharge per unit of
width); J – hydraulic gradient.
WU_KIOT- WRIE 21
GBU,CoET, WRIE By Koang G.
River Engineering River Characteristics
Here x denotes the correction factor to be taken from the diagram on Fig. (1.13). Average
velocity is expressed in a similar manner,
V 2.3 d
log 30.2 .x (1.34)
u* k ks
WU_KIOT- WRIE 24
River Engineering River Characteristics
WU_KIOT- WRIE 25
GBU,CoET,WRIE By Koang G.
River Engineering River Characteristics
4. Deltaic stage: is the last stage of the river just before it discharge into the sea.
The river is unable to carry its sediment load. As a result, It drops its sediments and
gets divided into channels on either side of the deposited sediment and form the
delta.
WU_KIOT- WRIE 26
GBU,CoET, WRIE By Koang G.
River Engineering River Characteristics
The behaviors of alluvial rivers depend to a large extent on the sediment carried by it. The
sediment carried by the river poses numerous problems, such as:
increasing of flood levels
sitting of reservoirs
silting of irrigation and navigation channels
splitting of a river into a number of interacted channels
meandering of rivers
Especially the meandering causes the river to leave its original course and adopt a new course.
An alluvial river usually has the following three stages:
1. flow in a straight reach
2. flow at bends
3. development of meanders
1. Flow in a straight reach: the river cross section is in the shape of a trough, with high
velocity flow in the middle of the section.
Since the velocity is higher in the middle, the water surface level will be lower
in the middle and higher at the edges. (see fig.)
Due to the existence of this transverse gradient from sides towards the center,
transverse rotary currents get developed. However,
straight reaches are very few in alluvial channels.
WU_KIOT- WRIE 27
GBU,CoET,WRIE by Koang G.
River Engineering River Characteristics
Process: when the flow moves around a bend, a centrifugal force is exerted upon the water,
which results in the formation of transverse slope of water surface from the convex edge to the
concave edge, creating greater pressure near the convex edge. To keep its level, water tends to
move from the convex side towards the concave side. However, the top most water surface
movement is prevented by the centrifugal force. Moreover, towards the bottom, the velocities
are much less than towards the top; and enough centrifugal force is not available to counteract
the tendency of water at the top to move inwards. Hence, the water dives in, from the top at the
concave end, and moves at the bottom towards the convex end. These rotary currents cause the
erosion of concave edge and deposition on the convex edge forming shoal on this edge. When
once the bend forms, it tends to make the curvature large and larger.
WU_KIOT- WRIE 28
GBU,CoET,WRIE By Koang G.