TIME SCALE FOR LOCAL SCOUR AT BRIDGE PIERS

By Bruce W. Melville1 and Yee-Meng Chiew2

ABSTRACT: The temporal development of clear-water local scour depth at cylindrical bridge piers in uniform
sand beds is considered. New data are presented and used to quantify the influence of flow duration on the
depth of local scour. An equilibrium time scale (t*) is defined. The data show that both t* and the equilibrium
scour depth (dse) are subject to similar influences of flow and sediment parameters, as might be expected because
they are inherently interdependent. A method is given for determination of the time for development of dse for
a given pier, sediment, and approach flow velocity and the concomitant estimation of the scour depth at any
stage during development of the equilibrium scour hole. The results show that the scour depth after 10% of the
time to equilibrium is between about 50% and 80% of the equilibrium scour depth, depending on the approach
flow velocity.
Downloaded from ascelibrary.org by Istanbul Universitesi on 07/22/13. Copyright ASCE. For personal use only; all rights reserved.

INTRODUCTION The shape of the flood flow hydrograph is important as well

as the flood duration. Typically, the flood duration determines
The process of local scour at bridge piers is time dependent. if the equilibrium live-bed scour will develop. After the flood
An equilibrium between the erosive capability of the flow and peak passes, the flow recedes. The duration of the recession
the resistance to motion of the bed materials is progressively period is also important. With flow recession, clear-water con-
attained through erosion of the flow boundary. In fine-grained ditions may prevail, which could induce additional scour, es-
materials (sands and gravels), the equilibrium or final depth pecially if near threshold conditions are maintained over a con-
of local scour dse is rapidly attained in live-bed conditions, but siderable period.
rather more slowly in clear-water conditions (Fig. 1). Clear- The time effects are significant when considering the some-
water scour occurs for mean flow velocities up to the threshold times poor correlation between scour depths observed in the
velocity for bed sediment entrainment, i.e., V ⱕ Vc, while live- field and those measured in the laboratory. In order to achieve
bed scour occurs for V > Vc. The maximum equilibrium scour equilibrium conditions in small-scale laboratory experiments
depth dse)max occurs at V = Vc. In armored cobble or cohesive of clear-water scour depth development at bridge foundations,
sediment bed streams, multiple flood events may be required it is necessary to run the experiments for several days. Data
before the maximum clear-water scour is reached. This may obtained after lesser times, say 10 to 12 h, can exhibit scour
take many years. The equilibrium scour depth in live-bed con- depths less than 50% of the equilibrium depth.
ditions fluctuates due to the effects of bed form migration. The In order to clarify the effect of time on the development of
dashed lines in Fig. 1 represent the temporal average scour depth of scour at cylindrical piers under clear-water condi-
depth under live-bed conditions. The diagram also shows the tions, the writers conducted several series of experiments in
time taken, te, for the equilibrium scour depth to develop. The which the depth of scour was monitored as the scour hole
equilibrium time, te, is the focus of this paper. It increases developed. The experiments were run for a long period of time
rapidly with flow velocity under clear-water conditions, but to ensure that equilibrium was reached. In this paper, the new
then decreases rapidly for live-bed scour. data are analyzed to determine the influence of time on the
Existing equations for depth of local scour at bridge piers formation of clear-water scour holes. A method to predict the
give the equilibrium depth and are therefore conservative re- time to reach the clear-water equilibrium scour depth is pre-
garding temporal effects. For the live-bed conditions that typ- sented.
ically pertain in floods, equilibrium scour depths are appro-
priate. However, where clear-water scour conditions exist, the Framework for Analysis
equilibrium depth of scour may be overly conservative. Peak
flood flows may last only a number of hours or a few days in The relation between the depth of local scour at a bridge
the field, and short floods have insufficient time to generate pier ds and its dependent parameters can be written
equilibrium depths. For example, bridge piers situated on the
floodplain may be wet for periods of less than one day during ds = f [Flood Flow (␳, ␯, V, y, g), Bed Sediment (d50, ␴g, ␳s, Vc)
a flood; typically, clear-water conditions pertain at such sites. Bridge Pier Geometry (b, Sh, Al), Time (t, te)] (1)
The actual scour may be only a small fraction of the equilib-
rium scour depth, which could take weeks to fully develop. where ␳ and ␯ = fluid density and kinematic viscosity, respec-
Johnson and McCuen (1991) developed an analytical model tively; V = mean approach flow velocity; y = mean approach
to simulate the temporal process of local scour at piers. The flow depth; g = acceleration of gravity; d50 and ␴g = median
model was applied to a hypothetical bridge pier using a gen- size and geometric standard deviation of the sediment particle
erated sequence of flood flows over a 75 year period, the in- size distribution; ␳s = sediment density; Vc = critical mean
dividual storms being of 24 h duration. At the end of the 75 approach flow velocity for entrainment of bed sediment; b =
year period, the scour depth was still increasing. pier width; Sh and Al = parameters describing the shape and
alignment of the pier; t = time; te = time for equilibrium depth
1
Assoc. Prof. in Civ. and Resour. Engrg., Univ. of Auckland, Auckland, of scour to develop; and f denotes ‘‘a function of.’’
New Zealand. Assuming constant relative density of sediment and the ab-
2
Sr. Lect. in Civ. and Struct. Engrg., Nanyang Technol. Univ., Singa- sence of viscous effects, an expression for the equilibrium
pore 2263.
Note. Discussion open until June 1, 1999. To extend the closing date depth of local scour at a cylindrical pier of diameter D in a
one month, a written request must be filed with the ASCE Manager of uniform sediment can be written using (1) as follows:

