IJSRC293_proof ■ 5 April 2020 ■ 1/13

International Journal of Sediment Research xxx (xxxx) xxx

55
Contents lists available at ScienceDirect 56
57
58
International Journal of Sediment Research 59
60
journal homepage: www.elsevier.com/locate/ijsrc 61
62
63
Original Research 64
65
1 Experimental assessment and prediction of temporal scour depth 66
2 67
3 Q4 around a spur dike 68
4 69
a, *
5 Q3 Manish Pandey , Manousos Valyrakis b, Meilan Qi c, Anurag Sharma d, 70
6
Ajay Singh Lodhi e 71
7 72
a
8 Department of Civil Engineering, National Institute of Technology, Warangal, India
b 73
School of Engineering, University of Glasgow, Glasgow, UK
9 c 74
School of Civil Engineering, Beijing Jiaotong University, Beijing, China
10 d
Department of Civil Engineering, Institute of Infrastructure Technology Research and Management, Ahmedabad, India 75
11 e
College of Agriculture, Waraseoni District, Balaghat M.P., India 76
12 77
13 78
14 a r t i c l e i n f o a b s t r a c t 79
15 80
16 Article history: Spur dikes are river training structures that have been extensively used worldwide for towards 81
17 Received 27 July 2019 enhancing flood control and the stability of embankments and riverbanks. However, scour around spur 82
Received in revised form dikes can be a major problem affecting their stability and hydraulic performance. The precise compu-
18 83
20 March 2020
19 tation of temporal scour depth at spur dikes is very important for the design of economical and safe spur
Accepted 23 March 2020 84
20 Available online xxx
dikes. This study focuses on experimentally assessing the temporal variation of scour depth around a
85
21 vertical wall spur dike and identifying the parameters, which mostly influence spur dike performance for
a channel bed surface comprised of sand-gravel mixtures. In the current study, the authors did physical
86
22 Keywords: 87
Spur dike
experiments in a flume based study to obtain new data, aimed at deriving a new predictive model for
23 88
Sand-gravel sediment mixture spur dike scour and comparing its performance to others found in the literature. It was found that the
24 dimensionless temporal scour depth variation increases with an increase in (i) the threshold velocity 89
Clear-water scour
25 Time-dependent scour depth ratio, (ii) the densimetric Froude number of the bed surface sediment mixture, (iii) the flow shallowness 90
26 (defined as the ratio of the approach flow depth, y, to the spur dike’s transverse length, l), and (iv) the 91
27 flow depth-particle size ratio. It is also concluded that the temporal scour depth variation in the sedi- 92
28 ment mixture is influenced by the non-uniformity of sediment and decreases with an increase in the 93
29 non-uniformity of the sediment mixture. A new mathematical model is derived for the estimation of 94
30 temporal scour depths in sand-gravel sediment mixtures. The proposed equation has been calibrated and
95
31 validated with the experimental data, demonstrating a good predictive capacity for the estimation of
96
temporal scour depth evolution.
32 97
© 2020 Published by Elsevier B.V. on behalf of International Research and Training Centre on Erosion and
33 98
Sedimentation/the World Association for Sedimentation and Erosion Research.
34 99
35 100
36 101
37 102
38 103
39 1. Introduction commonly used to divert flow from channel banks and protect 104
40 them against erosion. Generally, spur dikes are comprised of ver- 105
41 Scour and erosion processes at the bed surfaces and banks of tical built walls placed at an angle to the mean flow direction. Spur 106
42 natural rivers as well as built waterways (i.e. irrigation or naviga- dikes achieve their major function of protecting stream banks (i) 107
43 tion channels), are major challenges that need to be tackled effi- directly via diverting the flowing water away from the stream 108
44 ciently for proper water resources management. Scour processes in banks, and (ii) indirectly by reducing the flow velocity. Spur dikes 109
45 water bodies are the primary cause for geomorphic changes leading are amongst the most hydraulically and cost effective river training 110
46 to land loss, excessive sediment yield, and subsequent degradation structures that can be built at the channels' banks (Zhang et al., 111
47 of water quality. Spur dikes are man-made hydraulic structures, 2012), to safeguard their stability, maintain a channel’s navigation 112
48 capacity, as well as help maintain water quality and restore aquatic 113
49 habitat (Lodhi et al., 2016; Yazdi et al., 2010). 114
50 Flow structures around a spur dike may greatly vary, exhibiting 115
51 * Corresponding author. three-dimensional features in its vicinity. These complex flow 116
E-mail address: manishpandey3aug@gmail.com (M. Pandey).
52 117
53 118
https://doi.org/10.1016/j.ijsrc.2020.03.015
54 1001-6279/© 2020 Published by Elsevier B.V. on behalf of International Research and Training Centre on Erosion and Sedimentation/the World Association for Sedimentation 119
and Erosion Research.

Please cite this article as: Pandey, M et al., Experimental assessment and prediction of temporal scour depth around a spur dike, International
Journal of Sediment Research, https://doi.org/10.1016/j.ijsrc.2020.03.015
 IJSRC293_proof ■ 5 April 2020 ■ 2/13

