Wave Script Example

EXPERIMENTAL AND NUMERICAL
MODELLING OF WAVE-INDUCED CURRENT

AND WAVE TRANSFORMATION IN PRESENCE
OF SUBMERGED BREAKWATERS
by
Mojtaba Tajziehchi
A thesis submitted in fulfillment of the

requirements for the degree of
Doctor of Philosophy
School of Civil and Environmental Engineering

The University of New South Wales
Australia
September, 2006
ABSTRACT
Two dimensional experimental and numerical modeling of wave transmission and

wave-induced current over detached submerged breakwaters has been carried out in this
thesis.
Two preliminary 3D and a comprehensive series of 2D laboratory experiments have been

conducted in the wave basin and 3 m wide wave flume. The preliminary 3D
experimental tests qualitatively investigated the flow behavior behind a submerged
breakwater and confirmed the validity of the 2D tests. The 2D laboratory tests examined
wave breaking, reflection, transmission as well as wave-induced set-up and currents over
submerged breakwater/reef structures.
Different approaches to experimental data processing are examined in producing

reliable application of the 2D laboratory measurements. Sensitivity of wave
transmission coefficient, wave-induced set-up and wave-induced discharge over
submerged breakwaters to other dimensional and non-dimensional parameters are
comprehensively investigated. Previously published analytical/experimental studies for
predicting/calculating wave breaking, wave transmission, wave-induced set-up and
current are discussed and compared with the present experimental results. Improved
equations/models are presented.
Numerical modeling of the hydrodynamic effects of wave breaking and flow over a
submerged breakwater is investigated using Delft3D. The capability of the Delft3D
numerical model to simulate wave height transformation and wave-induced current over
submerged breakwaters is provided. Four different approaches using
combinations/options within the two main modules of Delft3D (SWAN and FLOW) are
tested in the numerical simulations and the results are compared to the laboratory
experimental data. Guidance is provided as to the most appropriate application of
WAVE/FLOW/ROLLER modules in Delft3D for the reliable prediction of discharge
and wave height over different width submerged breakwaters.
ii
ACKNOWLEDGMENT
I would like to express my sincere thanks to my supervisor, Associate Professor Ron

Cox for his continued support, expert advice and enlightening comments. I greatly
appreciate his effort to develop my research skills, critical thinking and academic
writing as intellectual assets of my academic life, I am deeply indebted to him.
Technical advice and supports of my co-supervisor Dr. Bruce Cathers throughout of this
study is mostly appreciated. I am also indebted to Dr. William Peirson for his valuable
advice and comments during this research.
I would like to extend my appreciation to staff members at Water Research Laboratory

for their support and enjoyable work environment. Special thanks to John Hart and John
Baird for their assistance in setting up the 3m wide flume and installing the instruments.
Caroline Hedges in Water Reference Library is thanked for her support and assistance
with library recourses and literature review.
I would also like to thank Drs. Maarten van Ormondth and Dirk Jan Walstra from
WL/Delft Hydraulics for their technical support and advice on running the numerical
models. Many thanks to my fellow research student at UNSW Mr. Ali Roufegarinejad
for his endeavor and assistance on drawing some figures of this thesis.
The Ministry of Science, Research and Technology of I.R Iran is mostly appreciated for
providing the scholarship and supporting this study.
Last but not least, I wish to thank my beloved wife Shiva and my lovely children,
Parand and Kian for their encouragement and patience throughout this study. I
gratefully dedicate this thesis to them and my dear parents that I am always indebted for
their encouragement and support.
iii
TABLE OF CONTENTS
Abstract.………………………………………………………………........................... ii
Acknowledgements...……………………………………………………………...........iii
Table of Contents………...…………………………………………..............................iv
List of Symbols………………………….......................................................................ix
List of Figures …………………………………………………………………..……. xii
List of Tables ……………………………………………………………………....... xxi
Chapter 1 INTRODUCTION ……………………………………………..…………… 1
1.1 Background ………………………………………………..…….……………….. 1
1.2 Purposes and Significance of Study …………………..…………………….…… 3
1.3 Objective …………………………………………………………………………..6
1.4 Layout of the Thesis ……………………………………………………………....7
Chapter 2 LITERATURE REVIEW …………………………………………………..10
2.1 Wave Transmission ………………………………………………………….…. 11
2.1.1 Experimental Studies …………………………………………………….…. 11
2.1.2 Analytical/Numerical Modelling of Wave Transmission ………..…………. 26
2.2 Set-up and Piling-up.………………………………………………….……..…...27
2.3 Wave-induced Current.…………………………………………………………..31
2.4 Wave-induced Set-up and Current over the Reef ………………….………..….. 32
2.5 Shoreline Response to Submerged Breakwaters ………………….……………. 36
2.6 Breaking Wave Roller ……………………………………………………….…. 39
2.7 Summary……………..……………………………………………………….…. 41
CHAPTER 3 EXPERIMENTAL SETUP AND DATA SAMPLING ……………….. 43
3.1 Wave Basin …………………………………………………………………..… 43
3.2 Wave Flume ……………………………………………………………….……. 44
3.2.1 Flume Layout ………………………………………………………….……. 44
3.2.2 Wave Generation and Paddle Calibration…..……………………………......45
3.3 Breakwater……………………………………………………………….………46
iv
3.4 Instrumentations …………………………………………………………..……. 47
3.4.1 Wave Probes ………………………………………………………………... 47
3.4.2 Water Surface Level Followers ………………………………………….…. 47
3.4.3 Acoustic Doppler Velocimeter (ADV) ……………………………………... 48
3.4.4 Monitoring ………………………………………………………………….. 49
3.4.5 Flow Direction Indicator …………………………………………………… 49
3.5 Test Conditions ………………………………………………………….……… 50
3.5.1 Three Dimensional Wave Basin Test ………………………………………. 50
3.5.2 Two Dimensional Testing in 3m Wide Flume …………………..…………. 50
3.6 Data Sampling ………………………………………………………………….. 51
3.6.1 Water Flow Direction …………………………………………………….… 51
3.6.2 Water Surface Level ..………………………………………………………. 51
3.6.3 Mean Water surface level …………………………………………………... 52
3.6.4 Horizontal Velocity ………………………………………………………… 52
3.7 Summary ……………………………………………………………………… 55
CHAPTER 4 EXPERIMENTAL DATA PROCESSING ………………………….… 56
4.1 Water Surface Fluctuation …………………………………………………….... 56
4.1.1 Harmonic Generation in the Flume without Breakwater ………………….... 57
4.1.2 Harmonic generation above the breakwater………………………..………...60
4.2 Incident Wave Height …………………………………………………………... 62
4.2.1 Method of Healy …….……………………………………………….……... 62
4.2.2 Method of Goda and Suzuki.………..………………………………………..64
4.2.3 Method of Mansard and Funke.……………………………………..………..66
4.2.4 Comparison of the Methods …………………….………………………….. 69
4.3 Wave Breaking ……………………………………………….………………… 70
4.3.1 Previous Studies of Wave Breaking over Shoals .………………………….. 71
4.3.2 Analysis of Wave Breaking .…….………………………………………….. 74
4.4 Wave-Induced Set-up ..…………………………………………………………. 75
4.4.1 Comparison of the Measurement Methods ……………………………..........79
v
4.5 Wave-induced Current ………………………………………………………….. 80
4.6 Conclusion ……………………………………………………………………… 81
CHAPTER 5 ANALYSIS AND DISUSSION OF EXPERIMENTAL RESULTS ….. 87
5.1 Wave Transmission ..…………………………………………………………….87
5.1.1 Transmitted wave height …………………………………………………… 87
5.1.2 Transmission Coefficient …..………………………………………………. 90
5.1.3 Effective Parameters for prediction of Kt …..……………………………… 90
5.1.4 Comparison with the Previous Models .……………………………………. 93
5.1.5 Proposed Empirical Model …………………………………………………. 95
5.2 Wave-induced Flow ……………………………………………………………. 98
5.2.1 Effectiveness of Dimensional Parameters ………………………………….. 99
5.2.2 Non-dimensional Analysis ……………………………………….………...102
5.2.3 Previous Models and Modification……..………………………………….. 104
5.2.3.1 Symonds et al. (1995) model………………………………………….105
5.2.3.2 Drei and Lamberti (1999) model ……………………………………. 108
5.2.3.3 Gourlay and Colleter (2005) model …………………………………..109
5.2.4 Modified Model for Discharge ..……………………………………………114
5.3 Wave-induced Set-up ..…………………………………………………………115
5.3.1 Sensitivity of Set-up to Other Parameters ……..……………….…………. 116
5.3.2 Comparison with Existing Models ……………………………….……….. 117
5.4 Summary and Discussion.....……………………………………………………118
CHAPTER 6 CONCEPTUAL DESCRIPTION OF WAVE AND FLOW MODULES
WITHIN DELFT3D NUMERICAL MODEL SYSTEM....………………………….122
6.1 Conceptual Description of Wave Model (SWAN)....…………………………..122
6.1.1 Wave dissipation in SWAN.….…………………………………………….124
6.1.1.1 Depth-induced Breaking dissipation....……………………………….124
6.1.1.2 Dissipation Due to Whitecapping.…………………………………….125
6.1.1.3 Bottom Friction Dissipation..…………………………………………126
6.1.2 Wave Propagation through Obstacles.………………………………………128
vi
6.1.3 Wave-induced forces....……………………………………………………..129
6.2 Conceptual Description of Flow Model.………………………………………. 130
6.2.1 Continuity Equation ………………………………………………………. 131
6.2.2 Momentum Equation ……………………………………………………… 132
6.2.3 Bottom Friction in FLOW….....…………………………………………….132
6.2.4 Wave-induced enhancement of bed shear stress ..………………………… 133
6.4 Roller Model. ...…………………………………………………………………134
CHAPTER 7 NUMERICAL MODELLING RESULTS.…..……………………….. 137
7.1 Model Setup …………………………………………………………………….138
7.1.1 Scale of the Model.………………………………………………………….138
7.1.2 Boundary Conditions ……………………………………………………….140
7.1.3 Physical Parameters ...………………………………………………………140
7.2 Wave Model….…………………………………………………………………141
7.2.1 Sensitivity Analysis….……………………………………………………...142
7.2.1.1 Empty flume …………………………………………………………142
7.2.2.2 Flume with Breakwater ………………………………………………145
7.2.2 Breaking Dissipation in SWAN …..………………………………………..146
7.2.3 Modified Dissipation Model in SWAN ……………………………………149
7.3 Flow Module ……………………………………………………………………154
7.3.1 FLOW-OBSTACLE model ..………………………………………………155
7.3.2 FLOW-WAVE Model ...……………………………………………………158
7.3.3 FLOW-ROLLER Model…………………………………………………….161
7.3.3.1 Breaking Parameter in ROLLER Model...……………………………162
7.4 Summary and Discussion……………………………………………………….168
Chapter 8 CONCLUSIONS ANS RECOMMENDATIONS…………………………172
8.1 Summary and Conclusion……………………………………………………….172
8.2 Discussion and Recommendations for Future Research..………………………176
References…………………………………………………………………………….178
vii
Appendix A: LIST OF TEST CONDITIONS AND VELOCITY MEASUREMENT
POINTS IN LABORATORY WAVE FLUME MODE......………………………….A.1
Appendix B: SAMPLES OF MATLAB PROGRAMS DEVELOPED IN THE
THESIS………………………………………………………….…………………….B.1
Appendix C: ATTRIBUTE FILES USED IN THE NUMERICAL MODEL………...C.1
viii
LIST OF SYMBOLS
Variable definition
A roller area
At area of breakwater cross section
a wave amplitude
ai incident wave amplitude
aR reflected wave amplitude
B breakwater crest width
Cbottom bottom friction coefficient
CD discharge coefficient of broad-crested weir
Cr Courant number
C2D 2D-Chézy coefficient
c wave velocity
cg wave group velocity
d Breakwater crest height
d50 median grain size
D dimensionless reef-top wave length
DW wave energy dissipation
Dr roller energy dissipation
Db depth-induced wave energy dissipation
Dtot total wave energy dissipation
Ew wave energy
Er roller energy
Fc nonlinearity parameter
Fx wave force in x- direction
Fy wave force in y- direction
Fξ depth-averaged wave-induced forcing in ξ direction
f wave frequency/friction coefficient
fwr wave friction factor
g gravitational acceleration
h water depth offshore the breakwater
he water depth at the reef edge
hf friction head loss
hs water depth over the breakwater (submergence depth)
H wave height
Hb breaking wave height
Hc critical incident wave height
Hi incident wave height
Hm maximum incident wave height
Hmax maximum wave height
Hmin minimum wave height
Hrms root mean squared wave height
HR reflected wave height
Hs significant wave height
Ht Transmitted wave height
Iw Wilmott number
Kp reef profile shape factor
ix
KR wave reflection coefficient
Kt wave transmission coefficient
k wave number
ks Nikuradse roughness length
Li incident wavelength
Lo deepwater wavelength
Lk deepwater wavelength corresponding to wave number k
Lp deepwater wavelength corresponding to peak period
mo zero moment of wave spectrum
Mk modified factor of Miche criterion
N Dimensionless set-up
N̂ action density of wave
n̂ Manning coefficient
Qb wave breaking probability
q wave-induced discharge (flux) per unit width
q* dimensionless wave-induced discharge
r friction parameter
R2 root mean squared
Ru potential run-up
S submergence ratio
S( f ) wave spectrum
Sds,br( f ) depth-induced breaking spectrum dissipation
Sds,w( f ) wave spectrum dissipation due to whitecapping
Sds,f( f ) wave spectrum dissipation due to bottom friction
Sk skewness
Sxx cross shore component of cross shore radiation stress
Sxy cross shore component of longshore radiation stress
Syx longshore component of cross shore radiation stress
~
s overall wave steepness
T wave period
U depth-averaged current velocity over the breakwater
UR= Ur Ursell number
uorb cross shore (horizontal) orbital velocity near the bottom
vx cross shore (horizontal) orbital velocity
Vx=Vr horizontal cross shore wave-induced current
α breakwater side slope
β distortion
δ piling-up lee side of the breakwater
δm momentum set-up
δs continuity set-up
Γ steepness dependent coefficient
γ=γb breaker index/parameter
γr wave reform index over the reef
η set-up (distance of water level from mean water level)
η maximum set-up over the reefs
ξ Iribarren number
κ free second harmonic wave number
ρ water density
x
τc bed shear stress due to flow
τm maximum bed shear stress
τw wave-induced bed shear stress
Θ=θ wave radiation angle
ω wave frequency
subscripts
b breaker conditions
c calculated value
i incident wave
m measured value
o deep water conditions
x x-direction
y y-direction
xi
LIST OF FIGURES
Figure 1.1 Salient and Tombolo formation behind offshore breakwaters.

Figure 2.1. Definition of submerged breakwater parameters and wave transmission.
Figure 2.2. Suggested curves for estimating wave transmission coefficient over low-
crested breakwaters.
Figure 2.3. Influence of breaking conditions on wave transmission coefficient.
Figure 2.4. Proposed wave transmission coefficient criteria.
Figure 2.5. Transmission coefficient of regular waves past over solid and rubblemound
breakwaters.
Figure 2.6. The effect of armour rock size on transmission coefficient.
Figure 2.7. Variation of transmission coefficient against breakwater crest width.
Figure 2.8. Breaker types and criteria for rectangular artificial reefs.
Figure 2.9. Prediction of transmission coefficient as a function of relative crest height.
Figure 2.10. Examples of mean water surface curves under wave action.
Figure 2.11. The influence of crest width on set-up.
Figure 2.12. Definition for energy flux theory for wave set-up on a submerged
breakwater.
Figure 2.13. Piling-up levels as function of net transmitted discharge.
Figure 2.14. Dimensionless set-up and current as a function of submergence ratio.
Figure 2.15. Comparison of theoretical and experimental values of discharge ratio.
Figure 2.16. Empirical relationship for preliminary prediction of shoreline response to

submerged structures.
Figure 2.17. Set-up of mean water level computed with and without the roller
contribution).
Figure 2.18. Intercomparison of computed longshore current velocities.
Figure 3.1. 24.5x16.0x0.7m wave basin used for preliminary 3D tests.
xii
Figure 3.2. Two piston type wave generator paddles applied in the wave
basin.
Figure 3.3 Layout of 2D laboratory test in 1.5x3.0x35m wave flume.
Figure 3.4. 1m wide internal channel located in the wave flume.
Figure 3.5. Accurate water follower gauge to control water level in the flume.
Figure 3.6 Combination of gravel beach and wave absorber panel.
Figure 3.7. Signal generator and data acquisition system in the wave flume.
Figure 3.8. Curves for estimating of paddle stroke to generate desired wave height.
Figure 3.9. Rectangular breakwater used in 3D tests.
Figure 3.10. Trapezoidal breakwaters used in the wave flume tests.
Figure 3.11. Capacitance wave probe utilized for water surface elevetion measurement.
Figure 3.12. Franklin Water Level Follower system for measuring mean water level
across the flume.
Figure 3.13. Standard 10 MHz down-looking ADV probe used to measure wave-
induced flow over the breakwater.
Figure 3.14. 250x250mm mesh of floats and weaving threads to investigate water flow
direction on the surface and bottom of the basin behind and between the break water in
3D test.
Figure 3.15. Long continuous rectangular breakwater in 3D preliminary test "TP-1".
Figure 3.16. Two segmented rectangular breakwater with 500mm gap in 3D preliminary
test "TP-2".
Figure 3.17. Flow direction at the water surface and bottom behind the rectangular
2000 mm crest length submerged breakwater in the 3D wave basin preliminary test.
Figure 3.18. Flow direction at the water surface and bottom behind the two segmented
rectangular 1000 mm crest length submerged breakwaters with 500 mm gap in the 3D
wave basin preliminary test.
Figure 3.19. Flume layout and wave probes positions for the tests without breakwater.
Figure 3.20. Wave flume layout and wave probes and pressure taps positions for test
with 3500mm breakwater crest width.
xiii
Figure 3.23. Horizontal and vertical velocity variations in X,Y and Z direction
measured by ADV at 40mm from top of breakwater.
measured by ADV at 40mm from top of breakwater.
Figure 4.1. Variation of wave profile at different stations of probes in the empty flume.
Figure 4.2. Spectrum of water level at the position of probe4.
Figure 4.3. Variation of dimensionless parameters in the test without breakwater for
harmonic generation and no higher harmonic circumstances.
Figure 4.4. Wave regime around a shelf and a rectangular barrier (Goda et al., 1999).
Figure 4.5. Spectrum for water surface elevation for the selected tests of T1 (B=3500
mm), T2 (B=1500 mm) and T3 (B=300 mm) in front, over and behind the breakwater.
Figure 4.6. Comparison of zero up-crossing and down-crossing methods for

calculating wave height from the traverse wave probe.
Figure 4.7. Comparison of smoothed and non-smoothed wave height for zero down-
crossing method.
Figure 4.8. Calculated incident and reflected spectrum using the method of Goda and
Suzuki (1976) for test T0-26.
Figure 4.9. Comparison of conventional method of Healy (1953) for calculation of

incident wave height with Goda & Suzuki (1976) method.
Figure 4.10. Calculated incident and reflected spectrum using the method of Mansard
and Funke (1980) for test T3-30.
Figure 4.11. Comparison of conventional method for calculation of incident wave

height with Mansard & Funke (1980) method.
Figure 4.12. Variation of inverse submergence ratio “Hi/hs” as a function of wave

steepness for breaking and non-breaking waves.
Figure 4.13. Variation of wave steepness in relation to the ratio of hs/L for breaking
and non-breaking waves.
xiv
Figure 4.14. Variation of wave height ratio against relative breakwater crest width for
breaking and non-breaking waves over the breakwater.
Figure 4.15. Calculated breaker-depth index “γb” as a function of dimensionless

breakwater crest width using different methods.
Figure 4.16. Wave profile shape of different progressive gravity waves (CERC, 2004).
Figure 4.17. (a) Variation of Ursell number and skewness of water surface elevation
for the test without breakwater, (b) variation of relative wave height and skewness of
water surface level for the test without breakwater (Test T0).
Figure 4.18. Measured wave-induced set-up over and behind the breakwater using
Franklin Water Level Follower and wave probes (B=3500 mm).
Figure 4.21. Current velocity profile definition over the breakwater.
Figure 4.22. Variation of velocity profile over the breakwater.
Figure 5.1 Variation of transmission coefficient against incident wave height.
Figure 5.2 Variation of transmission coefficient against wave period.
Figure 5.3 Variation of transmission coefficient against breakwater crest width.
Figure 5.4 Variation of transmission coefficient against submergence depth.
Figure 5.5 Variation of correlation coefficient between transmission coefficient "Kt"

and submergence depth "hs" (solid blue line), wave period "T" (dash red line) and
incident wave height "Hi" (dash-dot purple line) against breakwater crest width "B".
Figure 5.6 Variation of correlation coefficient between transmission coefficient " Kt "
and submergence depth " hs " (solid blue line), wave period "T" (dash red line) and
Breakwater crest width" B " (dash-dot purple line) against incident wave height " Hi "
Figure 5.7 Variation of transmission coefficient against different dimensionless

variables for the whole tests with breaking wave over the breakwater.
Figure 5.8 Comparison of measured transmission coefficient with previous

experimental models.
xv
Figure 5.9 The values of distortion and Root Mean Squared Error for previous
empirical models.
Figure 5.10 Variation of transmission coefficient against the ratio of breakwater crest
width to incident wave height for the tests with Iribarren numbers greater and smaller
than 3.
Figure 5.11 Comparison of the measured and calculated transmission coefficient using
the proposed models.
Figure 5.12 Statistics comparison of measured transmission coefficient with the

calculated values from the proposed models.
Figure 5.13 Comparison of the measured and calculated transmission coefficient using
the proposed model.
Figure 5.14 Variation of discharge over the breakwater as a function of incident wave
height for different breakwater crest width, water depth and wave period.
Figure 5.15 Wave-generated flow as a function of wave period for constant wave
height and different water level and breakwater crest width.
Figure 5.16 Variation of wave-induced discharge as functions of submergence depth

for given wave height and wave period.
Figure 5.17 Variation of discharge ratio against: a) submergence ratio, b) inverse of

breakwater crest width ratio, c) dimensionless wave height, d)wave steepness.
Figure 5.18 Envelop of maximum dimensionless discharge as a function of

submergence ratio for different ranges of breakwater width ratio.
Figure 5.19 Schematic definition of parameters used in Symonds’ et al. (1995) model.
Figure 5.20 Variation of calibrated friction term in the model of Symonds et al. (1995)
as functions of : a) Breakwater crest width ratio, b) Breakwater crest width.
Figure 5.21 Comparison of measured and calculated current velocity over the
submerged breakwaters/reefs using the calibrated method of Symonds et al. (1995).
Figure 5.22 Calculated statistics indicators for comparison the results of calibrated
method of Symonds et al. (1995) with the measured data.
Figure 5.23 Comparison of discharge ratio between measured by Drei and Lamberti
(1999) and the present experimental test. The red and blue dash line demonstrates the
average value of measured discharge ratio by Drei and Lamberti and the present study,
respectively.
Figure 5.24 Coral reef definitions.
xvi
Figure 5.25 Wave-induced flow over a horizontal ideal reef: (i) reef-rim control (ii)
Reef-top control.
Figure 5.26 Wave breaking conditions and reformation over the reef as a function of
nonlinearity parameter Fco.
Figure 5.27 variation of wave set-up as a function of M=3g0.5Hi2T/(64п (ή+hs)1.5).
Figure 5.28 Comparison of Gourlay and Colleter (2005) model with measured
laboratory data.
Figure 5.29 Variation of overestimating of model of Gourlay and Colleter (2005) as a

function of crest with ratio for S>0.7.
Figure 5.30 Comparison of amended model of Gourlay and Colleter (2005) with
measured laboratory data for dimensionless discharge.
Figure 5.31 Statistics indicator, Wilmott index (IW) and error function (ε) for
evaluating the goodness of Gourlay and Colleter (2005) model and the modified model.
Figure 5.32 Variations of maximum wave set-up over (behind) the breakwater as a
function of incident wave height for different breakwater crest width.
Figure 5.33 Maximum wave set-up over (behind) the submerged breakwater with
different crest width as functions of total water depth for given wave heights and period.
Figure 5.34 Variation of maximum wave set-up over (behind) submerged breakwater
with different crest width and water depth for a given wave height.
Figure 5.35 Comparison of measured maximum set-up with calculated values (in mm)
using the proposed equation of Diskin et al. (1970).
Figure 5.36 Comparison of measured and calculated maximum set-up over submerged
breakwater with two different breakwater crest width using the proposed model of
Gourlay and Colleter (2005).
Figure 5.37 Comparison of measured maximum set-up over submerged breakwater

with two different breakwater crest width with the calibrated model of Gourlay and
Colleter (2005). The reef shape factor KP was adopted 0.2 for the narrower breakwater
(B=1500 mm).
Figure 6.1 Variation of breaking probability Qb in Battjes and Janssen (1978) theory
against non-dimensional ratio of Hrms/Hm
Figure 6.2 Definition sketch for wave roller (after Dally and Brown, 1995).
Figure 7.1. Comparison of calculated wave transformation over submerged breakwater

in the model with 1m and 0.2 m grid size
xvii
Figure 7.2. A sample of computational grid used in the numerical model.
Figure 7.3. Wave height and wave energy dissipation along the flume without
breakwater due to bottom friction (Hasselmann et al., 1973) for test T1-27.
breakwater due to bottom friction (Collins, 1972) for test T1-27.
breakwater due to bottom friction (Madsen et al., 1988) for test T1-27.
Figure 7.6 Wave height and wave energy dissipation along the flume without
breakwater for test T1-27 due to steepness-induced breaking (whitecapping).
Figure 7.7. Wave height attenuation and wave energy dissipation along the flume
without breakwater.
Figure 7.8. Calculated wave height variation along the flume when depth-induced
wave breaking was considered in SWAN with and without bottom friction effects.
Figure 7.9. Calculated wave energy dissipation along the flume when depth-induced
wave breaking was considered in SWAN with and without bottom friction effects.
wave breaking was considered in SWAN with and without whitecapping effects.
wave breaking was considered in SWAN with and without whitecapping effects.
Figure 7.12. Wave height attenuation around submerged breakwater due to depth-
induced breaking.
Figure 7.13. Comparison of calculated transmission coefficient by SWAN model with

the measured one in laboratory experiments
Figure 7.14. Comparison of measured transmission coefficients with calculated using

