EROSION, SEDIMENTATION

AND THE RIVER BASIN

DIMAS, ARIANE SACHIKO
ASENTISTA, JUNE HYAME B.
GUADALUPE, HAIDE B.
MANALASTAS, ARVIN REY L.
Reporters
The erosion and transport of sediment by water is a key
process in shaping a river basin, and has important
economic and environmental consequences. Predicting
rates of erosion and transport of sediment and its
deposition on the floodplains or in reservoirs, lakes and
estuaries is one of the tasks of the hydrologist. It also
presents an introduction to geomorphology- the study
of the formation of the landscape- insofar as it is
influenced by water.
11.1 Physical Descriptors of Catchment Form
Stream order. Horton suggested a classification of
stream order as measure of the amount of branching
within a basin. A first-order stream is a small,
unbranched tributary. A second-order stream has only
first-order tributaries. A third-order stream has only
first- and second-order tributaries. The order of a
particular drainage basin is determined by the order of
the principal stream.
 LAW OF STREAM NUMBERS
𝒌−𝒖
𝑵𝒖 = 𝒓𝒃 eq.1
𝒖−𝟏
𝑳𝒖 = 𝑳𝟏 𝒓𝒆 eq.2
𝒖−𝟏
𝑨𝒖 = 𝑨𝟏 𝒓𝒂 eq.3
Drainage density. The total lengths of streams within a
catchment divided by the drainage area defines the
drainage density, the length of channels per unit area. A
high drainage density reflects a highly dissected basin,
which should respond relatively rapidly to a rainfall input,
while low drainage density reflects a poorly drained basin
with slow hydrologic responses. Low drainage densities are
observed where soil materials are resistant to erosion or
very permeable and where the relief is small.
 The average length of
Length of overland flow.
overland flow 𝐿𝑜 may be approximated by
𝟏
𝑳𝒐 = eq.4
𝟐𝑫
𝑺𝒄
Horton suggested that the denominator be multiplied by 𝟏− where
𝑺𝒈
𝑆𝑐 and 𝑆𝑔 are the average channel and the ground slopes, respectively.
Area relations. Data for a number of the larger
rivers of the world seem to perform to the
equation
𝟎.𝟔
𝑳 = 𝟏. 𝟐𝟕𝑨 eq.5
Basin shapes. The shape of the catchment affects the streamflow
hydrograph and pear-flow rates. Numerous efforts to develop a factor
which describes basin shape by a single numerical index have been
reported. Basins tend to the form of a pear-shaped ovoid, but
geologic controls result in many substantial deviations from this
shape. Horton suggested the dimensionless form factor 𝑅𝑓 as an
index of shape,
𝑨
𝑹𝒇 = eq.6
𝑳𝟐𝒃