冉 冊
Journals. The manuscript for this paper was submitted for review and
possible publication on July 22, 1997. This paper is part of the Journal dse V y d50 t
=f , , , (2)
of Hydraulic Engineering, Vol. 125, No. 1, January, 1999. 䉷ASCE, D Vc D D te
ISSN 0733-9429/99/0001-0059 – 0065/$8.00 ⫹ $.50 per page. Paper No.
16274. where the first three parameters on the right-hand side repre-
JOURNAL OF HYDRAULIC ENGINEERING / JANUARY 1999 / 59

J. Hydraul. Eng. 1999.125:59-65.

Downloaded from ascelibrary.org by Istanbul Universitesi on 07/22/13. Copyright ASCE. For personal use only; all rights reserved.

FIG. 1. Variation of Local Scour Depth with Flow Velocity and

Time

sent, respectively, the stage of sediment transport on the ap-

proach flow bed, termed the flow intensity; the depth of flow
relative to the width of the pier, termed the flow shallowness;
and the sediment median size relative to the pier width, termed
the sediment coarseness. Fig. 2 shows schematic diagrams in-
dicating the effects of these three parameters. The trends are
discussed in detail elsewhere, e.g., Melville (1997). Only a
brief description is included here.
Under clear-water conditions, the local scour depth in uni-
form sediment increases almost linearly with flow intensity to
a maximum at the threshold velocity. The maximum scour
depth is called the threshold peak. As the velocity exceeds the
threshold velocity, the local scour depth in uniform sediment
first decreases and then increases again to a second peak, these
changes being relatively small, but the threshold peak is not
exceeded providing the sediment is uniform. The second peak
occurs at about the transition flat bed stage of sediment trans-
port on the channel bed and is termed the live-bed peak. The
scour depth changes with flow intensity are explained in terms
of the balance between sediment input to and output from the
scour hole. For clear-water scour, there is no sediment input.
These trends are evident in many data sets, including those of
Chabert and Engeldinger (1956), Shen et al. (1966), Maza
(1968), Ettema (1980), and Chiew (1984).
Flow shallowness influences local scour depth when the
horseshoe vortex, the principal cause of scour, is affected by
the formation of the surface roller (or bow wave) that forms
at the leading edge of the pier. The two vortices have opposite
directions of rotation. In principle, so long as they do not in-
terfere with each other, the local scour depth is independent
of flow depth. In such cases, the local scour is said to occur
at narrow piers. With decreasing flow depth, the surface roller
becomes more dominant and renders the base vortex less ca-
pable of entraining sediment. Thus, the local scour depth is
reduced for shallower flows. In very shallow flows, the local
scour becomes independent of flow depth. The local scour is
said to occur at a wide pier in such cases. These trends are
evident in the laboratory data of many researchers, including
Chabert and Engeldinger (1956), Laursen and Toch (1956),
Laursen (1963), Hancu (1971), Bonasoundas (1973), Basak
(1975), Breusers et al. (1977), Jain and Fischer (1980), Ettema
(1980), Chee (1982), and Chiew (1984).
For uniform sediments, local scour depths are unaffected by
sediment size unless the sediment is relatively coarse. Labo-
ratory data show that the local scour depth is influenced by FIG. 2. Influence of (a) Flow Intensity; (b) Flow Shallowness;
sediment size when D/d50 < 50. For local scour at piers, Ettema and (c) Sediment Coarseness on Local Scour Depth

