B-11 Fourth International Conference on Scour and Erosion 2008

RIVER BANK EROSION AND SUSTAINABLE

PROTECTION STRATEGIES

Md.. Shofiul Islam1

1 Scientific Officer, River Research Institute (RRI)
(Faridpur 7800. Bangladesh)
E-mail:[email protected]

Bank erosion is the most common problem faced in river engineering practices in many countries,
especially in Bangladesh and this has been recognized as an awful threat to the society. So control of
erosion is very much important to save agricultural land, property and infrastructures like bridges, culverts,
buildings etc. located alongside the rivers. Treatment of bank erosion is usually expensive and is not
particularly susceptible to engineering analysis. Because diagnosis and prediction of bank erosion is often
difficult as it may be caused by a number of factors operating separately or together. Engineers play a
dominant role to design the optimum bank protection works, which sounds justified both technically and
economically. In this paper the probable causes, mechanisms and methods of predictions of bank erosion
and sustainable strategies of different bank protection measures with special attention to economic analysis
are briefly discussed.

Key Words : river bank erosion, sustainable strategies, bank protection works

1. INTRODUCTION

Engineers play important roles for poverty

alleviation in many ways, e.g. by designing, planning
and implementing infrastructures. As the government
fund is limited and there is always a shortage of
money, there should have limitations in spending
money for public investments because there may be
more hospitals, schools, roads, bridges and so on.
Fig.1 Typical Bank erosion along the Jamuna Right Bank
Due to the geometric location flood is the
recurring problem in Bangladesh, which causes
serious devastation to the property and life alongside source of sediment deposited at the downstream
the river. Bank erosion not only causes damage to the backwater areas. The area dominated by significant
immediate ‘blow-out’ site, but also responsible for sedimentation also simultaneously expands the
the loss or damage to valuable farmland, wildlife cross-sectional area by widening due to side bank
habitat, buildings, roads, bridges, and other public erosion. So, the riverbank erosion control or
and private structures and property. In economic riverbank protection in a sustainable manner is
point of view, mitigation of bank erosion in necessary to save the expected losses. Sustainability
Bangladesh has become an integral part of poverty may be ensured by proper diagnosis and predictions
reduction. The severity of bank erosion may be of bank erosion and treatment applied to select
realized from Fig.1. Having considered the extreme technically and economically justified protective
severity, bank erosion assessment is the prime measures.
objective prior to take sustainable protective
measures against erosion. Bank erosion is the major