For a circle 𝑅𝑓 = = 0.79; for a square with the outlet at the
4
midpoint of one side 𝑅𝑓 = 1; and for the square with the outlet at
one corner 𝑅𝑓 = 0.5.
 11.2 Descriptors of Catchment Relief
Channel slope. Typical channel profiles are concave upward. In addition all
but the very smallest basins contain several channels is difficult. Taylor
and Schwarz calculated the slope of a uniform channel having the same
length and time of flow as the main channel. Since the velocity is
proportional to square root of slope, the procedure used by Taylor and
Schwarz is equivalent to weighting channel segments by the square root
of their slope, which gives relatively less weight to the steep upstream
reaches of the stream.
σ𝒊=𝒏
𝒊=𝟏 √𝑺𝒕 𝟐
𝑹𝒔 = ( ) eq. 7
𝒏
Land slope. The slope of the ground surface is a factor in the
overland-flow process and hence a parameter of hydrologic
interest, especially on small basins where the overland-flow
process may be a dominant factor in determining hydrograph
shape.
The distribution of land-surface slope can be determined by
establishing a grid or a set of randomly located points over a map
of the catchment. The slope of a short segment of line normal to
the contours is determined at each grid intersection or random
point. The mean, median and variance of the resulting distribution
can be calculated.
Area-elevation data. An area-elevation curve can be distributed
by planimetering the area between contours on a topographic map
and plotting the cumulative area above (or below) a given
elevation versus that elevation.
Snowmelt computations in mountainous areas must usually
be made for elevation zones because both snow depth and
temperature are functions of elevation. Precipitation is
mountainous areas may sometimes be weighted by elevation in
calculations of basin average precipitation.
Aspect. The aspect of a slope is the direction toward which the
slope faces. Because of the influence of insolation on snowmelt,
aspect may be interest in dealing with snow. Precipitation
amounts are often influence by the aspect of a slope relative to
the direction of the wind. Customarily, aspect is used as a
characteristic of a particular point or at most of a specific
hillside. The distribution of aspect may be determined in
manner similar to the described for land slope. A grid or set of
random points is superimposed over a map of the catchment
and the bearing of a line normal to the contours at each
intersection is recorded.
 11.3 Hydraulic Geometry
Hydraulic geometry describes the character of the
channels of a basin: the variation of mean depth, top
width, and velocity at a particular cross section and
between cross sections. These relations apply to
alluvial channels, where the cross section is readily
adapted to the flows which occur, but are less reliable
where rock outcrops control the channel
characteristics.
The basic equations of hydraulic geometry are
𝒃
B = a𝒒 eq. 8
D = 𝒄𝒒 𝒇 eq.9
𝒎
v = 𝒌𝒒 eq.10
Table 11-1 present values of the exponents b, f and m determined by various
investigators
 11.4 Stream Patterns
When viewed in plan, stream channels may be describe as
meandering, braided or straight. A meandering stream flows in large, more or
less symmetrical loops or bend. The median length of the meandering
streams appears to be about 1.5 times the valley length [13]; i.e., the
sinuosity averages about 1.5. the wavelength of meanders ranges from 7-11
times the channel width, and the radius of curvature of the bend usually
ranges between 2 and 3 times the channel width. The amplitude of the
meanders or width of the meander belt varies considerably and seems to be
controlled more by the characteristics of the bank material than by other
factors. Amplitude usually ranges from 10 to 20 times channel width.
 A braided stream consists of many intertwined channels
(anabranches) separated by islands. Braided streams tend to
be very wide and relatively shallow with coarse bed material.
No formal statements about the geometry of braided streams
are possible. Few long straight channels exist in nature, but
many lack sufficient curvature to be called a meandering
stream. A straight stream is commonly defined as one with a
sinuosity of less than 1.25.
 Braided channels are usually found in reaches where the
banks are easily erosible-sandy material with little vegetal
protection. Bed material is relatively coarse and of
heterogeneous particle sizes. The slope of the braided reach is
greater than that of adjacent unbraided reaches. Hydraulically,
the braided reach is less efficient than the unbraided reach.
The total width of branches in a braided reach may be 1.5 to 2
times that of an undivided channel and the depth of flow is
correspondingly less. Braided is thus a way of dissipating
energy when stream slope steepens. Velocity increases that
would otherwise lead to erosion are thus avoided.
 An initially straight channel, either in a laboratory flume
or in the field, will usually develop meander as water flow
through if the bank material is erosible. A meandering channel
may be 1.5 to 2 times as long as a nonmeandering channel. Its
slope is correspondingly reduced, but head losses are
increased both because of the longer channel length and
because of the bend losses. Without these losses, velocities
would be higher, with corresponding tendency to downcut the
channel. Many meandering streams cannot downcut cannot
occur, some other device is required to dissipate the available
energy.
 11.5 Floodplains
The floodplain of a river is the valley floor adjacent to
the incised channel, which may be inundated during high
water. The river tends to swing back and forth across the
valley bottom, reworking the floodplain deposits and eroding
first one valley side and then the other. Floodplains are built
up primarily from deposition of sediment in the river channel
and deposition of fine sediments on the floodplain when
flooded. Additionally, organic material may accumulate in
cutoff meander loops (oxbow lakes).
 11. 6 The Erosion Process
Soil can be eroded, i.e., moved from its current location, by
the action of wind, water, gravity(landslide) and human activity,
water erosion may be viewed as starting with the detachment of
soil particles by the impact of raindrops [(16)]. The kinetic energy
of the drops can splash soil particles into the air. The splash and
overland-flow processes are responsible for sheet erosion, the
relatively uniform degradation of the soil surface. Sheet erosion is
difficult to detect except as the soil surface is lowered below old
soil marks on fence post, tree roots are exposed, or small pillars of
soil capped by stone remains.
 Raindrops vary in diameter d from 0.5 to 6 mm
(0.02 to 0.25 in) and terminal velocity v varies with
diameter from about 2 to 9 m/s (7 to 30 ft/s). Since
kinetic energy is proportional to d³v², the erosive power
of the largest drops may be 10,000 times that of the
smaller. This conforms with the observation that a few
intense storms account for most of the erosion. The
effect is augmented by the fact that overland flow is
more likely to occur during intense rains.
 Mass movement of soil, as either the slow
downward creep of the soil mass or the rapid collapse
of a slope (landslide) is an important mechanism
delivering soil to the streams in steep canyons with
unstable side slopes. Landslides occurring as the result
of earthquakes or the saturation of the slopes during
heavy rain may create temporary dams whose
subsequent overstopping and erosion may create flood
waves downstream.
 11.7 Factors Controlling Erosion
A cohesive soil will resist splash erosion more
readily than loose soils. Generally, splash erosion
increases with an increasing fraction of sand in the soil
because of the loss of cohesion. Splash erosion
decreases with an increasing percentage of water-
stable aggregates. A soil whose individual grains do not
tend to form aggregate will erode more readily than
one in which aggregates are plentiful.
 Rates of erosion are greater on steep slopes than
on flats slopes. The steeper the slope the more
effective splash erosion is in moving soil downslope.
Overland-flow velocities are also greater on steep
slopes, and mass movements are more likely to occur
in steep terrain. Length of slope is also important. The
shorter the slope length, the sooner the eroded
material reaches the stream, but this is offset by the
fact that overland-flow discharge and velocity increase
with length of slope.
 Land use is also important factor in fixing the rate of
erosion. Poor cropping practices or careless construction of
roads may greatly accelerate erosion. Removal of vegetation
by fire or lumbering may also increase the erosion hazard.
Proper soil conservation practices may greatly reduce erosion
losses. The universal soil loss equation [17] attempts to
combine all these factors, but it is difficult to express the
rainfall regime in a single index number, and field
determination of soil erosibility is not yet generally available.
Hence, this equation and others of a similar nature are, a best,
approximate.
 11.8 Suspended-Sediment Transport
Sediment moves in the stream as suspended sediment in the
flowing water and as bed load, which slides or rolls along the
channel bottom. A third term, saltation, is used to describe the
movement of particles which seem to bounce along the bed. The
process are not independent, for materials which appears as bed
load at on section may be suspension at another. Another useful
distinction is between bed-material load, represented by those
particles of grain size normally found in the stream bed. Wash load
consist of the fine material washed into the stream during rainfall
and which normally travels through the system without
redepositing.
The settling velocity of suspended particles in still
water is approximated by Strokes’ law:

𝟐 𝝆𝒈 −𝝆 𝒈𝒓𝟐
𝑽𝜹 = eq. 11
𝟗𝝁
Where 𝜌𝑔 and 𝜌 are densities of the particle and the liquid,
respectively,𝑟 is the radius of the particle, and 𝜇 is the absolute
viscosity of the water.
 The general two-dimensional nonequilibrium equation for
suspended-sediment transport is
𝝏𝒄𝜹 𝝏𝒄𝜹 𝝏𝝐𝒙 𝝏𝒄𝜹 𝝏𝝐𝒚 𝝏𝒄𝜹 𝝏𝟐 𝒄𝜹 𝝏𝟐 𝒄𝜹
𝒗 = 𝒗𝜹 + + + 𝜺𝒙 𝟐 + 𝜺𝒚 𝟐 eq. 12
𝝏𝒙 𝝏𝒚 𝝏𝒙 𝝏𝒙 𝝏𝒚 𝝏𝒚 𝝏𝒙 𝝏𝒚

Where 𝑐𝛿 is the sediment concentration for a particular size of

particle, 𝑣𝛿 is settling velocity, 𝜀 is a mixing coefficient, and 𝑥 and 𝑦
are longitudinal and vertical dimensions, respectively
 11.9 Bed-Material Transport
For many years analysis of bed-load transport has been based
on the classical equation of du boys [19]
𝑻𝒐
𝑮𝒇 = 𝒀 𝑻𝒐 − 𝑻𝒄 𝑒𝑞. 13
𝒘
Where 𝐺𝑓 is the rate of bed-load transport per unit width of
stream, 𝑌 is an empirical coefficient depending on the size and
shape of the sediment particles, 𝑤 is the specific weight of
water, 𝑇𝑜 is the shear at the streambed, and 𝑇𝑐 is the
magnitude of shear at which transport begins.
 Accuracy of instruments for bed-load measurement is so
uncertain that field comparison of bed-load measurement is
difficult. The validity of bed-load formulas is therefore quite
indefinite. Recent work on the bed-load problem has utilized
the concepts of turbulent flow and statistical variation of fluid
forces at a point. A widely used approach is that of Einstein
[21], who defines the intensity of bed-load transport as