SWAN (dissipation model of Battjes and Janssen, 1978 with γ=0.73).
Figure 7.15. Variation and effectiveness of different dimensionless parameters on

calculated breaker parameter (Battjes and Janssen, 1978 dissipation model).
Figure 7.16. Variation of derived calibrated breaking index against submergence ratio
hs/Hi for all experimental data (dissipation model of Battjes and Janssen, 1978).
Figure 7.17. Proposed model for calculating breaker parameter in SWAN (selected
experimental data, 0.5<hs/Hi ≤2.0).
xviii
Figure 7.18. Comparison of measured and calculated transmission coefficient Kt using
the proposed calibrated values of breaker index in dissipation equation of Battjes and
Janssen (1978).
Figure 7.19. Proposed linear model for calculating breaker index γ in wave dissipation
model.
Figure 7.20. The variation of calculated discharge over the time in FLOW module.
Figure 7.21. Calculated incident wave height variation along the flume for FLOW-
OBSTACLE model.
Figure 7.22. Calculated water level changes along the flume for FLOW-OBSTACLE
model.
Figure 7.23. Calculated wave-induced forces along the flume for FLOW-OBSTACLE
model.
Figure 7.24. Calculated depth averaged velocity in the flume and around the
submerged breakwater for FLOW OBSTACLE model.
Figure 7.25. Comparison of calculated and measured discharge over submerged

breakwater when FLOW-OBSTACLE model applied in the simulation.
Figure 7.26. Statistical comparison of measured values of discharge with the calculated
values from FLOW-OBSTACLE model for different ranges of breakwater crest width
(B/L).
Figure 7.27. Calculated wave-induced forces by SWAN in x- direction around the

breakwater using radiation stress gradients method and wave energy dissipation method.
Figure 7.28. Calculated depth-averaged velocity and discharge in the flume using
FLOW-WAVE-2 model.
FLOW-WAVE-2 model.
FLOW-WAVE-2 model.
Figure 7.31. Computed discharge using radiation stress gradients (FLOW-WAVE-1)

and wave energy dissipation.
Figure 7.32. Comparison of calculated and measured discharge using radiation stress
gradients (FLOW-WAVE-1) and wave energy dissipation (FLOW-WAVE-2) methods.
Figure 7.33 Deviation between calculated and measured discharge as a function of

breakwater crest width for radiation stress gradients and energy dissipation methods.
xix
values from FLOW-WAVE-1 and FLOW-WAVE-2 models.
Figure 7.35 Statistical comparison of measured values of discharge with the calculated
values from FLOW-WAVE-1 and FLOW-WAVE-2 models for different range of
breakwater crest width B/L.
Figure 7.36. Comparison of measured transmission coefficient with the calculated Kt

using Roelvink’s (1993) dissipation model and the proposed value of breaking
parameter γ=0.55.

calculated breaker parameter (Roelvink, 1993 depth-induced dissipation model).
Figure 7.38. Proposed model for calculating the calibrated breaker parameter in the
dissipation model of Roelvink (1993). Selected experimental data with 0.5<hs/Hi ≤2.0
was adopted during calibration processes.
Figure 7.39. Comparison of measured transmission coefficients Kt with the calculated

values using the calibrated breaker parameters (Equation 7.14).
Figure 7.40. Comparison of measured and calculated transmission coefficients Kt with

the using the calibrated breaker parameters in the Roelvink’s dissipation model.
Figure 7.41. Variation of calculated short-wave and roller energy by FLOW-ROLLER

model along the flume and over the submerged breakwater for test T1-26.


Figure 7.44. Computed depth-averaged current velocity along the flume and over the
submerged breakwater using FLOW-ROLLER model for Test T1-26.
Figure 7.45. Computed depth-averaged discharge per unit width along the flume and
over the submerged breakwater using FLOW-ROLLER model for Test T1-26.
Figure 7.46. Calculated discharge over the submerged breakwater using roller model
(FLOW-ROLLER) against the measured value from laboratory experiments.
values from FLOW-ROLLER models for different groups of crest width ratio.
Figure 7.48. Comparison of all applicable models in Deflt3D in predicting discharge

over submerged breakwaters for different relative breakwater crest width groups.
xx
LIST OF TABLES
Table 2.1 Test conditions of previous laboratory experimental research on submerged
breakwater.......................................................................................................................14
Table 3.1 Test conditions for 3D experimental test in the wave basin…………………50
Table 3.2 Test conditions in the 3m wave flume (2D)....................................................51
Table 3.3. A sample of user input parameters for data acquisition using ADV.............54
Table 4.1. Variation of Ursell number, wave steepness and relative wave height in the
tests without breakwater..................................................................................................58
Table 4.2 Calculated incident and reflected wave height for all test conditions using the
method of Healy (1953); See Appendix A of test conditions for each test number........63
Table 4.3 Calculated incident wave height and reflection coefficient for the test
without breakwater and with 300 mm breakwater crest width using three different
methods…………………………………………………………………………………68
Table 4.4 Statistical comparison of the methods of Goda and Suzuki (1976) and
Mansard and Funke (1980) with the conventional method to calculate incident wave
height from the measured wave spectrum in laboratory………………………………..69
Table 4.5 Summary of previous studies to calculate breaker-depth index for plain
beaches.............................................................................................................................71
Table 4.6 Difference of transition values of breaker type for plane slopes…………....72
Table 4.7 List of tests with breaking wave and comparison with predicted wave
conditions using previous proposed methods..................................................................77
Table 4.8 Measured maximum set-down and set-up from Franklin Water Level
Follower probes...............................................................................................................84
Table 4.9 measured wave-induced discharge over submerged breakwaters..................85
Table 5.1 The average transmitted wave height applying different methods and
calculated transmission coefficients for all test conditions with breaking wave over the
submerged breakwater.....................................................................................................89
Table 5.2 Correlation coefficient between Kt and other dimensionless variables...........93
Table 5.3 Test limitation for the previous proposed transmission coefficient formulas
and the present experimental data...................................................................................95
xxi
Table 5.4 Calibration coefficients calculated by nonlinear regression analysis for best
fitting of data and the corresponding statistical parameters for evaluation of the
models............................................................................................................................100
Table 5.5 Correlation coefficient between the measured discharge and other
independent dimensional parameter..............................................................................101
Table 5.6 Correlation coefficient between the measured dimensionless discharge and
other non-dimensional parameters.................................................................................104
Table 5.7 Calculated current velocity over the breakwater/reef using the calibrated
model of Symonds et al. (1995).....................................................................................107
Table 5.8 Calculated reef profile factor Kp for the reefs with different slope.............113
Table 7.1 Scaled test conditions used in the numerical model....................................139
Table 7.2 Wave and water depth condition for scaled model of tests T1-26 and T1-
27...................................................................................................................................143
Table 7.3 Test conditions applied in the numerical analysis by Johnson (2006)........149
Table 7.4 Values of derived calibrated breaker parameter γ for different selected test
conditions……………………………………………………………………………..152
Table 7.5 Calculated discharge “qc” and transmission coefficient “Kt” over submerged
breakwaters for the test series applied in FLOW-OBSTACLE model…….................157
Table 7.6 Calculated discharge and calibrated breaking parameter over the submerged
breakwaters for the test series applied in FLOW-WAVE model (scaled model).........161
Table 7.7 Statistics comparison of calculated and measured transmission coefficient

applying the proposed calibrated breaker parameter γ .................................................164
breakwaters for the test series applied in FLOW-ROLLER model...............................167
Table 7.9 Statistical parameters for the four models…………………………..…….168
Table A.1 Test conditions in wave flume without breakwater....................................A.1
Table A.2 Test conditions in wave flume with 3500mm crest with breakwater……...A.2
Table A.3 Test conditions in wave flume with 1500mm crest with breakwater...........A.3
Table A.4 Test conditions in wave flume with 300mm crest with breakwater.............A.4
xxii
CHAPTER 1
INTRODUCTION
1.1 Background
Wave breaking in the surf zone causes high turbulence, wave-induced surges and
currents that suspend and transport sediment resulting in coastal erosion. Various
coastal structures (e.g. breakwaters, seawalls, dikes and revetments) have been
constructed to reduce wave impacts and solve problems of coastal erosion.
Offshore breakwaters are constructed a distance from the beach to protect the shoreline
from erosion against storms. These kind of structures can retard erosion on existing
beaches in an environmentally friendly way, promoting natural sediment to form a new
beach and maintain a wide beach for storm damage reduction and recreation. Detached
(offshore) breakwaters have been successfully used around the world (as a single
structure, segmented or in combination with other measures) to protect beaches by (i)
reducing wave energy and rearranging waves and currents in the lee side of the
breakwater, (ii) redistributing sediment transport pattern in the protected area so as to
provide desirable beach features, (iii) preventing sedimentation in port access ways.
The first offshore breakwater was constructed in 1935 at Winthrop Beach, USA
(Chasten et al., 1993). The project which included 5 segmented offshore breakwaters
(91 m length each) to protect 625 meters of shoreline created permanent tombolos (see
Figure 1.1) along the beach. Pilarczyk and Zeidler (1996) indicated that 9.2 km of
beaches in USA and 9.4 km in Japan are protected using detached offshore breakwaters
and that the use is growing in USA and Europe.
Detached offshore breakwaters may be shoreline parallel or oblique and emerged or

low-crested (semi submerged or fully submerged). Emerged breakwaters have crest
level above the mean water level so that waves cannot overtop the crest (no wave
1
transmission over the structure) whilst frequent over topping and relatively high wave
transmission occurs across low-crested breakwaters. Crest level of fully submerged
breakwaters are permanently below the mean sea level and waves pass over the
breakwater with breaking. Higher energy is transmitted over submerged breakwaters
than semi-submerged structures.
One of the main advantages of employing submerged breakwaters in coastal

management is that the protective function can be fulfilled without spoiling landscape.
This is increasingly important in recreational and residential coastal developments. In
addition, submerged breakwaters are beneficial in providing fish breeding habitat and
shelter areas. These protective structures also allow water exchanges between the lee
side and sea side of the breakwater to maintain water quality at the beach side for
recreational purposes.
Well designed submerged breakwaters permit continued movement of longshore

transport in the lee of the structure thus reducing any adverse effects of downdrift
erosion. Providing a calm area (by incident wave dissipation), submerged breakwaters
can capture the sediment transport in the sheltered area to improve shorelines by
creating a salient at the beach (Figure 1.1). Moreover, construction of submerged
breakwaters can be successfully integrated with artificial sand nourishment projects to
restore beach profile (Lamberti and Mancinelli, 1996) – diminish nourishment losses
and extend the nourishment project life.
Submerged (low-crested) offshore breakwaters have been recently used as an effective

system for shore protection and renourishment in coastal management. Submerged
breakwaters of varying crest widths have been constructed in coastal areas. In tidal
environments and when frequent storm surges occur (where breakwater crest height is
increasingly submerged below the sea water level) narrow-crested structures are less
effective in shore protection. Broad-crested submerged breakwaters (artificial reefs) are
more effective in high submergence depth – however, proper cost-benefit studies should
be carried out in any consideration of a wide crested submerged breakwater.
In Europe, around 1200 single or segmented low-crested breakwaters have been

constructed to protect European coastlines (Lamberti et al, 2005). Different breakwater
2
geometries, submergence depths (water depth over the breakwater) and distance from
the shore line have been considered in the European. DELOS research project
(environmental DEsign of LOw crested coastal defence Structures). DELOS was
established in Europe aiming to “promote effective and environmentally compatible
design of low crested structures to protect European shore lines against erosion”
(Lamberti, 2005). 18 partners from 7 European countries were involved in the DELOS
project, thus indicates the significance of the application of these structures to coastal
engineering and management. The scientific results of engineering, socio-economic and
ecological studies of low crested breakwaters were published in a special issue of
Coastal Engineering, Volume 52, Issues 10-11, 2005.
Submerged breakwaters have also been used in USA coastlines. Narrow crested
prefabricated submerged breakwaters with triangle cross sections have been constructed
at lower central east coast of Florida, Palm Beach County, New Jersey Coast, Cape May
County and Monmouth County in USA (Stauble and Tabar, 2003). The objective of the
structures was wave height attenuation and maintaining a stable shoreline position
against storm waves. Field monitoring of the functionality of the narrow crested
breakwaters by Stauble and Tabar (2003) showed that using only the shore parallel
narrow crested structures without any other shore protection structures was ineffective
in stabilizing the shore line behind the breakwater.
1.2 Purposes and Significance of Study
The functionality of submerged breakwaters depends on the incident wave climate,

breakwater geometry (e.g. crest width and seaward slope), water depth over the
structure (submergence depth) and distance from the shoreline. Shoreline changes
behind submerged breakwaters are influenced by wave energy that reaches the beach
and the currents (pattern and magnitude) behind the structure. The functional design
knowledge of submerged breakwaters including their impacts on wave transmission,
currents, sediment processes and shoreline response is still developing.
Many numerical modelling and empirical approaches for estimating wave transmission
over submerged breakwaters have been developed (Johnson et al., 1951; Adams and
Sonu, 1986; Losada et al., 1996; d’Angremond et al., 1996; Seabrook and Hall, 1998;
3
Schlurmann et al., 2002; van der Meer et al. 2005). Most of the previous experimental
researches on wave transmission over submerged breakwaters have been carried out for
relatively narrow crested and semi submerged breakwaters. The results of experiments
being limited to the tested ranges of breakwater crest width. As indicated in the previous
section, narrow crested (fully) submerged breakwaters were found to be ineffective in
coastal protection in most field investigations. Therefore more investigations are needed
to extend the studies of wave transmission over submerged breakwaters with wide to
broad crest width due to their improved efficiency in protection of coastal area.
The gradient of wave-induced set-up behind submerged breakwaters causes water to

flow behind such structures. Wave induced set-up and current are dependent on mass
flux over the breakwater which is contributed by wave averaged flow generated by
radiation stress gradients, leakage of water through the porous structure of the
breakwater and mass transport by breaking wave rollers (Lesser et al. 2003).
Most of the existing numerical models apply the radiation stress formula provided by
Longuet-Higgins and Stewart (1964). In this formula wave set-up depends on incident
wave energy, wave direction, water depth and wave number in shallow water. However,
mass transport due to breaking wave rollers has not been applied in any hydrodynamic
models for submerged breakwaters.
Gourlay (1993, 1996a) conducted several experimental tests for measuring wave set-up
and mass flux over reefs. He observed that wave set-up on a horizontal coral reef
increases with wave height and wave set-up does not occur in large water depths which
allow waves to pass over the reef. He also showed that wave-induced flow across the
reef increases with increasing incident wave height and period in a similar manner to
wave set-up. In his research, only reefs with long crest width were tested due to research
focus being on broad coral reefs. The impact of crest width (shorter), surface roughness
and permeability of structure were ignored. Symonds et al. (1995) presented a
theoretical model for wave-driven flow over shallow reefs. They investigated the wave
set-up and cross-reef currents generated by waves breaking on the reef face as a
function of incident wave height and still water depth over the reef. They provided a
linear theoretical model demonstrating how the relative magnitudes of the currents and
set-up depend on the geometry of the reef and the magnitude of the forces including
4
radiation stress. The model verification is limited to the specific field measurement
conditions and the model has not been verified for the wider range of submerged
breakwater geometries, water levels and wave conditions tested in laboratory
experiments within this thesis.
A few numerical modelling studies have been carried out to simulate wave-induced
current behind submerged breakwaters. van der Biezen et al. (1997) applied
Unibest_TC and Delft3D-mor (2DV and 2DH morpho-dynamic models, respectively)
to simulate water level changes, current and beach profile evolution behind submerged
breakwaters. They found that the calculated current pattern behind the breakwater was
similar to 3D experimental testing (van der Biezen et al., 1996). The model currents
were however smaller than the measured velocities, probably because the contribution
of wave-induced mass flux over the breakwater was ignored in the numerical model.
Lesser et al. (2003) used Delft3D to investigate, numerically, the effects of varying
design parameters on performance of submerged breakwaters. They underlined the
importance of appropriate calculation of mass flux in the numerical model and indicated
that there is no adequate data on discharge passing over submerged breakwaters to
calibrate and verify the model.
Johnson et al. (2005) investigated with a numerical model, the wave evolution and
current around submerged breakwaters. They applied two phase-averaged models in the
numerical simulation, a wave model (LIMWAVE) and a Q3D nearshore circulation
model (LIMCIR) in addition to a phase-resolving approach based on a 2DH Bousinesq-
type model for wave evolution over permeable submerged breakwaters. The calculated
transmitted wave height using both phase-average and phase resolving models were
found to be in good agreement with experimental results if the wave breaking sub-
model was tuned for wave dissipation over the submerged breakwater. Johnson et al.
(2005) indicated that the phase-averaged models predict flow circulation pattern around
the breakwaters that was qualitatively in good agreement with the laboratory
measurements. However, the predicted values of current speeds and wave set-up inshore
of the breakwater were found to be higher than laboratory test results. Zanuttigh and
Lamberti (2006) conducted a series of experimental tests in a wave basin to investigate
wave transmission and current around submerged and semi submerged breakwaters.
They used the experimental data to validate a numerical model (MIKE21 PMS) for the
5
calculation of waves and currents in the vicinity of submerged breakwaters. Zanuttigh
and Lamberti (2006) found discrepancies among experimental and numerical results to
be related to the breaking model adopted in the numerical model (Battjes and Janssen,
1978). Battjes and Janssen’s model overestimated energy dissipation due to the abrupt
depth decrease.
In this thesis the hydrodynamic effects of submerged breakwaters are investigated,

experimentally and numerically. Comprehensive 2D laboratory experimental tests have
been carried out in 3 m wide wave flume in presence of impermeable submerged
breakwater/reef structures with wide ranges of breakwater crest width, water depth and
wave climate. A central 1 m wide channel was built such as to minimize/prevent flow
over the structure. Wave transmission, wave-induced set up and current (mass flux) over
and behind the breakwater were measured and analysed. The numerical model
“Delft3D” was applied to simulate waves and currents over and behind the structure.
Different approaches were investigated in the hydrodynamic modelling - models were
calibrated (improved with the experimental data) and guidelines given to the most
appropriate modelling approach to be used for various submerged breakwater
applications.
1.3 Objective
The main goal of this study was to investigate wave transformation and wave-induced
mass flux over a submerged breakwaters/reef. A comprehensive 2D experimental
laboratory test program has been conducted to explore the effect of different parameters
on wave transmission and current over submerged breakwaters. Numerical modelling
of wave deformation and current has also been carried out and calibrated to provide a
functional design tool to predict hydrodynamic effects of submerged breakwaters. The
specific aims towards the achievement of these goals are given as follows:
i- Setting up laboratory tests to measure wave transformation, wave-induced set-up

and wave-induced mass flux over and behind submerged breakwaters for a broad
range of geometric, water level and wave conditions.
6
ii- Analysis of test data to investigate wave breaking (breaker index) over
submerged breakwaters.
iii- Analysing data to examine the effectiveness of various parameters (dimensional

and non-dimensional) on wave transmission coefficient and mass flux.
iv- Comparison of measured transmission coefficient with previous empirical

models and developing a reliable improved relationship to calculate transmission
coefficient over submerged breakwaters over a wide range of conditions.
v- Comparison of measured wave-induced flux with previous analytical models

and developing a reliable improved relationships to calculate wave-induced flux over
reef type submerged breakwaters.
vi- Modifying the depth-induced breaking numerical dissipation models to improve

the calculation of breaker parameter over submerged breakwaters. The values of
calculated breaker parameter being used to calibrate the numerical models.
vii- Providing functional design advice as to the best application of Delft3D for
numerical modelling of hydrodynamic characteristics for different submerged
breakwater conditions.
1.4 Layout of the thesis
The thesis deals with the experimental and numerical modeling of wave-induced current
(mass flux) and wave transmission over submerged breakwaters. The main body of the
thesis is divided into seven chapters including this introduction. The introduction
provides an overview of the background, topics and motivation of the research. The
contents of the other chapters in this thesis are briefly presented as follows:
Chapter 2 introduces past research on submerged breakwater focusing on experimental

and analytical modeling of wave transmission, wave-induced set-up and wave-induced
mass flux over submerged breakwaters and reefs. Recent numerical investigations on
simulating wave transmission and current behind submerged breakwaters are also
7
reviewed briefly. In addition, a brief review of recent research on shoreline response to
submerged breakwaters is provided.
Chapter 3 details the experimental set up (2D and 3D) and equipment utilized in the
laboratory testing. The methods of data acquisition throughout the tests are also defined
in this chapter.
In Chapter 4, different approaches for data processing were examined in deriving the
most appropriate methods of producing reliable application of the laboratory
measurements. Water surface elevation (set-up and set-down), incident wave height,
wave-induced discharge and wave breaking over the submerged breakwater were
investigated in this chapter and the most appropriate methods are proposed to calculate
water level, incident wave height and wave breaking prediction from measured data.
Chapter 5 presents the analysis of the experimental results. Wave height transformation
over the submerged breakwater, wave-induced set-up and wave-induced current are
analysed and the effectiveness of several parameters (dimensional and dimensionless)
are examined for improved understanding of hydrodynamic processes. Previously
published analytical/experimental studies for calculating wave transformation, wave-
induced set-up and current are discussed and compared with the present experimental
results. Improvements to the more appropriate previous models are proposed.
Chapter 6 presents the conceptual description of the Delft3D numerical model package
applied in the thesis.
In Chapter 7 the capability of numerical modeling (Delft3D) to simulate wave height