60 / JOURNAL OF HYDRAULIC ENGINEERING / JANUARY 1999

J. Hydraul. Eng. 1999.125:59-65.

 (1980) explained that for larger values of the sediment coarse- TABLE 1. Experimental Data Plotted in Figs. 5 – 8
ness ratio, individual grains are large relative to the groove
Run D dse d50 y V te
excavated by the downflow, and erosion is impeded because
number (mm) (mm) (mm) (mm) (m/s) (min) V/Vc
the porous bed dissipates some of the energy of the downflow.
(1) (2) (3) (4) (5) (6) (7) (8)
The scour is said to occur in coarse sediment in such cases.
Attention is now focused on the time ratio t/te. It is postu- N1 70 94 0.96 200 0.231 3,142 0.688
lated (and confirmed by the data presented in this paper) that N2 70 77 0.96 70 0.245 2,703 0.654
because te and dse are inherently interdependent, both should N3 70 69 0.96 70 0.218 2,880 0.623
have similar dependence on the same set of parameters. For N4 70 79 0.96 70 0.231 2,817 0.671
example, te should depend on flow shallowness for lower val- N5 70 39 0.96 70 0.198 1,537 0.623
ues of y/D, but become independent of y/D for narrow piers. N6 50 60 0.96 70 0.231 2,610 0.657
Thus, the relation for te is written N7 50 60 0.96 200 0.231 3,090 0.649

冉 冊
N8 38 54 0.96 200 0.243 4,035 0.690
Vte V y d50 N9 38 45 0.96 200 0.231 1,741 0.660
=f , , (3) AU1 16 24 0.80 200 0.213 1,753 0.676
D Vc D D
Downloaded from ascelibrary.org by Istanbul Universitesi on 07/22/13. Copyright ASCE. For personal use only; all rights reserved.

