ISSN 0097-8078, Water Resources, 2017, Vol. 44, No. 2, pp. 226–236. © Pleiades Publishing, Ltd., 2017.
Original Russian Text © N.I. Alekseevskii, A.Yu. Sidorchuk, 2017, published in Vodnye Resursy, 2017, Vol. 44, No. 2, pp. 147–157.
HYDROPHYSICAL PROCESSES
Morphology and Dynamics of Active Bed Forms
in the Terek R. Channel
N. I. Alekseevskii† and A. Yu. Sidorchuk*
Moscow State University, Moscow, 119991 Russia
*Е-mail: aleksey.sidorchuk@geogr.msu.ru
Received September 4, 2015
Abstract—Active bed forms of three major classes are formed in Terek lower reaches during summer floods.
They include ripples, dunes of the first order, and dunes of the second order (from smaller to larger), which
commonly form an incomplete hierarchy. The morphology of the bed forms is essentially stochastic and can
be adequately described by probability distribution functions of bed form characteristics for some narrow
ranges of hydraulic flow characteristics. At the same time, the mean values of bed form morphometric characteristics (length, height, and asymmetry) show stable relationships with flow velocity and depth. The celerity of active dunes can be adequately described by a modified Snishchenko–Kopaliani formula. The channelforming sediments that move as bed forms in Terek lower reaches account for 7% of sediment transport rate
of all channel-forming sediments, a value near the lower limit for rivers with sand alluvium.
Keywords: active bed forms, classification, rivers with flood regime, channel-forming sediment transport rate
DOI: 10.1134/S0097807817010031
INTRODUCTION
Bed forms in river channel and large laboratory
flumes can be of two major types: with active motion
(active bed forms) and passive motion (passive bed
forms) [11]. Examples of studies of the morphology
and dynamics of passive bed forms are numerous, as
only passive bed forms can be measured in rivers with
short spring floods (often with an ice drift) and long
low-water periods. In large (and, the more so, in
small) laboratory flumes, passive bed forms are commonly formed because of the small relative flow
depths. The researchers believe that the bed forms on
river bed are always asymmetrical and have a triangular shape with a steep downstream and gentle
upstream slope, their motion being governed by bed
load, resulting in the formation of cross bedding in
alluvial strata (see, e.g., [9]).
Studies in tropic rivers with typically long spring
floods show active bed form generation in channels in
high-water periods with large relative depths. These
bed forms are asymmetrical symmetrical and have
ellipsoidal shape with convex upstream and downstream slopes; their motion is governed by the dynamics of the wavelike flow structure that generates the
bed form; the result of this motion is bedload sediment
transport and the formation of a wavy, almost horizontal lamination in the channel alluvium [3, 11,
16, 17].
† Deceased.
Active bed forms can also be studied in rivers with
high summer floods. In Russia, these are, primarily,
the rivers of the North-Caucasian and East-Siberian
types according to B.D. Zaikov [6]. A typical example
of rivers with the hydrological regime of North-Caucasian type is the Terek with its tributaries. The highest
floods occur in the lower reaches of the Terek (downstream of the Sunzha mouth), where summer rain
nourishment dominates (on the average, almost 33%
of the annual runoff, according to [7]). The water and
sediment discharges during such floods in some years
can be higher than those during spring floods. Data
on water and sediment discharges and channel transformations were collected during a summer flood in a
15-km reach of the Lower Terek (downstream of the
Sunzha R. mouth). These materials were only partly
published in [2, 10], though they contain very rare for
Russian rivers data on the morphology and dynamics
of active channel bed forms.
STUDY AREA
A series of seven bends with a mean step and amplitude of 1300 and 600 m, respectively, and the mean
channel width of 260 m was studied in a 15-km reach
of the Terek R. downstream of the Sunzha mouth
(Fig. 1a). The dynamics and morphology of the bed
forms were measured in a linear segment between two
downstream bends in this series (Fig. 1b). During the
construction of a new road (completed in 2012), a
steep bend was cut by a canal, reducing the length of
226
MORPHOLOGY AND DYNAMICS OF ACTIVE BED FORMS
227
(a)
(b)
1982
2014
(b)
12
10
13
11
9
8
6
2002
2
3
4 5
7
Longitudinal
profile 1
Longitudinal profile 2
500 m
Fig. 1. (a) Study area of 1982 in the Terek R. downstream of the Sunzha R. in SPOT 5 photograph of Sep. 2, 2006, and (b) positions of measurement profiles for studying the morphology and dynamics of channel bed forms. The SPOT image was presented
by MSU Geoportal.
this river segment to 11.5 km. The natural horizontal
transformations of these bends are also considerable—
the mean bank erosion rates over 50 years are ~3
m/year with local mean velocities reaching 10–15
m/year [2]. However, the banks of the linear channel
segments where bed form measurements were carried
out remained unchanged in 1982–2014 (Fig. 1b);
therefore, data of 1982 are still quite representative.
Channel shallows changed their positions from year to
year, but the type of relief configuration remained
unchanged, i.e., a linear channel with a middle bar
(commonly submerged) in the upstream part of the
segment; accordingly, two pool hollows are located
here, merging to form one in the downstream part of
the segment.