2 M. Pandey et al. / International Journal of Sediment Research xxx (xxxx) xxx

1 structures scale with the size of the spur dike and flow velocity, dimension (size of sieve opening) of grains with cumulative count 66
2 containing sufficient energy for the removal of bed material (Diplas at 84% and 16% finer, respectively. In a non-uniform sediment 67
3 et al., 2008; Valyrakis et al., 2010; Valyrakis et al., 2013), leading to mixture (parent bed material), the removal of finer sediment par- 68
4 the generation of local scour. The study of scour processes around ticles ceases with the development of an armor layer, comprised of 69
5 spur dikes is very important for river training and can provide coarser sediment (Kothyari et al., 2007). The geometric standard 70
6 practical guidelines that hydraulic engineers and practitioners may deviation for the armor layer is always smaller to that of its parent 71
7 use for their effective operation and management (Zhang et al., particle size distribution (Oliveto & Hager, 2002). Once formed, the 72
8 2018). armor layer represents a water-worked bed surface, which remains 73
9 Local scour is well defined in fluvial hydraulics as the removal of stable, not allowing the further removal of finer sediment. The 74
10 sediment particles from alluvial streams. It is an important same processes are observed during the dynamic evolution of scour 75
11 geomorphic process that affects virtually all hydraulic infrastruc- hole geometry, which gradually progresses until the geometry is 76
12 ture, ranging from bridge piers (Valyrakis et al., 2015) to embank- stable, indicating that equilibrium scour has been reached. Typi- 77
13 ments and riverbanks (Liu et al., 2017), and is recognized as the cally equilibrium scour conditions are assessed using a range of 78
14 primary cause for the failure of spur dikes (Kothyari et al., 2007; parameters including the mean approach flow features (velocity 79
15 Pandey et al., 2017). Local scour is classified into two categories; and depth of flow) and the coarser (armor layer) bed particles and 80
16 (i) Clear-water scour and (ii) Live-bed scour. For clear-water scour, parent bed material characteristics (Sui et al., 2010). 81
17 there is no sediment transport along the stream, other than a local To date, a growing number of laboratory and field studies, along 82
18 redistribution of sediment around and downstream the scoured with numerical simulations have been done towards estimating the 83
19 structure, and the threshold velocity ratio is always less than 1 (U/ maximum scour depth at equilibrium conditions (Garde et al., 1961, 84
20 Uc < 1), where U is the approach mean flow velocity and Uc is the 1963; Gill, 1972; Koken & Gogus, 2015; Kothyari et al., 2007; Kuhnle 85
21 threshold velocity for initiation of sediment entrainment. For live- & Alonso, 2013; Nagy, 2004; Nasrollahi et al., 2008; Pandey et al., 86
22 bed scour conditions, sediment is transported along the stream and 2015; Radan & Vaghefi, 2016; Vaghefi et al., 2009; Yagci et al., 87
23 the threshold velocity ratio is always greater than 1 (U/Uc > 1). 2016; among others). The temporal evolution of scour depth (dst) 88
24 Investigators mainly focus on the typically most commonly occur- also is an important factor for economical and realistic designs, as 89
25 ring clear-water scour processes. Fig. 1, shows an upstream view of well as for allowing the safe and efficient hydraulic operation of 90
26 a typical recent clear water spur dike failure, which occurred in the spur dikes, especially during periods of extreme weather. A hy- 91
27 Gaula river near Haldwani, Uttrakhand (India). Such failures due to draulic structure is safe if the estimated scour depth is less than the 92
28 clear water spur dike scour processes typically occur around the foundation depth, typically considering a certain safety factor (Guo, 93
29 world, and, thus, form the focus of the current study. 2012), otherwise it may be put at risk and should be closed for 94
30 Spur dike scour has been studied more extensively for uniform public safety during extreme hydrologic conditions. Knowledge of 95
31 sediment beds, compared to mixed sediment bed surfaces the time required for the development of equilibrium scour may 96
32 (Ezzeldin, 2019; Fazli, Ghodsian, & Neyshabouri, 2008; Koken & also permit assessing the need for repairs. However, despite its 97
33 Gogus, 2015; Kothyari & Ranga Raju, 2001; Kuhnle & Alonso, importance, relatively few studies have been done on the literature 98
34 2013; Liang et al. 2019; Mostafa et al., 2019; Nagy, 2004; Pandey for the prediction of temporal evolution of scour depth in sediment 99
35 et al., 2015; Vaghefi et al., 2009; Zhang et al., 2018; amongst mixtures (Kothyari et al., 2007; Oliveto & Hager, 2002). Kothyari 100
36 others). Even if a uniform bed surface is a good approximation for et al. (1992) developed a method to estimate the evolution of 101
37 many cases, most commonly streambeds are non-uniform, temporal scour depth by considering primary vortex structures 102
38 comprised of distinct particle sizes ranging from fine sand to shed past spur dikes. The time to reach equilibrium scour, is shorter 103
39 coarse gravel. By definition, non-uniform sediment, comprised of for uniform sediment compared to non-uniform sediment mixtures 104
40 distinct individual fractions in the particle size distribution, exhibit
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi (Kothyari et al., 2007; Oliveto & Hager, 2002). It has been stated by 105
41 a geometric standard deviation ðsg ¼ d84 =d16 Þ greater than 1.4 Kothyari and Ranga Raju (2001) that a pier or an abutment is similar 106
42 (Raikar & Dey, 2005), where sg is the geometric standard deviation to a vertical wall spur dike, as far as scour processes are concerned. 107
43 of particle size distribution, and d84 and d16 refer to the maximum They also stated that the effect of the developing boundary layer is 108
44 responsible for smaller scour geometries, for the case of spur dikes. 109
45 The foregoing studies demonstrate that the focus of the litera- 110
46 ture over the last two decades has been on assessing spur dike 111
47 equilibrium scour, for uniform sediment. Given their importance, 112
48 the current study aims to consider the importance of flow and 113
49 sediment parameters on assessing the dynamic temporal evolution 114
50 of clear-water, spur dike scour, for uniform, as well as for non- 115
51 uniform, sediment bed surfaces. The current study has two main 116
52 objectives: (i) to experimentally measure and evaluate the effect of 117
53 different parameters on the temporal evolution of scour depth at a 118
54 spur dike, and (ii) to develop a predictive model for the temporal 119
55 evolution of scour depth, for non-uniform sediment mixtures. 120
56 Herein, the results of 32 time-dependent laboratory scour experi- 121
57 ments, totalling 480 datasets, are used to derive and validate a 122
58 predictive model for the temporal evolution of spur dike scour. 123
59 124
60 2. Background theory 125
61 126
62 Parameters influencing the evolution of temporal scour depth 127
63 around spur dikes are, their geometry, the properties of the sedi- 128
64 Fig. 1. Typical scour around a spur dike in the Gaula river near Haldwani, Uttrakhand, ment at the bed surface, the mean flow parameters, and time (until 129
65 India. equilibrium scour depth has been reached). Time-dependent scour 130

Please cite this article as: Pandey, M et al., Experimental assessment and prediction of temporal scour depth around a spur dike, International
Journal of Sediment Research, https://doi.org/10.1016/j.ijsrc.2020.03.015
 IJSRC293_proof ■ 5 April 2020 ■ 3/13