AU2 25 25 0.80 200 0.213 2,695 0.676

and t* = Vte /D is termed the equilibrium time scale. N10 70 116 0.96 95 0.278 3,252 0.827
N11 70 118 0.96 95 0.278 4,194 0.827
Eq. (3) is now examined using laboratory data. The data are
N12 70 107 0.96 50 0.269 4,055 0.823
restricted to clear-water scour conditions, uniform sediments, N13 70 121 0.96 70 0.304 4,089 0.850
and cylindrical piers. The data are used to demonstrate that N14 50 89 0.96 70 0.298 3,215 0.845
the trends evident in the influence of flow intensity, flow shal- N15 38 77 0.96 200 0.278 4,230 0.817
lowness, and sediment coarseness on the equilibrium time N16 38 97 0.96 200 0.324 4,530 0.919
scale are similar to those shown in Fig. 2 for the equilibrium N17 50 91 0.96 70 0.317 4,163 0.901
scour depth, and that (3) is an appropriate functional relation AU3 16 32 0.80 200 0.250 1,740 0.827
for t*. AU4 25 48 0.80 200 0.250 2,895 0.827
AU5 16 38 0.80 200 0.294 2,065 0.957
AU6 25 64 0.80 200 0.294 2,690 0.957
Experiments
N18 70 133 0.96 200 0.271 4,260 0.783
The experiments were conducted in four different flumes, N19 70 107 0.96 77 0.271 3,100 0.779
three at the University of Auckland (AU) and one at the Nan- N20 70 91 0.96 70 0.251 4,210 0.700
N21 50 79 0.96 70 0.265 4,158 0.753
yang Technological University (N). Table 1 gives details of N22 70 25 0.96 70 0.179 1,338 0.519
the experiments. Uniform sands were used at each venue, with N23 50 26 0.96 200 0.197 1,425 0.527
sediment coarseness ranging from 20 to 222. For most of the N24 70 43 0.96 200 0.197 1,815 0.593
data, sediment coarseness exceeds 50. Flow intensity ranges N25 70 37 0.96 70 0.185 2,429 0.556
from 0.46 to 0.957, while flow shallowness ranges from 0.6 N26 70 35 0.96 70 0.185 1,433 0.556
to 12.5. To extend the range of data, Ettema’s (1980) mea- N27 50 29 0.96 70 0.198 1,768 0.587
surements of local scour depth at cylindrical piers are used to N28 70 7 0.96 70 0.165 300 0.467
supplement the new data. For Ettema’s data (shown in Tables N29 70 20 0.96 200 0.171 1,298 0.400
1 and 2, reference E), flow intensity ranges from 0.5 to 0.95, N30 70 28 0.96 200 0.185 1,300 0.468
N31 50 11 0.96 200 0.174 450 0.459
flow shallowness ranges from 0.197 to 21.05, and sediment
N32 38 4 0.96 200 0.174 200 0.441
coarseness ranges from 15 to 178.6. Graf (1995) published the AU7 16 10 0.90 200 0.175 320 0.530
results of three test runs of local scour development at cylin- AU8 25 17 0.90 200 0.175 1,235 0.530
drical piers. These data are also included in Table 1 (reference AU9 200 237 0.90 135 0.269 7,140 0.72
G). The mean velocities for Graf’s (1995) data were estimated AU10 200 209 0.90 120 0.291 5,520 0.79
using the semilogarithmic velocity distribution equation. AU11 200 109 0.90 120 0.228 3,780 0.62
AU12 200 59 0.90 230 0.186 840 0.46
TEMPORAL DEVELOPMENT OF SCOUR AU13 200 235 0.90 181 0.320 2,610 0.82
E1 28.5 58 1.9 600 0.571 1,990 0.9
Because equilibrium clear-water scour depth is reached as- E2 28.5 57 1.9 200 0.487 3,725 0.9
ymptotically with time, it can take a very long (perhaps infi- E3 28.5 48 1.9 100 0.434 2,184 0.9
nite) time for the equilibrium scour hole to form. Laboratory E4 28.5 41 1.9 50 0.381 2,455 0.9
E5 101.6 224 1.9 600 0.571 5,640 0.9
observations reveal that an apparent equilibrium scour hole
E6 101.6 223 1.9 400 0.540 5,582 0.9
may continue to deepen at a relatively slow rate long after E7 101.6 211 1.9 300 0.518 5,615 0.9
‘‘equilibrium’’ conditions were thought to exist. Consequently, E8 101.6 1.98 1.9 200 0.487 5,360 0.9
it is difficult to know when to stop an experiment, i.e., how E9 101.6 170 1.9 100 0.434 5,286 0.9
to define the time for development of the equilibrium scour E10 101.6 147 1.9 50 0.381 5,320 0.9
depth. To overcome this inherent difficulty, a new definition E11 101.6 69 1.9 20 0.311 5,625 0.9
for time to equilibrium is used here. Thus, te is defined as the E12 101.6 179 5.35 600 1.00 2,214 0.9
time at which the scour hole develops to a depth (the equilib- E13 101.6 178 5.35 400 0.939 1,498 0.9
rium depth, dse) at which the rate of increase of scour does not E14 101.6 165 5.35 300 0.894 1,780 0.9
exceed 5% of the pier diameter in the succeeding 24 h period, E15 101.6 148 5.35 200 0.831 1,765 0.9
E16 101.6 131 5.35 100 0.723 2,205 0.9
i.e. E17 101.6 102 5.35 50 0.615 2,216 0.9
d(dse) 0.05D E18 150 318 1.9 600 0.571 9,277 0.9
ⱕ (4) E19 50.8 115 1.9 600 0.571 3,660 0.9
dt 24 h E20 50.8 85 1.9 600 0.477 3,320 0.75
E21 50.8 16 1.9 600 0.318 580 0.50
For example, in an experiment with a pier diameter D = 100 G1 110 174 2.1 170 0.579 4,780 1.0
mm, if the increase of the scour depth is less than 5 mm in G2 100 195 2.1 232 0.609 6,200 1.0
24 h, equilibrium is assumed to have been reached. For the G3 150 259 2.1 232 0.609 6,280 1.0
new data, the values of dse and te in columns 3 and 7, respec-
JOURNAL OF HYDRAULIC ENGINEERING / JANUARY 1999 / 61

J. Hydraul. Eng. 1999.125:59-65.

 TABLE 2. Experimental Data Plotted in Fig. 9

Run D dse d50 y V te

number (mm) (mm) (mm) (mm) (m/s) (min) V/Vc
(1) (2) (3) (4) (5) (6) (7) (8)
E22 150 225 0.84 600 0.428 15,000a 0.95
E23 150 229 1.9 600 0.604 10,106 0.95
E24 150 207 5.35 600 1.06 3,802 0.95
E25 150 188 7.8 600 1.208 2,925 0.95
E26 101.6 227 0.8 600 0.399 10,000a 0.9
E27 101.6 227 0.84 600 0.405 10,500a 0.9
E28 101.6 224 1.9 600 0.572 5,640 0.9
E29 101.6 208 3.8 600 0.895 2,002 0.9
E30 101.6 179 5.35 600 1.0 2,214 0.9
E31 101.6 166 7.8 600 1.14 2,945 0.9
FIG. 4. Plot of Eq. (5) Indicating Temporal Development of Lo-
E32 150 315 0.8 600 0.399 15,000a 0.9
Downloaded from ascelibrary.org by Istanbul Universitesi on 07/22/13. Copyright ASCE. For personal use only; all rights reserved.