The depths were measured on July 22–August 30,
1982, by a PEL-3 depth sounder along two longitudinal profiles with lengths of 970 and 500 m fixed by
bank marks. The distances on the profiles were determined with the use of 13 transverse bank marks spaced
60–100 m apart. The measurements were made by
downstream drifting, with the measurement of the
mean surface water flow between the cross sections.
Twice a day (in the morning and evening), bed relief
was measured at the middle bar and in the right and
WATER RESOURCES
Vol. 44
No. 2
2017
joint pool hollows. Such measurement frequency was
enough to reliably identify the same large bed forms in
sound records made at different times and to evaluate
the velocity of their displacement. Hydraulic (mean
flow velocity and the depth over the bed form) and
morphometric (the total length, the lengths of the
upstream and downstream slopes, and the bed form
height) characteristics were measured for all bed
forms. The total of 57 measurements were processed
(5927 bed forms) on a long longitudinal profile 1 and
44 measurements (3481 bed forms), on a short
profile 2.
HYDROLOGICAL–HYDRAULIC
CONDITIONS OF THE LOWER TEREK
The Lower Terek in the study region has a drainage
area of 37400 km2, the mean water discharge of
294 m3/s, the mean maximal discharge of 970 m3/s;
the absolute recorded maximum of water discharge
was ~2000 m3/s during a flood in June 1914. The mean
annual suspended sediment transport rate is estimated
at 580 kg/s (18.3 million t/year), though it shows wide
variations; for example, the sediment transport rate in
2002 was in excess of 63 million t [5].
228
ALEKSEEVSKII, SIDORCHUK
The measurements of 1982 were carried out at a
peak of a flood (with water discharges of
570‒670 m3/s), which lasted from July 22 to August 3
and was followed by a slow decline with water discharge reaching 360 m3/s by August 30. The flow rate
varied both along the measurement profiles and over
time from 0.12 to 3.5 m/s; the depth varied from 0.4 to
10.3 m; and the Froude number, from 0.02 to 0.9.
Two types of sediments can be identified in the
lower Terek R.: transit sediments (with particle diameter <0.05 mm), which do not contribute to channel
relief transformation (they are absent in the channel
alluvium), and channel-forming sediments, of which
channel relief forms are composed. The transit sediments occur in suspension, while the channel-forming
sediments, both in the suspension and as bed forms.
The transport rate of transit sediments at flood peak in
cross-section number 12 reached 6000 kg/s at their
concentration of 8.9 kg/m3 and dropped to 370–
440 kg/s at concentration of 1.0–1.2 kg/m3 by the end
of the study. The temporal variations of transit sediment concentration were high: at flood peak in different measurement periods, at a narrow variation range
of water discharges (620‒670 m3/s), it changed
4.5 times, i.e., from 2 to 9 kg/m3, at lesser water discharges, its variations dropped to 2 times. At the same
time, the spatial variation of transit sediment concentration within the study reach was relatively low, averaging 20% (at a maximum of 60%). On the average,
50% of transit sediments are fractions with particle
diameter of 0.05‒0.01 mm; 20% have the size of
0.01‒0.005 mm; and 30%, finer than 0.005 mm.
Analysis of a vast body of data on the grain-size distribution of suspended sediments at different hydrological stations along the lower Terek shows that the content of transit sediments in their total concentration
averages ~60%. Therefore, their annual yield in the
study area is estimated at 11 million t. Transit sediments are generated on river basin; they are not deposited in the Lower Terek channel (only on its floodplain); and their relationship with the hydraulic characteristic of the channel flow is indirect.
Conversely, the concentration of channel-forming
sediments depends on the local hydraulic conditions
in the channel. The bottom concentration (at a height
of Δ above the bed) of suspended channel-forming
sediments of the ith fraction Ci(Δ) is determined by the
proportion of the sediments of the ith fraction of bed
alluvium pi, the settling velocity of the particles ωi, and
the flow velocity U. The bed alluvium in the studied
river reach is sand with a mean diameter of 0.4 mm: on
the average, 20% are coarse sands with diameter of
1‒0.5 mm; ~60% are medium-size sands with a diameter of 0.5‒0.25 mm; 20% are fine sands with a diameter of 0.25‒0.1 mm; and ~2% are fine sands with a
diameter of 0.1‒0.05 mm. The distribution of sediments over fractions is close to lognormal. For the
lower Terek, the following empirical relationship was
derived [2]:
()
3
(1)
C i ( Δ ) = 1.2 × 10 −5 U pi ,
ω
here, the value of Δ is determined by the height of the
fish-like sink-stone, on which the nozzle of the vacuum bathometer was installed, and amounts to 12–
15 cm. The distribution of the channel-forming sediment concentration over flow depth can be well
approximated by Rouse–Velikanov formula [4]. Both
time and space variations of the concentration of
channel-forming sediments and their proportion in
the total suspension concentration are wide. Thus, at
the flood peak of July 21, 1982, the proportion of
channel-forming sediments (with a diameter of
>0.05 mm), measured at 13 cross-sections, varied
from 16 to 56% in the section-averaged suspended
sediment concentration. Accordingly, the particlesize distribution of the suspended channel-forming
sediments also varies widely, though commonly dominating are sediments of the fraction 0.25‒0.1 mm (up
to 60% of channel-forming sediments).