M. Pandey et al. / International Journal of Sediment Research xxx (xxxx) xxx 3

1 depth (dst) around a vertical wall spur dike in non-uniform sedi-

66
2 ment mixtures can be written as; Rh 1=6 67
Fdi ¼ 1:65 D*  150 (10)
3   d50 68
4 dst ¼ f d50 ; sg ; rs Uc ; U; r; y; l; t (1) 69
5 In the foregoing equations (Eqs. (6)e(8)), Rh is the hydraulic 70
6 In Eq. (1), d50 ; sg ; Uc and rs are the sediment parameters; radius, Fdi is densimetric Froude number for incipient sediment 71
r; U and y are the flow parameters; l is the spur dike length scale;
1=3
7  72
and t refers to the time of the dynamically evolving scour process. entrainment, D* ¼ ðS  1Þg y2 d50 is the dimensionless parti-
8 73
9 Specifically, d50 is the median diameter of the bed surface material, cle size, and n is the kinematic viscosity of water. For sand-gravel 74
10 rs is the density of sediment, r is the density of water, y is the flow mixtures, d50 can be approximated by the median particle size of 75
depth. From Eq. (1), some non-dimensional parameters can be P
11 sediment mixture, d50 ¼ da with da ¼ pi di , where pi is the per- 76
12 considered, such as, the non-dimensional time function, T, the centage of sediment by weight corresponding to the size fraction di . 77
13 densimentric Froude number, Fd, and the densimetric Froude Further, S is the spur dike shape factor, S ¼ 1 for vertical wall 78
14 number for a sediment mixture, Fsm (Oliveto & Hager, 2002). These spur dikes; Ss is the submergence factor for non-submerged spur 79
15 are expressed as follows: dikes (taken as 1), Sca is the cascade factor, ¼ 1:25 for vertical wall 80
16  qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi. 1=3  spur dikes; and b ¼ l=B is the spur dike obstruction factor, with B 81
1=3
17 T ¼ t sg ðS  1Þgd50 l2 y (2) being the channel width. Eq. (6) has been tested for a range of 82
18 values of Fd-Fdb. Similar to Oliveto and Hager (2002), it was also 83
19 concluded that the temporal evolution of the scour depth is greatly 84
20 where S ¼ rs =r is the specific gravity of the solid grains and g is influenced by sg, for non-uniform mixtures. 85
21 gravitational acceleration. 86
22 . qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3. Experiments 87
23 Fd ¼ U ðS  1Þgd50 (3) 88
24 All experiments were done in a flume that has 24.0 m long, 1.0 m 89
25 wide, and 0.5 m deep. The working test section, has dimensions of 90
1=3
26 Fsm ¼ sg Fd (4) 4.0 m  1.0 m  0.4 m, starting 12 m from the flume’s inlet. All 91
27 experiments were done under clear water scour conditions. Fig. 2 92
Oliveto and Hager (2002) derived an equation to calculate the
28 shows a sketch of the flume, with its test section. 93
time-dependent scour depth, as given by Eq. (5), based on the
29 For each of the experimental runs, the working section of the 94
geometric standard deviation, playing an important role for non-
30 flume was fully filled with one of the two pre-prepared sediment 95
uniform sediment mixtures, and the densimetric Froude number,
31 mixtures, up to the desired bed surface level, defined by concrete 96
affecting the temporal evolution of scour depth. For non-uniform
32 blocks placed upstream and downstream the working section. The 97
sediment mixtures, they introduced the inception densimetric
33 first sediment mixture is comprised of 50% sand (d50s ¼ 0.27 mm, 98
particle Froude number, Fdi:
34 sg ¼ 1.22) and 50% gravel (d50g ¼ 2.7 mm, sg ¼ 1.21), and the 99
35
dst second comprises of 50% sand (d50s ¼ 0.27 mm, sg ¼ 1.22) and 50% 100
1=2 1:5
36 ¼ 0:068sg Fd log T (5) gravel (d50g ¼ 3.1 mm, sg ¼ 1.18), where d50s is the median 101
RL
37 diameter of sand and d50g is the median diameter of gravel. A 102
38 where RL ¼ (l2y)1/3, is the reference length. Eq. (5) applies for clear- linear profiler has been used to manually level the sediment bed 103
39 water conditions and when the densimetric Froude number is surface in the working section. Four vertical rectangular wall spur 104
40 greater than the densimetric Froude number at the sediment bed dikes, with different transverse length scales (of 6.0, 9.0, 11.5, and 105
41 inception conditions around the abutment (Hager & Oliveto, 2002). 14.0 cm), were used as physical models for the spur dikes. All 106
42 Kothyari et al. (2007) modified the approach of Oliveto and experiments were done for an unsubmerged spur dike case. The 107
43 Hager (2002) on the basis of the difference of Fd-Fdb, where Fdb is water supply into the flume was regulated by a valve, at the inlet 108
44 the densimetric Froude number for the inception of scour at spur tank upstream of the flume. A sharp crested, pre-calibrated weir 109
45 dikes. They proposed a time dependent scour depth equation, given was provided at the end of flume to measure the flow rate. Scour 110
46 by Eqs. (6)e(10): depth (dst) was measured with a point gauge at different time 111
47 intervals, i.e. t ¼ 1, 3, 5, 10, 15, 30 and 60 min up to several hours, 112
48 dst until an equilibrium scour geometry was reached (demonstrated 113
1=2
49 ¼ 0:272sg ðFd  Fdb Þ2=3 log T Fd > Fdb (6) by current invariance of any measurable scour depth with the 114
RL
50 passage of time). Each experiment in the current study was done 115
51 where Fdb is expended as: for 20 h. However, equilibrium time (te) varied in the range of 116
52 8e15 h. For checking the equilibrium scour, the scour depth was 117
"
1=6 #
53 Rh measured over time at the nose and junction of the spur dike 118
S=4
54 Fdb ¼ Fdi  1:26SSs Sca b s1=3
g (7) every 30 min after the equilibrium time with an accuracy of 119
d50
55 ±1 mm. 120
56 Before the start of each experimental run, the spur dike was 121
where Fdi is computed as:
57 fixed vertically at the right side of the channel wall (as viewed from 122
58
1=6 upstream), 14.0 m from the flume entrance (Fig. 2). The working- 123
Rh
59 Fdi ¼ 2:33D0:25
* D*  10 (8) section was perfectly leveled with respect to the longitudinal 124
60 d50 slope of the flume bed and then covered with a Perspex sheet. The 125
61 tail gate and inlet valve were used to obtain predetermined flow 126
62
1=6 conditions in the flume. Once the predetermined flow condition 127
1=12 Rh
63 Fdi ¼ 1:08D* 10 < D* < 150 (9) was established, the Perspex sheet was removed very carefully so 128
d50
64 that no scour occurred due to this process. This allowed the time 129
65 required for the scour depth (dst) to develop to be accurately 130

Please cite this article as: Pandey, M et al., Experimental assessment and prediction of temporal scour depth around a spur dike, International
Journal of Sediment Research, https://doi.org/10.1016/j.ijsrc.2020.03.015
 IJSRC293_proof ■ 5 April 2020 ■ 4/13

4 M. Pandey et al. / International Journal of Sediment Research xxx (xxxx) xxx

1 measured. For each test, hydraulic parameters and scour depth data It can be said that the scour rate at the nose of the spur-dike is 66
2 are listed in Table 1. For vertical wall elements like spur dikes and greater than at the upstream wall junction and it takes more time to 67
3 abutments scour always starts just downstream of the corner re- reach equilibrium, as shown in Fig. 4bec. The small scour rate at the 68
4 gion protruding into the upstream flow, than migrates first toward spur dike wall junction means it more quickly reaches the equi- 69
5 the upstream side (Oliveto & Hager, 2002). This process is done librium state. For an extreme case, if the flow is low enough and 70
6 over a very small time interval. In the current study, scour depth scour only occurs at the nose, the equilibrium time of scour at the 71
7 was measured at different time intervals at the upstream nose of wall junction is very low or should be zero. Fig. 4d shows the 72
8 the spur dike, as this is the expected location at the equilibrium variation between the ratios of temporal scour depths at the nose 73
9 scour condition where the maximum scour depth occurs (Kothyari and the wall junction and scour time, with respect to the equilib- 74
10 & Ranga Raju, 2001; Lodhi et al., 2016; Pandey et al., 2015). For the rium scour time. It can be clearly observed that scour progresses 75
11 current study, all experiments refer to clear water scour conditions. quickly in its initial stage (<6% scour time). All tests show similar 76
12 Table 1 also lists the threshold velocity and shear velocity for the rates of change of the scour depth, but for 6%e25% of the scour 77
13 bed surface particles, calculated using the same methodology as in time, the decreasing rate of the scour depths ratio increases with an 78
14 Sui et al. (2010). Fig. 3, shows a characteristic three-dimensional increase in the approach velocities (R5 to R8). However, at the final 79
15 visualization for the bed surface elevation, for one of the experi- stage of the scour (80%e100%), the scour depths ratios increase 80
16 mental runs, demonstrative of the scour hole geometry obtained at with increasing approach velocities (R5 to R8). For any particular 81
17 equilibrium scour conditions. test case, the scour depth ratio initially decreases rapidly, but after 82
18 25% of the equilibrium scour time, the rate of decrease of the scour 83
19 depth ratio becomes very small or constant, as can be seen in 84
20 4. Results Fig. 4d. 85
21 86
22 The time dependent scour depths with different sand-gravel 87
5. Discussions
23 mixtures were measured at the nose and the spur dike-wall junc- 88
24 tion. Fig. 4 shows the comparison of time dependent scour depth 89
The foregoing results are used to formulate a discussion on how
25 variation at the nose of the spur dike and spur dike-wall junction. 90
the temporal evolution of spur dike scour depth may be affected by
26 Fig. 4a shows that the temporal scour variation at the nose of the 91
flow and spur dike design parameters, including the effect of a non-
27 spur dike is always higher than at the spur dike-wall junction. From 92
uniform bed surface (comprised of mixed sediment). Then a re-
28 the measurements of the bed surface elevation (Fig. 4aeb), it can be 93
evaluation of recently presented equations is done, and the study
29 observed that 83%e89% of the equilibrium scour depth at the spur 94
concludes by proposing a new predictive model for spur dike scour
30 dike’s nose developed within 6% of the time needed to reach 95
depth, yielding demonstrably better results than for previous
31 equilibrium, while these values range from 75% to 90% for the spur 96
equations.
32 dike-wall junction. Fig. 4bec demonstrate the rate of change of bed 97
33 surface elevation, and clearly indicate that scour at the initial stage 98
34 develops rapidly (similar to Fig. 4aeb). However, after 40% of the 5.1. Influence of mean flow parameters and spur dike length on 99
35 equilibrium sour time, the rate of change of scour depth almost temporal scour depth variation 100
36 flattens, being particularly slow for the scour at the nose (compared 101
37 to 60% of the equilibrium scour time for the spur dike-wall junction Fig. 5aed, show the variation of dst vs. RL, dst/RL vs. y/l, dst/RL vs. 102
38 location). log (T) and dst/RL vs. Fd. The scour depth at the spur dike’s nose 103
39 104
40 105
41 106
42 107
43 108
44 109
45 110
46 111
47 112
48 113
49 114
50 115
51 116
52 117
53 118
54 119
55 120
56 121
57 122
58 123
59 124
60 125
61 126
62 127
63 128
64 129
65 Fig. 2. Layout top (plan) and side (elevation) views of the flume, showing the working section. 130