cal Scour Depth

E33 150 314 0.84 600 0.405 15,000a 0.9
E34 150 318 1.9 600 0.572 9,277 0.9
E35 150 308 5.35 600 1.0 5,370 0.9 Flow shallowness effects are depicted in Fig. 5, which is
E36 150 290 7.8 600 1.14 2,996 0.9 based on the new data. The data are divided into sets of dif-
a
Adjusted value. ferent ranges of flow intensity. For each flow intensity value,
the equilibrium time scale increases with y/D for shallow flows
and becomes independent of y/D for deep flows. The trend is
tively, in Table 1 are based on the foregoing definition. For similar to that shown in Fig. 2 for equilibrium scour depth.
Ettema’s (1980) data, his measured values of dse and te are Some of the scatter in the data is due to their having a range
used in the analysis. of sediment coarseness values. To further elucidate the influ-
Fig. 3 shows the new data for the temporal development of ence of flow shallowness on t*, Ettema’s (1980) data are plot-
the scour hole plotted as ds /dse versus t/te, with the flow inten- ted in Fig. 6, using the same axes as in Fig. 5, for two values
sity as a third parameter. In the figure, ds represents the scour of sediment size and a constant flow intensity value of 0.9.
depth at a particular time, t. The dimensionless parameters that The data defining the lower curve in Fig. 6 have D/d50 ⬃ 19,
comprise the axes of Fig. 3 constrain the curve for each set of while most of the data defining the upper curve have higher
data to pass through the point (1, 1), as shown. The figure values of D/d50. The upper curve (V/Vc = 0.9) matches well
shows a family of curves with increasing flow intensity ratio with that shown in Fig. 5 for V/Vc = 0.95.
in the direction of increasing scour depth (or decreasing time). All of the data from Table 1 for which 1.0 > V/Vc > 0.9,
The data indicate that 50% of the equilibrium scour depth i.e., those data having no influence of flow intensity, are plot-
(0.5dse) is attained in a time varying from 0.1% to 10% of te, ted in Fig. 7 to define the influence of flow shallowness. Et-
depending on the flow intensity. Similarly, 80% of the equi- tema’s data for d50 = 5.35 mm are excluded because they con-
librium scour depth is developed in a time varying from only tain an apparent strong influence of sediment coarseness, as
5% to almost 40% of the time to equilibrium. The data dem- shown by Fig. 6. The data are plotted using logarithmic axes
onstrate the significance of time in the measurement and es- to be consistent with similar plots for equilibrium scour depth,
timation of scour depth. For example, an experiment stopped given in Melville (1997). An envelope curve is fitted to the
after 4 h for which equilibrium time is 3 days would have data. The curve resembles that presented by Melville (1997)
reached only about 50% to 80% of the equilibrium scour for local equilibrium scour depth. It follows that both the equi-
depth. librium time scale and the equilibrium scour depth are char-
The data in Fig. 3 are well represented by the following acterized by similar dependence on flow shallowness. The ap-
equation: parent limit to the influence of flow shallowness on t* occurs

再 冏 冉 冊冏 冎 at y/D ⬃ 6, although there is a minimal effect of flow shal-

1.6
ds Vc t
= exp ⫺0.03 ln (5) lowness for y/D exceeding about 3. The curve in Fig. 7 has
dse V te the following equation:

冉冊
which is plotted in Fig. 4. The modulus sign in (5) is necessary 0.25
to ensure that the argument of the exponential function is neg- y y y
t* = 1.6 ⫻ 106 ⱕ 6; t* = 2.5 ⫻ 106 >6 (6a,b)
ative. D D D

FIG. 3. New Laboratory Data Showing Temporal Development FIG. 5. New Laboratory Data Showing Relation between t* and
of Scour Depth y/D

62 / JOURNAL OF HYDRAULIC ENGINEERING / JANUARY 1999

J. Hydraul. Eng. 1999.125:59-65.

 FIG. 9. Laboratory Data by Ettema (1980) Showing Relation
between t* and D/d50 for Constant Flow Intensity, V/Vc = 0.9
FIG. 6. Laboratory Data by Ettema (1980) Showing Relation
Downloaded from ascelibrary.org by Istanbul Universitesi on 07/22/13. Copyright ASCE. For personal use only; all rights reserved.