On the average, the proportion of channel-forming
sediments in the total suspended sediment concentration is ~40%; and their annual runoff in the study area
is estimated at 7.3 million t.
SAND-WAVE RELIEF
IN THE LOWER TEREK CHANNEL
The lower reaches of the Terek, as well as many
other rivers with sand alluvium, shows a complex hierarchic structure of channel sand-wave forms. Up to six
levels of sand-wave hierarchy can be identified. Large
sand waves—alternating bars and middle bars—with a
mean length of 600 m commonly consist of mediumsize sand-waves, both larger, with a mean length of
230 m, and smaller with a mean depth of 90 m, modeling the larger ones. The large and medium-size sand
waves are inundated during floods, while at low discharges, they rise over the surface and determine the
configuration of the low-water channel. This complex
is superimposed by small and smallest sand-waves,
which can include up to three hierarchic levels: the
largest second-order dunes (dunes_2), the smaller
first-order dunes (dunes_1), and the smallest sandwaves—megaripples. The small and smallest sandwaves almost always are under water. The repeated
measurements in the summer of 1982 were focused on
this type of bottom bed forms.
Classification of Bed Forms
The common hierarchy of small and smallest bed
forms in rivers with sand alluvium is as follows [1, 12,
14]: dunes_2 superimpose on the medium-size sandwaves; superimposed on dunes_2 are smaller dunes_1,
which, in their turn, are modeled by megaripples. The
WATER RESOURCES
Vol. 44
No. 2
2017
MORPHOLOGY AND DYNAMICS OF ACTIVE BED FORMS
229
Flow direction
July 21, 1982, 15.10
July 21, 1982, 18.50
July 22, 1982, 10.15
July 22, 1982, 19.20
500 cm
July 23, 1982, 08.10
300
100
July 23, 1982, 16.10
Marks
Distance
from mark 5
9
10
400
11
500
13
12
600
700 m
Fig. 2. Morphology and dynamics of active dunes_1 in the Lower Terek in the profile 1 (marks 9–13) during flood rise at conventional level of 800 cm. The distances are measured from mark 5.
situation during floods in the lower Terek is different.
Full three-member bed form hierarchy can be generated at some reaches of the river channel; the above
classification is applicable to this case. In other
reaches of the channel (or in the same reach, but in
other time), no full hierarchy will form: it can be double-member or absent at all. Such incomplete structure of the bed form relief most often generates in the
Lower Terek (Fig. 2); therefore, to classify bed forms
in such structures requires the following additional
analysis of channel relief.
It is reasonable to assign the bed forms of the Lower
Terek special indices. In the case of three-member
hierarchy, we denote the largest bed forms as bed
forms_003, on which bed forms_023 are superimposed, with bed forms_123 on the top level. In the case
of double-member hierarchy, bed forms_020 will be at
the lower level and bed forms_120 at the upper level.
WATER RESOURCES
Vol. 44
No. 2
2017
In the case of single-member channel relief structure,
the index bed forms_100 will be used.
Theoretical studies into the origin of the bed form
relief of river channel, carried out by a method of small
initial perturbations, showed that, for megaripples, the
dimensionless wave number krD = 2πD/Lr is related to
the Froude number Fr = U gD by an inverse proportion relationship [8, 12]:
kr D = 1 ,
Fr
(2)
where Lr is the length of the megaripples, U is flow
velocity, D is its depth, measured as the depth over the
bed form crest plus half the bed form height, g is acceleration due to gravity. For the boundary between the
small and medium-size sand-waves, a formula was
obtained [10]:
230
ALEKSEEVSKII, SIDORCHUK
N
_003 _023 _23
Dunes_2
35 70 140
Dunes_1
Megaripples
30 60 120
25 50 100
20 40 80
15 30 60
10 20 40
5 10 20
0
–1.5
–1.3 –1.1
–0.9 –0.7 –0.5 –0.3 –0.1 0 0.1 0.2 0.3 0.4 0.5 log (kD × Fr)
–1.4
–1.2
–1.0 –0.8
–0.6 –0.4 –0.2
Bed form_003
Bed form_023
Bed form_123
Fig. 3. Histograms and plots of functions approximating the values of log(2πDFr/L) for bed forms_003, _023, and _123.
0.3
2g
(3)
k mD = 2 exp ( − 2.0Fr ) ,
C 0
here C0 is the coefficient in Chezy formula, km is the
wave number for medium-size sand-waves. In the diagram 2πD/L ~ Fr, a domain of small sand-waves is
located between these two lines, where theoretical calculations give no additional boundaries.
The empirical data on the three-member hierarchical structure of dunes and megaripples in the river
channel, collected in rivers with sand alluvium [12],
confirm the theoretical relationships (2) and (3). The
results of measurements for bed forms in the Lower
Terek, which form a complete three-member hierarchy (bed forms _123, _023, and _003), imposed on the
diagram 2πD/L ~ Fr, also confirm the earlier identified domains, despite the quite explainable stochastic
character of their boundaries. The probability densities for the logarithms of coefficient a = 2πDFr/L for
bed forms of these classes (Fig. 3) can be well approximated (Kolmogorov–Smirnov test 0.03–0.05) by the
normal distribution (i.e., the distribution of coefficient a is lognormal).