Please cite this article as: Pandey, M et al., Experimental assessment and prediction of temporal scour depth around a spur dike, International
Journal of Sediment Research, https://doi.org/10.1016/j.ijsrc.2020.03.015
 IJSRC293_proof ■ 5 April 2020 ■ 5/13

M. Pandey et al. / International Journal of Sediment Research xxx (xxxx) xxx 5

1 Table 1 66
2 Experimental conditions and equilibrium spur dike scour hole, obtained from the experiments done in the current study. 67
3 Run l (m) y (m) U (m/s) d50s (mm) d50g (mm) Fd U/Uc te (min) dse (m) 68
4 R1 0.140 0.112 0.41 0.27 2.7 2.66 0.90 480 0.149
69
5 R2 0.140 0.105 0.35 0.27 2.7 2.27 0.77 660 0.111 70
6 R3 0.140 0.1 0.31 0.27 2.7 2.01 0.68 720 0.095 71
7 R4 0.140 0.09 0.28 0.27 2.7 1.82 0.61 840 0.072 72
R5 0.115 0.112 0.41 0.27 2.7 2.66 0.90 540 0.128
8 73
R6 0.115 0.105 0.35 0.27 2.7 2.27 0.77 690 0.091
9 R7 0.115 0.1 0.31 0.7 2.7 2.01 0.68 780 0.076 74
10 R8 0.115 0.09 0.28 0.27 2.7 1.82 0.61 840 0.057 75
11 R9 0.090 0.112 0.41 0.27 2.7 2.66 0.90 600 0.104 76
12 R10 0.090 0.105 0.35 0.27 2.7 2.27 0.77 660 0.078 77
R11 0.090 0.1 0.31 0.27 2.7 2.01 0.68 780 0.063
13 R12 0.090 0.09 0.28 0.27 2.7 1.82 0.61 870 0.051
78
14 R13 0.060 0.112 0.41 0.27 2.7 2.66 0.90 630 0.074 79
15 R14 0.060 0.105 0.35 0.27 2.7 2.27 0.77 720 0.058 80
16 R15 0.060 0.1 0.31 0.27 2.7 2.01 0.68 810 0.045 81
R16 0.060 0.09 0.28 0.27 2.7 1.82 0.61 900 0.038
17 82
R17 0.140 0.112 0.41 0.27 3.1 2.49 0.84 510 0.127
18 R18 0.140 0.105 0.35 0.27 3.1 2.13 0.71 600 0.096 83
19 R19 0.140 0.1 0.31 0.27 3.1 1.89 0.63 660 0.074 84
20 R20 0.140 0.09 0.28 0.27 3.1 1.70 0.57 810 0.057 85
21 R21 0.115 0.112 0.41 0.27 3.1 2.49 0.84 570 0.107 86
R22 0.115 0.105 0.35 0.27 3.1 2.13 0.71 660 0.078
22 R23 0.115 0.1 0.31 0.27 3.1 1.89 0.63 720 0.059
87
23 R24 0.115 0.09 0.28 0.27 3.1 1.70 0.57 840 0.047 88
24 R25 0.090 0.112 0.41 0.27 3.1 2.49 0.84 600 0.086 89
25 R26 0.090 0.105 0.35 0.27 3.1 2.13 0.71 660 0.068 90
R27 0.090 0.11 0.31 0.27 3.1 1.89 0.63 750 0.054
26 91
R28 0.090 0.13 0.28 0.27 3.1 1.70 0.57 870 0.041
27 R29 0.060 0.12 0.41 0.27 3.1 2.49 0.84 600 0.058 92
28 R30 0.060 0.11 0.35 0.27 3.1 2.13 0.71 720 0.043 93
29 R31 0.060 0.13 0.31 0.27 3.1 1.89 0.63 780 0.037 94
30 R32 0.060 0.12 0.28 0.27 3.1 1.70 0.57 870 0.032 95
31 96
32 97
33 98
34 99
35 100
36 101
37 102
38 103
39 104
40 105
41 106
42 107
43 108
44 109
45 110
46 111
47 112
48 113
49 114
50 115
51 116
52 117
53 118
54 119
55 120
56 121
57 122
58 123
59 124
60 125
61 Fig. 3. Side view of a typical equilibrium scour hole around the spur dike and the corresponding topographical representation of the measured bed surface elevation (see inset at 126
62 top left). 127
63 128
64 129
65 130

Please cite this article as: Pandey, M et al., Experimental assessment and prediction of temporal scour depth around a spur dike, International
Journal of Sediment Research, https://doi.org/10.1016/j.ijsrc.2020.03.015
 IJSRC293_proof ■ 5 April 2020 ■ 6/13

6 M. Pandey et al. / International Journal of Sediment Research xxx (xxxx) xxx

1 66
2 67
3 68
4 69
5 70
6 71
7 72
8 73
9 74
10 75
11 76
12 77
13 78
14 79
15 80
16 81
17 82
18 83
19 84
20 85
21 86
22 87
23 88
24 89
25 90
26 91
27 92
28 93
29 94
30 95
31 96
32 97
33 98
34 99
35 100
36 101
37 102
38 103
39 104
40 105
41 106
Fig. 4. (a) Temporal variation of scour depth at the nose ( R5, R6, R7, R8) and junction ( R5, R6, R7, R8) of the spur dikes studied herein, for a range of flow
42 107
conditions, (b) percentage scour variation with time at nose of spur dike, (c) percentage scour variation with time at spur dike-wall junction and (d) ratio of the scour depth at the
43 nose and the wall junction. 108
44 109
45 110
46 measured at time t (dst) is found to consistently increase with As the scour hole enlarges, mainly with the removal of fine sedi- 111
47 increasing reference length (RL), transverse length of the spur dike ment, the developing armor layer slowly raises the effective 112
48 and approach flow depth, as shown in Fig. 5a. Similar results were threshold bed-shear stresses, which reduces the growth of the 113
49 obtained by Oliveto and Hager (2002), for vertical wall abutments. scour-hole, (Raikar & Dey, 2005). The rate of temporal scour depth 114
50 Fig. 5a shows that the temporally evolving scour depth, increases variation increases with the y/d50 ratio, as can be seen in Fig. 5c. 115
51 with increasing U/Uc and reaches maximum for U/Uc ¼ 0.80e0.90. Fig. 5c indicates that the temporal scour depth variation becomes 116
52 Fig. 5b, shows the influence of the ratio of the approach flow approximately constant after log T ¼ 6.5. 117
53 depth to the transverse length of the spur dike (y/l) on the A number of previous studies, have demonstrated the strong 118
54 dimensionless temporal scour depth variation. The ratio of y/l is dependence of scouring on mean flow as well as sediment prop- 119
55 also known as flow shallowness (Lanca et al., 2013). For a particular erties (Ezzeldin, 2019; Kothyari & Ranga Raju, 2001; Pandey et al., 120
56 spur dike length, the temporal scour depth increases with flow 2018; Yagci et al., 2016; Zhang et al., 2018). Hence, it is necessary 121
57 shallowness. The effect of the flow shallowness is more noticeable to relate the temporal scour depth evolution to the particle densi- 122
58 for the case of smallest transverse length of the spur dike. The rate metric Froude number, which is a function of the approach mean 123
59 of temporal scour depth variation increases with a decrease in the velocity and particle size, i.e. dst/RL z f(Fd)  f(T), where f(Fd) is the 124
60 transverse length of spur dike. function of densimetric Froude number and f(T) is the time func- 125
61 Fig. 5c, shows the effect of dimensionless time parameter (T) on tion. Fig. 5d shows the variation of dst/RL vs Fd. 126
62 dst/RL with different ranges of y/d50. For different ranges of y/d50, Fig. 5d shows the influence of the densimentric particle Froude 127
63 Fig. 5c clearly illustrates that the variation of temporal scour depth number (Fd) on the maximum dimensionless temporal scour depth 128
64 increases with an increase in T. Fig. 5c, also shows that the temporal (dst/RL) variation. Fig. 5d, clearly indicates that the ratio dst/RL in- 129
65 scour depth variation increases with the increase in the ratio y/d50. creases with increasing Fd. Fd, mainly depends on the approach 130