between t* and y/D

Comparison of the two curves drawn in Fig. 6 indicates that
an influence of sediment coarseness is apparent, because t*
values are reduced for D/d50 = 19. An analysis of three other
sets of data [given by Ettema (1980); details in Table 2] for
apparent trends in the influence of sediment coarseness on t*
is shown in Fig. 9. These data should not be affected by flow
intensity or flow shallowness, because V/Vc ⬃ 1 and y/D > 4,
respectively. For the five points identified by the asterisk in
Table 2, the te values tabulated by Ettema (1980) have been
reduced before being included in the plot, because his raw data
(plots of ds against time) indicate that equilibrium had been
attained earlier. The reduced values are shown in Table 2. The
trend depicted in Fig. 9 is consistent with that shown for the
effect of sediment coarseness in Fig. 2. The limit to sediment
coarseness influence on t* occurs at D/d50 ⬃ 100, according
FIG. 7. Influence of Flow Shallowness on t* to the plot. The curve plotted in Fig. 9 has the following equa-
tion:

冉冊
0.21
In Fig. 8, all of the data from Table 1 for which y/D > 2.8, D D
i.e., those data having no influence, or minimal influence, of t* = 9.5 ⫻ 105 ⱕ 100 (8a)
d50 d50
flow shallowness, are plotted to show the effect of flow inten-
sity. Again, Ettema’s data for d50 = 5.35 mm are excluded D
because they contain an apparent strong influence of sediment t* = 2.5 ⫻ 106 > 100 (8b)
d50
coarseness, as shown by Fig. 6. The remaining scatter in the
plotted data is partly due to their having a range of values of In spite of the quite good correlation demonstrated in the data
sediment coarseness. The diagram shows that the equilibrium in Fig. 9, (8) is not recommended to define the function for
time scale increases rapidly with flow intensity for clear-water the influence of sediment coarseness, because other data in
scour, attaining a maximum value at the threshold condition Ettema (1980) do not show consistent trends. Further study is
for sediment transport. At higher live-bed flows, t* is expected needed before definitive results for D/d50 effects can be ob-
to rapidly decrease again, as shown by the dashed line. For tained. Significantly, the condition D/d50 > 100 is often satis-
clear-water scour conditions, the relation between the equilib- fied in the field, rendering sediment coarseness effects rela-
rium time scale and the flow intensity is similar to that for tively minor in practice.
equilibrium scour depth (Fig. 2). It follows that flow intensity
affects t* and dse /D in a similar manner. The straight line fitted SCOUR DEPTH ESTIMATION
to the data in Fig. 8 has the following equation:

冉 冊
The new relations for time effects presented here can be
V V
t* = 4.17 ⫻ 106 ⫺ 0.4 0.4 < ⱕ 1.0 (7) used with existing equations for local scour depth estimation,
Vc Vc as demonstrated below for the method given in Melville
(1997). That method, with the addition of a time factor and
simplified to apply to cylindrical piers founded in uniform sed-
iments, can be written
ds = KyD KI Kd Kt (9)
where the Ks are empirical expressions, given in Appendix I,
accounting for the various influences on scour depth. The time
factor, Kt, is given by (5), because that equation gives the ratio
of the scour depth at a particular time t to that at te, i.e.
ds
Kt = (10)
dse
Application of (5) requires knowledge of te. Eq. (6) and (7)
give the effects of flow shallowness and flow intensity on t*,
FIG. 8. Influence of Flow Intensity on t* respectively. It is convenient to combine these effects in one
JOURNAL OF HYDRAULIC ENGINEERING / JANUARY 1999 / 63

J. Hydraul. Eng. 1999.125:59-65.

 equation and to write the resulting relation in terms of te, giv- TABLE 3. Example of Local Scour Depth Estimation
ing Scour depth Percentage development

te(days) = 48.26 冉
D V
V Vc 冊
⫺ 0.4
y
D
>6 (11a)
1h
Flood peak duration
(1)
(m)
(2)
0.77
(%)
(3)
28

冉 冊冉 冊
0.25
D V y y 3h 1.07 38
te(days) = 30.89 ⫺ 0.4 ⱕ6 (11b) 6h 1.29 46
V Vc D D 12 h 1.53 55
1 day 1.78 64
where D and V must be expressed in a consistent system of 2 days 2.05 73
units, e.g., meters and meters per second, respectively. Eq. (11) 1 week 2.49 89
is restricted to 1 ⱖ V/Vc ⱖ 0.4. The lower limit (V/Vc = 0.4) 2 weeks 2.68 96
is similar to that suggested by other researchers for the incip- Equilibrium (29.4 days) 2.79 100
ient velocity for local scour at cylindrical piers, e.g., Chiew
(1995). For convenience, the first equation in (11) is plotted
Downloaded from ascelibrary.org by Istanbul Universitesi on 07/22/13. Copyright ASCE. For personal use only; all rights reserved.