transformation and wave-induced current over submerged breakwaters is provided. Four
different approaches using combinations/options within the two main modules of
Delft3D (SWAN and FLOW) are tested in the numerical simulations and the results
were compared to the laboratory experimental data. Guidance is provided as to most
appropriate WAVE/FLOW/ROLLER modules to predict discharge and wave height
over different width submerged breakwaters.
Chapter 8 concludes this body of work with main summaries and recommendations for
future research.
8
Publications and seminars presented during the completion of this thesis are provided in
Appendix D.
9
Figure 1.1. Salient and Tombolo formation behind offshore breakwaters (Kamphuis,
2000).
CHAPTER 2
LITERATURE REVIEW
During the past three decades, detached submerged breakwaters have been increasingly
used as an effective system for shore protection and renourishment in coastal
management. A submerged breakwater is a barrier structure which is constructed
offshore of the beach so that its crest is at or below the still water level. These structures
dissipate wave energy reaching the shoreline by causing the wave to break when waves
pass over the crest of the submerged breakwater. One of the main advantages of
employing submerged breakwaters in coastal management is that the protective function
can be fulfilled without spoiling visual landscape. This is increasingly important in
recreational and residential coastal developments. In addition, submerged breakwaters
can be beneficial in providing fish breeding habitat and shelter areas.
Historically, as reported (Soper, 1940 and Beach Erosion Board, 1940), a submerged
breakwater was used for the first time in United States as an aid in the control of beach
erosion. A vertical wall type barrier was constructed with sheet-steel piling system so
that the top of the piling was at about still-water level. The impact of submerged
breakwaters on beach profile depends on wave transmission over and wave-induced
current behind the structure. The effects of breakwater submergence, wave climate,
distance of structure to shoreline and crest width of breakwater on wave/current
hydrodynamics and morphological seabed and shoreline changes are not still well
understood.
Laboratory experimental testing, numerical modelling approaches and field studies

continue to be developed for improved understanding of hydrodynamic and
morphodynamic effects of this kind of structure in the coastal zone. The previous
laboratory experimental research on submerged breakwaters is summarized in Table
2.1. The table clearly indicates test conditions (wave climate, breakwater type and
10
geometry, water depth, wave flume/basin dimensions, etc) of previous laboratory
studies.
This chapter presents historical investigations and research of hydrodynamic studies on

submerged breakwaters including wave transformation/transformation, wave-induced
set-up and current across submerged breakwaters or reefs.
2.1 Wave Transmission
Various experiment laboratory studies and numerical models have been developed for
predicting wave transmission passing across low-crested breakwaters.
2.1.1 Experimental Studies
The first reliable quantitative data on the effect of submerged breakwaters on wave
action has been reported by Beach Erosion Board (1940). In these studies, a wide
variety of structure type and geometry were investigated in order to better understand
the damping effect of vertical walls and barriers with triangle and trapezoidal cross-
sections. Tests on a trapezoidal section have also been made by Stucky and Bonnard
(1937). Both sets of the former experiments were carried as two-dimensional, wherein
the wave crests were parallel to the barrier axis. Morison (1949) examined rectangular
breakwaters placed on both a horizontal and a sloping bed. The general effectiveness of
rectangular submerged breakwaters in reducing wave action was also studied in the
laboratory by Johnson et al. (1951). The effectiveness of various parameters of
breakwater geometry and wave climate on wave damping was investigated in a series of
model studies in a wave channel (Table 2.1). They found that for a given relative barrier
dimension (B/Lo = the ratio of breakwater top width to deepwater wave length) and
relative depth (d/h = the ratio of breakwater crest height above the bed to water depth),
the flatter (longer period) waves were less affected by the breakwater with small wave
damping action being observed (for definition of the parameters see Figure 2.1a,b).
Johnson et al. also showed that a wide barrier has a better damping action than narrow
barriers, especially for steeper waves.
11
Goda et al. (1967) conducted a series of laboratory experiments with regular waves on
vertical and composite overtopping breakwaters in a wave channel which was separated
by two parallel walls within a wider wave basin in order to maintain water level the
same offshore and behind the breakwaters (Table 2.1). They found that the transmission
coefficient is mostly governed by the submergence ratio (hs / Hi) whilst the
dimensionless wave characteristics such as Hi/Lo and H i /h did not affect the relation
between submergence ratio and wave transmission coefficient. Goda et al. (1967) also
observed that transmission coefficient decreases with increasing crest width of the
breakwater. The transmitted waves in their experiments were found to compose many
wave trains having varying periods of T, T/2, T/3, … (higher harmonics) travelling with
their own celerities. The following empirical equation was proposed by Goda et al.
(1967) for calculating transmission coefficient “Kt” over mounded vertical wall and
composite breakwaters:
⎡ π ⎛ h ⎞⎤
K t = 0.5 ⎢1 − sin ⎜⎜ β − s ⎟⎟⎥ (2.1)
⎣ 2α ⎝ H i ⎠⎦
in which hs is distance of still water lever from breakwater crest, α and β are variables -
α=2.0 and β has the value of 0.1, 0.3 and 0.5 for high, medium and low mound
breakwaters respectively. Re-analysing the data with the correction of values of incident
and reflected wave height calculation, Goda (1969) proposed revised values of the
parameters α=2.2 and β=0.4 for a vertical wall breakwater. He found the range of β, at
0.1 to 0.35 for composite breakwater depending upon the mound height.
2D laboratory tests in the wave tank were carried out by Dick and Brebner (1968) to
investigate the verification of Dean (1945) theory for wave transmission over thin
barriers. For calculating transmitted wave height, they measured the waves having the
same frequency as the incident wave:
T
2 2
η 2 (t ) dt
T ∫0
Ht = (2.2)
where η is the displacement of water surface behind the barrier. Dick and Brebner
(1968) confirmed Dean’s theory for calculating wave reflection and transmission at a
12
thin barrier. Dick and Brebner (1968) also found that transmission coefficient is lower
for permeable breakwater rather than solid one and “the energy transmitted being
between 36 to 64 percent has been transferred to frequencies higher than the incident
wave”.
Seeling (1980) conducted a series of laboratory tests with a smooth and impermeable
overtopping breakwater using both regular and irregular waves (Table 2.1). He revealed
that submergence ratio (hs /Hi) is the most important parameter controlling transmission
coefficient. He verified the empirical model of Goda (1969) for regular waves with a
best fit of the variables α=2.6 and β=0.15. The method of Goda (1969) was also found
acceptable as an estimation of transmitted wave height distribution for irregular waves.
Changes in spectral shape were found small for irregular conditions whilst as much as
90% of the transmitted wave energy for regular waves appeared at higher order
harmonics.
Abdul Khader and Rai (1980) carried out a series of two dimensional laboratory tests on
smooth and impermeable submerged breakwaters with different shape and geometry
which are provided in Table 2.1. They observed that the dissipation of wave energy is
very dependent on relative crest hight (d/h). Abdul Khader and Rai (1980) found that
both rectangular and trapezoidal breakwaters are more effective in wave dissipation
with relatively high wave steepness rather than low steepness. However, trapezoidal
breakwaters are effective even for low range of d/h due to the shoaling impacts caused
by sloping sides of the breakwater.
Low-crest rock armoured breakwaters were experimentally examined by Allsop (1983)

to study the stability of the structure, number of waves overtopping and transmission
coefficient. The tests were carried out under random wave attack with different climate
(height and period, see Table 2.1). He modified Goda’s empirical model (Goda, 1969)
for wave transmission coefficient by substituting “Kt - 0.5” instead of Kt and introduced
R* instead of hs /Hi to fit the model to measured data:
(K t − 0.5) = 0.5⎡⎢1 − sin π (β − R ∗ )⎤⎥ (2.3)

⎣ 2α ⎦
13
Table 2.1 Test conditions of previous laboratory experimental research on submerged breakwater.
Flume/Basin
Water depth Wave condition Breakwater Breakwater dimension
References dimension
type
h (cm) (m) r/ir Hi (cm) Tp (sec) B (cm) d (cm) tan α
Johnson et al. (1951) - 0.33 R - - plywood 10,22,33 - 0
Goda et al. (1967) 50,35 20x30 3-30 0.8-2.76 wall 40,90 20-70 ∞
Diskin et al. (1970) 18 1.4x0.6x26 R 10.6-16.1 1.22-1.62 rubble 12-14 15,21 -
Seeling (1980) hs=-21 to 42 - Ir-r 0.08-0.177 0.91-3.46 rubble 30,40 33,66,75 var
Abdul Khader & Rai
- 0.9x0.9x30 R 4.7-13.1 - plywood 12-36 0.15-0.97(d/h) 0-1:2
(1980)
Allsop (1983) Var. 1.5x3x42 Ir 5.5-19.4 0.4-1.71 rubble 13.9 16.6-22.2 1:2
Amniti et al. (1983) - 0.8x1.5x20 R 6.25-15.6 1.0,1.5,2.0 var var 6.25,9.4,12.5 1:1-var
Powell & Allsop (1985) hs = -7.9 to 18.6 Ir 9-22.9 1.39-2.3 rubble 14-30 25-66 var
Adams & Choule (1986) Var. 24x37 Ir 3.8-8.2 - rubble 22-48 h-d=3.6-(-3.6) 1:2
Ahrens (1987) 25-30 1.2(0.61)x46x42.7 Ir 2.25-18.2 1.45&3.6 rubble 5.6-9.0 17-35 1:1.5
van der Meer (1988) 40 - Ir 7.5-19.2 1.96-2.6 rubble 30 -12.8 to 9.4 1:2
Gómez and Valdés
hs = 6- 13 - R - 1.5-3.5 rubble 0.8-1.2 -
(1990)
Daemen (1991) hs =-19.6 to 57 1.2x1.0x50 Ir 0.049-0.148 0.99-2.88 rubble 0.34 0.4 1:1.5
Petti & Roul (1992) hs =6 0.8x0.8x50 Ir 8.6-14.1 1.2-1.53 impermeable 24 14 1:3.5,1:1.5
Chiaia et al. (1992) 30 1.2x1.0x45 Ir 1.58-2.63 - rubble 60 25 1:1.5
Davies & Kriebel (1992) hs =5.1,0,-5.1 Basin(0.61x16.61) r-ir 0.9-1.81 - PVC-rubble 15.2 10.2,15.2,20.3
De Later (1996) hs =10 0.6x14x28 R 8,10,12 1.55 rubble 16 30 0.67
Gourlay (1996a) 32,37,42 6(3)x30 R 2.8-20.4 0.9-2.2 mortar 15 32 1:6
Groenewoud et al. (1996) 40 1.0x0.8x32 r-ir 10,13.3,6.7 1.29-2.07 rubble 16 30 0.67
hs =0,5,10,15,20 1.2x1.0x47 5,10,15,20 1.2,1.5,2.0 Rubble 30,250
Seabrook (1997) r-ir - var
hs =3.2,6.3,12.6 1.2x25x30 3.2,6.3,9.5 0.95-1.98 rubble 19,38,95
Rivero et al. (1997) hs =38 5x3x100 r-ir 25,37.5,50 2.5,3,3.5,4 rubble 61 112 1:2
Loveless et al. (1998) 400-650 1.1x1.5x15 50-200 4.5-11.2 20-80 35,50
r-ir rubble 1:2
29.7,32,368 36x23 5,6.6,7.35 0.83,1.1,1.7 0.14,35 21.5,12.5
Drei & Lamberti (1999) 40 0.8x0.8x48 r-ir 4,8,16 0.92-2.5 plywood - 40,41,43.45.5
14
Table 2.1 (Continued) Test conditions of previous laboratory experimental research on submerged breakwater
Flume/Basin Breakwater
Water depth Wave condition Breakwater dimension
References dimension type
h (cm) (m) r/ir Hi (cm) Tp (sec) B (cm) d (cm) tan α
Bleck & Oumeraci (2001) 70 2.0x100 r-ir 8-12-16-20 1.1-6.0 plywood 50,100 40,50,60 0
Vidal et al. (2001) hs =-0.05,0,0.05 0.8x0.6x24 r-ir 5, 10, 15 1.6,2.4,3.2 rubble 25,100 30,35,40 1:2
Schlurmann et al. (2002) 62,72,82 2.0x100 r-ir 8-20 1.0-5.0 plywood 100 52 0
Roul & Faedo (2002) 10,15,20 1.2x1.0x36 Ir 2.4-15.5 1.1-2.33 rubble 20 10,15,20,25,30 1:2
Calabrese et al.
100-170 7.0x5.0x300 - 60-100 3.5,4.5,6.5 rubble 100 130 1:2
(2002,2003)
Gironella et al. (2002) - 1.5x3.0x35 Ir - - rubble 122.5 158.5 1:2
Melito and Melby (2002) 20,50,70 2x3(0.9)x74.6 Ir 3.5-22.4 1.12,1.88,2.62 Core-loc 24.3 40,80 1:1.5
Zanuttigh & Lamberti
hs =0,7,-3 1x12x18 r/ir 4-12.1 0.7-1.97 rubble 20,60 20 0.5
(2003)
Garcia et al. (2004) 30,35,40 0.8x0.6x24.05 r/ir 5,10,15 1.6,2.4,3.2 rubble 25,100 24 0.5
Notes: Wave conditions tested being “r” regular or “ir” irregular waves.
15
where
12
∗ h ⎛ s ⎞
R = s ⋅⎜ ⎟ (2.4)
H s ⎝ 2π ⎠
in which Hs is significant wave height at breakwater location in absence of the structure

and s is the wave steepness (Hi/Lo). Powell and Allsop (1985) provided empirical curves
(as design guidelines) for estimating wave transmission coefficient over low-crested
permeable breakwater with 40% porosity (Figure 2.2).
A three-dimensional laboratory hydraulic model study was conducted by the Offshore

Technology Corporation (Adams and Choule, 1986) to investigate the effectiveness of
the existing submerged breakwater constructed for the protection of Santa Monika pier.
Data analysis of Adams and Choule confirmed that the curves provided by Tanaka
(1976) may be used as design tools for predicting transmitted wave height passing
across a submerged breakwater. However, some care must be used in applying the
curves of Tanaka (1976) which are known to underpredict the transmission coefficient.
Ahren (1987) investigated the performance of low-crested breakwaters by conducting a

wide series of experimental tests (Table 2.1) at the US Army Engineer Waterways
Experiment Station’s Coastal Engineering Research Centre. The aim of the laboratory
tests was to study stability of rouble-mound reef-type breakwaters and their effects on
wave transmission and reflection. He proposed the following expression for wave
transmission coefficient “Kt”:
1.0 hs
Kt = 0.592
for < −1.0 (2.5)
⎛H A ⎞ Hs
1.0 + ⎜ s 2t ⎟
⎜L d ⎟
⎝ p 50 ⎠
1.0 hs
Kt = C2
for > −1.0 (2.6)
⎛d ⎞
C1
⎛ At ⎞ ⎡ ⎛ ⎞ ⎛ A3 2 ⎞⎤ Hs
1.0 + ⎜ ⎟ ⎜ ⎟ exp ⎢C3 ⎜ hs ⎟⎟ + C4 ⎜ 2t ⎟⎥
⎝h⎠ ⎜ hL ⎟ ⎜ ⎜ d 50 L p ⎟⎥
⎝ p ⎠ ⎢⎣ ⎝ H s ⎠ ⎝ ⎠⎦
where the constant C1=1.188, C2=0.261, C3=-0.592 and C4=0.00551. The parameters Hs
indicates the significant wave height, At is area of breakwater cross section, d50 is
16
dimension of stone, Lp is wave length corresponding to peak wave period, d is reef crest
height, h is still water level and hs is the distance between water surface and the crest
reef (positive above the crest reef). Ahren (1987) found that the reflection coefficient is
not much dependent on dimensionless submergence (hs/Hs) and can be calculated by the
following equation:
⎡ ⎛ h ⎞ −1
⎛ ⎞⎤
K R = exp ⎢C1 ⎜ ⎟ + C ⎛⎜ d ⎞⎟ + C3 ⎛⎜ At ⎞⎟ + C4 ⎜ − hs ⎟⎟⎥ (2.7)
⎜ H
⎢⎣ ⎜⎝ L p ⎟ 2
⎠ ⎝h⎠ ⎝d ⎠ ⎝ s ⎠⎥⎦
where C1= - 0.6774, C2= - 0.293, C3= - 0.0860 and C4= 0.0833.
Analysing enhanced experimental data (Table 2.1), Gómez Pina and Valdés (1990)
reported that the wave transmission coefficient has a decreasing oscillatory trend as a
function of relative crest width (B/Lo). They also suggested that for both breaking and
non-breaking wave conditions, considering the parameter “ξ · (B/hs)” gives reasonable
results for predicting transmission coefficient “Kt” (Figure 2.3) where ξ is Iribarren
number that can be calculated by knowing the onshore face slope of breakwater “α” and
incident wave steepness:
tan α
ξ= (2.8)
H i Lo
in which Hi is incident wave height and Lo is incident wave length in deepwater.
The stability and wave transmission of rubble mound low-crested breakwaters (reef-
type) were investigated by van der Meer (1988, 1990 and 1991). He presented a new
formula for wave transmission coefficient using previous laboratory data (Seeling,
1980; Daemrich and Kahle, 1985; Ahrens, 1987; Ahrens, 1989; van der Meer, 1988 and
Daemen, 1991; see Table 2.1 for test conditions). van der Meer (1990) proposed criteria
for the wave transmission coefficient based on relative free board variations (Figure
2.4):
Kt = 0.8 for 1.13 < hs/H i< 2.0
Kt = 0.46 + 0.3 hs/Hi for 0.2 < hs/Hi< 1.13 (2.9)
17
Kt = 0.1 for -2.0 < hs/Hi < -1.2
Daemen (1991) conducted two dimensional tests (as part of Master’s thesis) in a 50 m
length, 1.0 m width and 1.2 m depth wave flume at Delft Hydraulics with irregular
waves (Table 2.1). van der Meer (1991) and van der Meer and Daemen (1994) proposed
an empirical formula for low-crested rubble-mound breakwaters, taking into account the
crest width and height, wave steepness and incident wave height. They found a linear
relationship between the wave transmission coefficient “Kt” and relative crest hight
hs/d50:
K t = a(− hs d 50 ) + b (2.10)
where
a = 0.031(H i d 50 ) − 0.24 (2.11)
and the coefficient “b” for conventional breakwaters is defined by:
b = −5.42(H i Lo ) + 0.0323(H i d 50 ) − 0.0017(B d 50 ) + 0.51

1.84
(2.12)
and for reef type (gradually deforming) breakwaters:
b = −2.6 (H i Lo ) − 0.05 (H i d 50 ) + 0.85 (2.13)
in which hs is water depth over the breakwater d50 is nominal diameter of armour rock,
Hi is significant incident wave height and Lo is wave length of the incident wave height
in deep water. The validity of the above transmission coefficient depends on the range
of various dimensionless parameters used in the tests. The formula is valid for 1 < Hi/d50
< 6 and 0.01 < Hi/Lo < 0.05. The range of variation of transmission coefficient Kt in the
test for conventional breakwaters is 0.075 < Kt < 0.75 and for reef type breakwaters is
0.15 < K t < 0.6.
Davies and Kriebel (1992) carried out a series of experimental tests on solid and
rubblemound breakwater models using both regular and irregular tests (Table 2.1). They
proposed a new parameter of freeboard ratio “(hs+Ru)/Hi” for better description of wave
18
transmission past a reef breakwater at all values of freeboard where Ru is the potential
run-up as defined by Ahrens and McCartney (1975) as a function of Iribarren number ξ:
Ru aξ
= (2.14)
H i 1 + bξ
in which a and b are empirical coefficients which have the values of a=0.775 and
b=0.361, as proposed by Gunbak (1979) for rubble mound breakwaters. Davies and
Kriebel (1992) conducted more than 250 two-dimensional tests in 0.61m wide channel
separated from the wave basin with two plexiglass walls. Two types of breakwater
(rubble-mound and solid) were tested with irregular and regular waves respectively.
They found that transmission coefficient does not significantly differ between the solid
and rubble-mound breakwater in the case of crest height below still water level. The
tests also indicated that the Bulk Number ( B = At d 502 ) does not seem to be a primary
parameter in determining wave transmission for submerged breakwater and trivial
difference in transmission coefficient observed between the rubblemound and solid
(smooth) breakwater, particularly for low steepness waves (Figure 2.5).
Data sets collected from other experimental test (Seeling, 1980; Allsop, 1983; Daemrich
and Kahle, 1985; Powel and Allsop, 1985; van der Meer, 1988; Daemen, 1991) were
reanalysed by d’Angremond et al. (1996) to find a more precise expression for
transmitted wave height passing over permeable and impermeable submerged
breakwaters:
−0.31
h ⎛ B ⎞
K t = 0.4 s + ⎜⎜ ⎟
H i ⎝ H i ⎟⎠
(
× a 1 − e −0.5ξ ) (2.15)
where a=0.64 and 0.8 for permeable and impermeable breakwaters respectively with
limits 0.075< K t < 0.80 for both conditions and ξ is the Iribarren parameter that was
defined in Equation 2.8. Some data with extremely steep or breaking waves (Hi/Li> 0.6
and Hs/ h> 0.54) were discarded in their study.
In total, almost 800 two and three dimensional laboratory tests (Table 2.1) were
performed with irregular waves by Seabrook and Hall (1997, 1998) to investigate
19
transmitted wave height passing over submerged rubble mound breakwaters. They
found that the formula provided by Ahrens (1987) and van der Meer (1991) are not
suitable for calculating Kt particularly when crest width is large. Seabrook and Hall
(1998) observed that the relative submergence and crest width are most important in
determining transmission coefficient. They provided an improved design equation for
transmission coefficient Kt at submerged rubble mound breakwater considering the
effect of crest width:
⎛ −0.56⎛⎜⎜ hs ⎞⎟⎟−1.09⎛⎜ H i ⎞⎟ ⎛B h ⎞ ⎛ h H ⎞⎞
K t = 1 − ⎜ e ⎝ i ⎠ ⎝ ⎠ + 0.047⎜⎜ ⋅ s ⎟⎟ − 0.067⎜⎜ s ⋅ i ⎟⎟ ⎟
H B
(2.16)
⎜ ⎟
⎝ ⎝ Lo d 50 ⎠ ⎝ B d 50 ⎠ ⎠
The following ranges were recommended for applying the proposed equation:
B hs
0≤ ⋅ ≤ 7.08
Lo d 50
(2.17)
h H
0 ≤ s ⋅ i ≤ 2.14
B d 50
Seabrook and Hall (1997) also found that the armour stone size has little effect on wave
transmission (Figure 2.6) and the effect of crest width on transmission coefficient is
evidenced in transfer of wave energy to higher harmonic frequencies as the waves pass
across the breakwater (Figure 2.7).
Large-scale experimental modelling was conducted by Rivero et al. (1997) at Maritime

Engineering Laboratory (LIM), Catalonia University of Technology (UPC), Spain (see
Table 2.1 for test conditions). The experiments include measurement of free surface
elevation and velocity components of regular and irregular waves passing across a
submerged breakwater. The results of measured transmitted wave height revealed that
both formulas provided by van der Meer (1990) and Daemen (1991) give
underestimated values of transmission coefficient Kt. They also observed the presence
of low frequency motions shoreward of the breakwater and these were more significant
for smaller freeboards.
20
Reanalysing a wide range of data sets, Gironella and Sánchez-Arcilla (1999) provided a
model for predicting wave transmission and reflection coefficient in the presence of a
submerged breakwater. These expressions involved the ratio of “hs/Lo” (submergence
over deepwater wavelength) and Iribarren number “ξ ” as expressed in Equation 2.8. For
transmission coefficient, they used multilinear regression to fit an equation to the data:
⎛ h ⎞ h
K t = C1 ⋅ ⎜⎜ ξ ⋅ s ⎟⎟ + C 2 ⋅ s + C3 (2.18)
⎝ Lo ⎠ Lo
where the empirical coefficient of C1, C2 and C3 were determined as 6.43, 14.63 and
0.52, respectively, which leads to a level of correlation of R2=0.98. The proposed
equation is valid in the range of experimental test values, which were:
hs Hi
3.2 ≤ ξ ≤ 5.5 ; 0≤ ≤ 0.04 ; 0.015 ≤ ≤ 0.04
Lo Lo
Bleck and Oumeraci (2001) studied the wave damping and spectrum evolution for a
rectangular submerged breakwater by 2D experimental modelling in the wave flume
(Table 2.1). The tests were run with both regular and irregular waves with constant
water depth, different wave conditions and breakwater dimensions. They realized that
the relative water depth value “hs/Hi” is the best parameter for describing wave
transmission and the following empirical expression was provided with a coefficient of
determination R2=0.9:
⎛ h ⎞
K t = 1.0 − 0.83 ⋅ exp⎜⎜ − 0.72 ⋅ s ⎟⎟ (2.19)
⎝ Hi ⎠
They also proposed the following expression for predicting spectrum evolution of
transmitted waves:
k m−1
T−10,t = ⋅ T−10,i (2.20)
K t2
where Kt is transmission coefficient and minus first moment coefficient “km-1” is

expressed by the relative water depth “hs/Hi”:
21
⎛ h ⎞
k m−1 = 1.0 − 1.17 ⋅ exp⎜⎜ − 0.55 ⋅ s ⎟⎟ (2.21)
⎝ Hi ⎠
and the wave periods T01 and T-10 are defined with:
T01 =
m0
=
∫ S ( f )df (2.22)
m1 ∫ f S ( f )df
∫ f S ( f )df
−1
m
T−10 = −1 = (2.23)
m0 ∫ S ( f )df
The subscripts “t ” and “i ” correspond to transmitted and incident waves, respectively.