Please cite this article as: Pandey, M et al., Experimental assessment and prediction of temporal scour depth around a spur dike, International
Journal of Sediment Research, https://doi.org/10.1016/j.ijsrc.2020.03.015
 IJSRC293_proof ■ 5 April 2020 ■ 7/13

M. Pandey et al. / International Journal of Sediment Research xxx (xxxx) xxx 7

1 66
2 67
3 68
4 69
5 70
6 71
7 72
8 73
9 74
10 75
11 76
12 77
13 78
14 79
15 80
16 81
17 82
18 83
19 84
20 85
21 86
22 87
23 88
24 89
25 90
26 91
27 92
28 93
29 94
30 95
31 96
32 97
33 98
34 99
35 100
36 101
37 102
38 103
39 104
40 105
41 Fig. 5. Influence of mean flow parameters and spur dike length on the temporally varying scour depth, at the spur dike’s nose (dst): (a) dst vs. RL, (b) dst/RL vs. y/l, (c) dst/RL vs. log (T) 106
42 and (d) dst/RL vs. Fd. 107
43 108
44 109
velocity and particle size. It has already been explained that the spur dike, Fig. 6aed clearly show that the maximum temporal scour
45 110
scour depth variation increases with increasing approach flow ve- depth increases with an increase of the sediment mixture’s Froude
46 111
locity. Thus, it is observed that the maximum temporal scour depth number, which indicates that the temporal scour depth increases
47 112
variation increases with a decrease in the median diameter of the with decreasing sediment non-uniformity. It has also been
48 113
sediment mixture. observed, development of an armor layer in the scoured region
49 114
begins causing in an exposure of coarser gravel sizes due to
50 115
washing out of the finer gravel particles. The scour rate at its initial
51 5.2. Influence of sediment mixture characteristics on the temporal 116
stage (i.e. before 60 min) is very high. Up to 80% of the scour occurs
52 scour depth 117
before this time, as can be seen in Fig. 6aed. The scour rate with
53 118
time was very slow after this initial stage, because of the formation
54 The current study mainly deals with a non-uniform sediment 119
of a protective armor layer. Once a stable armor layer forms around
55 bed and non-uniform sediment is well characterized by the 120
the spur dike, no further removal of sediment occurs in the scour
56 geometric standard deviation of particle size distribution. It has 121
hole.
57 been observed by the authors that the temporal scour depth 122
58 variation for sediment mixtures is influenced by the non- 123
59 uniformity of sediment. Hence, the effect of sediment mixture 5.3. A re-assessment of the performance of past models 124
60 on dst was analyzed in terms of the sediment mixture Froude 125
61 number ðFsm ¼ s1=3 Fd Þ. Fig. 6aed shows the effect of Fsm on Measured values of the scour depth (dst) at predefined time 126
62 temporal scour depth evolution at different times (t ¼ 1, 15, 60, intervals (t), at the location of maximum scour depth at the up- 127
63 and 120 min). stream nose of the spur dike, are shown in Fig. 7, for different 128
64 Sediment’s non-uniformity has a significant effect on temporal transverse lengths of the spur dike and flow conditions. In the 129
65 scour depth evolution. For a fixed time and transverse length of the current study, the time dependent scour depth has been also 130

Please cite this article as: Pandey, M et al., Experimental assessment and prediction of temporal scour depth around a spur dike, International
Journal of Sediment Research, https://doi.org/10.1016/j.ijsrc.2020.03.015
 IJSRC293_proof ■ 5 April 2020 ■ 8/13

8 M. Pandey et al. / International Journal of Sediment Research xxx (xxxx) xxx

1 calculated using the equations given by Oliveto and Hager (2002) 66

2 and Kothyari et al. (2007), as shown in Fig. 7aeh. These estimated
67
dst l
3 values of time dependent scour depths also are compared with ¼ f Fsm ; ;T (11) 68
4 observed values from the experiments presented in the current RL d50 69
5 study. It is observed that the model predictions, using either of the It has been stated by previous researchers that the element 70
6 previous models, are in a good agreement with the obtained Reynolds number plays a minor role for scour processes (Lim & 71
7 experimental data, for the larger transverse lengths of the exam- Chiew, 2001; Sui et al., 2010). The armor-layer development 72
8 ined spur dikes (e.g., l ¼ 11.5 and 14.0 cm e see Fig. 7aeb). For the processes and thickness are typically not explicitly considered 73
9 shorter spur dike transverse lengths, both equations exhibit an as a function of temporal scour depth variation. The importance 74
10 increasing difference between observed and computed values of of the different parameters for the temporal scour depth vari- 75
11 time-dependent scour depths (Fig. 7ced, g-h). ation, included in Eq. (11), has been demonstrated in the 76
12 foregoing sections. Using these parameters, an empirical rela- 77
13 tion has been derived to calculate the temporal scour depth 78
14 5.4. A new model for temporal scour depth 79
variation, as given by Eq. (12), employing a nonlinear regres-
15 sion method: 80
16 Theoretically, the temporal scour depth around a vertical wall 81
17 spur dike has been briefly explained in the foregoing sections. Eqs. 82
(1)e(4) show the dependent parameters of the temporal scour

18 dst l 0:28 83
depth evolution relations. In a sand-gravel sediment mixture, the ¼ 0:02ðFsm Þ1:68 ðlog TÞ 12
19 RL d50 84
20 characteristics of the parent bed material and armor-layer material 85
21 are the most important to consider (Sui et al., 2010). Using Buck- Eq. (12) offers estimates of the temporal evolution of scour 86
22 ingham’s pi theorem, Eq. (1) may be written in dimensionless form depth around the nose of a vertical wall spur dike under clear-water 87
23 as: scour conditions. 88
24 89
25 90
26 91
27 92
28 93
29 94
30 95
31 96
32 97
33 98
34 99
35 100
36 101
37 102
38 103
39 104
40 105
41 106
42 107
43 108
44 109
45 110
46 111
47 112
48 113
49 114
50 115
51 116
52 117
53 118
54 119
55 120
56 121
57 122
58 123
59 124
60 125
61 126
62 127
63 128
64 129
65 Fig. 6. Influence of sediment mixture in terms of Fsm on temporal scour depth variation. 130

Please cite this article as: Pandey, M et al., Experimental assessment and prediction of temporal scour depth around a spur dike, International
Journal of Sediment Research, https://doi.org/10.1016/j.ijsrc.2020.03.015
 IJSRC293_proof ■ 5 April 2020 ■ 9/13