In the case of incomplete hierarchy, the bed forms
with indices _100, _120, _103, and _023 can be
referred to certain class only with some probability.
Based on the size and hydraulic characteristics of the
flow (by the value of coefficient a), the bed forms_100
and _120 can be referred to either megaripples or
dunes_1, or dunes_2; the bed forms_103 are both
megaripples and dunes_1; the bed forms_020 are
mostly dunes_1. As a first approximation, the regional
classification of such bed forms can be based on histograms of the logarithms of the coefficient a = 2πDFr/L
for bed forms_123, _023, and _003 (Fig. 3). The number of megaripples nr in the group of N different bed
forms, combined by the value of log a (in some specified range), can be evaluated as:
N _123
(4)
,
N _123 + N _ 023 + N _ 003
and they can be separated from this group by random
choice. To determine the number of dunes_1 in such
group, N_123 is replaced by N_023 in the numerator
of formula (4), and for dunes_2, by N_003. Here,
N_123, N_023, and N_003 are the numbers of bed
forms of appropriate types in measurements on the
Lower Terek in this range of loga values, according to
histograms in Fig. 3.
Formula (4) was applied to data arrays with bed
form characteristics of the Lower Terek, resulting in
that the bed forms in the incomplete hierarchies were
also classified as megaripples, dunes_1, and dunes_2.
The probability densities for the logarithms of the
coefficient a = 2πDFr/L for the entire complex of bed
forms are also well approximated by a normal distribution with the means and standard deviations close to
those obtained only for bed forms in three-member
hierarchic complexes (Table 1).
nr = N
Morphology of Megaripples and Dunes
The results of processing measurement data on the
Lower Terek show (Table 1; Fig. 4), that the megaripple length in the Lower Terek for the mode of lognormal probability density function can be described by
the formula
k r D = 0.98 ,
Fr
(5)
Lr = 6.4DFr.
(6)
or
WATER RESOURCES
Vol. 44
No. 2
2017
MORPHOLOGY AND DYNAMICS OF ACTIVE BED FORMS
231
Table 1. Characteristics of the normal distribution function for logarithms of the coefficient a = 2πDFr/L for bed forms
in the Lower Terek
Bed forms
Number of bed forms
Mean value of log(a)
Standard deviation
Bed forms in three-member hierarchic complexes
825
–0.074
0.243
305
–0.538
0.211
158
–0.936
0.22
Combined bed form array after classification by formula (4)
5840
–0.009
0.196
1741
–0.451
0.195
575
–0.865
0.187
Megaripples
Dunes_1
Dunes_2
Megaripples
Dunes_1
Dunes_2
In the case of dunes_1, the coefficient in formula
(5) is 0.35, and for dunes_2, it is 0.14 (Fig. 4). The
appropriate coefficients in formula (6) are 17.7 and
46.0. A linear relationship appears at Fr < 0.6, and, in
more kinetic flows, the relationship becomes exponential.
For the probability density functions constructed
for bed form lengths within narrow ranges of flow
velocities (ΔU = 0.5 m/s) and depths (ΔD = 0.5 m), the
distribution is still lognormal. As was the case with the
dunes and megaripples in other rivers [12], the Lower
Terek shows a close relationship between the standard
deviation σ and the mean values Lav. Earlier, a linear
relationship was found to exist between these variables, i.e., the coefficient of variation σ/Lav kept constant whatever the combination of flow velocity and
depth. Data on the Terek, obtained for a wider range
of hydraulic characteristics of the flow, show that the
linear relationship between σ and Lav is still true here
with reliability R2 = 0.94. The coefficient of variation
is 0.43 for megaripple lengths and 0.28 for dunes_1; in
the case of dunes _2, the available data are not enough
to obtain a reliable relationship.
During the flood of July–August 1982, most bed
forms in the Lower Terek were isometric in the longitudinal section. For megaripples, the ratios of the
upper and lower slopes A = Lu/Ll, in 75% of cases fall
within 0.5‒2.0; and, in 30% of cases, within 0.8‒1.25.
For dunes_1, 64% of A values lie within 0.5‒2.0; and
23%, within 0.8‒1.25; for dunes_2, these ranges contain 68 and 22% skewness values. The probability distribution for the logarithms of skewness A is described
by Laplace function (Fig. 5) with a distinct mode at
values logA ~ 0 (A ~ 1). Within narrow ranges of flow
velocities and depths, a dependence of bed form asymmetry σ on the mean values Аav can be constructed.
This dependence is nonlinear and it can be approximated by a power function with R 2 = 0.52‒0.58:
q
σ = rAav
.
(7)
For megaripples, r = 0.8, q = 1.6; for dunes_1 – r = 1.0
and q = 2.0. Therefore, for bed forms with skewness
WATER RESOURCES
Vol. 44
No. 2
2017
Kolmogorov–Smirnov test
0.025
0.04
0.056
0.03
0.05
0.051
values less than unit (the upper slope is shorter than
the lower slope), the coefficient of variation of skewness is less than that for bed forms with skewness
greater than unit (the upper slope is longer than the
lower).