𝑮𝒇 𝝆 𝒍
∅=
𝒘 𝝆𝜹 − 𝝆 𝒈𝒅𝟑
𝑒𝑞. 14
And the 𝑓𝑙𝑜𝑤 𝑖𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦 as
𝝆𝜹 − 𝝆 𝒅
𝝋= eq. 15
𝝆 𝒔𝑹
Where 𝑤 is the specific weight of water, 𝜌 is the density of
water, 𝜌𝛿 is the density of the bed material, 𝑑 is the grain
diameter, 𝑠 is the channel slope, and R is the hydraulic radius.
Einstein divided the total hydraulic radius in portions
depending on grain roughness 𝑅1 and on bedforms, 𝑅 𝑛 . In
practice R is frequently taken as the total hydraulic radius of
the channel or the mean depth. An empirical relation between
∅ 𝑎𝑛𝑑 𝜑 (fig. 11-9) permits the solution of eq.
 11.10 Sediment Measurement
Early suspended-sediment observations were made with
open bottles or complex grab samplers, which failed to
provide adequate data for a number of reasons. A good
sampler must cause minimum disturbance of streamflow,
avoid errors from short-period fluctuations in sediment
concentration, and give results which can be related to
velocity measurements. These requirements seem to be met
in a series of sampler [22] designed at the Iowa Hydraulic
Laboratory under the sponsorship of several federal agencies.
Figure 11-10 The DH-48 depth- integrating hand sampler for use on small
streams.
 The collected samples are filtered and the
sediment dried. The ratio of dry weight of sediment to
total weight of the sample is the sediment
concentration, usually expressed in parts per million or
milligrams per liter. Other analyses which may be
performed include determination of grained-size
distribution, fall velocity and occasionally, heavy-
mineral or chemical analysis. The latter tests may be
useful in tracing the original source of the sediment.
 11.11 Sediment-Rating Curves
Sediment measurements, like current-meter measurements, give
only occasional samples of the sediment discharge. A sediment-rating
curve relating suspended-sediment discharge and water discharge (fig.
11-11) is commonly used to estimate sediment load on days when no
measurements are available. Areal distribution of runoff may be a
factor if different portions of the basin are more prolific sediment
sources than others. Sediment-rating curves should be used with
caution and where possible applied to small and relatively
homogeneous basins.
Figure 11-11. Sediment-rating curve for the Powder River at Arvada, Wyoming
 11.12 Sediment Yield of a Catchment
The average annual sediment production from a
catchment is dependent on many factors such as climate, solid
type, land use, topography, and the presence of reservoirs.
Langbein and Schumm used data from a number of basins to
construct the curve of Fig. 11-12, which relates average annual
sediment production per unit area to mean annual
precipitation. Maximum production rates occur at about
300mm (12 in) of mean annual precipitation because such
areas usually have little protective vegetal cover.
 Fleming [30] utilized data from over 250 catchments
about the world to derive relations (eq. 11-16) and table 11-3
for mean annual suspended load 𝑄𝑠 in tons as a function of
mean annual discharge in cubic feet per second for various
vegetal covers:
𝑸𝒔 = 𝒂𝑸 𝒉 𝑒𝑞. 16
Errors of ±50 percent may be accepted from these relations.
 For catchments without sediment records the relations
of Table 11-3 may be viewed as offering an order-of-
magnitude estimate of sediment yield. If possible, such
estimates should be compared with sediment data on similar
basins in the same region.
 11.13 Sediment Simulation
Negev [31] developed a model for suspended-load
transport and tested it with good results. In this model the
amount of soil splash R is made a function of the hourly
precipitation amount i
𝒋
R= 𝑲𝟏 𝒊 eq. 17
The transport of splash residue S is a function of the residue in
storage on the ground surface 𝑅𝑠 and the overland flow rate 𝑞0
𝒌
S= 𝑲𝟐 𝑹𝟐 𝒒𝒐 eq. 18
And the sediment washed from impervious areas E is
E= 𝑲𝟑 𝑹 eq. 19
The total wash load then becomes
W= R + S + E eq. 20
Unless overland flow occurs, the only source of wash
load is from impervious areas and is very small.
 Gully erosion G is also related to overland flow
𝒎
G= 𝑲𝟒 𝒒𝟎 eq. 21
Negev divided the sediment from gully erosion into two
portions. The first B has particle size substantially the same as
the bed-material lead in the stream
B= ( i- 𝑲𝟓 )G eq. 22
And the second portion, which he called interload I,