M. Pandey et al. / International Journal of Sediment Research xxx (xxxx) xxx 9

1 66
2 67
3 68
4 69
5 70
6 71
7 72
8 73
9 74
10 75
11 76
12 77
13 78
14 79
15 80
16 81
17 82
18 83
19 84
20 85
21 86
22 87
23 88
24 89
25 90
26 91
27 92
28 93
29 94
30 95
31 96
32 97
33 98
34 99
35 100
36 101
37 102
38 103
39 104
40 105
41 Fig. 7. Comparison of the obtained experimental data of the temporal scour depth variation with the predictions calculated by Oliveto and Hager (2002) and Kothyari et al. (2007), 106
42 for a range of spur dike lengths and flow rates. ( observed, Oliveto and Hager (2002), Kothyari et al. (2007) for U ¼ 0.41 m/s; observed, Oliveto and Hager (2002), 107
Kothyari et al. (2007) for U ¼ 0.35 m/s; observed, Oliveto and Hager (2002), Kothyari et al. (2007) for U ¼ 0.31 m/s; observed, 2 Oliveto and Hager (2002), Kothyari
43 108
et al. (2007) for U ¼ 0.28 m/s).
44 109
45 110
46 111
47 112
48 5.4.1. Assessment of the model’s performance 5.4.2. Sensitivity analysis 113
49 The proposed approach Eq. (12) also is compared with the A sensitivity analysis is undertaken to classify the most 114
50 approach of Oliveto and Hager (2002) and Kothyari et al. (2007) and critical non-dimensional input parameter of the proposed rela- 115
51 the results for each of the approaches are assessed in terms of fits of tion and to check the dependency of different parameters on dst/ 116
52 the observed and estimated scour depths to the line of perfect RL, as listed in Table 2. The sensitivity analysis is done by using 117
53 agreement, as shown in Fig. 8. the average values of all non-dimensional values of independent 118
54 Fig. 9, shows the variation between the total number of per- and dependent parameters from the experimental data. An 119
55 centage data and percentage error. For Oliveto and Hager (2002) assumption has been made for this analysis, i.e. the errors in 120
56 and Kothyari et al. (2007), approximately 55% of the datasets each input variable are independent. The average values of non- 121
57 were found to be within ±25% error, whereas all datasets for the dimensional input parameters Fsm, l/d50, and log (T) for the 122
58 proposed relation were found to be within ±20% error lines. Both datasets used in this analysis are 1.39, 64.75, and 6.5, 123
59 previous studies proposed their approaches after analyzing the respectively. 124
60 non-uniform sediment data, but they considered densimetric par- If a percentage error DJ in the output is known as the difference 125
61 ticle Froude number rather than densimetric mixture Froude between values of output computed for inputs U and UþDU, then 126
62 number. the percentage error might be estimated as the absolute sensitivity 127
63 (V ¼ DJ/DU). Here the output is J ¼ dst/RL and input U ¼ Fsm, l/d50, 128
64 129
65 130

Please cite this article as: Pandey, M et al., Experimental assessment and prediction of temporal scour depth around a spur dike, International
Journal of Sediment Research, https://doi.org/10.1016/j.ijsrc.2020.03.015
 IJSRC293_proof ■ 5 April 2020 ■ 10/13

10 M. Pandey et al. / International Journal of Sediment Research xxx (xxxx) xxx

1 66
2 67
3 68
4 69
5 70
6 71
7 72
8 73
9 74
10 75
11 76
12 77
13 78
14 79
15 80
16 81
17 82
18 83
19 84
20 85
21 86
22 87
23 88
24 89
25 90
26 91
27 92
28 93
29 94
30 95
31 96
32 97
33 98
34 99
35 100
36 101
37 102
38 103
39 104
40 105
41 106
42 Fig. 7. (continued). 107
43 108
44 109
and T. The error also can be expressed in a relative form £ ¼ DJ/J. et al., 2007; Oliveto & Hager, 2002) have been evaluated for their
45 110
The error DJ in output is fundamentally the deviation sensitivity accuracy using 480 flume experimental datasets. For an estimation
46 111
with DU being the error. The relative sensitivity can be expressed of the difference between calculated and experimental non-
47 112
¥ ¼ (UDJ)/(JDU) (Ahmad, 2013). dimensional temporal scour depth (dst/RL), the discrepancy ratio
48 113
The sensitivity analysis is completed by changing the each input (DR, described by White et al., 1973) is used as an error degree and
49 114
parameter by ±10%. The outcomes of sensitivity analysis are listed is described as: Q1
50 115
in Table 2, which indicates that Fsm is the most sensitive parameter
51 116
followed by log T, and l/d50. The relative sensitivity (¥) of Fsm is

52 dst computed 117

nearly 1.7 and 6.8 for log T and l/d50, respectively; for a þ10% in-
53 DR ¼ log (13) 118
crease in U. However, for a 10% decrease in U, the relative sensitivity dst observed
54 119
(¥) of Fsm is nearly 1.7 and 5.8 for log T and l/d50, respectively. The
55
computation accuracy of the proposed equation greatly depends on For DR ¼ 0, the computed (dst/RL) is identical to the experimental 120
56 (dst/RL). For negative (positive) values of discrepancy ratio, the 121
Fsm, followed by log T, and l/d50. Thus, the temporal variation of
57 computed value of the temporal scour depth evolution is smaller 122
scour depth at the nose of the spur dike, greatly depends on the
58 (larger) than the experimental value. Accuracy is described as the 123
densimetric mixture Froude number. The detailed results of the
59 frequency of cases for which the DR is within a suitable range for 124
sensitivity analysis are listed in Table 2.
60 the total number of data, as shown in Fig. 10. It cab be noted from 125
61 Fig. 10, that the data frequencies within DR ¼ ±0.01 are 104, 116, 126
62 5.4.3. Comparative statistical analysis and 198 for Oliveto and Hager (2002), Kothyari et al. (2007), and the 127
63 In order to check the accuracy of the proposed relation, it has proposed equation, respectively. It can be said that the proposed 128
64 been compared the experimental values with existing relations. equation is more accurate to calculate the temporal depths of scour 129
65 The proposed Eq. (12) and previously proposed equations (Kothyari compared to existing relations. 130

Please cite this article as: Pandey, M et al., Experimental assessment and prediction of temporal scour depth around a spur dike, International
Journal of Sediment Research, https://doi.org/10.1016/j.ijsrc.2020.03.015
 IJSRC293_proof ■ 5 April 2020 ■ 11/13

M. Pandey et al. / International Journal of Sediment Research xxx (xxxx) xxx 11

1 66
2 67
3 68
4 69
5 70
6 71
7 72
8 73
9 74
10 75
11 76
12 77
13 78
14 79
15 80
16 81
17 82
18 83
19 84
20 85
21 86
22 87
23 88
24 89
25 90
26 91
27 92
28 93
29 94
30 95
31 96
32 97
33 98
Fig. 8. Calculated versus observed values of temporal scour depths at the nose of spur dike using: (a) the proposed approach; (b) Oliveto and Hager’s (2002) equation and using
34 99
Kothyari et al.’s (2007) equation.
35 100
36 101
37 102
38 103
39 104
40 105
41 106
42 107
43 108
44 109
45 110
46 111
47 112
48 113
49 114
50 115
51 116
52 117
53 Fig. 9. Variation of total number of percentage data versus percentage error.
118
54 119
55 120
Fig. 10. Variation of experimental data frequency and DR for selected equations.
56 121
57 Table 2 122
58 Outcome of the sensitivity analysis. 123
59 Percentage change U DU DJ V £ ¥ 124
Table 3
60 125
DU is þ10% increase in U Fsm 0.139 0.129 0.928 0.176 1.765 Outcome of the statistical analysis.
61 log (T) 0.649 0.073 0.112 0.100 0.999
126
62 Relations CC MAE MSE RMSE MAPE 127
l/d50 6.475 0.019 0.003 0.026 0.260
63 DU is 10% decrease in U Fsm 0.139 0.129 0.871 0.166 1.655 Proposed Equation (Eq. (12)) 0.98 0.029 0.001 0.037 0.011 128
64 log (T) 0.649 0.073 0.109 0.097 0.971 Oliveto and Hager (2002) 0.81 0.124 0.024 0.154 0.042 129
l/d50 6.475 0.019 0.003 0.029 0.287 Kothyarie et al. (2007) 0.78 0.159 0.051 0.224 0.054
65 130