The isometry of bed forms is an important indicator to the active stage of their motion. Also typical of
this stage is the convex shape of both upper and lower
slopes, which is most distinct in large dunes (Fig. 2).
The relative height of the megaripples h/D changes
with flow velocity. At velocities of 0.1‒0.5 m/s, the
bed of the channel is nearly flat; at the velocity of 1.3–
1.5 m/s, h/D reaches a maximum; with a further
increase in the velocity, the relative height of the
megaripples decreases. This regularity is not strong for
the medium values of h/D (Fig. 6a), but it can be
clearly seen in the upper envelope of the field of points.
The relative height of dunes_1 increases with flow
velocity increasing to 2.5‒2.7 m/s, while at a further
increase in the velocities, the bed forms show a tendency toward erosion. Dunes_2 keep increasing their
height at flow velocities of ~3 m/s (Fig. 6a). These regularities are clearly seen for both medium and maximal values of h/D. Thus, an increase in flow velocity
causes flattening of bed forms; in this case, the megaripples require lesser flow velocities than the dunes_1
do; the dunes_2 show no such effect within the measured velocity range.
Within narrow flow velocity ranges, the probability
densities of the relative heights of all types of megaripples are well described by gamma distribution
(Fig. 6b). The dependence of standard deviation σ of
the relative bed form height on their mean values
(h/D)av is linear, and the coefficient of variation is 0.52
for megaripples and 0.55 for dunes_1.
In shallow areas, a drop in river water level may
cause a short-time increase in flow velocity up to
2.5‒3.0 m/s in the conditions of low flow depth. In
this case, the bed forms will be completely washed out
to give place to a flat bed of the second stage. In some
cases, such regime in shallow zones leads to the forma-
232
ALEKSEEVSKII, SIDORCHUK
kD = 2πD/L
10
kd1D = 0.35/Fr
Megaripples
Dunes_1
krD = 0.98/Fr
kd2D = 0.14/Fr
1
Dunes_2
0.1
0.01
0.1
1
Er
Fig. 4. Correlation of the dimensionless wave number kD = 2πD/L for megaripples, dunes_1 and dunes_2 in the Lower Terek
with the Froude number.
tion of antidunes, which have concave slopes on both
upstream and downstream sides (Fig. 7).
Dune Dynamics
The bed form celrity was determined by repeated
measurements within the longitudinal profile 1 in a
160-m-long segment between marks 10 and 12
(Figs. 1, 2). This is a single pool hollow with uniform
flow and steady depths under the bed forms; therefore,
the mean hydraulic characteristics and the celerities
were evaluated for individual bed forms over the time
of their passage through this segment (Table 2).
A histogram of the logarithms of coefficient a
shows these bed forms to belong to the class of
dunes_1. Commonly, for a part of their existence time,
they are disturbed by megaripples (bed form_120),
while in the other part of the time, they are simple
dunes_1 (bed form_100). In some cases, dunes_1 disturb the surface of large dunes_2 (Fig. 2 gives the situation as of the morning of July 22, 1982). In all cases,
these are active bed forms, and their shape does not
show any significant unidirectional changes (skewness) over the movement time [11].
The celerity of dunes_1 (Cg, m/h) is determined by
the hydraulic characteristics of the flow and can be
WATER RESOURCES
Vol. 44
No. 2
2017
MORPHOLOGY AND DYNAMICS OF ACTIVE BED FORMS
N
2400
2200
2000
1800
1600
1400
1200
1000
800
600
400
200
0
Laplace (x; –0.0737; 0.2184)
N = 4990, Mean = –0.074, StdDv = 0.31
233
(а)
h/D
0.50
0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
Dunes_2
Dunes_1
Megaripples
–1.2 –0.8 –0.4
0 0.2 0.4 0.6 0.8 1.0 1.2
–1.0 –0.6 –0.2
log (Lu/Ld)
0
N
700
Fig. 5. Histogram and a plot of approximating function for
the skewness of Lower Terek megaripples.
600
0.5
1.0
1.5
2.0
2.5
(b)
3.0
U, m/s
N = 5770, Mean = 0.1, StdDv = 0.056
Gamma (x/0.0298; 3.3415)/0.0298
500
satisfactorily described by an expression similar to the
Snishchenko–Kopaliany formula [13]:
3
(8)
C g = bU .
D
The mean value of the empirical coefficient b in formula (8) is 1.91 s3/(m h). The coefficient tends to
increase with the simplification of bed form hierarchy,
simple dunes_1 commonly move faster than those disturbed by megaripples:
b = b1 + b2 p.
(9)
Here, p is the proportion of time t/T, where t is the
time within which the dunes_1, when passing a segment 160 m within time T, are simple bed forms; the
coefficients b1 and b2 are 1.8 and 0.6 s3/(m h), respectively. Formula (9) indicates only a tendency, the scatter of points in this relationship is high; though, in
some cases, this tendency is very clear (Fig. 3.13 in
[12]).
The comparison of data on the motion of active
dunes_1 for the Lower Terek with the measured celerities of active dunes in the Niger R. (Table 1 in [11])
shows them to be in satisfactory agreement (Fig. 8).
For the combined data, the mean value of coefficient
b equals 1.9 (at the Cg expressed in m/h) with the reliability R 2 = 0.82. In addition, this data can be satisfactorily (with a reliability R2 = 0.71) approximated by
Snishchenko–Kopaliani formula with coefficient
0.016 (with Cg expressed in m/s).