Reanalysing data and keeping the structure of the equations, Bleck and Oumeraci
(2002) modified the equations for transmitted wave height and spectrum for rectangular
impermeable artificial reefs as:
H t = [1.0 − 0.92 exp(− 0.83 hs H i )]⋅ H i (2.24)
T01,t = [1.0 − 0.39 exp(− 0.58 hs H i )]⋅ T01,i (2.25)
T−10,t = [1.0 − 0.24 exp(− 0.63 hs H i )]⋅ T−10,i (2.26)
Bleck and Oumeraci (2004) provided a modified breaking index criterion of Miche
(1944) for waves passing over a rectangular artificial reef (Figure 2.8):
Hi ⎛ 2π ⎞
= 0.142 tanh ⎜⎜ M k ⋅ ⋅ hs ⎟⎟ (2.27)
Li ⎝ Li ⎠
where “Mk” is a modified Miche criterion (Miche, 1944) - found to be equal to 0.74
during the tests.
Some experimental tests were performed by Roul and Faedo (2002) in the wave flume
of the IMAGE Department of University of Padova, Italy. Test conditions are presented
in Table 2.1. The aim of the experiments was to better understand the performance of
22
rubble-mound submerged breakwaters in terms of wave transmission, reflection and
overtopping under breaking waves conditions. They applied the dimensionless
freeboard “(hs+Ru)/Hi” as proposed by Davies and Kriebel (1992) and developed an
empirical expression for wave transmission coefficient:
2
⎛ h + Ru ⎞ ⎛ h + Ru ⎞
K t = − 0.0928⎜⎜ s ⎟⎟ + 0.1862⎜⎜ s ⎟⎟ + 0.1176 (2.28)
⎝ Hs ⎠ ⎝ Hs ⎠
where all parameters in the equation have been defined previously.
Large-scale laboratory model tests on rubble mound submerged breakwaters were

carried out at “Grosser WellenKanal” of Hannover, Germany (Calabrese et al., 2002
and 2003; see Table 2.1 for test conditions). The tests were run with irregular and
broken waves with the aim of verifying, at large scale, the existing formulas for wave
transmission and set-up. Calabrese et al. (2002) found that it is more convenient to
make submergence depth “hs”, non dimensional by the crest width “B”, rather than
incident wave height. d’Angremond et al. (1996) equation (Equation 2.15) was found
more reliable than other previous methods for estimating wave transmission coefficient.
They developed a predictive expression for transmission coefficient using data obtained
from the large-scale experimental tests:
hs
Kt = a ⋅ +b (2.29)
B
Calabrese et al. (2002) found an exponential formula for the intercept “b” in the above
equation rather than power function form as proposed by d’Angremond et al. (1996):
B
− 0.0845
b =α ⋅e Hi
(2.30)
where
α = 1 − 0.562 ⋅ e −0.0507ξ (2.31)
23
and the angular coefficient “a” was expressed as a function of relative crest width
“B/Hi”:
B
0.2568
a = β ⋅e Hi
(2.32)
in which β is a scale parameter where water depth has been included:
Hi
β = 0.6957 ⋅ − 0.7021 (2.33)
h
Calabrese et al. (2002) formula was calibrated in the test range of:
− 0.3 ≤ hs B ≤ 0.4
1.06 ≤ B Hi ≤ 8.13
0.31 ≤ H i h ≤ 0.61
3.0 ≤ ξ ≤ 5.20
Melito and Melby (2002) conducted 2D experimental wave flume tests (Table 2.1) to
investigate the wave run-up and transmission coefficient of a single CORE-LOC armour
layer low-crested breakwater. The experiment was restricted to a unique breakwater
crest width and the influence of crest width was not taken into account, although this
parameter has substantial effect on wave transmission. They proposed the following
curves for estimating transmission coefficient of CORE-LOC armour layer submerged
breakwater (Figure 2.9):
Kt = 0.95 for 1.0 < hs/Hi
Kt = 0.56 + 0.39(hs/Hi) for -1.30 < hs / Hi < 1.0 (2.34)
Kt = 0.05 for hs / Hi < -1.3
van der Meer et al. (2004) collected data of more than 2300 tests (from numerous
researches) in developing a 2D empirical wave transmission coefficient formula for
rubble mound low-crested breakwaters. They improved the formula provided by van der
Meer (1991) and d’Angremond et al. (1996) considering the value of crest width ratio
24
“B/Hi”. van der Meer et al. (2004) found that for rubble mound structures with B/Hi < 10
the equation provided by d’Angremond et al. (1996) is still applicable and for B/Hi > 10,
a simple modification of Equation 2.15 (d’Angremond et al., 1996) provides more
reasonable results of transmission coefficient. They proposed the following modified
equation for low crested rubblemound breakwater when B/Hi > 10:
−0.65
⎛ B ⎞
h
K t = 0.35 s + 0.51⎜⎜
Hi
⎟⎟ (
⋅ 1 − e −0.41ξ ) (2.35)
⎝ Hi ⎠
van der Meer et al. (2004) also found that considering the maximum of Kt regardless the
ratio of B/Hi (as proposed by d’Angremond et al., 1996) may give inaccurate results and
proposed an upper limit of Kt as a linear function of B/Hi:
K tu = −0.006 B H i + 0.93 (2.36)
The lower limit of transmission coefficient Kt was kept constant as proposed by

d’Angremond et al. (1996).
For smooth and impermeable structures, van der Meer et al. (2004) provided the
following equation to be used for transmission at smooth low-crested breakwaters with
minimum and maximum transmission coefficient Kt, 0.075 and 0.8 respectively:
K t = 0.3
hs
Hi
(
+ 0.75 1 − e −0.5ξ ) for ξ<3 (2.37)
−0.31
h ⎛ B ⎞
K t = 0.3 s + ⎜⎜ ⎟
H i ⎝ H i ⎟⎠
(
∗ 0.75 1 − e −0.5ξ ) for 3 ≥ ξ (2.38)
They also investigated the influence of wave angle “Θ ” on transmission coefficient.

van der Meer et al. (2004) indicated that transmitted wave height in case of rubble
mound structures is not sensitive to incident wave angle, whereas it is influenced by
wave angle in smooth breakwaters where the equations may be improved by a cosine
function. For example:
25
⎡ ⎤
h
H
( )
K t = ⎢0.3 s + 0.75 1 − e −0.5ξ ⎥ (cos Θ )
23
for ξ < 3 (2.39)
⎣ i ⎦
with the following limitations:
1<ξ<3 0o ≤ Θ ≤ 70o 1 < B/Hi < 4
2.1.2 Analytical/Numerical Modelling of Wave Transmission
A numerical model was presented by Kobayashi and Wurjanto (1988) to predict

monochromatic wave reflection and transmission over an impermeable submerged
breakwater. They modified the proposed model of Kobayashi et al. (1987) based on the
finite-amplitude shallow water equations for predicting wave reflection and run-up on
rough or smooth impermeable slopes. The model provided moderately good prediction
of transmission coefficient over submerged impermeable breakwaters in comparison
with the limited data measured in laboratory by the authors.
Rojanakamthorn et al. (1990) developed a mathematical model for the computation of

wave transformation over a permeable submerged breakwater. They applied the
equation of waves on a porous layer as a two-dimensional elliptic equation analogous to
the mild slope equation. Comparison of the model with the 2D experiments conducted
by Rojanakamthorn et al. (1990) showed generally good agreement between the
numerical model and experiment. However, the calculated transmitted wave height was
slightly smaller than the measured one.
Based on the VOF (Volume Of Fluid) method, a two-equation k-ε turbulence model was
applied by Shen et al. (2004) to simulate the propagation of cnoidal waves over a
submerged bar (breakwater). The VOF method has the capability to simulate free-
surface flow. The calculated water surface elevation around a breakwater was compared
with the 2D experimental data reported by Ohyama et al. (1995). The results showed
relatively good agreement between calculated and measured water surface however,
some discrepancy was observed due to higher harmonics generation over the breakwater
in the laboratory test. Shen et al. (2004) indicated that a more refined turbulence model
might be needed to address the differences.
26
Garcia et al. (2004) applied a numerical model named COrnell BReaking waves And
Structures (COBRAS) to calculate water surface elevation and flow in the presence of
permeable low-crested breakwaters for regular breaking waves. The COBRAS model
solves the 2DV Reynolds Averaged Navier-Stokes (RANS) equation that was firstly
provided by Lin and Liu (1998). The model is based on the composition of the
instantaneous velocity and pressure fields into mean and turbulent components”, Garcia
et al. (2004). The results of wave height envelope and water surface around the
breakwater, the pressure field inside the rubble and the velocities on the seaward slope
were compared with data enhanced from 2D experimental tests carried out by Vidal et
al. (2001). The comparison showed that the model reproduces the measured quantities
with good agreement. The model was also proven to be a powerful tool in examining
the near-field flow characteristics around submerged breakwaters. However, the
computed values of flow were not compared with the experimental data in the paper.
Lara et al. (2006) extended the application of the model provided by Garcia et al. (2004)
for random wave interaction with a submerged permeable breakwater. They reported
that the model gives good results of wave height envelope, mean water level, spectral
shape and pressure inside the breakwater in comparison with the data measured by
Vidal et al. (2001).
Wave modelling in the presence of submerged breakwaters was carried out by Johnson
(2006) using MIKE 21 PMS which is a refraction/diffraction model based on the
parabolic approximation to the mild slope equation. He found that applying the depth-
limited breaking dissipation model of Battjes and Janssen (1978) in MIKE 21 PMS
reproduces higher energy dissipation than experimental data (Zanuttigh and Lamberti,
2003) over a submerged breakwater. The breaker parameter in the Battjes and Janssen’s
dissipation model was used as calibration factor by Johnson (2006) and a simple
relationship between breaking parameter and submergence ratio was provided. The
transmission coefficient obtained from the calibrated model was compared satisfactorily
with the laboratory measured data.
2.2 Set-up and Piling-up
Breaking waves passing over a submerged breakwater causes water to flow into the
protected area (onshore) of the breakwater. On the other hand, the difference in mean
27
water level inside and outside of the protected zone results in water flowing out of this
area. Piling-up will occur when a quasi-equilibrium between inflow and out-flow is
reached. Few experimental, theoretical or numerical studies have been undertaken that
provide a good understanding of set-up and the behaviour of longshore current behind
Longuet-Higgins (1967) derived a simple analytical formula for calculating mean sea
level difference between the two sides of a submerged breakwater “δ” exposed to non-
breaking waves:
δ=
(H + H r2 ) ⋅ k i
i
2
−
H t2 ⋅ k t
(2.40)
8 sinh 2k i h1 8 sinh 2k t h2
where “H” is wave height, “k” is wave number and the subscripts “i”, “r” and “t” stand
for incident, reflected and transmitted waves, respectively. “h1” and “h2” are water
depth corresponding to sea-side and shore-side of the breakwater. Measurements of 2D
laboratory measurements by Dick (1968) showed that the equation provided by
Longuet-Higgins (1967) greatly underestimates the mean sea level difference “δ ” .
Diskin et al. (1970) conducted a series of 2D laboratory tests in the wave flume with
low-crested and submerged rubble mound breakwaters exposed to regular waves (Table
2.1). They tried to measure water surface elevation by using piezometers across the
bottom of the flume (Figure 2.10). Diskin et al. fitted a curve to points of relative set-up
ratio (δ/Hi) versus submergence ratio (hs /Hi) and proposed the following simple
equation for calculating piling-up behind low-crested breakwaters:
δ ⎛ ⎛ hs ⎞ ⎞⎟
2
⎜
= 0.6 exp − ⎜⎜ 0.7 + ⎟ (2.41)
Hi ⎜ ⎝ H i ⎟⎠ ⎟
⎝ ⎠
The equation is valid in the range of the tested values of submergence ratio (-2.0 < hs /Hi
< 1.5). It should be noted the test conditions were such that the piled-up water behind
the breakwater could return offshore only by passing over and through the structure. In
this case no longshore current can develop behind the breakwater.
28
An extensive series of 2D and 3D experimental tests were carried out by Debski and
Loveless (1997) and Loveless et al. (1998) at the University of Bristol, UK. Details of
test conditions are provided in Table 2.1. They measured wave-induced set-up, velocity,
current and transmitted wave height in both wave flume and wave basin for permeable
and impermeable breakwaters. Using a pump to maintain water levels the same behind
and in front of the breakwater in the 2D flume tests, Debski and Loveless (1997) could
measure the quantity of flow necessary to suppress wave set-up. Velocity measurements
were made using an ADV (Acoustic Doppler Velocimeter) both inside and around the
tested breakwater. Wave-induced set-up was found to be significantly smaller in 3D
tests as the water level gradient behind the breakwater generates longshore flow which
may result in beach erosion. Beach response to submerged breakwaters was investigated
in the wave basin (3D) specifically for the Elmer Beach project. Loveless et al. (1997)
found that set-up increases with breakwater crest width (Figure 2.11). They provided
the following empirical equation from the data collected in their 2D laboratory
experiments to predict the set-up behind low-crested breakwaters:
δ = 0.125
H i Lo
T h
(
⋅ ⋅ exp − 20(hs d )
2
) (2.42)
where Hi is incident wave height, T is wave period, Lo is wave length in deepwater, h is

water depth at the offshore toe of the breakwater and d is the breakwater crest height.
The experiments also demonstrated that set-up is significantly influenced by the

porosity (influenced by the rock size d50) and the crest width “B ” of the breakwater.
Loveless et al. (1998) improved the original Equation 2.42 by linear regression analysis
and provided the following expression for calculating set-up behind a permeable
submerged breakwater:
δ (H i Lo hT )
2
= ⋅ exp[20(hs d )] (2.43)
B 8 gd 50
Calabrese et al. (2003) developed a model based on a simple hypothesis to calculate

wave-induced set-up behind submerged breakwaters. They assumed that wave set-up is
29
dominated by the amount of momentum released by the breaking waves on the
breakwater “δm” plus a further contribution called “continuity set-up”, δs:
δ = δm + δs (2.44)
Applying the momentum equation and assuming that the surf zone extends from the
breaker point on the sea-face slope to the inshore toe of the breakwater (linear
increasing of set-up), Calabrese et al. (2003) provided the following equation for the
momentum set-up “δm”:
δ m = 0.5[− b + (b 2 + 4c )]
0.5
(2.45)
where
b = (2h − A) (2.46)
⎡⎛ x + B ⎞ h + hs ⎤
A = ⎢⎜⎜1 + b ⎟⎟ ⋅ d − b ⋅ xb ⎥ (2.47)
⎣⎝ Ls ⎠ Ls ⎦
c=
3
8
(
⋅ H i2 ⋅ 1 − K t2 ) (2.48)
in which “Kt” is wave transmission coefficient, hb is wave breaker depth over the
offshore slope, hs is submergence depth, d is breakwater crest height and Ls is the
effective crest width of the breakwater (Figure 2.12).
The continuity set-up “δs” may be determined by assuming uniform return flow:
q2
δs = 10 3
⋅ Beq (2.49)
f 2 ⋅ Rc
where
1 g
q= ⋅ H i2 ⋅ (2.50)
8 h
30
Beq = B + d ⋅ cot g α (2.51)
Calabrese et al. (2003) had to use an unrealistic friction factor “ f = 20”, as calibration
coefficient, to obtain good agreement between the model and experimental data from
large scale 2D wave flume laboratory tests (Calebrese et al., 2002 and 2003; see Table
2.1) .
2.3 Wave-induced Current
Drei and Lamberti (1999) examined behaviour for narrow-crested impermeable

submerged breakwaters in wave flume experiments (Table 2.1). They measured wave
transmission and reflection, wave-induced set-up and overtopping discharge under
regular and irregular wave attack. The method of set-up measurement behind the
structure involved averaging of data collected from 8 wave probes located close to the
onshore beach after 80 seconds of wave attack. Drei and Lamberti (1999) found that
wave set-up at submerged breakwaters is higher than provided by set-up theory for
beaches because the breaker index at the submerged breakwater is greater. Drei and
Lamberti found the average non-dimensional set-up “δ/Hi” to be about 0.49 for all
tested submergence depths whilst for only zero submergence it was found to be 0.8 in
contrast to the value 0.3 found by Loveless et al. (1998). The reason for the high
discrepancy was ascribed to there being no return flow through the impermeable
breakwater used in the Drei and Lamberti’s tests. Pumping water from behind the
breakwater to offshore and maintaining water level, Drei and Lamberti (1999) provided
more reliable data on the free overtopping discharge over semi-submerged breakwater,
it being evaluated by the following proposed simple equation:
q = 0.4 H i1.5 ⋅ g 0.5 (2.52)
Roul et al. (2004) studied the piling up discharge of overtopping rubble mound
breakwaters and controlled set-up behind the breakwater by pumping piled water from
onshore to offshore side in the wave flume (see Table 2.1 for test conditions). They
found that set-up decreases linearly with increasing net transmitted discharge (Figure
2.13) and they thence proposed that the two points of maximum set-up with no pumping
and discharge with zero set-up may be connected with a straight line. The other values
31
of set-up and discharge can be related to other parameters such as breakwater length and
gap width.
2.4 Wave-induced Set-up and Current over the Reef
A number of two and three dimensional laboratory experiments were conducted

(Gourlay, 1993, 1996a, 1996b) at The University of Queensland, Australia, with the aim
of understanding better the behaviour of some natural reefs in Australia. Wave
transformation, wave induced set-up and current over the reef were measured by wave
probes, piezometer and propeller miniflowmeter, respectively (for more details see
Gourlay, 1996a and Table 2.1). Two dimensional tests showed that both set-up and
current increases with incident wave height and for constant wave conditions, set-up
increases with decreasing submergence depth while current increases with submergence
depth. The results also revealed that set-up is sensitive to wave period with set-up
increasing with wave period up to a certain value of set-up and then becoming constant.
Gourlay (1993) proposed two dimensionless representations for set-up and current:
η
N= (2.53)
T gH i
q
q* = (2.54)
gH i3
where η demonstrates the maximum set-up over the reef (~ δ ).
Plotting dimensionless discharge “q* ” and dimensionless set-up “N ” against

submergence ratio (S=(hs+δ)/Hi) showed that for small submergence ratio (S <0.7) the
relative set-up decreases rapidly as S increases and beyond that it decreases at a slower
rate (Figure 2.14). Gourlay (1993) indicated that this criteria for submergence ratio is
related to wave breaking process since for low submergence (S < 0.7) waves break at the
reef edge, though it breaks over the reef for high submergence (S > 0.7). Gourlay
(1996a) observed that set-up occurs over the reef when incident wave height “Hi” is
approximately greater than 0.4(h s + δ ) which is the threshold value of wave height over
the reef (Hm) with gently sloping reef-top (Gourlay, 1994).
32
Symonds (1994) and Symonds et al. (1995) solved the momentum and continuity
equations analytically, assuming linear friction, and derived dimensionless equations for
calculating wave-induced set-up and current velocity in three zones in the presence of
reefs: offshore the surf zone, surf zone (over the slope) and the reef top. They assumed
set-up is zero at the onshore side of the reef and offshore of the surf zone. The following
equation was provided for calculating current velocity over the reef top:
3
U = − gγ r2 tan α
(1 − R1 )(1 − R2 ) × hs (2.55)
2 R1 (1 − R2 ) + R2 (1 − R1 ) r
where
hs
R1 = (2.56)
hb
and
B
R2 = (2.57)
B + (hb − hs ) tan α
in which U is integrated current velocity over the reef, γr is wave deformation index
over the reef (γr =Hr/hs), Hr is the maximum wave height over the reef, hs is water depth
over the reef, hb is wave breaking depth over the slope, B is reef length, β is reef face
slope and r is the friction term in the momentum equation and determined by (r = cf .uw)
where uw is the wave orbital velocity and may be calculated from linear wave theory.
Friction factor cf is of order 0.01 for sandy bottom (Longuet-Higgins, 1970) and 0.1 to
0.2 for coral reefs (Nelson 1996). In applying the model to John Brewer reef,
Townsville, Australia, Symonds et al. (1995) adopted wave deformation index γr =0.35
(based on site investigations of Hardy, 1993), an adjusted reef width B=300 m and
friction coefficient r= 0.25 to obtain the best fit between the model and observed
currents from the field. It was noted by Symonds et al. that these calibration factors are
not unique and other values of B and r may be found to fit the model to the observed
data.
33
Hearn (1999) developed an analytical hydrodynamic model using the depth-averaged
steady state momentum equation and both linear and quadratic friction theory to
calculate wave-driven flow across a coral reef. He introduced a current depth coefficient
to describe the current changes against other relevant parameters, e.g. water depth.
Hearn compared the calculated current depth coefficient with data obtained from two
natural reefs in Australia and USA. He found that the coefficients are sensitive to the
form of friction law on the reef top and good agreement was obtained with a linear
friction law.
Massel and Brinkman (2001) provided an analytical approach to model wave-induced

set-up and current over shoals and corral reefs. They assumed that the sea face of the
reef is not steep (1:80 to 1:30) and long enough for the development of wave breaking.
Wave height over the reef/shoal was considered constant as a fraction of water depth
(submergence). The model indicated how the geometry of the reef affects wave-induced
current and set-up. Nevertheless, the model has not been verified with experimental or
field data.
For reefs with low submergence (S < 0.7) Gourlay and Colleter (2005) applied the
energy balance equation and found the equation of discharge over the reef the same as
flow over broad crested weirs:
q CS 2
q∗ = = (2.58)
gH i3 (S + C F )
2
where C = (2 3) C D and F = f .B 8 H i .
1 .5
in which q* is dimensionless discharge or flux ratio, CD is discharge coefficient of