Please cite this article as: Pandey, M et al., Experimental assessment and prediction of temporal scour depth around a spur dike, International
Journal of Sediment Research, https://doi.org/10.1016/j.ijsrc.2020.03.015
 IJSRC293_proof ■ 5 April 2020 ■ 12/13

12 M. Pandey et al. / International Journal of Sediment Research xxx (xxxx) xxx

1 The performance of the proposed Eq. (12) and previously pro- using a range of non-dimensional parameters Fsm, l/d50, and T, 66
2 posed (Kothyari et al., 2007; Oliveto & Hager, 2002) equations has empirically found to affect the non-dimensional temporal scour 67
3 been also checked using the several statistical indices. These indices depth evolution at the nose of the spur dike. The performance of 68
4 are known as way to evaluate the degree of the agreement between this model also is compared with the approaches of Oliveto and 69
5 computed and experimental values of dst/RL. If X is the experimental Hager (2002) and Kothyari et al. (2007). 70
6 value of dst/RL and X0 is the corresponding calculated value of dst/RL. It is found that the proposed model, in addition to having 71
7 Ahmad (2013) defined these statistical indices as: wider applicability to non-uniform bed surfaces, also better 72
8 matches the experimentally observed values of temporal scour 73
9 Coefficient of correlation; depths, largely due to the consideration of the densimetric 74
P P P 0
10 n XX 0  X X (14) Froude number. The performance of the proposed model is found 75
CC ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiqffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
11 P 2 P 2 P 02 P to be superior to the existing temporal scour depth equations, 76
12 n X  ð XÞ n X  ð X 0 Þ2 particularly for sand-gravel mixtures, as their comparison to the 77
13 flume experiments reveals. Statistically, the proposed equation 78
14 1Xn  
 shows good agreement between experimental and computed 79
15
Mean absolute error; MAE ¼ Xi  Xii  (15)
80
n i¼1 values, and a better predictive capacity compared to existing
16 models. Q2 81
17 1Xn  0 2 82
18 Mean square error; MSE ¼ i¼1
Xi  Xi (16) 83
n
19 List of notation 84
20 sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 85
Pn  0 2
21 i¼1 Xi  X
ı 86
Root mean square error; RMSE ¼ (17)
22 n B Channel width 87
23 d50g Median diameter of sediment mixture 88
 0
24 100Xn Xi  Xi  d50g Median diameter of gravel 89
25 Mean absolute percentage error; MAPE ¼ d50s Median diameter of sand 90
n i¼1 jXi j
26 d16 Particle size at 16% finer 91
(18)
27 d84 Particle size at 84% finer 92
P
28 In aforementioned Eqs. (14)e(18), n is the total number of da ¼ pi di Median particle size of sediment mixture 93
29 experimental data sets (n ¼ 480 in this study). The values of CC, di Particle size corresponds to pi 94
30 MAE, MSE, RMSE, and MAPE for each equation are listed in Table 3. dst Scour depth at time t 95
31 The value of CC of the proposed equation is higher than for the D* ¼ ððS  1Þg=v2 Þ1=3 d50 Non-dimensional particle size 96
32 others, and the values of MAE, MSE, RMSE, and MAPE are lower as 97
Fd ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
U ffi Densimetric Froude number
33 compared to existing relations. These results mean the proposed ðS1Þgd50 98

1=6
34 equation statistically performs better than the others. However, the 99
35 Fdi ¼ 2:33D0:25*
Rh
d
Inception densimetric particle Froude 100
proposed equation has an edge over the relations proposed by 50

36 Oliveto and Hager (2002) and Kothyari et al. (2007). It is apparent number 101
"
1=6 #
37 that the better representation of sediment non-uniformity by in- 102
S=4
38 clusion of the densimetric mixture Froude number, results in better Fdb ¼ Fdi  1:26SSs Sca b Rh
d50
s1=3
g Densimetric Froude 103
39 estimates and a higher predictive ability for the proposed equation number for inception 104
40 (Eq. (12)). of scour 105
41 106
1=3
42 Fsm ¼ sg Fd Froude number of sediment mixture 107
43 6. Conclusions g Gravitational acceleration 108
44 l Transverse length of spur dike 109
45 The temporal variation of scour depth around a vertical wall n Total number of datasets 110
46 spur dike has been studied experimentally at the upstream nose of N Spur dike shape factor (Oliveto & Hager, 2002) 111
47 the spur dike in a sand-gravel mixture. The influence of different pi Percentage of sediment by weight 112
48 parameters on the temporal scour depth evolution has been dis- Q Flow rate 113
49 cussed in the current study. It is observed that the dimensionless Rh Hydraulic radius 114
50 temporal scour depth variation increases with increase in U/Uc, Fd, RL¼(l2y)1/3 Reference length 115
51 y/l, y/d50, and T and decreases with an increase in particle size. S Specific gravity 116
52 A comparison of the measured rates of temporal evolution of t Time 117
53 scour depth at the spur dike’s nose and its junction with the wall te Equilibrium time of scour 118
54 also is undertaken. It is found that the temporal scour depth vari- T Non-dimensional time function 119
55 ation remains always less compared to the spur dike’s nose (around y Approach flow depth 120
56 8e30% less). U Approach mean velocity 121
57 Furthermore, the temporal scour depth variation in sediment Uc Threshold velocity for particle entrainment 122
58 mixtures is influenced by the non-uniformity of sediment and in- r Density of water 123
59 creases with increases in densimetric Froude number of the sedi- rs Density of sediment 124
60 ment mixture. The temporal scour depth increases with decreases n Kinematic viscosity of water 125
61 in non-uniformity of sediment. sg Geometric standard deviation of particle sizes 126
62 A new temporal scour depth equation for estimation of S Spur dike shape factor (Kothyari et al., 2007) 127
63 maximum scour depth around spur dikes in sand-gravel mixtures, Sca Spur dike cascade factor 128
64 which takes into account the non-uniformity of sediment, has been Ss Spur dike submergence factor 129
65 proposed. The proposed model is derived from dimensional theory b ¼ l=B Spur dike obstruction factor 130

Please cite this article as: Pandey, M et al., Experimental assessment and prediction of temporal scour depth around a spur dike, International
Journal of Sediment Research, https://doi.org/10.1016/j.ijsrc.2020.03.015
 IJSRC293_proof ■ 5 April 2020 ■ 13/13