in Fig. 10, which gives te values where flow shallowness ef- te. From (11) or Fig. 10, te = 29.4 days. Finally, ds can be
fects are unimportant ( y/D > 6). For y/D < 6, te values from found using (5) with t = 6 h. Thus, ds = Kt dse = 0.416 dse =
Fig. 10 need to be multiplied by 0.64(y/D)0.25, as shown. Fig. 1.29 m. Hence, the scour depth after 6 h is about 46% of the
10 can be used as an alternative to (11). equilibrium depth, which would take 29.4 days to develop in
In sands that do not form ripples (d50 larger than about 0.6 this case. Local scour depths at other times are shown in Table
mm), the time for the equilibrium scour to develop (te) is a 3.
maximum at the threshold velocity (V = Vc) and when y/D >
6. The maximum value te)max, is given by
CONCLUSIONS
D
te)max(days) = 28.96 (12)
V This study is limited to local scouring at cylindrical bridge
piers in uniform sand beds. The following conclusions are
where D and V must be expressed in a consistent system of drawn from this study:
units, e.g., meters and meters per second, respectively. Where
flow intensity effects (V/Vc < 1) or flow shallowness effects
(y/D < 6) are present, te values are reduced from those given 1. The equilibrium depth of scour at a bridge pier under
by (12). clear-water conditions is approached asymptotically as
shown in Fig. 3.
2. Scour depths after 10% of the time to equilibrium vary
Example
between about 50% and 80% of the equilibrium depth,
To demonstrate the new method to allow for time effects in depending on the flow intensity (Fig. 3).
scour depth estimations, an example is presented. It is required 3. The equilibrium time scale (t* = Vte /D) for development
to determine the local scour depth at a cylindrical pier. The of a clear-water local scour hole at a bridge pier is a
following data are available: function of flow intensity (V/Vc), flow shallowness (y/D),
and sediment coarseness (D/d50). The influences of flow
• Pier diameter, D = 2 m intensity, flow shallowness, and sediment coarseness on
• Uniform sand size, d50 = 1.0 mm t* and dse /D are demonstrated, using laboratory data, to
• Flow depth, y = 2.5 m be similar. Recommended functions for the effects of
• Mean velocity (at flood peak), V = 0.5 m/s flow shallowness and flow intensity are given by (6) and
• Peak flood duration, t = 6 h (7).
• Mean velocity at sediment threshold, Vc = 0.8 m/s 4. Eq. (11) can be used to estimate the time taken for equi-
librium scour depth to develop. When used in conjunc-
tion with (5) (or Fig. 4) and a suitable equation for equi-
A procedure for determining Vc is given in Melville (1997). librium scour depth prediction, the new equations allow
First, determine dse using Melville (1997). From (13b), KyD estimation of local scour depth at any stage throughout
= 4.47 m for D/y = 0.8. For D/d50 = 2,000, Kd = 1.0, from the development of a scour hole. Knowledge of the de-
(15b). From (14a) for V/Vc = 0.625, KI = 0.625. Hence, dse = sign flood hydrograph is necessary to assess the time of
KyD KI Kd = 2.79 m. The second step involves determination of scouring (t). The relations given have not been confirmed
at prototype scale and, in that sense, are unproven. Read-
ers are urged to test the equations against any suitable
large-scale/prototype information; the writers would ap-
preciate hearing about the outcomes of such examina-
tions of the proposed relations.

APPENDIX I. SCOUR DEPTH ESTIMATION FACTORS

ACCORDING TO MELVILLE (1997)

Flow Depth – Pier Width

KyD = 2.4D for D/y < 0.7 (13a)

KyD = 2(yD)0.5 for 0.7 < D/y < 5 (13b)

FIG. 10. Design Chart for Estimation of te KyD = 4.5y for D/y > 5 (13c)