400
300
200
100
0
0.04
0.10
0.16
0.22
0.28
0.34
0.40
h/D
Fig. 6. (a) The relative height of bed forms versus flow
velocity and (b) probability distribution function for the
relative height of megaripples.
In such calculations, it is convenient to determine
the mean values of the hydraulic characteristics of
flow and, for bed form morphometry, the heights and
celerities, not for individual bed forms, but for all bed
forms in the area under consideration for the specified
time interval (Table 3). It can be seen that, during the
flood in July–August 1982, the unit sediment transport rate in the form of bed forms in the pool hollow,
on the average, is 7.4% of the transport rate of susсm
750
Flow direction
Edge
650
550
SEDIMENT TRANSPORT RATE
The unit sediment transport rate qs3 in the form of
bed forms is evaluated by the formula
(10)
q s3 = 0.5hgC g.
WATER RESOURCES
Vol. 44
No. 2
2017
0
20
40
60
80
100 m
Fig. 7. Antidunes in shallows in the Lower Terek channel
during flood recession.
234
ALEKSEEVSKII, SIDORCHUK
Table 2. Morphology and dynamics of individual dunes_1 of the Lower Terek on profile 1 between marks 10–12 (N is the
number of measurements with the particular dune_1, T is the duration of observations of the motion of this dune_1, U is
the average flow velocity over the dune in this period, D is the mean depth of the flow over the dune + 1/2 of dune height,
L is mean dune length, h is the mean dune height, C is the mean dune celerity, p is the part of the time when the dune was
not disturbed by megaripples)
Observation period
Dune
number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
beginning
end
date
time
date
July 22
July 22
July 22
July 23
July 23
July 25
July 26
July 26
July 27
July 28
July 28
July 29
July 29
July 30
July 31
Aug. 1
Aug. 1
Aug. 3
Aug. 4
Aug. 4
Aug. 6
Aug. 6
Aug. 7
Aug. 9
Aug. 10
Aug. 11
Aug. 12
Aug. 14
10:30
10:30
17:10
8:10
16:10
9:00
13:40
13:40
9:00
9:45
9:45
10:40
10:40
13:20
19:20
12:15
12:15
9:25
13:40
13:40
9:45
19:20
9:45
18:35
8:00
16:10
10:00
13:00
July 22
July 23
July 23
July 24
July 24
July 25
July 26
July 26
July 27
July 28
July 29
July 29
July 30
Aug. 1
Aug. 3
Aug. 3
Aug. 3
Aug. 3
Aug. 6
Aug. 7
Aug. 7
Aug. 7
Aug. 9
Aug. 10
Aug. 12
Aug. 14
Aug. 13
Aug. 14
N
T, h
U, m/s
D, m
L, m
H, m
C, m/h
p
2.25
2.51
2.62
2.29
2.53
2.53
2.36
2.48
2.73
2.48
2.29
2.24
2.02
2.07
2.08
2.14
2.19
2.2
2.02
2.04
2.05
2.08
1.96
2.31
2.05
1.93
1.72
1.77
3.9
4.6
4.57
5.22
5.2
4.75
7.48
7.65
6.22
5.17
5.3
5.35
5.84
6.44
7.33
7.01
6.36
5.93
6.72
6.19
5.71
5.92
6.32
6.44
5.99
6.12
6.5
6.35
60.84
71.03
81.92
46.94
84.89
113.4
45.64
114.43
104.28
18.32
46.92
37.57
115.34
68.31
32.72
40.8
33.09
67.33
40.05
44.89
60.47
54.83
43.81
53.99
68.61
68.44
48.04
43.69
2.77
3.1
3.53
2.47
2.77
2.45
2.47
2.4
2.3
1.4
1.52
1.3
2
2.13
2.48
2.35
1.62
3.05
2.31
2.52
2.52
1.23
1.7
1.96
3.39
2.51
1.67
3
6.49
6.22
5.23
5.41
6.1
6.16
4.7
3.59
5.26
7.5
3.37
5.14
3.06
2.14
2.74
2.69
2.74
3.52
2.23
2.36
2.53
2.75
2.5
3.73
2.54
1.66
1.67
3.91
0.33
0.25
0.25
0.75
0.33
0
0.33
0.5
0
0.67
0
0
0
0
0.5
1
0.5
0
0.38
0.1
0.6
0
0.33
0.6
0.13
0
0
0.5
time
19:20
8:10
16:10
13:20
15:50
12:30
18:40
17:10
19:10
17:50
18:50
18:50
17:10
15:25
9:25
15:25
15:25
15:25
19:20
16:07
16:07
16:07
18:35
16:15
10:00
18:55
19:05
18:55
3
4
4
4
3
2
3
2
3
3
5
2
4
4
6
6
6
2
8
10
5
3
3
5
8
8
3
2
8.8
21.7
23.0
29.2
23.7
3.5
5.0
3.5
10.2
8.1
33.1
8.2
30.5
50.1
62.1
51.2
51.2
6.0
53.7
74.4
30.4
20.8
56.8
21.7
50.0
74.8
33.1
5.9
pended channel-forming sediments qs2, and the total
transport rates of channel-forming sediments account
for >70% of transit sediment transport rate. These are
the characteristics of the conditions in the pool for the
given flood. On the average for the entire channel of
the Lower Terek, the part of the channel-forming sediments amounts to 40% of all suspended sediments
and, clearly, it is less than that in the pool hollow
during spring flood. The proportions of the transport
rates of the channel-forming sediments moving in different manner is more stable, as they are determined
by the same hydrological characteristics of the flow
and these transport rates satisfy the linear relationship