broad-crested weir, f is friction factor of flow in open channel, B is the width of the reef,
and Hi is incident wave height. Gourlay and Colleter (2005) also used energy equation
between the point of maximum set-up over the reef and onshore reef edge for higher
submergence (0.7< S ), a condition in which flow over the reef is subcritical, the flow is
open channel and flux may be determined as follows:
34
q S3
q* = =M PT (2.59)
gH i3 ⎛ 32 3 *⎞
⎜ S + KP ⋅ q ⎟
⎝ 2 ⎠
where
⎡ S2 q ∗3 ⎛ Hi ⎞
12
⎤
PT = ⎢1 − K R2 − 4πγ r2 − 16π 2 ⎜⎜ 2
⎟⎟ ⎥ (2.60)
⎢⎣ D S ⎝ gT ⎠ ⎥⎦
and
⎞⎛ 8 K P ⎞
⎛ 3
M =⎜ ⎟⎜⎜ (
⎟⎟ gH i .T 2 )
12
(2.61)
⎝ 64π ⎠⎝ f . B ⎠
in which D is dimensionless reef-top wave length (g1/2T/hs1/2), Kp is reef profile shape

factor (see Gourlay, 1996b) and γr (=Hr/hs) is wave deformation index over the reef as
already discussed. Wave reflection coefficient is ignored due to small value for barriers
with high submergence. They adopted the coefficients γr =0.4, KP=0.8, D=12.5 and
f=0.025 and compared the calculated values of flux ratio “qT* ” with the measured
laboratory experimental “qE* ” data from Gourlay (1996a), (Figure 2.15).
Johnson et al. (2005) investigated numerically the wave evolution and current around
submerged breakwaters. They adopted two approaches, phase-averaged and phase-
resolving, to simulate wave, mean water level changes and current in vicinity of
submerged breakwaters. Two phase-averaged models were applied in their numerical
simulation, a wave model (LIMWAVE) and a Q3D nearshore circulation model
(LIMCIR) which have been developed at Universitat Politecnica de Catalunya, Spain
and MIKE 21 PMS for wave and MIKE 21 HD for flow modeling (for more details see
Johnson et al., 2005). The phase-resolving approach was based on a 2DH Bousinesq-
type model for wave evolution over permeable submerged breakwaters, developed at
Aristotle Univeresity of Thessaloniki, Greece. The results of all models were compared
with 3D wave basin experimental tests (Lamberti et al., 2003; see Table 2.1). The
calculated transmitted wave height using both phase-average and phase resolving
models was found to be in good agreement with experimental results if the wave
35
breaking sub-model was tuned for wave dissipation over the submerged breakwater.
Johnson et al. (2005) indicated that the phase-averaged models predicted the flow
circulation pattern around the breakwaters that was qualitatively in good agreement with
the laboratory measurements. However, the predicted values of current speeds and wave
set-up inshore of the breakwater were found to be higher than the laboratory test results.
2.5 Shoreline Response to Submerged Breakwaters
Aminti et al. (1983) conducted a series of tests in a wave flume with both fixed and
moveable bed to investigate the hydrodynamic and morphodynamic effects of
breakwater parameters (Table 2.1). Different specifications of breakwater were applied
in the experiments aimed at a better understanding of submerged breakwaters as a shore
protection structure. The changes of equilibrium beach profile for any experimental
barrier tested were found small. However, the presence of the barrier seemed to be
effective in reducing oscillation of the beach under various wave attacks.
Two dimensional experimental tests (Table 2.1) were carried out by Chiaia et al. (1992)
to study the morphological efficiency of submerged breakwaters. They also observed
that submerged breakwaters could not reduce the erosion of the shoreline and that the
beach profile reached almost the same position both with and without the breakwater.
The main deficiency of the morphological experimental modelling in 2D wave flume is
that the effects of long shore current and circulation on shoreline changes has not been
taken into account behind the submerged breakwater.
Browder and Dean (1996) monitored the hydrodynamic and morphodynamic effects of
a 594m shoreline parallel precast concrete submerged breakwater in Palm Beach,
Florida, USA. The results of 3-year monitoring revealed that the erosion of the beach
behind the structure was 2.3 times higher than before breakwater construction. The
reason was found to be increasing longshore currents due to ponding of water trapped
behind the structure.
Tomasicchio (1996) compared the results of a field study and laboratory tests for two
different defense systems at Osita, Roma, Italy, where the maximum spring-tide is about
30 cm. The first protective system was the combination of a submerged barrier (reef)
36
and beach nourishment for 3 km stretch. The second protective system was a 1 km long
rubble mound submerged breakwater. He indicated that although both systems have a
low environmental impact, well designed rubble mound submerged breakwaters are
able to attenuate incident wave energy and reduce sediment transport to protect
shoreline with lower costs.
González et al. (1999) proposed an equilibrium beach profile model for perched
beaches. They modified Dean’s (1977) model by considering wave reflection offshore
and onshore of the submerged breakwater as an important process in beach profile
change. They ignored the effect of the wave breaking over the submerged breakwater as
in most applications they considered the crest width is much smaller than the incident
wave length. The model showed that the effectiveness of the protected beach shoreline
depends on the dimensionless submergence (hs/h) so that the minor and maximum
efficiency is achieved for hs/h > 0.5 and for hs/h < 0.1, respectively. A good agreement
was observed between the model and 2D experimental model data of Chatham (1972)
and Sorensen and Beil (1988).
An equilibrium beach profile model for reef-protected beaches has also been provided
by Muñóz-Pérez et al. (1999). Contrary to the González et al. (1999) model, they
assumed that wave energy dissipates uniformly along the submerged reef and waves
break on the shelf. The model was proved against field measured data from reef-
protected beaches in Spain and Australia. Muñóz-Pérez et al. (1999) indicated that no
equilibrium beach profile is possible within a distance of 10 to 30 hs from the reef edge,
where hs is the submergence over the reef.
Hanson and Kraus (1990) applied the 1-line shoreline change GENESIS numerical
model to predict shoreline response to transmissive offshore obstacles. Hanson and
Kraus (1990) investigated the influence of varying wave climates and transmissivity on
shoreline change. They found that shoreline advances with decreasing wave height and
increasing wave period. They also observed that a salient extends seaward with
decreasing transmissivity. Comparing the numerical model and the available field data
(Hanson and Kraus, 1990), the authors provided criteria to discriminate the formation of
either a salient or tombolo. The numerical model GENESIS is not able to calculate
37
piling up and the amount of water passing over the submerged breakwater. Therefore,
the effect of longshore current behind the structure on shoreline has been ignored.
Groenewoud et al. (1996) studied the results of experimental testing in a wave flume
and a wave basin (Table 2.1) to better understand beach response to submerged
breakwaters. In the case of a submerged breakwater with gaps between segmented
structures, 2D experimental testing cannot give reasonable results since the breakwater
sections create perched beaches. 3D tests showed that a significant amount of sand was
removed from behind the structures and transported offshore through the gaps.
Groenewoud et al. (1996) found that the “exposure ratio” (the ratio of gap width to the
sum of the breakwater length and gap) has a significant influence on shoreline response
to submerged breakwaters.
van der Biezen et al. (1997) analysed the results of extensive laboratory experimental
tests carried out in six European Universities (laboratories) which focused on
hydrodynamic and morphodynamic effects of submerged breakwaters. 2D vertical tests
showed that after a long time the retreat of the shoreline in the presence of a submerged
breakwater is smaller than that observed without a breakwater with the offshore
sediment transport blocked by the breakwater. However, there was significant sediment
loss through gaps and segmented breakwaters were found to have a negative effect on
the beach profile development. van der Biezen et al. (1998) also undertook numerical
Delft3D modelling which indicated significantly larger erosion near the water line and
the shoreward toe of the breakwater than the laboratory measured profile (De Later,
1996). In the numerical modelling, van der Biezen et al. (1998) applied non-linear
shallow water equations with unrealistic values of bottom friction to find better
agreement between the model and the laboratory tests. Comparison of measured and
calculated current behind the breakwaters revealed that there was a lower current in the
numerical model since mass transport of water over the submerged breakwater due to
wave breaking was not taken into account in the model.
The morphodynamic effects of varying a number of design geometric parameters of

submerged breakwaters were investigated numerically by Lesser et al. (2003). They
employed a modified version of Delft3D with some changes to the wave module
(SWAN) for wave transmission over the submerged breakwater. The model results
38
revealed that small changes in design parameters (e.g. gaps) can have a substantial
effect on performance of submerged breakwaters. Moreover, inappropriate design of
submerged breakwaters may lead to considerable erosion around the structure and at the
shoreline. Lesser et al. (2003) calibrated the hydrodynamic numerical model by
adjustment of the roughness coefficient at the top of the submerged breakwater.
Ranasinghe et al. (2006) showed that the effect of submerged breakwaters (artificial
reefs) on erosion or accretion of shoreline depends on the distance of the structure
offshore the beach. They conducted a series of 3D wave basin laboratory tests
associated with 2DH numerical modelling using MIKE 21 CAMS to simulate
hydrodynamic and morphodynamic effects of submerged structures. They also found
that the predominant incident wave angle and the submergence depth of the breakwater
significantly affect the magnitude of shoreline response - these parameters cannot,
however, govern the mode of shoreline evolution (accretion or erosion). An empirical
relationship between shoreline change ratio (maximum shoreline change to the
breakwater length “Y/B”) and dimensionless offshore breakwater distance (breakwater
offshore distance to natural surfzone width “Sa/SZW ”) was proposed by Ranasinghe et
al. (2006) as a preliminary engineering tool to predict shoreline response to submerged
structures (Figure 2.16).
Turner (2006) proposed an easily-applied technique to assess shoreline evolution in

presence of nearshore breakwaters and reefs. He used odd-even function analysis to
discriminate the principle modes of shoreline changes (erosion or accretion) in the
vicinity of detached structures. The applicability of the proposed odd-even function
analysis was successfully verified by comparison of the predicted shoreline with the
field observation from image processing of shoreline changes immediately after
construction of a submerged reef structure at the northern Gold Coast, Queensland,
Australia (Turner, 2006).
2.6 Breaking Wave Roller
Svendsen (1984a, b) first considered the effect of the breaking wave roller on radiation
stress in the surf zone. He assumed that the roller propagates shoreward with the same
velocity as wave celerity. Adopting the roller area (A=0.9H2, where H is wave height)
39
from Duncan (1981), Svendsen proposed the following equation for radiation stress in
the surf zone including wave roller:
⎛3 h⎞
S xx = ρ gH 2 ⎜ Bo + 0.9 ⎟ (2.62)
⎝2 L⎠
where
2
1 T⎛ η ⎞
T ∫0 ⎝ H ⎠
Bo = ⎜ ⎟ dt (2.63)
in which T is wave period, H is wave height and η is the vertical distance of wave crest
from mean water surface (MWS).
Dally and Brown (1995) developed a new numerical model for calculating the breaking
wave roller area. They used depth-integrated, time averaged energy balance which
contains contributions for both wave motion and the roller. The frictional dissipation of
the roller surface was considered as the rate of energy dissipation in the energy
equation. Numerical solution of the governing equation for the creation and evolution of
the roller area revealed that the dissipation coefficient given by the angle of inclination
of the roller operates as the main parameter for calibration of the model. The proposed
roller model was applied by Dally and Brown (1995) to calculate wave-induced
longshore current over plane beaches.
Reniers and Battjes (1997) conducted a series of 3D experimental tests in a large wave
basin with regular and irregular (random) waves to measure wave transformation, wave-
induced set-up and the cross-shore distribution of longshore current velocity on a
sloping bed. They also provided a numerical model for random wave transformation,
set-up and longshore current velocity with the concept of surface roller model being
included in the equations. Comparison of the numerical model and experimental tests
indicated that the model underestimated set-up and longshore current without the roller
concept - including the effect of the surface roller in the equations showed improved
and reasonable results both for set-up and longshore current (Figure 2.17 and 2.18).
Lesser et al. (2003) also indicated that wave roller effects should be included in the
numerical model to make it more reliable in predicting hydrodynamic and
40
morphological effects of submerged breakwaters. Without the wave roller effects,
unrealistic friction parameters need to be adopted as calibration factor in the model for
fitting to experimental or field data.
2.7 Summary
The functional design knowledge of submerged breakwaters including their impacts on

wave transmission, currents, sediment processes and shoreline response is still
developing.
Many numerical modelling and empirical approaches for estimating wave transmission
over submerged breakwaters have been developed (Johnson et al., 1951; Adams and
Sonu, 1986; Losada et al., 1996; d’Angremond et al., 1996; Seabrook and Hall, 1998;
Schlurmann et al., 2002; van der Meer et al. 2005). Most of the previous experimental
research on wave transmission over submerged breakwaters has been carried out as a
subset of investigations into relatively narrow crested semi submerged breakwaters. The
results of these previous experiments are limited to the tested ranges of breakwater crest
width. As indicated in Chapter 1, narrow crested (fully) submerged breakwaters have
been found to be ineffective in coastal protection in most field applications. Therefore
more investigations are needed to extend the studies of wave transmission over
submerged breakwaters with wide to broad crest width due to their improved efficiency
in protection of coastal area.
The gradient of wave-induced set-up behind submerged breakwaters causes water to

flow along the shoreline. Only a few experimental, theoretical or numerical studies have
been undertaken that provide an understanding of set-up and the behaviour of longshore
current behind submerged breakwaters (Longuet-Higgins, 1967; Diskin et al. ,1970;
Debski and Loveless ,1997; Calabrese et al. ,2003).
However, wave induced set-up and current are dependent on mass flux over the
breakwater which is contributed by wave averaged flow generated by radiation stress
gradients, leakage of water through the porous structure of the breakwater and mass
transport by breaking wave rollers. Drei and Lamberti (1999) examined flow behaviour
for narrow-crested impermeable submerged breakwaters in wave flume experiments.
41
Pumping water from behind the breakwater to offshore and maintaining water level,
Drei and Lamberti (1999) provided reliable data on the free overtopping discharge over
semi-submerged and narrow-crested breakwater. Gourlay (1993, 1996a, 1996b)
conducted several experimental tests for measuring wave set-up and mass flux over
wide crested reefs. Based on the laboratory experiments of Gourlay (1996a, 1996b),
Gourlay and Colleter (2005) provided an analytical approach to calculate wave-induced
set-up and current over wide reefs. The model can predict wave-induced flux passing
over wide reefs (B/Lo > 0.21) as a function of incident wave climate, breaking criteria
and reef geometry and roughness. Symonds et al. (1995) provided a linear theoretical
model demonstrating how the relative magnitudes of the currents and set-up depend on
the geometry of the reef and the magnitude of the forces including radiation stress. The
Symonds et al. (1995) model verification is limited to some specific field measurements
(John Brewer Reef) - the model has not been verified for the wider range of submerged
breakwater geometries, water levels and wave conditions tested in laboratory
experiments within this thesis.
A few numerical modelling studies have been carried out to simulate wave-induced
current behind submerged breakwaters. (van der Biezen et al. ,1997; Johnson et al.
,2005 ;Zanuttigh and Lamberti,2006;). Zanuttigh and Lamberti (2006) found
discrepancies among experimental and numerical results to be related to the wave
breaking model adopted in the numerical model. Therefore, a modified/calibrated wave
breaking model needs to be developed to calculate wave energy dissipation passing over
submerged breakwater. Moreover, most of the existing numerical models apply the
radiation stress formula provided by Longuet-Higgins and Stewart (1964). In this
formula wave set-up depends on incident wave energy, wave direction, water depth and
wave number in shallow water. However, mass transport due to breaking wave rollers
has not been applied in any hydrodynamic models for submerged breakwaters.
Finally, the reliability of any numerical morphological model for predicting seabed and
shoreline response in the vicinity of submerged breakwaters depends on the initial
prediction of hydrodynamic effects (wave transformation, breaking and mass flux) at
the submerged breakwater. The hydrodynamic effects are the focus of this thesis.
42
Hi
hs
( submerged)
-hs
(semi-submerged)
d
d50
Figure 2.1. Definition of submerged (semi-submerged) breakwater parameters and wave

transmission (after Wamsely et al, 2002).
-hs/Hmo
Figure 2.2. Suggested curves for estimating wave transmission coefficient over low-
crested breakwaters (from Daemen, 1991).
ξ
(B/hs)*ξ
Figure 2.3. Influence of breaking conditions on wave transmission coefficient (Gómez

Pina and Valdés, 1990).
-hs/Hi
Figure 2.4. Proposed wave transmission coefficient criteria by van der Meer (1990) as a
function of submergence ratio (Daemen, 1991).
-(hs+Ru)/Hi
Figure 2.5. Transmission coefficient of regular waves past over solid and rubblemound
breakwaters applying new freeboard parameter “(hs+Ru)/Hi” (Davies and Kriebel, 1992).
Figure 2.6. The effect of armour rock size on transmission coefficient (Seabrook, 1997).
Figure 2.7. Variation of transmission coefficient against breakwater crest width “B”
(Seabrook, 1997).
Figure 2.8. Breaker types and criteria for rectangular artificial reefs (after Bleck and
Oumeraci, 2004).
-hs/Hi
Figure 2.9. Prediction of transmission coefficient as a function of relative crest height

(Melito and Melby, 2002).
Figure 2.10. Examples of mean water surface curves under wave action (Diskins et al.,
1970).
Figure 2.11. The influence of crest width on set-up where for Model 2 B=8.0 m, for
Model 2a B=12m and for Model 2b B=4.0m, Rc=-hs, submergence depth (Loveless and
Debski, 1997).
hb hs
h
d
Figure 2.12. Definition for energy flux theory for wave set-up on a submerged
breakwater (Calabrese et al., 2003).
Figure 2.13. Piling-up levels as function of net transmitted discharge qR (Roul et al.,
2004).
Figure 2.14. Dimensionless set-up and current as a function of submergence ratio
(Gourlay, 1996a).
Figure 2.15. Comparison of theoretical and experimental values of discharge ratio “q* ”
(Gourlay and Colleter, 2005).
0.8
shore-normal (a)
0.6 oblique
0.4
Y/B
0.2
0.0
-0.2
-0.4
0 1 2 3 4
Sa / SZW
0.8
higher crest(0.5m)
higher crest (0.5m) (b)
------- lower crest (1.0m)
0.6 lower crest (1.0m)
0.4
Y/B
0.2
0.0
-0.2
-0.4
0 1 2 3 4
Sa / SZW
Figure 2.16. Empirical relationship for preliminary prediction of shoreline response to

submerged structures. (a) Normal waves (solid line) and oblique waves (dashed lines), (b)
normal waves, higher submergence (solid line) and lower submergence (dashed line). (Y
is magnitude of maximum shoreline change, B is structure length, Sa is offshore distance
of the structure and SZW is natural surf zone width); (after Ranasinghe et al. , 2006).
Figure 2.17. Set-up of mean water level computed with (solid line) and without (dashed
line) the roller contribution, compared with measurements (Reniers and Battjes, 1997).
Figure 2.18. Intercomparison of computed longshore current velocities without mixing

(solid line), without a roller (dash-dotted line), without a roller and without mixing
(dashed line); (Reniers and Battjes, 1997).
CHAPTER 3
EXPERIMENTAL SETUP AND DATA SAMPLING
Two preliminary 3D and a comprehensive series of 2D laboratory experiments have

been carried out in the wave basin and 3 m wide wave flume at Water Research
Laboratory (WRL), School of Civil and Environmental Engineering, The University of
New South Wales. The purpose of the preliminary 3D experimental test was to
qualitatively investigate the flow pattern behind the submerged breakwater as described
in section 3.1. The broad range of 2D Laboratory tests undertaken are to examine wave
induced-setup and current over submerged breakwater/reef structures concurrent to
with wave transformation and reflection with the objective of providing a more
thorough understanding of hydrodynamics behavior in the vicinity of broad crested
submerged breakwaters/reefs. Limitations in the existing research have been outlined in
Chapter 2. The objective of this chapter is to describe in detail the experimental setup,
equipment and the method of data acquisition throughout the tests.
3.1 Wave Basin
The preliminary tests were carried out in the 24.5x16.0x0.7m wave basin at Water
Research Laboratory (Figure 3.1). Two moveable hydraulic piston type paddles (each
6.0 m wide) were in the wave basin side by side to give a 12 m paddle to generate
regular and irregular waves (as indicated in Figure 3.2). The bottom of the basin was
made of concrete with slope of 1:50 which was previously used for physical modelling
of shoreline response to submerged breakwater as reported by Ranasinghe et al. (2006).
The wave paddle actuator system and signal generator used in this test are the same as
the wave flume (for more details see Section 3.2).
Water depth in the wave basin was controlled by an automatic water level system
installed in the north western corner of the basin. A point gauge was used to activate an
electronic valve which controlled flow through a 1 inch pipe.
43
3.2 Wave Flume
3.2.1 Flume Layout
A comprehensive set of laboratory testing was carried out in the 3m wide, 1.5m deep
and 30m long wave flume (Figure 3.3). The horizontal bottom and the side walls of the
flume were constructed with concrete (relatively smooth).
One of the disadvantages of 2D wave flume experiments for submerged breakwaters (as
indicated by Loveless et al., 1998) is the creation of offshore directed return flow
(undertow) over the submerged breakwater due to piling-up of water at the end of the
flume and wave set-up shoreward of the submerged breakwater. In such circumstances,
the offshore current disturbs measurements of onshore flow due to wave breaking and
the magnitude of onshore and offshore flow over the submerged breakwater is not a
good representation of flow over the structure. The return flow also agitates the wave
breaking process. In addressing these problems, an internal channel with dimension of
1m wide and 12m long was constructed at the centre of the 3 m flume, 6.45m from the
paddle toward the end of the flume, to ensure that the wave-induced current returns
through the channels on both sides of the internal channel and not over the submerged
breakwater (Figure 3.4). Therefore, return flow, due to shoreward set-up, was
diminished above the submerged breakwater and set-up over and behind the structure
remained as a function of breakwater top friction. Moreover, measurements of
shoreward directed free flow overtopping the structure were more reliable.
The internal channel side walls were made from plywood panels except in the region of
breakwater location, at one side of the channel, which was made from transparent
Perspex. The aim of using a Perspex wall in this area was to provide a clear
visualization panel to monitor wave transformation and current pattern over the
breakwater during the tests. The flume bed in the internal channel was kept horizontal
along the flume. All measurements were made in the internal channel where submerged
breakwaters were located. Offshore water depth within any test was maintained at 300,
400 and 450 mm by an automatic electronic flow valve controlled by accurate water
level sensor (Figure 3.5).
44
Two wave energy absorber systems were employed in the flume to minimise wave
reflection in the flume. The 1V:6H sloping beach was constructed with random size
gravel stones at the end of the main flume (Figure 3.6). The distance of the beach (toe of
the gravel) from the end of the internal channel was kept 2.20m during all tests. The
reason for adopting such a relative steep beach is the limitation in the length of the
flume. An additional wave absorbing system was employed which was composed of
perforated plastic mesh covered by a synthetic 20 mm geotextile mat. The geotextile
mat absorbers were located in front of the beach in the main flume (3m wide) and also
at the end of both side channels (1m wide) with the slope of 1:1 where waves
propagated toward the beach with high energy and without encountering any obstacles
(breakwaters). The wave absorbers in the side channels were located in the line of
onshore toe of the breakwaters in the main channel (4200 mm offshore the beach). The
effectiveness of the wave absorbers was investigated for all test conditions in the flume
without any breakwater.
3.2.2 Wave Generation and Paddle Calibration
One 3m wide and 1.5 m high hydraulic actuated piston type paddle in the flume was
used to generate monochromatic waves in the flume (Figure 3.3). The paddle was able
to generate both monochromatic and irregular waves based on the signal generator
system. The ATO Sweep Function Generator-F74 was operated and set on the desired
monochromatic wave frequency throughout the tests. The signal generator provided the
appropriate signal voltages to the hydraulic ram actuator to produce the desired wave
period and stroke. The system of signal generator and actuator are shown in Figure 3.7.
Before starting the test, the paddle was calibrated to generate the desired test wave
climates. The design curves provided by Gilbert et al. (1970) were used for preliminary
estimation of the paddle stroke. The operating conditions of a wave generator are
defined by determination of the dimensionless variable η:
h
η= (3.1)
gT 2
where T is the desired wave period, g is the acceleration due to gravity and h is the
depth of water in the flume. The amplitude of the paddle stroke is then calculated by
45
defining the dimensionless variable G using the curve corresponding to W=1.0 (piston
type wave generator) and required η in Figure 3.8:
a
G= (3.2)
s
in which a indicates the desired wave amplitude and s is the amplitude of the paddle
stroke. The paddle was calibrated for the empty wave flume (without breakwater). The
stroke of the paddle was found to be almost linear with the actuator voltage settings.
Generated wave heights were measured to establish a basic relationship to actuator
voltage settings for the different test wave period. Nevertheless, the measured incident
wave heights in each test conditions were individually measured and applied in data
analysis.
3.3 Breakwater
Two different rectangular and three trapezoidal breakwaters were constructed and tested
in the wave basin and wave flume, respectively. Both types of breakwaters were
constructed with smooth plywood material.
Two segmented 1000x384x50 mm rectangular breakwaters with 500 mm gap and one
continuous breakwater with the same geometry but 2000 mm length were placed on the
bottom of the slope in the wave basin (Figure 3.9).
Three different submerged breakwater geometries with 3500, 1500 and 300 mm crest
width (in line of flume and wave direction); 300mm height and 1000 mm cross width
were located on a fixed horizontal bottom of the internal channel. The breakwaters were
constructed with horizontal top and slopes of 1V:2H on both ends (Figure 3.10). The
breakwaters were fixed on the bottom of the internal flume so that the shoreward toes of
all types of the breakwaters were located at the distance of 16.00 m onshore of the wave
paddle. The locations of 3 mm diameter holes (taps) on the top and shoreward slope of
the breakwaters depend on the breakwater crest width which are shown in Figure 3.10.
The purpose of the taps over the breakwater is measurement of mean water level across
the structure as described in sections 3.4.2 and 3.6.2.
46
3.4 Instrumentations
3.4.1 Wave Probes
Wave height and water level changes along the flume were measured by 10 capacitance
wave gauges (Figure 3.11). The capacitance wave guage/probe measures the change in
capacitance of the conductor as the air-water interface fluctuates with the passage of
waves. The voltage outputs of the wave probes are converted to water levels using
calibration factors for each wave probe. To calibrate the wave probes, several different
water levels (3 or more) were measured against the corresponding output voltages from
the probes. The voltage conversion factor would be determined as the slope of the best
fit linear curve to the height-voltage diagram. Examining the variation of calibration
factor for each probe during the time indicated that calibration of the probes is needed
every two weeks to achieve errors less than ±1.0mm.
The probes were manufactured by Manly Hydraulic Laboratory (MHL), The New South
Wales Department of Public Works and Services. The specifications at manufacture
indicates scale factors of the wave capacitance probe units as ±5V per ±200mm.
3.4.2 Water Surface Level Followers
Three Franklin Water surface Level Followers, Mk 4, (called FWLF hereafter) mounted
above 100mm diameter cylindrical dash pots were employed to accurately measure
mean water surface level at 7 points along the flume. The three dash pots could be
switched to 7 pressure tappings over the breakwater and along the bottom of the flume.
Each tapping was connected by 10 mm plastic tubing to the cylinder dash pots via a set
of control valves. The 100mm diameter cylinders were able to dissipate the fluctuation
of excess water pressure due to incident waves in the flume. Figure 3.12 shows the
water level measurement system applied in the wave flume.
The FWLF was developed by A.L. Franklin Pty Ltd. The instrument senses the water
level electronically by make and break switching at the water surface which activates a
DC motor that drives a lead screw to move the sensor in contact with the water surface.
The measured water surface is displayed with the mechanical digital readout in the
machine and also the voltage is transmitted to the multi channel data logger system.
47
FWLF is able to measure water surface level with the accuracy of 10-1 mm. The
instruments were calibrated several times during the testing series.
3.4.3 Acoustic Doppler Velocimeter (ADV)
A SonTek 10MHz down-looking Acoustic Doppler Velocimeter (ADV) was used to