316
2. CAUSES OF BANK FAILURE rate of bank retreat. Among them, piping due to
seepage is an important one.
The riverbank undergoes to erosion by hydraulic
and geo-technical instability. Hydraulic instability is Bank erosion phenomenon in large-scale rivers is
caused by scour at the toe of a marginally stable bank, a cyclic process of the following four sub-processes:
flood propagation and flood recession, debris and
vegetation, removal of bank vegetation, detachment 1) Steepening of bank slopes caused by erosion of
of coarse sediment by wave action, secondary current lower part of the slope and riverbed near the
etc. Besides, constricted bridge crossings or other slope toe.
encroachments that involve acceleration and 2) Slip failure of the steepened bank slope by losing
concentration of flood flows tends to cause ‘back its stability.
eddies’ or reverse circulation downstream, which can 3) Movement of failed bank material toward
sometimes erode huge embankments into river bends. riverbed along the slip surface.
Local bank protection and river training works
designed to protect against bank erosion at one point 4) Removal of failed bank material on lower part of
or reach of a river often provoke accelerated bank the slope and riverbed near the slope toe.
erosion elsewhere. The shearing of bank material by Erosion that steepens bank slope occurs locally
hydraulic action at high discharges is a most effective because of sediment continuity, in other words,
process, especially on non-cohesive banks and because channel widening by longitudinally uniform
against bank projections. Large scale eddying erosion must raise riverbed at the equivalent amount
induced by bank irregularities can enlarge existing of sediment produced by bank erosion. The local
embankments and increase the amplitude of steepening of bank slope is considered to take place
projections, which become more susceptible to due to flow concentration following channel plan
subsequent attack. Geo-technical instability is caused forms and bed topographies.
by detachment of more coarse grained layers in any
given alluvial bank, by water flowing out of the bank 4. FACTORS TO CONSIDER IN BANK
face, termed as ‘piping’ or ‘sapping’ (Hagerty and EROSION ASSESSMENT
Hamel, 1989). Cohesive banks particularly
susceptible to seepage force and piping mechanisms The following factors should consider when
that may so lower the internal resistance of the bank erosion is assessed for planning protection
material as to induce failure. Whereas the lower bank works:
is eroded by hydraulic action, the upper bank is less
affected by flow forces but fails because of 1. Hydraulics (stage-discharge, flow structures, flow
undercutting which produces different types of resistance, maximum near-bank velocity, distribution
cantilever action in the cohesive material. River stage of shear stress, secondary currents and turbulence,
drawdown following floodplain inundation water level variations).
contributes to riverward creep and/or sliding of
alluvium as does riverward seepage and consequent 2. Morphology (Riverbed deformation by computing
piping. bed shear stress, bed topography, channel planform,
migration of bar and bed shear stress).
3. BANK EROSION & MECHANISM
3. Sediment transport (suspended sediment, bed load,
The traditional perception is that banks fail by wash load).
basal scour. The shear stress associated with the
water flow along the bank corrodes the toe of the 4. Stability of banks and riparian structures
bank, which steepens the bank, makes the slope (vegetation, bank angle, critical bank height is a
unstable, and, eventually, gives rise to structural function of bank angle, tension crack depth).
failure of subaqueous and upper banks. This
perception is the basis of several mathematical
models of bank erosion in which rate of bank retreat 5. Soil properties (size, gradation, stratification of
is assumed to be directly related to near-bank flow bank sediment, bulk density, friction angle, cohesion
velocity and shear stress (Darby and Thorne, 1996; etc.).
Throne and Osman, 1988). Several studies suggest
that there are other hydraulic factors than shear stress
exerted on banks, which may also significantly affect