64 / JOURNAL OF HYDRAULIC ENGINEERING / JANUARY 1999

J. Hydraul. Eng. 1999.125:59-65.

 Flow Intensity Laursen, E. M. (1963). ‘‘Analysis of relief bridge scour.’’ J. Hydr. Div.,
ASCE, 89(3), 93 – 118.
KI = V/Vc for V/Vc < 1; KI = 1 for V/Vc > 1 (14a,b) Laursen, E. M., and Toch, A. (1956). ‘‘Scour around bridge piers and
abutments.’’ Bull. No. 4, Iowa Highways Research Board, Ames, Iowa.
Maza, J. A. (1968). ‘‘Socavacion en cauces naturales.’’ A. J. Miguel-
Particle Size Rodriguez, translator, Rep. No. 114, School of Engineering, University
of Auckland, Auckland, New Zealand.
Kd = 0.57 log(2.24D/d50) for D/d50 < 25 (15a) Melville, B. W. (1997). ‘‘Pier and abutment scour — an integrated ap-
proach.’’ J. Hydr. Engrg., ASCE, 123(2), 125 – 136.
Kd = 1 for D/d50 > 1 (15b) Shen, H. W., Schneider, V. R., and Karaki, S. S. (1966). Mechanics of
local scour. U.S. Department of Commerce, National Bureau of Stan-
APPENDIX II. REFERENCES dards, Institute for Applied Technology, Washington, D.C.

Basak, V. (1975). ‘‘Scour at square piles.’’ Rep. No. 583, Devlet Su isteri APPENDIX III. NOTATION
genel mudurlugu, Ankara, Turkey.
Bonasoundas, M. (1973). ‘‘Strömungsvorgang und kolkproblem.’’ Rep. The following symbols are used in this paper:
No. 28, Oskar von Miller Institu, Technical University, Munich, Ger-
Downloaded from ascelibrary.org by Istanbul Universitesi on 07/22/13. Copyright ASCE. For personal use only; all rights reserved.

many (in German).

Breusers, H. N. C., Nicollet, G., and Shen, H. W. (1977). ‘‘Local scour
Al = parameter describing per alignment;
around cylindrical piers.’’ J. Hydr. Res., Delft, The Netherlands, 15(3), b = width of pier;
211 – 252. D = diameter of cylindrical pier;
Chabert, J., and Engeldinger, P. (1956). Etude des affouillements autour d50 = median size of particle size distribution;
des piles des ponts. Laboratoire d’Hydraulique, Chatou, France (in ds = local scour depth at time t;
French). dse = local scour depth at equilibrium;
Chee, R. K. W. (1982). ‘‘Live-bed scour at bridge piers.’’ Rep. No. 290, g = acceleration of gravity;
School of Engrg., University of Auckland, Auckland, New Zealand. Kd = particle size factor;
Chiew, Y. M. (1984). ‘‘Local scour at bridge piers.’’ Rep. No. 355, Dept. KI = flow intensity factor;
of Civ. Engrg., University of Auckland, Auckland, New Zealand.
Chiew, Y. M. (1995). ‘‘Mechanics of riprap failure at bridge piers.’’ J.
Kt = time factor;
Hydr. Engrg., ASCE, 121(9), 635 – 643. KyD = flow depth-pier width expression (m);
Ettema, R. (1980). ‘‘Scour at bridge piers.’’ Rep. No. 216, School of Sh = parameter describing pier shape;
Engrg., University of Auckland, Auckland, New Zealand. t = time;
Graf, W. H. (1995). ‘‘Local scour around piers.’’ Annu. Rep., Laboratoire te = time to develop equilibrium scour depth;
de Recherches Hydrauliques, Ecole Polytechnique Federale de Lau- t* = equilibrium time scale = Vte/D;
sanne, Lausanne, Switzerland, B.33.1 – B.33.8. V = mean approach flow velocity;
Hancu, S. (1971). ‘‘Sur le calcul des affouillements locaux dams la zone Vc = mean approach flow velocity at threshold condition for
des piles des ponts.’’ Proc., 14th IAHR Congress, Vol. 3, International sediment movement;
Association for Hydraulic Research, Delft, The Netherlands, 299 – 313.
Jain, S. C., and Fischer, E. E. (1980). ‘‘Scour around bridge piers at high
y = mean approach flow depth;
flow velocities.’’ J. Hydr. Div., ASCE, 106(11), 1827 – 1842. ␯ = kinematic viscosity of water;
Johnson, P. A., and McCuen, R. H. (1991). ‘‘A temporal spatial pier scour ␳ = density of water;
model.’’ Transp. Res. Rec. 1319, Transportation Research Board, ␳s = sediment density; and
Washington, D.C., 143 – 149. ␴g = standard deviation of sediment particle size distribution.

JOURNAL OF HYDRAULIC ENGINEERING / JANUARY 1999 / 65

J. Hydraul. Eng. 1999.125:59-65.