measure the velocity profile over the submerged breakwater. The ADV uses the
Doppler principle effect to measure the velocity of water with the accuracy of 10-1
cm/sec, within the velocity range of 3 to 250 cm/sec using a sampling rate between 1
and 25 Hz. It is reported that the lowest range, 3 cm/sec, will yield good results for
flows down to about 0.1 cm/sec (SonTek, 1997).The instrument includes the signal
processing hardware and the 10 MHz ADV probe. The signal processing hardware acts
as signal generator and data processor for the ADV to makes velocity measurements.
The 10 MHz probe consists of probe and signal conditioning module which is
connected to the processor with high frequency cable (Figure 3.13). The probe
comprises three acoustic receivers oriented at 120o, a central transmitter and a stem
connecting to signal conditioning module. The ADV is able to measure velocity in three
components at a single point (sampling volume) in the Cartesian (XYZ) co-ordinate
system. The positive Z-axis is identified upward along the mounting stem and the
positive X-axis and Y-axis are defined from acoustic transmitter to the marked acoustic
receiver (red painted) using a right hand co-ordinate system, respectively. The sampling
volume is the volume of water with the dimension of approximately 0.3 cm3 in which
the ADV makes velocity measurements and located 5 cm from the acoustic transmitter.
Therefore, the down-looking ADV is able to measure velocity with high accuracy in
water depths more than 5cm. Nevertheless, when working very near the boundary it
may not be able to use a high velocity range. As indicated in SonTek (1997), the ADV
is designed to provide reliable performance for years with minimal maintenance and no
re-calibration is required. The signals are sent to a laptop equipped with ADV software
for data processing and analysis.
The accuracy of the ADV velocity data depends on the speed of sound and probe
geometry. The speed of sound is calculated by the software using the user defined input
data of water salinity and temperature. Probe geometry is calibrated by manufacturer
with the measured velocity accuracy of ±1.0%.
48
For each sampling, nine values are recorded by ADV: three velocity components, signal
strength and correlation coefficient for each of three receivers. The main interest in the
measurements is the velocity components in three direction X,Y,Z. The other measured
data are used as data quality remarks. Signal strength is a measure of the strength of the
return reflection from the water. Using the ADV software, signal strength can be
accessed as signal amplitude in counts (1 count ~ 0.43 dB) or as signal to noise ratio
(SNR) in dB. As SNR decreases, the noise in ADV velocity measurements will increase
due to insufficiency of particulate matter or high number of air bubbles in the water.
The recommended value of SNR is at least 15 dB. However, the ADV can operate
reliably with the SNR as low as 5 dB for mean current measurements. Correlation
coefficient is a quality parameter that is a direct output of the Doppler velocity
calculations. It may be used to monitor data quality during collection and to edit data in
post-processing. The proposed value of correlation is reported by the manufacturer to be
between 70 and 100%. However, for mean velocity measurements, correlation values as
low as 30% can be used. The ADV can also measure the distance of the control volume
from the bottom (boundary) with the accuracy of ±1 mm.
3.4.4 Monitoring
All the tests were monitored with a video camera installed on the top of both the basin
and the 3m flume with the observations being recorded on video tape and digital camera
for further analysis.
3.4.5 Flow Direction Indicator
Meshes of weaving threads and floats with dimension of 250x250mm were employed to
identify flow direction at the bottom and surface of water behind the submerged
breakwater in the preliminary 3D test in the wave basin (Figure 3.14). The threads and
floats were tied to the steel chicken wire which was fixed on the bottom of the wave
basin behind the breakwater. Dye injection was made in both 3D and 2D tests to trace
wave-induced flow pattern over and behind the breakwater, qualitatively.
49
3.5 Test Conditions
3.5.1 Three Dimensional Wave Basin Test
Two tests were conducted in the wave basin with the same wave conditions and
different breakwater geometry. Wave height and wave period in both tests were the
same, 30mm and 1.41 sec, respectively. The 2080mm continuous long breakwater was
used in the first test (TP-1) with 20mm water depth over the breakwater (Figure 3.15).
The second test (TP-2) was made using two separate 1040mm long breakwaters with a
500mm separation gap and the same height and submergence (Figure 3.16). Table 3.1
summarizes the test conditions in 3D physical modeling.
Table 3.1 Test conditions for 3D experimental test in the wave basin
Test Conditions Test TP-1 Test TP-2
Water depth at the paddle (mm) 270 270
Wave type regular regular
Wave height (mm) 30 30
Wave period (sec) 1.41 1.41
Breakwater length (mm) 2000 2x1000
Breakwater width (mm) 384 384
Breakwater height (mm) 50 50
Gap size (mm) N/A 500
Water depth over the breakwater (mm) 20 20
3.5.2 Two Dimensional Testing in 3m Wide Flume
A total of105 tests were performed under different combinations of 12 monochromatic

wave climates, 3 different trapezoidal breakwater crest widths in the wave flume with 3
different water depths. 29 tests were also carried out in the flume with the same water
depth and almost similar wave conditions without breakwater in the flume to examine
the generated wave heights and beach reflection. The range of tested variations in water
depth, breakwater crest width, wave period and target wave heights are presented in
Table 3.2. The list of test numbers and corresponding test conditions (breakwater crest
width, water depth, wave climate) are presented in Appendix A.
50
Table 3.2 Test conditions in the 3m wave flume (2D).
Breakwater width (mm) 3500 1500 300 -
Water depth (mm) 350 400 450 -
Target wave height (mm) 50 100 150 -
wave period (sec) 0.71 1.00 1.49 2.00
3.6 Data Sampling
3.6.1 Water Flow Direction
During the 3D tests in the wave basin the directions of flow at the bottom and surface
behind the breakwater was observed using weaving threads and floats as described in
section 3.4.5. All observations were recorded by a video recorder and digital camera.
Some examples of flow direction, indicated by threads and floats, are shown in Figure
3.17 and Figure 3.18 for the test with continuous and segmented breakwater,
respectively. The flow direction was also manually recorded by hand sketch and
compared to images.
3.6.2 Water Surface Level
The fluctuations of water surface level in the 2D flume were measured by capacitance
probes at 10 different points along the flume to calculate wave heights and water surface
fluctuations. The location of the probes and their distances from each other in the flume
depend on the breakwater crest width employed in the test. Figures 3.19 to 3.22 indicate
the location of the probes in each test condition (with and without breakwater). One of
the probes (Probe No. 2, P2 in the figure) was mounted on a traverse trolley moving
approximately at the velocity of 1/T cm /sec (T is the wave period) along the internal
channel, offshore of the breakwater. The trolley was passing through the flume for a
distance of minimum half a wave length to cover maximum and minimum water surface
for incident and reflected wave height analysis.
The sampling rate of water surface elevation using wave probes was adopted so that a
minimum of 10 samples were collected during one wave period. Two sampling
frequencies (10 and 20 Hz) were adopted during the tests depending on the wave period.
51
The length of water level sampling depends on the time of travelling of the trolley along
half or one wave length in the internal channel and varies between 1024 and 3800
samples for each measurement for short and long waves, respectively. The output
signals of wave probes (in Volts) were transferred to the multi channel data logger
system. Data from each probe was stored in EXCEL spreadsheets in individual columns
for each channel (probe) to be processed. Collected raw voltage data were converted to
water level using the corresponding calibration factor in EXCEL spreadsheet. Still
Water surface Level (SWL) in the calm flume was also measured with wave probes in
each test before wave generation.
3.6.3 Mean Water surface level
Mean Water Levels (MWL) were measured with high accuracy (10-1 mm) using three
Franklin Water Level Follower (FWLF) systems as indicated in section 3.4.2. The
locations of the taps along the flume depends on the crest width of the breakwater used
in the test and are shown in Figure 3.19 to 3.22 associated with wave probe locations.
Mean water level was measured in two steps for each test. At each step, the water
surface elevation was measured at 3 points and by turning the control valves the
measurement was undertaken at three other points. Each measurement step was
undertaken after stabilizing water surface level in the cylinders (more than 15 minutes
after wave running or control valves opening). The output signals of water surface (in
Volts) were transferred to the multi channel data logger system. All measurements of
wave height and mean water surface level were undertaken simultaneously. Data were
stored in the same EXCEL spreadsheet the same as the measured data from wave
probes. The voltages were transferred to distance by multiplying data with
corresponding calibration factor for FWLF. Still water surface level was also measured
before running the paddle in the calm flume.
3.6.4 Horizontal Velocity
The Horizontal velocity over the depth was measured by Acoustic Doppler Velocimeter
(ADV). Velocity profile over the depth was measured at several distances above the top
of the breakwater or bottom of the flume. The position of velocity measurement along
the flume depends on the breakwater geometry. The velocity measurement locations are
52
shown in Figure 3.19 to Figure 3.22 and indicated in Appendix A. In some cases
measurements of velocity were carried out at two different points to compare and verify
the results. The number of velocity samplings at each point and the depth of sampling
volume were varied depending on water depth and wave height of the test. The
measurements were undertaken so that the distance of the acoustic transmitter from the
bottom (breakwater or flume) was more than 50 mm to ensure that sampling volume is
above the bed through the tests. For the tests with 400 and 450 mm water (correspond to
100 and 150 mm submergence), velocity was measured mostly over the structure. In the
tests with 350 mm water depth (only 50 mm submergence depth) most of the
measurements were carried out just behind the breakwater (lee side) due to the
limitation in water depth over the breakwater. A sampling rate of 20 Hz for 1024
samples per measurement was adopted. The temperature of water was measured before
each test using a fixed thermometer located in the flume and introduced to the software
for measuring the speed of sound in the water. The ADV velocity range setting were
varied between 30-250 cm/sec depending on the turbulence of the flow over and behind
the breakwater. Higher values of velocity range were adopted for more turbulent flow
with strong wave breaking over the breakwater. Table 3.3 shows a sample of user input
parameters for data acquisition which is extracted from *.CTL output file.
The signal output of the probe was transferred to the signal processing unit and the
processed data were received and stored in the file with the extension of *.ADV using
ADV software version 4.0 (ADV.exe). The ADV files were transferred to 4 editable
ASCII format files with the commands in the software. The ASCII files include data of
velocity (*.VEL), signal to noise ratio (*.SNR), correlation (*.COR) and input data
(*.CTL). All data with ASCII format were transferred to EXCEL spreadsheets for data
processing and analysing. Figure 3.23 and 3.24 show some samples of measured
velocity in horizontal (X and Y) and vertical (Z) directions at a fixed point over the
breakwater in the tests with non-breaking and breaking waves. The positive X direction
is along the flume in the same direction as wave propagation. As indicated in the
figures, the values of velocity in Y direction (perpendicular to wave propagation
direction) are negligible in comparison with the X component of velocity “vx”.
Moreover, the magnitude of Y and Z component of velocities are almost zero in both
tests with and without breaking whilst it has a positive value in X direction for the test
53
with wave breaking over the breakwaters (onshore flow). This matter will be
comprehensively described in the following chapters.
Table 3.3. A sample of user input parameters for data acquisition using ADV.
\
Data File --------------- t1-34p31.ADV
File Size --------------- 23114 bytes
Start Date/Time --------- 01/09/1990 22:46:42
File Comments: Test t1-34

point 3, lower
Probes in File:
Probe 1 - Serial Number: A324
Sampling Rate ----------- 20.000 Hz

Sampling Interval ------- 0.050 s
# of Samples in File ---- 1024
Length of Time Series --- 51.200 s
Temperature ------------- 12.00 øC
Salinity ---------------- 1.50 ppt
Speed of Sound ---------- 1456.86 m/s
Velocity Range ---------- 30 cm/s
A/D Data ---------------- NOT PRESENT
Compass/Tilt Data ------- NOT PRESENT
Temperature Data ------- NOT PRESENT
Pressure Data ------- NOT PRESENT
Ext Sensor Data ------- NOT PRESENT
Ext pressure Data ------- NOT PRESENT
Ctd Data ------- NOT PRESENT
Velocity Coordinates ---- PROBE (X, Y, Z)
Distances to Boundary
Probe # from probe tip from sampling volume
---------------------------------------------------
1 5.84 cm 1.88 cm
---------------------------------------------------
54
3.7 Summary
Two preliminary 3D and a comprehensive series of 2D laboratory experiments have been

conducted in the wave basin and 3 m wide wave flume respectively. The preliminary 3D
experimental tests qualitatively investigated the flow behavior behind a rectangular
submerged breakwater and confirmed the validity of two dimensional testing.
The 2D laboratory tests examined wave breaking, reflection, transmission as well as

wave-induced set-up and currents over submerged breakwater for a wide range of test
combinations of water depth, wave climate and breakwater crest width (Table 3.2). An
internal channel with dimension of 1m wide and 12m long constructed at the centre of
the 3 m flume, ensures that the wave-induced current returns through the channels on
both sides of the internal channel and not over the submerged breakwater. Therefore,
return flow, due to shoreward set-up, was diminished above the submerged breakwater
and set-up over and behind the structure remained as a function of breakwater top
friction. This experimental method ensured that measurements of shoreward directed free
flow overtopping the structure are reliable.
Wave height and water level changes along the flume and over the breakwater were
measured by 10 capacitance wave gauges. Three Franklin Water surface Level Followers
(FWLF) were also employed to accurately measure mean water surface level ( +/- 0.1
mm) at 7 points along the flume.
A SonTek down-looking Acoustic Doppler Velocimeter (ADV) was used to measure the
velocity profile over the submerged breakwater with the accuracy of 0.1 m/sec.
All the tests were monitored with a video camera installed on the top of both the basin
and the 3m flume with the observations being recorded on video tape and digital camera
for further analysis.
55
Figure 3.1. 24.5x16.0x0.7m wave basin used for preliminary 3D tests.
Figure 3.2. Two piston type wave generator paddles applied in the wave
basin.
Traverse wave probe Fixed wave probe
ADV
Wave generator paddle Plywood panel Perspex panel Wave absorber Gravel beach
6
1
B hs
h d
Pressure taps connected to FWLF
SECTION
Wave generator paddle Wave absorber Wave absorber
Side channel
Main channel Submerged breakwater Gravel beach (1:6)
B
Side channel
PLAN
Figure 3.3. Layout of 2D laboratory test in 1.5x3.0x35m wave flume.

Figure 3.4. 1m wide internal channel located in the wave flume.
Figure 3.5. Accurate water follower gauge to control water level in the flume.
Figure 3.6 Combination of gravel beach and wave absorber panel.
Figure 3.7. Signal generator and data acquisition system in the wave flume.
Figure 3.8. Curves for estimating of paddle stroke to generate desired wave height
(Gilbert et al. 1971).
(a)
(b)
Figure 3.9. Rectangular breakwater used in 3D tests. a) Two segmented breakwater,

b)Continuous long breakwater.
1
2
Figure 3.10. Trapezoidal breakwaters used in the wave flume tests (B=3500mm for
test T1, B=1500mm for test T2 and B=300mm for test T3).
Figure 3.11. Capacitance wave probe utilized for water surface elevetion
measurement.
Figure 3.12. Franklin Water Level Follower system for measuring mean water level
across the flume.
Figure 3.13. Standard 10 MHz down-looking ADV probe used to measure wave-
induced flow over the breakwater (SonTek, 1997).
Figure 3.14. 250x250mm mesh of floats and weaving threads to investigate water
flow direction on the surface and bottom of the basin behind and between the break
water in 3D test.
Shore
Structure
Figure 3.15. Long continuous rectangular breakwater in 3D preliminary test "TP-1".

Shore
Structure
Figure 3.16. Two segmented rectangular breakwater with 500mm gap in 3D

preliminary test "TP-2".
1 2 3 4
Incident wave direction
Weaving threads
For bottom direction
1 2
Floats for surface

direction
3 4
Figure 3.17. Flow direction at the water surface and bottom behind the rectangular
2000 mm crest length submerged breakwater in the 3D wave basin preliminary test.
1 2 3 4
Incident wave direction
Weaving threads
For bottom direction
1 2
Floats for surface

direction
3 4
Figure 3.18. Flow direction at the water surface and bottom behind the two
segmented rectangular 1000 mm crest length submerged breakwaters with 500 mm
gap in the 3D wave basin preliminary test.
Traverse wave probe
Wave generator paddle Fixed wave probe

Plywood panel Wave absorber Gravel beach
P1 P3 P2 P4 P9 P10 6
1
Test T0
Figure 3.19. Flume layout and wave probes positions for the tests without breakwater (test T0).
All measurements in mm.
Traverse wave probe
ADV
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 6
1
T1 T2 T3 T4 T5 T6 T7
Test T1
Figure 3.20. Wave flume layout and wave probes (P) and pressure taps (T) positions for test with 3500mm breakwater crest width (test T1).
All ADV profile positions vary - see Table 4.9, Chapter 4 (All measurements in mm).
Traverse wave probe
ADV
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 6
1
T1 T3 T5 T7
T2 T4 T6
Test T2
Figure 3.21. Wave flume layout and wave probes and pressure taps positions for test with 1500mm breakwater crest width (test T2).
Traverse wave probe
ADV
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 6
1
T2 T4 T6
T3 T5
Test T3
Figure 3.22. Wave flume layout and wave probes (P) and pressure taps (T) positions for test with 300mm breakwater crest width (test T3).
30
(a)
vx (cm/sec)
-30
30 31 32 33 34 35 36 37 38 39 40
time (sec)
30
(b)
vy (cm/sec)
-30
30 31 32 33 34 35 36 37 38 39 40
time (sec)
30
(c)
vz (cm/sec)
-30
30 31 32 33 34 35 36 37 38 39 40
time (sec)
measured by ADV at 40mm from top of breakwater for the test with B=1500mm,
hs=150mm, Hi=43 and T=1.49 sec (T2-41, non-breaking). Positive direction of X is
along the flume from the paddle to the beach. a) mm, non-breaking. b) Hi=152 mm,
breaking.
60
(a)
30
vx (cm/sec)
-30
-60
30 31 32 33 34 35 36 37 38 39 40
time (sec)
60
(b)
30
vy (cm/sec)
-30
-60
30 31 32 33 34 35 36 37 38 39 40
time (sec)
60
(c)
30
vz (cm/sec)
-30
-60
30 31 32 33 34 35 36 37 38 39 40
time (sec)
measured by ADV at 40mm from top of breakwater for the test with B=1500mm,
hs=150mm, Hi=152 and T=1.49 sec (T2-43, breaking). Positive direction of X is
along the flume from the paddle to the beach.
CHAPTER 4
EXPERIMENTAL DATA PROCESSING
In the previous chapter, the details of experimental set-up and measurement methods
were discussed. An important part of experimental studies is processing of the data to
enhance its interpretation. Different approaches to data processing are examined in this
chapter to discover the most appropriate methods producing reliable application of the
laboratory measurements. Water surface elevation (set-up and set-down), incident wave
height, wave-induced discharge and wave breaking over the submerged breakwater are
investigated in this chapter and the most appropriate methods were adopted to calculate
water level, incident wave height and wave breaking prediction.
4.1 Water Surface Fluctuation
Water surface elevation data collected by wave probes was examined for better
understanding of the characteristics of the generated wave fields in the flume.
Propagation of waves generated by sinusoidal motion of wave generator paddle in the
flume is not ideally in the constant form expected from the regularity of the paddle
motion- the form of the wave slowly changes as it travels along the flume. The
evolution of wave form along the flume depends on wave steepness and relative water
depth (the ratio of water depth to wavelength). Moreover, the abrupt change in water
depth over the breakwater produces some higher harmonic waves and a periodic
behaviour of the reflection and transmission coefficient was observed. Higher harmonic
free waves can contaminate the study of monochromatic waves in the flume. Multer
(1970) found experimentally that the run-up on a steep slope might change by a factor
of two depending on the phase difference between the primary and the highest
secondary waves. Therefore, specific consideration of the harmonic generation is
needed to obtain reliable measurement of incident, reflected and transmitted wave
heights.
56
4.1.1 Harmonic Generation in the Flume Without Breakwater
Free secondary harmonic waves will always exist when waves are generated by a
sinusoidal moving wave maker. The primary and secondary waves travel at different
speeds in the channel and the resulting surface profile may display the presence of
secondary waves, depending on the distance from the wave maker. Figure 4.1 shows a
sample of wave profiles at different stations in the wave flume without breakwater. The
water surface variations were measured at the same time. As indicated in the figure, the
main wave is superimposed with a smaller wave, which is travelling down the flume
with smaller speed. Biesel and Suquet (1951) addressed this phenomenon for the first
time and following them, Morison and Crooke (1953), Le Mehaute et al. (1968) and
Iwagaki and Sakai (1970) reported harmonic generation trends in their experiments.
Madsen (1971) solved the second order Stokes equation for water surface profile
provoked by a sinusoidal moving paddle. He developed the following equation for
water surface profile that consists of a Stokes second order progressive wave and a free
second harmonic wave:
η = −a sin (kx − ω t ) − a p (2 ) cos 2(kx − ω t )