317
5. BANK EROSION PREDICTIONS 1975).
It is seen in the previous study that correlation
The prediction of future rates and direction of bank exists between the dimensional parameters as
erosion along a river is difficult problem that arises in Eh Ω
=k (4)
many engineering applications. In natural rivers, the w γb
best guide to future patterns of bank erosion is a local
Measurements on the Beatton River in the United
study of past patterns. Topographic maps and
States clearly show that bank erosion rates are
satellite images of several years, supplemented by
strongly controlled by bend curvature.
local witnesses, are usually the best sources of
information. However, the satellite images, due to
The lateral erosion of cohesive riverbank is given
their scale and resolution limitations, gives
by Arulanandan et al. (1980) and may be expressed
qualitative results to some extent. Although bank
as (Osman and Thorne, 1988)
erosion is quite a complicated process, over the years
R(τ − τ c )∆t
a number of methods were developed to predict the ∆w = ; R = 0.0022τ c e −0.13τ c (5)
bank erosion rates. One of the methods is related to ντ c
2D mathematical model to compare bank erosion on in which ∆w = bank erosion distance (m) in one bank;
the basis of local geometry, flow and sediment
τc = critical shear stress (dynes/cm2) for cohesive
processes (Mosselman, 1992 and DHI, 1996 as
soils; ν = soil unit weight (KN/m3) ; τ = average shear
described in DELFT/DHI, 1996). Other method
stress (dynes/cm2) and ∆t = computational time
estimates the yearly bank erosion rate on the basis of
interval in minutes.
(1) overall channel parameters (discharge, bank
material characteristics) and (2) local geometry For turbulent flow, shear stress (τ) is expressed
(Hickin and Nanson, 1984). In general, bank erosion by
(E) is the function of bed and bank material
du du
properties, geometry of the river, flow τ =µ + ρε (6)
characteristics. dy dy
du
Hickin and Nanson (1983) stated that the bank where, is the velocity gradient at depth y, u is the
dy
erosion rate, E is likely to depend on many variables
as local mean velocity at distance y from the boundary,
E = f(Ω, γb, h, Rc, w) (1) µ is dynamic viscosity, ρ is density of water, ε is eddy
viscosity.
in which Ω = stream power per unit bed area ; γb =
opposing force per unit boundary area resisting
Prediction of bank erosion rate along Jamuna
erosion (co-efficient of resistance to lateral migration,
N/m2 analogous to Manning’s n) dependent largely River, Bangladesh
The bank line migration of the Jamuna left bank
on bank strength; h = bank height; Rc = bend radius or
downstream of Bahadurabad calculated from SPOT
radius of curvature; and w = channel width.
images indicate that the bank erosion rates vary along
the bend and the maximum erosion occurred at the
Stream power is the rate of potential energy
downstream end of the bend (DELFT/DHI, 1996).
expenditure per unit length of channel discharge and
The maximum yearly bank erosion rate at the river
expressed mathematically by
Ω = γQi (2)
where, γ (= ρg) is the specific weight of water, Q is
discharge, i is slope. Eq.1 may be expressed by a
non-dimensional form as
Ω ⎛ h Rc ⎞
E=k ⎜ ; ⎟ (3)
γ b ⎝ w′ w ⎠
Rc
in which, k is a constant, used as a ratio to assess
w
bank erosion rate or the amount of hydraulic stress
Fig.2 Comparison of bank erosion rate
placed on the outside of the river bend (Bagnold,
estimates in relation to near bank velocity
1960; Ippen and Drinker, 1962; Hickin and Nanson,

318
bend downstream of Bahadurabad was about 800m, bank failure by slumping. The study revealed that the
which is classified as an extreme event of bank longest overall bank retreat has occurred near
erosion (Klaassen and Masselink, 1992). Within less Nakalia. In this reach the bank line has shifted
than three years the maximum retreat was westwards about 1.6 km since 1975. This has caused
approximately 2 km. The observed bank erosion rates serious loss of households as well as previous
in the river bend were compared with the different breaching of the South of Kaitala pump house. Bank
predictive methods. The method of computing bank erosion rates were computed from the digital satellite
erosion rate using near bank flow velocity by imagery by superimposing the successive bank lines.
Mosselman (1992) and DHI (1996) has been Channel migration and bank erosion rates were
compared with the observed bank erosion rate (Fig.2). computed using the method of Hickin and Nanson
This figure shows that the bank erosion rates (1975). Bank erosion rates averaged around
increases linearly with the near bank velocity as per 120m/year, (locally up to 250 m/year) during periods
Mosselman (1992), while bank erosion rates and near of aggressive bend development (1997-2000) and
bank velocity shows power relationship between about 50 m/year between 1975-2000.
them as per DHI. It is envisaged that DHI prediction
was complied with observed, where the rate was
moderate in Kamarjani location. However, the rate is
very high at the downstream of Bahadurabad, where,
DHI prediction is not accurate. It was concluded that
probably the flow structure in the bend and the effect
of bank erosion product in changing the planform is
very important. Klaassen and Masselink (1992)
noticed that applying the method of Hickin and
Nanson (1983) for predicting the bank erosion rate in
the Jamuna River yields very less bank erosion rate
than the observed erosion rate.