(11)
q s3 = 0.074q s2
with approximation reliability R 2 = 0.73. Now, with
the annual yield of suspended channel-forming sediments estimated at 7.3 million t, the annual sediment
yield in the form of bed forms will be 0.54 million t or
~0.3 million m3. As the alluvium in the Lower Terek is
not coarse, and the turbulence and the mean flow
velocity are high, a large amount of channel-forming
sediments can be transported in suspension. ThereWATER RESOURCES
Vol. 44
No. 2
2017
MORPHOLOGY AND DYNAMICS OF ACTIVE BED FORMS
C1D, m2/h
80
70
60
C1D = 1.7U3.1
R2 = 0.77
50
fore, the sediment yield in the form of bed forms, estimated at 7.4% of the yield of channel-forming sediments in suspension, seems to be close to the lower
boundary for such relationships in rivers with sand
alluvium. For example, in the Lower Lena at Kyusyur
station, this ratio is close to 100%, i.e., the amounts of
channel-forming sediments transported as bed forms
and as suspension are nearly equal [15].
40
CONCLUSIONS
Active dunes in the Terek
30
20
10
0
Active dunes
in the Niger
0.5
1.0
1.5
235
2.0
2.5
3.0
U, m/s
Fig. 8. Correlation of the celerity of active dune_1 in the
Lower Terek and the Niger with the flow velocity and
depth.
Active bed forms of three main classes—megaripples, dunes_1, and dunes_2—form in the Lower Terek
even during small summer floods with water discharges of 600–700 m3/s. The analysis of their morphology and dynamics generally confirms the major
relationships obtained for bed forms in other rivers
with sand alluvium. Data on the Lower Terek show
these can be applied to a wider range of hydraulic flow
characteristics, i.e., velocities, depths, and Froude
numbers.
The hierarchy of active bed forms in the Lower
Terek channel during floods is rarely complete (threemember). Two-member hierarchy, if any, is common.
This complicates the classification of bed forms.
Table 3. Daily averages of flow hydraulic characteristics (flow velocity U and depth D) and morphodynamic characteristics
of dunes_1 (height h and celerity Cg) in the 160-m-long segment of profile 1 between marks 10 and 12 during the flood of
July 22–August 10, 1982 in the pool hollow on the Lower Terek; calculated by formula (10) specific (per unit channel width)
transport rates of channel-forming sediments moving as bed forms qs3 are compared with the measured transport rates of
transit sediments qs1 and the transport rates of suspended channel-forming sediments qs2 calculated by formula (1) and
Rouse–Velikanov curve
Date
July 22
July 23
July 24
July 25
July 26
July 27
July 28
July 29
July 30
July 31
Aug 1
Aug. 2
Aug. 3
Aug. 4
Aug. 5
Aug. 6
Aug. 7
Aug. 9
Aug. 10
U, m/s
D, m
h, m
Cg, 105 m/s
qs3, m2/s
2.2
2.3
2.2
2.3
2.3
2.0
1.9
2.2
2.1
2.0
2.1
1.9
2.0
2.0
2.3
2.0
1.8
1.8
1.8
4.5
5.3
5.5
7.3
4.7
5.9
6.2
6.7
6.6
5.9
6.2
6.3
6.2
6.0
6.3
5.9
6.3
6.1
6.2
2.9
2.3
2.1
2.1
1.7
2.0
1.6
2.4
2.4
1.9
2.0
2.7
2.3
1.8
2.6
2.8
1.6
1.5
2.5
115.1
153.0
127.7
185.6
146.2
89.8
75.5
86.8
85.9
68.5
60.8
59.6
73.7
63.5
132.5
69.4
45.8
31.7
68.0
0.00162
0.00173
0.00134
0.00194
0.00120
0.00088
0.00060
0.00110
0.00105
0.00065
0.00061
0.00078
0.00084
0.00056
0.00167
0.00098
0.00040
0.00024
0.00087
WATER RESOURCES
Vol. 44
No. 2
2017
qs1, m2/s qs2, m2/s (qs2 + qs3)/qs1, % qs3/qs2, %
0.012
0.028
0.027
0.076
0.035
0.026
0.030
0.030
0.018
0.013
0.012
0.035
0.012
0.016
0.016
0.011
0.011
0.010
0.018
0.0143
0.0212
0.0162
0.0242
0.0187
0.0099
0.0094
0.0217
0.0145
0.0107
0.0127
0.0090
0.0104
0.0102
0.0237
0.0110
0.0068
0.0055
0.0064
134.5
82.9
64.6
34.1
57.5
40.9
33.4
74.9
86.8
90.4
108.8
28.1
89.6
67.0
156.1
105.7
67.2
56.3
39.7
11.4
8.2
8.3
8.0
6.4
8.9
6.3
5.1
7.2
6.1
4.8
8.6
8.1
5.4
7.0
8.9
5.8
4.3
13.5
236
ALEKSEEVSKII, SIDORCHUK
The morphology of active bed forms is essentially
stochastic and can be adequately described by probability distribution functions of bed form characteristics
for some ranges of hydraulic flow characteristics. The
distribution functions are asymmetric and can be best
described by lognormal or gamma distribution. The
first two moments of these distributions are commonly interrelated, making the distributions oneparameter.