(4.1)
cos (κx − 2ω t )
(2 )
+ aL
(2)
where superscript corresponds to the second order term and a is the amplitude of the
first order (harmonic) wave that is related to the amplitude of the wave-maker, S:
tanh kh
a=S⋅ (4.2)
n1
in which h is water depth, n1 is the ratio of group wave celerity to wave celerity (= cg /c)
and k and κ is the wave number of the main and the secondary waves, respectively.
The components of amplitude of second harmonics waves (ap(2) and aL(2)) in Equation
(4.1) can be found by the following equations:
3 a 2 L2
= ⋅
( 2)
ap (4.3)
16π 2 h 3
57
1 2 coth kh tanh κh ⎛ 3 n ⎞
= ⋅ ⋅ ⎜⎜ − 1 ⎟⎟
( 2)
aL a 2
(4.4)
2 h n2 ⎝ 4 sinh kh 2 ⎠
where
1⎛ 2κh ⎞
n2 = ⎜⎜1 + ⎟ (4.5)
2 ⎝ sinh 2κh ⎟⎠
The free second harmonic wave in the flume travels at a slightly slower speed than the
main Stokes wave since the corresponding wave number is greater than twice the main
wave number (κ >2k). Therefore, the resulting amplitude of the free second harmonic
wave a(2) varies with distance from the wave generator:
a (2) = a p [ (2)2
+ aL
(2)2
− 2a p a L
(2) (2)
⋅ cos (κ − 2k )x ]
1
2
(4.6)
Galvin (1968) showed that long waves generated at one end of a horizontal channel
develop into forms that are not periodic in space. Buhr Hansen and Svendsen (1974)
indicated that the main reason for the irregularities in the generated waveform in the
flume is that the paddle cannot produce the variation of the particle motion, which
corresponds to a progressive wave of constant form. They experimentally found that the
waves generated by a sinusoidal piston motion can be described as a second order
Stokes wave superimposed by a free second harmonic wave for the range of 0.1 < h/L <
0.65. They revealed that Stokes’ theory is not applicable to the prediction of wave
pattern in long waves ( h L < 0.1 ).
In the experiments involving sinusoidal long waves (h/L<0.1), a free second harmonic
wave will always exist (Flik and Guza, 1980). For waves of small amplitude, these
secondary waves are relatively small and may be negligible in most cases. The primary
and secondary waves travel at different speeds and the resulting water surface profile
may display the presence of secondary waves, depending on the distance from the
paddle. The existence of these secondary waves may be the cause of a number of
problems in laboratory tests for relatively long waves and the corresponding results
analysis. Therefore, understanding of the nature of secondary waves is important and
58
generation of monochromatic long waves free of secondary waves is highly desirable in
experimental laboratory tests.
In the present experimental study, monitoring of water surface elevation in the main
channel without breakwater was carried out by four fixed and one traversing probe as
discussed in Chapter 3 (Section 3.2). Spectral analysis of the recorded signals from the
probes, using Fast Fourier Transformation (FFT) method, was employed to investigate
the behaviour of water surface elevation and waves travelling along the flume. The
shape of the spectrum of the water surface elevation in the empty flume (no breakwater)
for some selected test conditions is shown in Figure 4.2. The figure corresponds to
probe No. 4, which was located at almost 13.00 m landward of the wave paddle.
The water level spectrums revealed that higher harmonic wave generation took place in
some specific test conditions. Variations of dimensionless parameters such as wave
steepness, relative water depth and Ursell parameter (Ursell, 1953) within different tests
with various wave and water depth conditions were examined to find the best indicators
for higher harmonic generation in the empty main channel (Table 4.1). Figure 4.3
demonstrates the variations of the dimensionless parameters in various test conditions.
The filled black and hollow circle points in the figure correspond to tests for which
harmonic generation and no harmonic generation were observed in the flume,
respectively. Figure 4.3 confirms that Ursell parameter is the best indicator for
prediction of harmonic wave generation in the flume while it is also good definition for
applicability of higher order nonlinear wave theories. In the present test ranges (without
breakwater), higher harmonic generation was observed mostly for the tests with Ursell
number greater than 10:
2
L H
U R = i 3 i ≥ 10 (4.7)
h
in which Li is wavelength, Hi is incident wave height and h is water depth in the

channel. Close attention needs to be paid to data analysis of tests with Ursell number
greater than 10 due to the likely generation of higher harmonics in the flume for such
tests.
59
Table 4.1. Variation of Ursell number, wave steepness and relative wave height in the tests without
breakwater (Test T0).
Test
Test No. h(mm) T(sec) L(mm) Hi KR h/L UR Hi/L Hi/d
range
T0-1 1 350 0.71 792 47.2 0.16 0.44 0.69 0.06 0.13
T0-5 2 350 1.00 1470 56.9 0.13 0.24 2.87 0.04 0.16
T0-6 3 350 1.00 1470 95.1 0.17 0.24 4.79 0.06 0.27
T0-9 4 350 1.49 2596 41.4 0.14 0.13 6.50 0.02 0.12
T0-10* 5 350 1.49 2596 81.7 0.20 0.13 12.83 0.03 0.23
T0-11* 6 350 1.49 2596 128.2 0.14 0.13 20.15 0.05 0.37
T0-13* 7 350 2.00 3631 42.7 0.08 0.10 13.13 0.01 0.12
T0-14-1* 8 350 2.00 3631 93.4 0.15 0.10 28.74 0.03 0.27
T0-14-2* 9 350 2.00 3631 121.3 0.16 0.10 37.30 0.03 0.35
T0-15-1* 10 350 2.00 3631 121.4 0.14 0.10 37.33 0.03 0.35
T0-15-2* 11 350 2.00 3631 173.2 0.21 0.10 53.27 0.05 0.49
T0-17 12 400 0.71 785 47.4 0.17 0.51 0.46 0.06 0.12
T0-21 13 400 1.00 1500 62.5 0.10 0.27 2.20 0.04 0.16
T0-22 14 400 1.00 1500 96.8 0.09 0.27 3.40 0.06 0.24
T0-25 15 400 1.49 2729 42.5 0.13 0.15 4.94 0.02 0.11
T0-26* 16 400 1.49 2729 87.3 0.13 0.15 10.15 0.03 0.22
T0-27* 17 400 1.49 2729 113.6 0.12 0.15 13.21 0.04 0.28
T0-29 18 400 2.00 3859 24.1 0.11 0.10 5.61 0.01 0.06
T0-30* 19 400 2.00 3859 98.2 0.06 0.10 22.85 0.03 0.25
T0-31* 20 400 2.00 3859 153.9 0.08 0.10 35.82 0.04 0.38
T0-33 21 450 0.71 786 45.0 0.10 0.57 0.31 0.06 0.10
T0-37 22 450 1.00 1520 43.1 0.14 0.30 1.09 0.03 0.10
T0-38 23 450 1.00 1520 97.6 0.08 0.30 2.47 0.06 0.22
T0-41 24 450 1.49 2842 52.4 0.09 0.16 4.65 0.02 0.12
T0-42* 25 450 1.49 2842 117.5 0.14 0.16 10.42 0.04 0.26
T0-43* 26 450 1.49 2842 147.6 0.14 0.16 13.09 0.05 0.33
T0-45 27 450 2.00 4066 42.6 0.05 0.11 7.73 0.01 0.09
T0-46* 28 450 2.00 4066 101.8 0.08 0.11 18.48 0.03 0.23
T0-47 29 450 2.00 4066 155.1 0.11 0.11 28.14 0.04 0.34
* Harmonic higher frequencies observed
4.1.2 Harmonic generation above the breakwater
Harmonic generation owing to abrupt changes in water depth is relatively well known
and some experimental, numerical approaches have been developed (e.g. Chiang et al.,
1969; Massel, 1983; Goda et al., 1999) to investigate the processes. When waves
approach a submerged breakwater, part of the wave energy is reflected by the barrier
whilst on the top of the breakwater harmonic generation with energy transfer to higher
harmonics is observed. Goda et al. (1999) revealed that progressive and retrogressive
waves exist on the top of the breakwater and form a partial standing wave system
(Figure 4.4). The higher harmonics generated over the step (breakwater) are transmitted
60
to the deeper water behind the breakwater as free waves. Longuet-Higgins (1977) found
that the second harmonic wave amplitude might be as large as approximately 45% of
the incident wave amplitude. Massel (1983) applied second order wave theory and
proposed a numerical model for higher harmonic generation (second order) over the
step. The model showed that the amplitude variation of the second harmonic over the
step is almost 6 times higher than in front of the step and higher steps produce higher
harmonic amplitudes since higher nonlinear interactions exit in the shallower water.
Losada et al. (1997) conducted a series of 2D experimental tests to examine harmonic

evolution of non-breaking monochromatic waves as they propagate over porous and
impermeable submerged breakwaters. They found that the effective water depth over
the breakwater provides information to evaluate the potential of harmonic generation.
Driscoll et al. (1992) observed significant harmonic generation over the breakwater
during the tests with single sine wave condition. They also reported the existence of
spatial amplitude modulation down-wave of the obstacle, which was related to
harmonic evolution (as wave reflection was trivial in the flume).
The spectra of water surface elevation in front (seaside), over and behind (landward) the
breakwater with different crest width for some selected spectrums are shown in Figure
4.5. The Figure corresponds to the test with breaking wave (T*-5) and non-breaking
waves (T*-38 and T*-45) passing over the breakwater with different crest width. The
shape of wave spectra offshore the breakwater showed similar trends of harmonic
generation to those noted in the flume without breakwater. However, higher harmonic
waves were observed seaward of the breakwater in a few tests (e.g. T1-29, T1-45, T2-
29, …) which did not occur in the flume without the breakwater. The reason for
existence of higher harmonic waves in the flume may be found in retrogressive higher
harmonic generated waves due to abrupt changes in the water depth (the breakwater)
along the progressive non-breaking wave path. Higher harmonic frequencies were
generated over and behind the breakwater in the tests with both no-breaking and
breaking waves. Investigation of the wave spectrums in more detail showed resonance
in the flume (over the breakwater) for tests T2-29 and T3-29 (breakwater crest widths of
1500 mm and 300mm, water depth h=400mm, wave hight Hi~50mm and period T=2.0
sec). These tests (T2-29, and T3-29) were discarded from further data analysis.
61
4.2 Incident Wave Height
Water surface elevation data, collected by wave probes along the wave flume, were
used to calculate incident, reflected and transmitted wave height. Three different
methods, the conventional traversing method of Healy (1953), Goda and Suzuki (1976)
and Mansard and Funke (1980) were implemented to calculate incident and reflected
wave height and the results were compared to adopt the most appropriate and reliable
method for ongoing analysis and interpretation. A number of MATLAB scripts were
developed to read water surface elevation data from EXCEL environment and calculate
incident/reflected wave height using the three different methods (See Appendix B for
the codes).
4.2.1 Method of Healy
When the wave generator paddle is first set in motion, primary incident wave travels
toward the beach and there is a primary reflected wave of amplitude usually a small
fraction of incident wave. The primary reflected wave is reflected from the paddle and
travels toward the shore as secondary incident wave and so on. The higher reflection
waves have progressively smaller amplitudes and the steady state is almost attained
after a few reflections (Ursell et al. 1960). In the presence of a highly dissipative beach,
the secondary reflected wave is usually negligible. In the middle region of the flume,
water level consists of the sum of incident and reflected waves and the relative phase of
primary, secondary and higher-order waves depend on the effective length of the
channel such that the wave height is not the same at all locations along the flume. The
water surface oscillates around a mean value over a length of 0.5L. Ursel et al. (1960)
reported that the maximum and minimum wave heights do not repeat exactly at
successive maxima and minima of the wave-height envelope and there is an uncertainty
in the wave height (average 3% for their test range).
The incident and reflected wave height (Hi and HR, respectively) can be defined by
measuring the maximum and minimum of the water surface envelope seaward of the
breakwater using the wave probe installed on the traverse trolley (Healy, 1953- Section
3.4):
62
H max + H min
Hi = (4.8)
2
H max − H min
HR = (4.9)
2
where Hmax and Hmin are maximum and minimum wave height, extracted from the water
surface envelop. Therefore, the reflection coefficient KR can be determined as:
HR H − H min
KR = = max (4.10)
Hi H max + H min
The MATLAB script “Inciwav.m” was developed (see Appendix B) to read data of
water surface elevation from corresponding EXCEL file for each test and calculate the
maximum and minimum of the water surface envelope seaside of the breakwater. Both
zero down crossing and zero up crossing methods were applied for calculating wave
height. Figure 4.6 represents the envelope of water surface using both zero down and up
crossing methods for the selected test. Small differences between water surface
envelopes were found between the two methods. The average wave heights extracted
from zero up crossing and zero down crossing methods were applied to calculate
incident and reflected wave height thus enhancing the results.
The Savitzky-Golay filtering method (Orfanidis, 1996) was applied to smooth the
measured water surface fluctuation. Savitzky-Golay filters are optimal in the sense that
they minimize the least-squares error in fitting a polynomial to each frame of noisy data.
The polynomial order k =0 (linear) and the frame size f =5 were adopted for smoothing
5 neighbourhood data with a line. The comparison of smoothed and non-smoothed
calculated water surface envelope is shown in Figure 4.7 for some selected tests. The
incident and reflected wave height were calculated in the MATLAB script using
equations 4.8 and 4.9 and the averaged smoothed water surface envelop. The summary
of calculated incident and reflected wave height, applying Healy (1953) method, are
presented in Table 4.2 for all test conditions.
The calculated reflection coefficient for the tests without breakwater revealed that the
reflection coefficient of the flume varies between 0.05 and 0.14.
63
4.2.2 Method of Goda and Suzuki
Goda and Suzuki (1976) provided a technique to resolve the incident and reflected wave
heights by analysing the amplitudes of Fourier components of recorded data from two
wave gauges separated by a fixed distance Δ l. For regular waves, the incident and
reflected wave amplitudes may be calculated by:
aI =
1
2 sin kΔl
[
⋅ ( A2 − A1 cos kΔl − B1 sin kΔl ) + (B2 + A1 sin kΔl − B1 sin kΔl )
2 2
] 1
2
(4.11)
aR =
1
2 sin kΔl
[
⋅ ( A2 − A1 cos kΔl + B1 sin kΔl ) + (B2 − A1 sin kΔl − B1 sin kΔl )
2 2
]
1
2
(4.12)
where A1, B1, A2 and B2 are the amplitudes of fundamental frequency of recorded wave
profile for probe 1 and 2, which can be calculated by Fourier analysis (FFT). Δl
indicates the distance between the two probes and k is the wave number.
The main frequency is considered in the resolution although it is noted that the actual
wave profile usually contained some higher harmonics. Goda and Suzuki indicated that
the method cannot be used around the condition of Δl/Lo=n/2 where n=0, 1, 2…. This
relates to the probes distance equal to an integer number of half wavelengths. They
proposed the best range of probes space of 0.05< Δl/Lo<0.45 and tests not be carried out
for 0.45< Δl/Lo<0.6.
The measurement of two progressive wave probes in front of the breakwater, using the
method of Goda and Suzuki, were made in the tests without breakwater (T0) and the
tests with 300mm breakwater crest width (T3) within the proposed range of probes
space. A sample of calculated incident and reflected spectrum in comparison of the total
spectrum, applying the method of Goda and Suzuki (1976) is defined in Figure 4.8. The
comparison of calculated incident wave height using the conventional method of Healy
(1953) and the method of Goda & Suzuki in Figure 4.9 reveals that more than 80% of
data are located inside the ±10% confidence zone and the method of Goda and
Suzuki(1976) agrees well with the conventional method to calculate incident wave
64
height for monochromatic wave circumstances (R2=0.97). The points beyond the ±10%
confidence interval belong to the tests in which higher harmonic generation observed
due to higher value of Ursell number in the flume or non-breaking waves passing over
the abrupt step (submerged breakwater). The values of calculated incident and reflected
wave height using the method of Goda and Suzuki are presented in Table 4.3. No
discussion on wave reflection has been made in this thesis as it is not of primary interest
to this study.
Table 4.2 Calculated incident and reflected wave height for all test conditions using the method of
Healy (1953); See Appendix A of test conditions for each test number.
without breakwater B=3500 mm B=1500 mm B=300 mm
Test No. Hi (mm) KR Test Hi KR Test Hi KR Test No. Hi KR
T0-1 47 0.16 T1-1 49 0.19 T2-1 45 0.19 T3-1 52 0.29
T0-5 - - T1-5 50 0.07 T2-5 62 0.09 T3-5 69 0.10
T0-6 - - T1-6 - - T2-6 104 0.09 T3-6 110 0.09
T0-9 41 0.14 T1-9 50 0.17 T2-9 44 0.22 T3-9 43 0.31
T0-10 - - T1-10 100 0.07 T2-10 99 0.04 T3-10 96 0.11
T0-11 - - T1-11 - - T2-11 160 0.10 T3-11 148 0.14
T0-13 43 0.08 T1-13 56 0.17 T2-13 62 0.19 T3-13 47 0.24
T0-14 - - T1-14 106 0.24 T2-14 115 0.20 T3-14 83 0.40
T0-15 121 0.14 T1-15 - - T2-15 146 0.21 T3-15 99 0.41
T0-17 - - T1-17 50 0.23 T2-17 58 0.19 T3-17 52 0.33
T0-21 63 0.10 T1-21 56 0.05 T2-21 57 0.06 T3-21 58 0.14
T0-22 97 0.09 T1-22 - - T2-22 113 0.11 T3-22 101 0.11
T0-25 43 0.13 T1-25 53 0.25 T2-25 50 0.29 T3-25 40 0.44
T0-26 87 0.13 T1-26 102 0.14 T2-26 107 0.14 T3-26 104 0.26
T0-27 114 0.12 T1-27 145 0.12 T2-27 136 0.12 T3-27 132 0.25
T0-30 98 0.06 T1-30 102 0.22 T2-30 106 0.16 T3-30 118 0.24
T0-31 154 0.08 T1-31 - - T2-31 137 0.18 T3-31 171 0.30
T0-33 45 0.10 T1-33 42 0.34 T2-33 56 0.25 T3-33 44 0.24
T0-37 43 0.14 T1-37 55 0.19 T2-37 44 0.14 T3-37 37 0.16
T0-38 98 0.08 T1-38 97 0.21 T2-38 86 0.09 T3-38 90 0.17
T0-41 52 0.09 T1-41 57 0.28 T2-41 43 0.27 T3-41 72 0.32
T0-42 - - T1-42 100 0.07 T2-42 121 0.23 T3-42 115 0.14
T0-43 148 0.14 T1-43 155 0.14 T2-43 152 0.20 T3-43 133 0.25
T0-45 43 0.05 T1-45 51 0.25 T2-45 57 0.40 T3-45 43 0.45
T0-46 102 0.08 T1-46 95 0.51 T2-46 102 0.56 T3-46 92 0.28
T0-47 - - T1-47 - - T2-47 156 0.08 T3-47 147 0.1
65
4.2.3 Method of Mansard and Funke
Mansard and Funke (1980) have developed the estimation of incident wave height
applying simultaneous measurement from three fixed wave probes. They assumed that
the irregular wave can be described as a linear superposition of an infinite number of
discrete components that travel at their own individual phase velocities. Applying a
least square analysis for decomposing the measured spectra into incident and reflected
spectra, they found that the incident and reflected spectra may be defined by:
z I ,k =
1
[B1,k ⋅ (R1k + iQ1k ) + B2,k ⋅ (R2 k + iQ2 k ) + B3,k ⋅ (R3k + iQ3k )] (4.13)
Dk
and
z R ,k =
1
[B1,k ⋅ (R1k − iQ1k ) + B2,k ⋅ (R2 k − iQ2 k ) + B3,k ⋅ (R3k − iQ3k )] (4.14)
Dk
where:
( )
Dk = 2 sin 2 β k + sin 2 γ k + sin 2 (γ k − β k )
R1k = sin β k + sin γ k
2 2
Q1k = sin β k ⋅ cos β k + sin γ k ⋅ cos γ k