Failure due to piping in non-cohesive riverbank Fig.3 Bank shifting at Nakalia-Pechakhola

have been investigated experimentally by Odgaard et (1973-2001)
al.1989 in the laboratory. Relations between pressure River Research Institute,RRI study: A study
gradient and seepage rate, rate of erosion in the on the Right Bank erosion of the Jamuna River
non-cohesive soil has been established. Physical downstream of the confluence of Hurasagar and
models using fine sand gives indication of induced Jamuna based on satellite image analysis indicates
bank erosion as envisaged in the Jamuna Bridge the prominence of bank erosion in the study area
Study (RRI, 1998b). So, the laboratory method may (Fig.3). The total maximum bank erosion during
also be employed to predict bank line shifting and in 1975-2001 was estimated about 1.4 km towards west,
that case use of lightweight materials as bank which is comparable to the FAP study (RRI, 2002).
sediment gives precious result. The type of bank erosion often encountered in
meandering rivers is also present in the Jamuna River,
6. ESTIMATION OF BANK EROSION especially at Kamarjani lication (DELFT/HDI,
ALONG JAMUNA RIVER AT NAKALIA 1996).
PECHAKHOLA AREA
7. BANK PROTECTION MEASURES
FAP 24 study: FAP 24 (DELFT/DHI, 1996)
studied the Riverbank erosion of the Jamuna River in Numerous methods are available to control bank
details and reported the important factors causing erosion. The bank protection measures have been
erosion are flow, sediment transport, channel categorized as structural, non-structural and
geometry and bed topography, vegetation, ground biological protection measures. However, structural
water level and their spatial variation and bank measures are taken for long-term protection.
material properties. The flow exerts shear stresses Different measures are described in short as follows:
that can remove particles from the bank either via (a) Structural measures
'peeling off' or via mass movement. The flow in a
river bend attacks the toe of the riverbank, removing Generally, two major types of structural measures are
the sediment from the toe, resulting in an practiced, namely, hard material protection and
over-steepening of the riverbank and causing the barrier across the river.
4

319
Hard material protection flow velocity reduction by proper arrangements of
This method is known as ‘resistive bank the barrier. Some of the examples of barriers are: i)
stabilization method’. It works by resisting the force Groynes deflect flow away from the bank and reduce
of the stream. This is a discontinuous bank protection near-bank flow velocity by dissipating flow energy.
method; bank scalloping is expected between hard Groynes are commonly of RCC and earth-boulder
points. The main aim of this method is to protect mix types ii) Spurs are commonly of earthen, RCC or
bank toe by boulder, stone, cc blocks etc. It resists wooden logs types. Spurs are solid or permeable and
erosive flow of the stream and stabilizes the toe of the submerged or non-submerged types. iii) Vanes are
bank. Success depends on the ability of stone to constructed at the river bends to redistribute flow
launch into the scour hole. The weight of the stone velocity. iv) Submerged bend way weirs placed at
resists the geo-technical failure. This technique is the upstream of bend redirects the water flowing
fairly suitable for continuous bank protection, for along the eroding bank at an angle perpendicular to
small radius and/or high degree of curvature bends. the weir. When the weirs are angled upstream water
For toe protection by placing stones, boulders, C.C. is directed away from the outer bank and toward the
blocks, riprap etc. bank shaping is important. inner bend. The stream’s strongest secondary current
Grading banks to more gradual slope does bank (helical flow) in the bend is broken up. Fig.5 shows
shaping. Toe of the concave bank posed to higher rate different types of barrier across the river.
of erosion and as a result convex bank needs less
protection than concave banks. Some examples of Inner bank
hard material protections are: i) Revetment C.C.
blocks are placed on the slope to protect the bank
continuously from erosion and geo-textile filters are ne
roy
placed underneath the blocks to protect from seepage fg
ieso
r
failure. ii) Guide bunds constructed at bridge Se
Outer bank
crossings to protect bridge abutments, its upstream
and downstream areas from erosion. iii) Boulders
Graded boulders are placed at the toe of bank to (a) Groynes
reduce shear stress and near-bank flow velocity. iv)
Brick matressing Brick mattresses are placed on the Outer bank
graded bank slope. Fig.4 shows different types of Series of spurs
hard material protection.