The arithmetic or geometric means of the morphometric characteristics of bed forms show stable relationships with hydraulic flow characteristics, in some
cases (as for megaripple lengths), such empirical relationships agree with theoretical calculations (formulas
(5), (6)).
The celerity of active dunes_1 can be acceptably
described by Snishchenko–Kopaliani formula or,
somewhat better, by its modification—formula (8).
The dynamics of bed forms in the hierarchic system
depends on the complexity of this system: the more
complex the hierarchy, the slower the motion of the
bed forms basic for this hierarchy.
The part of the channel-forming sediments moving
as bed forms in the Lower Terek is ~7% of the transport rate of all channel-forming sediments, i.e., it is
close to the lower limit for such relationships in rivers
with sand alluvium.
ACKNOWLEDGMENTS
This study was implemented under State funded
program “The Evolution and Transformation of Erosion–Channel Systems under Changing Environment
and Human impact.”
REFERENCES
1. Alekseevskii, N.I., Bed load transport at a developed
structure of channel relief, Meteorol. Gidrol., 1990,
no. 9, pp. 100–105.
2. Alekseevskii, N.I. and Sidorchuk, A.Yu., Morphology
and dynamics of channel relief in the Lower Terek, in
Zemel’nye i vodnye resursy (Land and Water Resources),
Moscow: Izd Mosk. Univ., 1990, pp. 87–95.
3. Berkovich, K.M. and Sidorchuk, A.Yu., Dynamics of
channel relief of the Lower Niger, Izv. Akad. Nauk
SSSR, Ser. Geogr., 1983, no. 4, pp. 66–72.
4. Velikanov, M.A., Dinamika ruslovykh potokov (Channel
Flow Dynamics), vol. 2 Nanosy i ruslo (Sediments and
the Channel), Moscow: Gostekhizdat, 1955.
5. Gorelits, O.V., Zemlyanov, I.V., Pavlovskii, A.E., and
Illaev, T.S., Channel deformations and runoff redistribution within the Terek River mouth reach after the
catastrophic flood of the year 2002, Water Resour.,
2006, vol. 33, no. 6, pp. 628–636.
6. Zaikov, B.D., Srednii stok i ego raspredelenie v godu na
territorii SSSR (Average Runoff and Its Inter-Year Distribution in USSR Territory), Moscow: Gidrometeoizdat, 1946.
7. Lur’e, P.M., Vodnye resursy i vodnyi balans Kavkaza
(Water Resources and Water Balance of the Caucasus),
St. Petersburg: Gidrometeoizdat, 2002.
8. Mikhinov, A.E., Instability of sand waves in a deformable channel, Meteorol. Gidrol., 1983, no. 11, pp. 84–91.
9. Rukhin, L.B., Osnovy litologii (Fundamentals of
Lithology), Leningrad: Gostoptekhizdat, 1961.
10. Sidorchuk, A.Yu., Morphology and dynamics of channel relief forms in Terek mouth area, in Kaspiiskoe more.
Gidrologiya ust’ev rek Tereka i Sulaka (Caspian Sea.
Hydrology of the Terek and Sulak River deltas), Moscow: Nauka, 1993, pp. 89–102.
11. Sidorchuk, A.Yu., Evaluating bedload transport rate in
a river bed taking into account data on the active and
passive bed-forms dynamics, Water Resour., 2015,
vol. 42, no. 1, pp. 38–51.
12. Sidorchuk, A.Yu., Struktura rel’efa rechnogo rusla
(Relief Structure of River Channels), St. Petersburg:
Gidrometeoizdat, 1992.
13. Snishchenko, B.F. and Kopaliani, Z.D., On the celerity
of bed forms in rivers and laboratory experiments, Tr.
GGI (Gos. Gidrol. Inst., State Hydrological Institute),
1978, no. 252, pp. 30–37.
14. Alekseevskiy, N., Movement of bed forms and sediment yield of rivers, IAHS Publication, 2004, no. 288,
pp. 395–403.
15. Alekseevskiy, N. and Sidorchuk, A., Total sediment
yield of the Lena River, Eastern Siberia, Hydrological
and Geochemical Processes in Large Scale River Basins.
Manaus, 1999, p. 26.
16. Best, J., The fluid dynamics of river dunes: a review and
some future research directions, J. Geophys. Res., 2005,
vol. 110, no. 4, p. 1–21.
17. Vionnet, C., Marti, C., Amsler, M., and Rodriguez, L.,
The use of relative celerities of bedforms to compute
sediment transport in the Paraná River, IAHS Publ.,
1998, no. 249, pp. 399–406.
Translated by G. Krichevets
WATER RESOURCES
Vol. 44
No. 2
2017