M. Pandey et al. / International Journal of Sediment Research xxx (xxxx) xxx 13

1 References Mostafa, M. M., Ahmed, H. S., Ahmed, A. A., Abdel-Raheem, G. A., & Ali, N. A. (2019). 41
Experimental study of flow characteristics around floodplain single groyne.
2 42
Journal of Hydro-environment Research, 22, 1e13.
3 Ahmad, Z. (2013). Prediction of longitudinal dispersion coefficient using laboratory 43
Nagy, H. M. (2004). Maximum depth of local scour near emerged vertical wall spur
and field data: Relationship comparisons. Hydrology Research, 44(2), 362e376.
4 Diplas, P., Dancey, C. L., Celik, A. O., Valyrakis, M., Greer, K., & Akar, T. (2008). The
dike. Alexandria Engineering Journal, 43(6), 819e829. 44
5 Nasrollahi, A., Ghodsian, M., & Neyshabouri, S. A. A. S. (2008). Local scour at 45
role of impulse on the initiation of particle movement under turbulent flow
permeable spur dikes. Journal of Applied Sciences, 8(19), 3398e3406.
6 conditions. Science, 322(5902), 717e720. 46
Oliveto, G., & Hager, W. H. (2002). Temporal evolution of clear-water pier and
Ezzeldin, R. M. (2019). Numerical and experimental investigation for the effect of
7 permeability of spur dikes on local scour. Journal of Hydroinformatics, 21(2),
abutment scour. Journal of Hydraulic Engineering, 128(9), 811e820. 47
8 Pandey, M., Ahmad, Z., & Sharma, P. K. (2015). Estimation of maximum scour depth 48
335e342.
near a spur dike. Canadian Journal of Civil Engineering, 43(3), 270e278.
9 Fazli, M., Ghodsian, M., & Neyshabouri, S. A. A. S. (2008). Scour and flow field around 49
Pandey, M., Ahmad, Z., & Sharma, P. K. (2017). Scour around impermeable spur
a spur dike in a 90 bend. International Journal of Sediment Research, 23(1),
10 dikes: A review. ISH Journal of Hydraulic Engineering, 24(1), 25e44. 50
56e68.
11 Pandey, M., Sharma, P. K., Ahmad, Z., & Karna, N. (2018). Maximum scour depth 51
Garde, R. J., Subramanya, K. S., & Nambudripad, K. D. (1961). Study of scour around
around bridge pier in gravel bed streams. Natural Hazards, 91(2), 819e836.
12 spur-dikes. Journal of the Hydraulics Division, ASCE, 87(6), 23e37.
Radan, P., & Vaghefi, M. (2016). Flow and scour pattern around submerged and non-
52
Garde, R. J., Subramanya, K., & Nambudripad, K. D. (1963). Closure of ‘Study of scour
13 submerged T-shaped spur dikes in a 90 bend using the SSIIM model. Inter- 53
around spur-dikes’. Journal of the Hydraulics Division, ASCE, 88, 167e175.
14 Gill, M. A. (1972). Erosion of sand beds around spur dikes. Journal of the Hydraulics
national Journal of River Basin Management, 14(2), 219e232. 54
Raikar, R. V., & Dey, S. (2005). Clear-water scour at bridge piers in fine and medium
15 Division, ASCE, 98(9), 1587e1602. 55
gravel beds. Canadian Journal of Civil Engineering, 32(4), 775e781.
16 Guo, J. (2012). Pier scour in clear water for sediment mixtures. Journal of Hydraulic 56
Sui, J., Afzalimehr, H., Samani, A. K., & Maherani, M. (2010). Clear-water scour
Research, 50(1), 18e27.
17 Hager, W. H., & Oliveto, G. (2002). Shields' entrainment criterion in bridge hy-
around semi-elliptical abutments with armored beds. International Journal of 57
Sediment Research, 25(3), 233e245.
18 draulics. Journal of Hydraulic Engineering, 128(5), 538e542. 58
Vaghefi, M., Ghodsian, M., & Salehi Neyshaboori, S. A. A. (2009). Experimental study
19 Koken, M., & Gogus, M. (2015). Effect of spur dike length on the horseshoe vortex 59
on the effect of a T-shaped spur dike length on scour in a 90 channel bend.
system and the bed shear stress distribution. Journal of Hydraulic Research,
20 53(2), 196e206.
Arabian Journal for Science and Engineering, 34(2), 337. 60
21 Valyrakis, M., Diplas, P., & Dancey, C. L. (2013). Entrainment of coarse particles in 61
Kothyari, U. C., Garde, R. C. J., & Ranga Raju, K. G. (1992). Temporal variation of scour
turbulent flows: An energy approach. Journal of Geophysical Research: Earth
22 around circular bridge piers. Journal of Hydraulic Engineering, 118(8), 1091e1106. 62
Surface, 118(1), 42e53.
Kothyari, U. C., Hager, W. H., & Oliveto, G. (2007). Generalized approach for clear-
23 water scour at bridge foundation elements. Journal of Hydraulic Engineering,
Valyrakis, M., Diplas, P., Dancey, C. L., Greer, K., & Celik, A. O. (2010). Role of 63
24 instantaneous force magnitude and duration on particle entrainment. Journal of 64
133(11), 1229e1240.
Geophysical Research: Earth Surface, 115(F2).
25 Kothyari, U. C., & Ranga Raju, K. G. (2001). Scour around spur dikes and bridge 65
Valyrakis, M., Michalis, P., & Zhang, H. (2015). A new system for bridge scour
abutments. Journal of Hydraulic Research, 39(4), 367e374.
26 monitoring and prediction. In Proceedings, 36th IAHR World Congress (Vol. 28), 66
Kuhnle, R., & Alonso, C. (2013). Flow near a model spur dike with a fixed scoured
27 June 28 eJuly 03, 2015, The Hague, The Netherlands. 67
bed. International Journal of Sediment Research, 28(3), 349e357.
^go, J. P., & Cardoso, A. H. (2013). Clear-water White, W. R., Milli, H., & Crabbe, A. D. (1973). Hydraulics Research Station Report No.
28 Lança, R. M., Fael, C. S., Maia, R. J., Pe
INT 119, Wallingford, U.K.. Sediment transport: An appraisal of available methods
68
scour at comparatively large cylindrical piers. Journal of Hydraulic Engineering,
29 Yagci, O., Celik, M. F., Kitsikoudis, V., Kirca, V. O., Hodoglu, C., Valyrakis, M., & Kaya, S. 69
ASCE, 139(11), 1117e1125.
30 Liang, C., Abbasi, S., Pourshahbaz, H., Taghvaei, P., & Tfwala, S. (2019). Investigation
(2016). Scour patterns around isolated vegetation elements. Advances in Water 70
Resources, 97, 251e265.
31 of flow, erosion, and sedimentation pattern around varied groynes under 71
Yazdi, J., Sarkardeh, H., Azamathulla, H. M., & Ghani, A. A. (2010). 3D simulation of
32 different hydraulic and geometric conditions: A numerical study. Water, 11(2), 72
flow around a single spur dike with free-surface flow. International Journal of
235.
33 Lim, F. H., & Chiew, Y. M. (2001). Parametric study of riprap failure around bridge
River Basin Management, 8(1), 55e62. 73
Zhang, H., Nakagawa, H., & Mizutani, H. (2012). Bed morphology and grain size
34 piers. Journal of Hydraulic Research, 39(1), 61e72. 74
characteristics around a spur dyke. International Journal of Sediment Research,
35 Liu, D., Valyrakis, M., & Williams, R. (2017). Flow hydrodynamics across open 75
27(2), 141e157.
channel flows with riparian zones: Implications for riverbank stability. Water,
36 9(9), 720.
Zhang, L., Wang, H., Zhang, X., Wang, B., & Chen, J. (2018). The 3-D morphology 76
37 evolution of spur dike scour under clear-water scour conditions. Water, 10(11), 77
Lodhi, A. S., Jain, R. K., & Sharma, P. K. (2016). Influence of cohesion on scour around
1583.
38 submerged dike founded in clayesandegravel mixtures. ISH Journal of Hy- 78
draulic Engineering, 22(1), 70e87.
39 79
40 80

Please cite this article as: Pandey, M et al., Experimental assessment and prediction of temporal scour depth around a spur dike, International
Journal of Sediment Research, https://doi.org/10.1016/j.ijsrc.2020.03.015