C.C. blocks
MWL
Concave bank Convex bank

(b) Permeable spurs

(a) Revetment
Fig.5 Barriers across the river
Brick matresses
Fukuoka (1989) discussed that vanes are used to
change the velocity distribution of current entering
MWL
into the bend in meandering channel as uniform as
possible by reducing lateral bed gradient. The
principle of bank erosion protection by using vanes is
(b)Brick Matressing
that the secondary current due to the centrifugal force
is offset by secondary current due to the vanes. It is
more rational to consider that vanes deform the flow
Fig.4 Hard material protections field locally with resultant reduction of scour of the
riverbed near the outer bank.
Barrier across the river
(b) Non-structural measures
This method is known as ‘redirectional bank
Non-structural measures are taken against
stabilization method’. The objective is to redirects
short-term protection. Some examples of
flow and energy of stream flow away from eroding
non-structural measures are described below:
bank. In this method, erosion is controlled through

320
i) Flow area increase by dredging: Shallow area of spaced nominally 1 m apart. iv) Bandallings:
the channel is dredged and the area of flow is Bandals are placed on the principles that bandals
increased, which reduces flow velocity. ii) Flow provide partial lateral and vertical obstruction to the
diversion at the upstream of the problem area by approach flow and induce fewer disturbances to the
channelization: The upstream approach of the river flow. The key issues of bandals for the control
problem area is re-channelized by dredging by pass of water and sediment are non-uniform vertical
channels and flow is regulated in the mid stream distribution of suspended sediment. Within the lower
channel. iii) Geo-bag dumping: Geo-bags are half of the flow depth, major portion of the sediment
dumped on the slope of the bank to arrest the bank on flow is concentrated. v) Crisscross porcupines:
temporary basis. Fig.6 shows different types of Bamboo or wooden porcupines are placed across the
non-structural measures. river, which sometimes work to dredge and stabilize
Geobag
the bank by depositing sediments. vi) Log hard
points: The log hard points are composed of log
bundles 20 m in length buried 10 m into the bank
MWL
with approximately 25 tons of stones placed around
the toe of the structure to protect against scour. Fig.7
shows different types of biological protections.
(a) Dumping Geo-bag

HFL
MWL
Top view Piles
Area to be
excavated
(a) Wooden Piling

(b) Flow area increase by dredging

Fig.6 Non-structural bank protection measures

(b) Crisscross porcupines

(c) Biological Protections

There are techniques other than traditional Fig.7 Biological protections
approaches those are practiced all over the world,
known as, bioengineering practices. The widely used
biological protection measures are: 8. PHYSICAL SCALE MODEL NEEDS IN
i) Bank vegetation Vegetation directly protects BANK PROTECTION WORKS
banks from erosion by reducing the near bank shear
stresses. Larger vegetation deflects flow. Vegetation Due to its numerous advantages, scale model is a
offers the additional benefit of modifying soil widely practiced reliable tool to optimize alternative
properties, increasing soil strength due to the structures and design parameters, particularly for
reinforcing properties of roots and lowering pore river training and bank protection works. Scale
water pressures. ii) Wooden piling: Wooden logs are model study on protection of Kazipur area from the
piled along the bank toe to arrest erosion. iii) Willow erosion of Jamuna River is discussed here to justify
posts: This technique is the means of controlling the argument.
stream bank erosion through the systematic Kazipur is situated on the right bank of Jamuna
installation of posts to stabilize eroding banks. It River suffered from flooding and continuous bank
lowers floodwater velocity on and near the eroding erosion. To mitigate the flood problem, Bangladesh
bank. Planting large willow cuttings (10 to 30 cm dia; Water Development Board (BWDB) constructed a
2 to 4 m long) has been widely practiced in the flood embankment along the right bank. But due to
United States for halting bank erosion and restoring continuous bank erosion, this flood embankment was
riparian zones (Watson et al. 1997; Shields et al. also under threat and partly had washed out. The
1995). Willow posts are emplaced along the stream shifting of the bank line was a continuous process
bank from the water edge landward using 3-5 rows and from the past history of bank line shifting it was