represents material finer than about 95 percent of the bed-
material load:
I = G-B = 𝑲𝟓 𝑮 eq. 23
The total suspended-load transport 𝑞𝑠 is then
𝒒𝒔 = 𝑲𝟔 𝑰𝒔 𝒒𝒏 + 𝑲𝟕 𝒒𝒓 eq. 24
Where q is the mean daily streamflow and 𝐼𝑠 is the quantity of
interload in storage in the streambed calculated by
maintaining a running balance of the input from erosion and
the outflow of suspended interload material.
 The procedure involves a number of coefficients
which must be found by calibration. The exponent j can
be assumed as 3.0, and k and m as 2.5. Negev assumed
n = r = the slope of a sediment-rating curve for the
station. The coefficient 𝐾5 can be estimated from
particle-size data. The overland flowrate is calculated
with a flow-simulation model.
 11.14 Reservoir Sedimentation
The rate at which the capacity of a reservoir is
reduced by sedimentation depend on (1) the quantity
of sediment inflow, (2) the percentage of this inflow
trapped in the reservoir, and (3) the density of the
deposited sediment. The quantity of sediment inflow
may be estimated by any methods discussed in Secs.
11.11 to 11.13, or if data are available, by reference to
mean-annual-yield data per unit area from similar
basins in the region.
 Table 11-4 presents some selected values of sediment
yield derived from reservoir surveys. These data are generally
obtained b surveying the reservoir with sounding lines or
echo-sounding equipment and are published periodically [32].
 By comparing sediment accumulation in reservoirs with
estimates of sediment inflow based on measured sediment
transport, Brune [33] derived a relationship between reservoir
trap efficiency, the percent of incoming sediment retained in the
reservoir, and the ratio of reservoir capacity to mean annual
water inflow. Trap efficiency must increase with the residence
time of the sediment-laden water in the reservoir. Figure 11-13,
which is modified from Brune [33] can be used to estimate the
fraction of the sediment inflow which is trapped. As the reservoir
is filled with sediment, the trap efficiency will decrease so that it
may be necessary to make the computation for several time
intervals with appropriate adjustment of the trap efficiency.
 The volume occupied by the sediment in the reservoir
will depend on the specific weight of the deposited material.
The specific weight varies with the kind of sediment and the
age of the deposits. Older sediments have more time to
consolidate and are under a superimposed load from the
more recent deposits. Lane and Koelzer [34] found that dry
specific weight 𝑤𝑡 at time t can be defined by
𝒘𝒕 = 𝒘𝟏 + K log t eq. 25
where 𝑤1 is the initial specific weight and K is a consolidation
coefficient (Table 11-5).
 If a mixture of materials is present in the sediment, a weighted
average specific weight should be calculated. Equation 25 applies to
each annual accumulation sediment, and the average weight of the
total sediment accumulation at time t must be found by integrating
from year 1 to year t. Table 11-6 presents average specific weights
after 50 years used by the U.S. Soil Conservation Service for general
design purposes.
Illustrative example 1:
A reservoir has a capacity of 6x106 𝑚3 and a drainage area
of 200 𝑘𝑚2 . Streamflow averages 350mm of runoff per year and
sediment production is estimated an average in-place density of
1500 kg/𝑘𝑚3 , how long will it take to reduce the reservoir
capacity to 1x106 𝑚3 ?
Annual inflow= 0.35x200x𝟏𝟎𝟔 = 70x𝟏𝟎𝟔 𝒎𝟑
𝟏𝟏𝟎𝟎𝒙𝟏𝟎𝟎𝟎𝒙𝟐𝟎𝟎
Annual sediment inflow= = 𝟏𝟑𝟑, 𝟎𝟎𝟎 𝒎𝟑
𝟏𝟓𝟎𝟎
We will calculate the time it takes to fill successive increments of one
million cubic meters of storage.
Note that if the mean value of the trap efficiency had been
used and a single increment of 5x106 𝑚3 considered, the time
would have been
6
5x10
= 𝟒𝟗. 𝟓 𝒚𝒆𝒂𝒓𝒔
133,000 𝑥 0.76
The difference in two answers is undoubtedly less than the
uncertainty in the input data.
Illustrative example 2:
A sediment consists of 5 percent sand, 43 percent silt,
and 52 percent clay. What is the in-place density of this
material after 30 years assuming the sediment is always
submerged? From equation 25, the first line of Table 11-5, and
the sediment composition data:

w= 0.05 x 93 + 0.43(65 + 5.7 log 30) + 0.52(30 + 16 log 30)

w= 0.05 x 93 + 0.43(65 + 5.7 x 1.477) + 0.52(30 + 16 x 1.47)
w= 4.65 + 31.57 + 27.89
w= 64.11 𝒍𝒃ൗ𝒇𝒕𝟑