321
easily understood the devastation nature of bank example, will be when the difference between the
erosion. To save the valuable land and properties benefit and the cost of protection (net-benefit) is
BWDB had decided to construct a T-head groyne to maximum. Conceptually, if the structure is
divert the flow towards the midstream. A scale model constructed larger than the optimum design the
study was conducted at River Research Institute ‘net-benefit’ will be reduced and if the structure is
(RRI) to find out the effectiveness of this hard smaller, the net-benefit also will be reduced. So,
measure and also find out an appropriate hydraulic over-design or under-design both have the similar
structure and its location to stop the bank erosion. shortcomings. At one extreme, if the costs actually
But after a series of test runs in the scale model, three exceed the benefits we have so over designed the
groynes were recommended instead of single groyne structure that the design would be criticized under
along with the optimum design parameters of the any sensible policy of public spending. On the other
groynes, size of the riprap and length of the falling hand, if the costs are less than the benefits we have
apron (Details may be foun in RRI, 1996). But under designed the structure to a similar extent. But
considering huge cost, series of spurs were proposed this may be inevitable if there are no public funds.
by the client and then tested in the scale model. Based The cost of protection is assumed to have a fixed
on the test results, seven spurs were recommended to component and to then increase proportionally with
protect 5 km area (Details may be found in RRI, the cube of the height. The annual value of a
1998a) as shown in Fig.8. The type of model spur is protection may be estimated from the
shown in Fig.9. damage-probability relationship based on a
probability of exceedence of a certain height, h
(details presented in Ahmed, 1994). A plausible cost
model as proposed by Ahmed (1994) is given by,
C = a +bh3 (7)
where, C is the cost, a is the fixed cost element, b is a
constant of variable cost component and h is the
height of protection work.
Fig.10 shows a typical cost-benefit relationship
when probability of exceeding a height is h. Based
on simulation results, Ahmed (1994) proposed
optimum and sub-optimum design strategies for
selection of bank protection works as shown in
F ig .8 L o c a tio n o f sp u rs a t K a z ip u r Fig.11.
35
b a se d o n p h y sic a l m o d e l stu d y Cost of construction
30
Benefit achieved
25
(In million T aka)
Cost and Benefit

20

15

10

0
0 1 2 3 4 5 6 7 8
Height of protection, h (m)

Fig.10 Benefit-cost relationship when Pr(X>h)

Fig.9 Scale model run showing a spur at Kazipur

16
In physical scale models, recent bathymetries are 14
Net-benefit (In million taka)

used and the future bank line shifting is considered to 12

recommend suitable bank protection works. For site 10

8
of structures, length and extent of bank erosion is 6 Optimum strategy
prerequisite. The study suggests carrying out scale 4 Sub-optimum strategy

model investigations prior to choice a technically 2

0
sound and economically justified protective 0.01 0.03 0.05 0.07 0.09 0.11 0.13 0.15 0.17 0.19 0.21
measures. Estimated Probability (X>h)

9. ECONOMIC ANALYSIS Fig.11 Comparison of optimum strategy and

The optimum design of a bank protection work, sub-optimum strategy with estimated probability
the height of a revetment along riverbank for of exceedence

322
10. CONCLUSIONS National conference on Hydraulic. Engg. (Micheal A. Ports
eds.), pp.118-123.
7. Hickin E, J. and Nanson, G. C. (1984): Lateral Migration
Before design of appropriate bank protection Rate of River Bends, Journal of Hydraulic Eng. Asse. Vol.
measures, bank erosion should be assessed by 110, No.11, November, pp 1557-1567.
available predictive methods and tools using required 8. Hickin, E.J. and Nanson, G.C. (1983) Channel migration
and incision on the Beatton River, Journal of Hydraulic Eng.
hydraulic data, morphological data, sediment data,
ASCE, vol. 109, No. 3, Mar, pp. 327-337.
soil properties and bank characteristics. Near bank 9. Hickin, E. J. and Nanson, G. C. (1975): The character of
flow velocity and shear stresses are two governing channel migration on the Beatton River, Northeast British
parameters to consider while predicting bank erosion. Columbia, Canada, Geological Society America Bulletuin,
To take appropriate bank protection measures, Vol. 86, pp. 487-494.
10. Ippen, A.T. and Drinker, P.A. (1962): Boundary shear stress
on-site assessment of the bank erosion and failure
in curved trapezoidal channels, Journal of Hydraulic Eng.
should be considered, as there might have several Div., HY 5, pp.143-179.
mechanisms those needs due attention and further 11. Klaassen, G. J. and Masselink, G. (1992): Planform changes
research. Physical scale model study has been proved of a braided river with fine sand as bed and bank material,
to be a useful tool for sound engineering judgment in Proc. Fifth Intl. Symp. on River Sed., Karlsruhe,
pp.459-471.
selecting technically feasible structures. The 12. Odgaard, A. J., Jain, S. C. and Luzbetak, D. J. (1989):
recommendations based on physical model study Hydraulic mechanisms of river bank erosion, In
results should be timely and properly implemented Proceedings of the 1989 National conference on Hydraulic.
and post-construction maintenance should be done so Engrg. (Micheal A. Ports eds.), pp.112-117.
as to make it sustainable. As the flood intensity and 13. RRI (2000): Study of bank erosion at Nakalia Pechakhola,
Bera, Pabna from the erosion of the Jamuna River, Research
frequency varies temporally, no structure should be Report, Faridpur, Bangladesh.
implemented based on assumptions or experiences 14. RRI (1998a): Scale model study for the protection of
from the other projects. Finally, economic analysis Kazipur area from the erosion of Jamuna river (additional
must be done to select cost-effective structures. tests), Final Report, Faridpur, Bangladesh.
15. RRI (1998b): Physical model study to investigate the extent
of bank erosion along east and west bank at the downstream
REFERENCES of the Jamuna bridge, Final Report, Faridpur, Bangladesh.
16. RRI (1996): Scale model investigation for the protection of
Kazipur area from the erosion of Jamuna river, Draft Final
1. Ahmed, A. M. M. M (1994): Economic benefit of Report, Faridpur, Bangladesh.
improving the accuracy and precision of flood predictions, 17. Shields, Jr., F.D., Bowie, A.J. and Cooper, C.M. (1995):
M.Sc. Eng. Thesis (unpublished), University of Control of stream bank erosion due to bed degradation with
Newcastle-upon-Tyne, England. vegetation and structure, Water Research Bulletin, Vol.
2. Bagnold, R.A. (1960): Some aspects of the shape of river 31(3), pp. 475-489.
meanders, US Geological Survey: Professional Paper 282E, 18. Thorne, C. R. and Osman, A. M. (1988): River bank
pp.135-144. stability analysis, II: Applications, J. of Hydraulic
3. Darby, S.E. and Thorne, C. R. (1996): Numerical Engineering, ASCE, 114 (2), pp.151-172.
simulation of widening and bed deformation of straight 19. Watson, C. C., Abt S. R. Derrick, D. (1997): Willow posts
sand bed channels, I: Model development, J. of Hydraulic bank stabilization, Journal of the American Water
Engineering, ASCE, Vol. 122 (4), pp.184-193. Resources Association, Vol. 33 (2), pp. 293-300.
4. DELFT/DHI (1996): River Survey Project, Final Report –
Annex 5, Morphological Characteristics, Dhaka,
Bangladesh, pp. 10.1-10.22.
5. Fukuoka, S. (1989): Groins and vanes developed basing
upon a new concept of bank protection, Proceedings of the
1989 National conference on Hydraulic. Eng. (Micheal A.
Ports eds.), pp. 224-229.
6. Hagerty, D. J. and Hamel, J. V. (1989): Geo-technical
aspects of riverbank erosion) In Proceedings of the 1989

323