R 2 k = sin γ k ⋅ sin (γ k − β k ) (4.15)
Q 2 k = sin γ k ⋅ cos(γ k − β k ) − 2 sin β k
R3k = − sin β k ⋅ sin (γ k − β k )
Q3 k = sin β k ⋅ cos(γ k − β k ) − 2 sin γ k
in which B1k, B2k and B1k are the Fourier coefficients of probes 1, 2 and 3 for frequency
k/T. βk and γk are relative distance terms determined by measuring the distances X12
and X13 between probe 2 and 1 and probe 3 and 1 where probe 1 is the closest to the
paddle:
2π ⋅ X 12
βk =
Lk
(4.16)
2π ⋅ X 13
γk =
Lk
66
in which Lk is the wave length corresponding to the wave with wave number k.
Based on laboratory test data with monochromatic waves, Mansard and Funke (1980)
recommend the following criteria of probe spacings for greater accuracy:
X 12 = Li / 10
Li / 6 < X 13 < Li / 3
(4.17)
X 13 ≠ Li / 5
X 13 ≠ 3Li / 10
They also recommended that the probes be located at least one wavelength away from
any reflective structures.
The incident and reflected spectrum using the method of Mansard and Funke (1980)
were calculated by the developed spreadsheet in MATLAB. Figure 4.10 represents a
sample of calculated incident and reflected spectrum for the test with 300 mm crest
width, 400 mm water depth, 117 mm wave height and 2.0 sec wave period (T3-30).
The comparison of calculated incident wave height using the conventional method of
Healy (1953) and Mansard and Funke (1980) is shown in Figure 4.11. The evaluation
being for Test T3 (B=300 mm) for which the probe spacings were mostly in the ranges
proposed by Mansard and Funke (1980) (Equation 4.17). The figure reveals that the
method is less sensitive to the tests with higher harmonic generation in comparison of
Goda and Suzuki’s method. It is also seen that the method is less sensitive to probe
spacing than was indicated by Mansard and Funke (1980). The values of calculated
incident and reflected wave height using the method of Mansard and Funke (1980) are
presented in Table 4.3 for both Tests T3 (B=300 mm) and T0 (no breakwater) results.
As mentioned in the previous section, the reflection coefficient “KR” is not of primary
concern in this study.
67
Table 4.3 Calculated incident wave height and reflection coefficient for the test without breakwater and with 300 mm breakwater crest width using three
different methods. The shaded areas correspond to tests for which the probe spaces are beyond the recommended distances (See Appendix A of test conditions).
No breakwater (T0) B=300 mm (T3)
Test
series KR Hi (mm) KR Hi (mm)
Healy Goda Mansard Healy Goda Mansard Healy Goda Mansard Healy Goda Mansard
1 0.16 0.68 0.56 47 84 72 - - - - - -
5 0.13 0.14 0.18 57 49 54 0.09 0.92 1.06 62 310 59
6 0.17 0.12 0.18 95 85 93 0.09 0.96 0.65 104 830 98
9 0.14 0.13 0.12 41 42 42 0.22 0.17 0.65 44 48 54
10 0.20 0.14 0.17 82 79 81 0.04 0.09 0.28 99 93 85
11 0.14 0.11 0.16 128 118 125 0.10 0.16 0.09 160 117 131
13 0.08 0.11 0.18 43 43 42 0.19 0.29 0.46 62 59 58
14 0.15 0.16 0.26 93 84 81 0.20 0.36 0.44 115 79 99
15 0.14 0.16 0.27 121 103 98 0.21 0.40 0.42 146 88 126
17 0.17 0.41 0.38 47 49 59 0.19 0.65 2.75 58 90 39
21 0.10 0.06 0.11 62 64 65 0.06 0.98 0.90 57 427 107
22 0.09 0.05 0.14 97 96 97 0.11 0.85 0.50 113 517 122
25 0.13 0.10 0.09 42 45 46 0.29 0.32 0.65 50 53 68
26 0.13 0.15 0.13 87 85 91 0.14 0.10 0.49 107 105 128
27 0.12 0.15 0.14 114 108 120 0.12 0.03 0.07 136 128 125
29 0.11 0.12 0.12 24 25 25 - - - - - -
30 0.06 0.05 0.06 98 90 89 0.16 0.23 0.18 106 80 111
33 0.10 0.43 0.27 45 51 51 0.25 0.89 2.83 56 106 47
37 0.14 0.04 0.05 43 47 47 0.14 0.66 1.08 44 69 79
38 0.08 0.06 0.12 98 97 99 0.09 0.51 0.96 86 165 137
41 0.09 0.08 0.07 52 54 54 0.27 0.25 0.39 43 45 45
42 0.14 0.11 0.10 117 117 124 0.23 0.13 0.32 121 113 116
43 0.14 0.15 0.14 148 147 171 0.20 0.07 0.20 152 136 136
45 0.05 0.07 0.07 43 43 43 0.72 0.81 0.90 68 50 163
46 0.08 0.05 0.07 102 91 92 0.79 0.58 0.82 104 94 193
47 0.11 0.07 0.11 155 129 135 0.08 0.19 0.21 156 119 143
68
4.2.4 Comparison of the Methods
The quantitative comparison of the previous methods with the conventional method of
Healy for incident wave height calculation was investigated by comparing the three
statistical parameters of bias or distortion (β), Root Mean Squared Error (RMSE) and R2:
⎛ Xc ⎞
N
∑ ⎜⎜ X
i =1 ⎝
⎟⎟
m ⎠i
β= (4.18)
N
12
⎡ N ⎛ Xc − Xm ⎞ 2⎤
⎢ ∑ ⎜⎜ ⎟⎟ ⎥
i =1 ⎝ Xm ⎠i ⎥
RMSE = ⎢ (4.19)
⎢ N ⎥
⎢ ⎥
⎢⎣ ⎥⎦
2
⎛ N ⎞
⎜ ∑ (X m − X m )i ( X c − X c )i ⎟
R 2 = ⎝N i =1 ⎠ (4.20)
∑ (X m − X m )i ∑ (X c − X c )i
N
2 2
i =1 i =1
where N represents the number of data in each group and Xc and Xm are respectively
calculated and measured values of the desired quantity. The bar sign indicates the
average of the corresponding parameter. The values of the statistical indicators in Table
4.4 indicate that both methods of Mansard and Funke and Goda and Suzuki are able to
calculate incident wave height with reasonable accuracy in comparison to the
conventional method of Healy when wave probes are located at the appropriate distance
from each other.
Table 4.4 Statistical comparison of the methods of Goda and Suzuki (1976) and Mansard and
Funke (1980) with the conventional method to calculate incident wave height from the measured
wave spectrum in laboratory for test T3 (B=300 mm).
2
Method β RMSE R
Goda 0.96 0.09 0.97
Mansard 0.94 0.09 0.97
69
As indicated in Figure 4.9, the method of Goda and Suzuki (1976) predicted the
incident wave height to be less than that using the method of Healy (1953), especially
for tests in which higher harmonics waves were generated in the flume. In Goda and
Suzuki’s method, only the fundamental frequency is considered to calculate incident
wave height. On the other hand, the method of Mansard and Funke (1980) is shown
(Figure 4.11) to be less sensitive to noise and deviations from the linear theory (higher
harmonic wave generation). Figure 4.11 indicates that fewer errors occur for Mansard
and Funke method when the incident wave is contaminated by higher harmonic waves.
Hydrodynamic effects of submerged breakwaters are governed by breakwater geometry

and progressive wave climate including wave energy. The methods of Healy (1953) and
Mansard and Funke (1980) are judged more applicable to calculate incident
monochromatic wave height from the recorded water surface. In both methods of Healy
(1953) and Mansard and Funke (1980), total spectrum (total energy) of the wave is
taken into account which is more realistic with respect to the amount of energy
approaching the breakwater in the experiments. Finally, data collected from the traverse
probe associated with the conventional method (Equation 4.8) were applied in this study
to calculate incident wave height.
4.3 Wave Breaking
One of the most significant phenomena in coastal dynamics is breaking of waves

passing across a beach or a shoal (e.g. reef or submerged breakwater). The prediction of
nearshore current, water level (set-up and set-down), sediment transport and beach
evolution in shallow water all depend on estimation of breaking wave properties such as
the breaker-depth index (the ratio of wave height to water depth at breaking point),
breaker type, plunging distance and wave reformation. In practice, the breaker-depth
index (H/h)max is usually used as the criteria for wave breaking. Stokes (1880) first
derived the limiting angle of wave steepness, α=120°, as a distinction between breaking
and non-breaking waves. For the finite water depth “h”, Miche (1940) obtained the limit
of wave steepness, numerically:
⎛ Hi ⎞
⎜⎜ ⎟⎟ = 0.142 tanh (kh ) (4.21)
⎝ Li ⎠ max
70
where Hi is the wave height, Li is the wave length and k is the wave number.
Many studies have been undertaken to explain breaking wave properties for plain
beaches with mild slope. The summary of the methods for calculating breaker-depth
index for plain beaches are indicated in Table 4.5. However, limited research has been
conducted to explain wave breaking process due to a submerged barrier/breakwater.
Table 4.5 Summary of previous studies to calculate breaker-depth index for plain beaches (Dack,
2004).
Model Limitations Limiting Wave
Source Water Model
Bed Slope Height
Depth
all
(H/L)b = 0.142 tanh 2π(h/L)b (H/L)max = 0.142
horizontal depths
Miche (1944)
bed shallow
(H/h)b = 0.88 (H/L)max = 0.88
water
horizontal deep
Michell (1893) (H/h)b = 0.142 (H/L)max = 0.142
bed water
McCowan horizontal shallow
(H/h)b = 0.78 (H/h)max = 0.78
(1891) bed water
(H/h)b = 1/βb
where; βb = 0.92 for m ≥
all
Galvin (1969) all slopes 0.07 and (H/h)max = 0.72
depths
βb =1.40 – 6.85m for m
≤0.07
Collins and Wier all
all slopes (H/h)b = 0.72 + 5.6m (H/h)max = 0.72
(1969) depths
all
Goda (1970) all slopes Hb/Ho = f((H/L)o , tan θ) -
depths
2
(H/h)b = b – a(Hb/gT )
-
all where; a = 43.75(1 – e
Weggel (1972) all slopes 19tanθ (H/h)max = 0.78
depths )
-19tan θ
b = 1.56/(1+e )
horizontal all
Nelson (1985) (H/h)lim = [Fc/(22 + 1.82Fc)] (H/h)lim ≤ 0.55
bed depths
0 > cot(θ) < shallow (H/h)max = 0.55
Nelson (1987) (H/h)max =0.55
0.01 water + 0.88exp(-0.012cot(θ))
4.3.1 Previous Studies of Wave Breaking over Shoals
In the case of submerged barrier, Smith and Kraus (1991) conducted a series of
laboratory tests to investigate properties of wave breaking over bars and artificial reefs.
They applied triangle bar shapes with seaward angles 5°, 10°, 15°, 20°, 30° and 40° and
found different breaking characteristics over bars and plane slope beaches. Smith and
Kraus found that the breaker-type transition values expressed in terms of the surf-
71
similarity parameters are lower than those proposed by Battjes (1975) for plane slopes
as shown in Table 4.6.
Table 4.6 Difference of transition values of breaker type for plane slopes
(Battjes, 1975) and barred profile beaches (Smith & Kraus, 1991).
range of similarity number “ξo”
breaker type
Battjes (1975) Smith & Kraus (1991)
surging or collapsing 3.3 < ξo 1.2 < ξo
plunging 0.5 < ξo < 3.3 0.4 < ξo < 1.2
spilling ξo < 5.0 ξo < 4.0
ξo = tan a /(Ho/Lo) ,deepwater similarity (Iribarren) parameter.
0.5
They also observed that the breaker-depth index “γb” is greater in barred profile than the
values reported by previous authors and proposed the following values for seaward bar
slope flatter than 10°:
γ b = 0.41+ 0.98ξ o for 0.3 ≤ ξ o ≤ 0.85 (4.22)
γ b = 1.45 − 0.22ξ o for 1.6 ≤ ξ o ≤ 3.5 (4.23)
Hara et al. (1992) investigated the breaking wave characteristics of a solitary wave
incident to a submerged step and trapezoidal breakwater, numerically. They presented
the modified surf similarity parameters “ξo′ ” and “ξo" ” for the step and symmetric
trapezoidal submerged breakwater, respectively:
ξ o′ = (d h )0.1 (H i h )
0.4
; step (4.24)
ξ o ″ = [ B h + (d h) 3.5 tan α ]0.2 (d h) ( H i h) 0.4 ; breakwater (4.25)
where Hi is incident wave height, h is water depth at the toe of the breakwater/step, d is
the breakwater/step height, B is the breakwater crest width and α indicates the seaward
slope angle of the step or breakwater. They proposed the following expression for
critical incident wave height “Hc” in order to estimate whether the incident wave height
breaks over the step/breakwater or not:
(
H c h = 1.012 − 1.063 (d h )
0.46
) ; 0.2 ≤ d h ≤ 0.6 ; step (4.26)
72
H c h = 0.952 − 0.591λ0.76 ; 0.2 ≤ λ ≤ 1.4 ; breakwater (4.27)
in which
λ = [(B h ) + (d h ) 2 tan α ]0.4 (d h ) (4.28)
Hara et al. (1992) developed the expression for calculating breaker-depth index “γb” for
the step and breakwater using the proposed modified similarity parameters:
( )
γ b = 5.885 − 5.09 ξ o′ 0.133 ⋅ (h hs ) ; 0.2 ≤ ξ o′ ≤ 1.1 ; step (4.29)
γ b = − 0.463ξ o′′ 0.133 +1.039 ; 0.2 ≤ ξ o′′ ≤ 2.0 ; breakwater (4.30)
Nelson (1994) showed that the upper limit value of breaker-depth index (H/h) over a
horizontal bed is 0.55 and developed the following formula (based on laboratory
experiments) for wave breaking criteria over horizontal bed such as reefs:
⎛ Hi ⎞ Fc
⎜ ⎟ = (4.31)
⎝ h ⎠ max 22 + 1.82 Fc
where
2.5
⎛ g ⎞⎟
0.5
⎛H ⎞
Fc = ⎜ i ⎟ ⋅ ⎜⎜ T (4.32)
⎝ h ⎠ ⎝ h ⎟⎠
Bleck and Oumeraci (2002) proposed a modified Miche criterion to find whether the
wave breaks over rectangular solid submerged breakwater or not:
Hi ⎛ 2π ⎞
= 0.142 tanh⎜⎜ M k ⋅ ⋅ hs ⎟⎟ (4.33)
Li ⎝ Li ⎠
where “Mk” is modification factor of Miche criterion, proposed as 0.74 by Bleck and
Oumeraci based on a set of experimental tests with rectangular submerged breakwater.
73
4.3.2 Analysis of Wave Breaking
The evolution of the waves passing over a submerged breakwater in flume tests was
processed using data collected from wave probes over and behind the structure as well
as visual analysis of the recorded images and video throughout the tests. The tests with
breaking waves (including spilling and plunging) were separated from those with non-
breaking waves for further analysis. Although the experiments were not aimed to
investigate wave breaking phenomena over the breakwater, the results provide
reasonable conceptional representations of the breaking process. Analysis of the results
showed that the minimum value of Hi/hs for breaking waves increases with decreasing
breakwater crest width. The value of breaker-depth index (γb ~ Hb/hs), was found to be
0.42, 0.5 and 0.58 for breakwater crest widths of 3500, 1500 and 300 mm, respectively
(Figure 4.12). This is consistent with the calculated values of breaker index (Table 4.7)
using Hara et al.’s (1992) relationships (Equation 4.26 to 4.30).
The modified breaking criterion of Miche (1944) was also found appropriate to predict
wave breaking over submerged breakwaters. Figure 4.13 indicates that the modified
Miche criterion (Equation 4.33) is suitable to discriminate breaking and non-breaking
waves passing over submerged breakwaters. However, the modification factor “Mk” was
found to be 0.63, that is 15% less than the value of 0.74 proposed by Bleck and
Oumeraci (2002) for rectangular breakwaters.
The knowledge of wave breaking over steep slopes and narrow submerged barrier
(abrupt water changes) has not been well developed yet and further complementary and
detailed studies are needed in this field. In the absence of this research, a breaking
criteria model is however needed to assist in better understanding and modelling of
other phenomenon such as wave-induced current and set-up in presence of submerged
breakwaters. Comparison of the breaker index calculated for each flume test condition
with that predicted using the methods of Nelson (1994), Hara et al. (1992) and modified
Miche are provided in Table 4.7. As indicated in Table 4.7 the calculated values of γb do
not change considerably within the three different methods. Comparison of observed
wave conditions passing over the submerged breakwaters throughout the experiments
with the predicted conditions using the above mentioned methods reveals that the
modified Miche criteria, with modification factor M=0.63 provides the best prediction
74
of wave breaking. Figure 4.14 shows the variation of the relative wave height (Hi/hs)
against the breakwater crest width ratio (B/Lo) for all test conditions (breaking and non-
breaking). The dotted line on the figure shows the minimum value of Hi/hs (and the
limit of breaking) for increasing breakwater crest width ratio. The reduction in critical
Hi/hs with increasing B/Lo for B/Lo < 1.0 is consistent with the work of Hara et al.
(1992), Hattori and Sakai (1994) and Iwata et al. (1996) who reported that the breaker
index reduces with breakwater crest width ratio. Based on the data presented, Hi/hs is
seen to have a value of 0.4 for 1 < B/Lo< 2.5. This value is consistent with the work of
Gourlay (1994). It is likely that the value remains at 0.4 for B/Lo > 2.5 but more
discriminatory test results of higher B/Lo are needed.The narrower crest breakwaters
(B=1500, 300 mm) operate as a barrier rather than a reef and the length of breaking is
not sufficient to fully develop wave breaking (saturated breaking) over the breakwater
while it might be fully developed over the wide crested structures (B=3500 mm). The
variation of calculated breaker-depth index “γb”, applying the three previous mentioned
methods, with respect to the breakwater crest width ratio (B/L) is shown in Figure 4.15.
As indicated in the figure, the calculated breaker-depth index from both Miche and
Nelson formulas are less affected by breakwater crest width. The method of Hara et al.
(1992) predicts γb to more realistically reduce with increasing crest width. The
correlation coefficient between the crest with ratio “B/L” and “γb” is 0.64, 0.23 and 0.20
for the methods of Hara et al., Nelson and Modified Miche, correspondingly. The
expression provided by Nelson (1994) may be applicable for the tests with larger
breakwater crest width (B/L ≥ 1.0) since Equation 4.31 was extracted from the
laboratory experimental data for waves over horizontal bed. Therefore, Hara et al.’s
method seems to be more appropriate to predict the breaker-depth index over
submerged breakwater with different crest width. The formula for calculating breaker
index over steps (Equation 4.29) is more applicable for the wide crested breakwater
(B/L ≥ 1.0) whilst Equation 4.30 is applicable to the prediction of γb over submerged
breakwaters with B/L ≤ 1.0.
4.4 Wave-Induced Set-up
Changes of mean water level due to shoaling, refraction and dissipation of waves occurs
in the surf zone. Longuet-Higgins and Stewart (1964) provided the following
75
momentum balance equation to calculate water level set-up for wave height Hi and
water depth h:
dS xx dη
+ ρg (h + η ) =0 (4.34)
dx dx
where radiation stress Sxx for the wave with radiation angle Θ can be defined:
⎛3 ⎞ 1
S xx = ⎜ p ⎟ EW + EW ⋅ p cos 2Θ (4.35)
⎝2 ⎠ 2
in which
1⎛ 2kh ⎞
p = ⎜⎜1 + ⎟ (4.36)
2 ⎝ sinh 2kh ⎟⎠
and the wave energy EW can be defined in terms of wave height Hi:
1
EW = ρ g H i2 (4.37)
8
Few experimental, theoretical and/or numerical studies have examined the water level set-up behind
submerged breakwaters-this was discussed in Chapter 2. In the current experimental testing, mean water
surface elevation and wave set-up were measured using both Franklin Water Level Follower (FWLF) and
wave probes at different positions along the flume as described in Chapter 3. MWL was calculated from
data enhanced from FWLF and wave probes by subtracting the average of data at each measurement point
from the average of analogous data from the calm flume (no waves):
n n
∑ηi ∑η calm i
η = i =1
− i =1
(4.38)
N N
76
Table 4.7 List of tests with breaking wave and comparison with predicted wave conditions using previous proposed methods.
Nelson (1994) Hara et al. (1992) Modified Miche
test No. h(mm) hs(mm) T(sec) B(mm) Hi(mm)
γb condition γb condition γb condition
t1-1 350 50 0.71 3500 48.8 0.53 breaking 0.50 breaking 0.55 breaking
t1-25 400 100 1.49 3500 52.6 0.54 non-breaking 0.51 breaking 0.56 non-breaking
t2-21 400 100 1.00 1500 56.9 0.52 breaking 0.53 non-breaking 0.55 breaking
t2-25 400 100 1.49 1500 49.5 0.54 non-breaking 0.52 non-breaking 0.56 non-breaking
77
Table 4.7 (Continued) List of tests with breaking wave and comparison with predicted wave conditions using previous proposed methods.
Nelson (1994) Hara et al. (1992) Modified Miche
test No. h(mm) hs(mm) T(sec) B(mm) Hi(mm)
γb condition γb condition γb condition
t2-38 450 150 1.00 1500 86.3 0.51 non-breaking 0.54 non-breaking 0.54 breaking
78
where ηi and ηcalmi represent water surface level data measured by FWLF or wave
probes during the active and calm flume (no waves), respectively. N is the number of
samples so that it is an integer factor of multiplication of sampling frequency to wave
period.
4.4.1 Comparison of the Measurement Methods
The profile of long waves propagating toward a shoal or breakwater gradually loses
their asymmetry. The nonlinear interaction between the breakwater and waves are more
dominant in the shallow water and higher order (nonlinear) wave theories are more
appropriate to simulate the variation of water surface elevation. In the Airy linear wave
theory, water surface (wave profile) is symmetrical, vertically and horizontally. Higher
order wave theories indicate that the water surface is not horizontally symmetric. The
second order wave profile is more peaked at the crest and flatter at the trough than the
first order theory (Figure 4.16) and higher order theory shows more asymmetry in the
wave profile. The rate of nonlinearity or the order of the wave theory may be defined by
parameters such as wave steepness (H/L), relative wave height (H/h) and Ursell number
(Ur). The higher the value of these parameters the higher nonlinear reaction of waves
with the bottom and hence, the higher asymmetry in the wave profile. Due to
asymmetry of wave profile in tests with high Ursell number, the average of measured
water surface levels from wave probes in the flume testings does not produce reliable
values of mean water level.
The asymmetry of water surface can be identified by calculating the Skewness (Sk) of
the measured water surface elevation from the wave probes:
3
1 N
Sk = ∑ (η i − η ) /η o3 (4.39)
N i =1
where
1 N
ηo = ∑ (ηi − η )2 (4.40)
N i =1
79
Calculation of skewness of water surface elevation for the tests without breakwater (T0)
indicated that the variation of skewness had similar trends for both Ursell number and
relative wave height. Figure 4.17 indicates that the skewness for the tests with higher
Ursell number and relative wave height is greater than those with lower Ursell number
(Ur< 20) due to nonlinearity performance of the large waves.
Moreover, when a long wave passes over the submerged breakwater, it may break
depending on wave steepness and breakwater crest width (see Section 4.3). In the case
of relatively high submergence or narrow crested barrier, waves may cease breaking and
reform as a result of increase in water depth behind (shoreward) the breakwater and
propagate toward the shoreline. However, in the flume test measurements of water
surface elevation along the breakwater and behind it, waves do not have adequate
distance within the flume area to fully reform - the partially broken wave has a
complicated profile, which is not symmetrical.
Consequently, the average of measured water surface elevation (enhanced from wave
probes) cannot be an appropriate representation of Mean Water Level (MWL)
specifically in the tests with nonlinear wave behaviour (high value of Ursell number)
due to water depth reduction over the submerged breakwater. The comparison of
measured MWL using Franklin Water Level Follower and wave probes for some
selected tests are shown in Figures 4.18 through 4.20. As discussed in Chapter 3, the
locations of the pressure taps for measuring water level by FWLF are not always at the
same position of the wave probes. However, linear interpolation of water level between
the points gives reasonable results for comparing both methods. The values of
maximum set-down and set-up over and onshore of the submerged breakwater are
defined in Table 4.8 for all test conditions.
4.5 Wave-induced Current
The discharge flux q per unit width (at a given location) due to wave breaking over the
submerged breakwater was calculated by integration of the measured current velocity
over the water depth. Current velocity in x- direction (positive from the paddle to the
beach) at various points above the bed (breakwater top) was calculated by time
averaging the measured horizontal velocity over a duration of 50 seconds with sampling
80
at 20 Hz. Point velocities were calculated from the measured horizontal velocity of 25,
30, 50 and 70 incident waves for wave periods of 2.0, 1.49, 1.0 and 0.71 seconds,
respectively. The total discharge flux q in cm2/sec was calculated by integration of the
measured current velocities over the depth (Figure 4.21), using trapezoidal integration
rule:
N −1
⎛ ⎞
q = 1 × Δz i ⎜Vx1 + 2∑Vx i + Vx N ⎟ (4.41)
2 ⎝ i =2 ⎠
where Vx is the horizontal current velocity in x- direction at the point i, Δzi is the height
of the segment and N is the number of measurement points over the depth. It is assumed
that the horizontal velocity is zero at the bottom and on the water surface (still water
level), it has the value of the velocity of the highest measured point. Figure 4.22 shows
samples of measured horizontal velocity over the breakwater for different incident wave
height. Due to low value of set-up/set-down with respect to the still water depth, the
water level changes over the velocity measurement point has been ignored. The values
of measured discharge flux and the measurement locations over the breakwater for
breaking wave condition are presented in Table 4.9.
4.6 Conclusion
Experimental data measured by different type of probes in the laboratory were

processed in this chapter to investigate wave deformation, incident wave height, water
level changes, wave-induced discharge and wave breaking along the flume and over the
submerged breakwater. Wave deformation was examined in the flume with and without
the breakwater. Spectral analysis of the recorded water surface level in the flume
without breakwater, using Fast Fourier Transformation (FFT) technique, revealed that
higher harmonic waves were generated for the test with Ursell number greater than 10.
The same trend was also observed in the flume offshore of the submerged breakwater.
Moreover, the abrupt changes in water depth over the breakwater produce some higher
harmonic waves and a periodic behaviour of the reflection and transmission coefficient
was observed. Higher harmonic waves were also observed seaward of the breakwater in
the few tests that did not exist in the flume without the breakwater. The reason for
existence of the free harmonic waves offshore the breakwater was found to be
81
retrogressive higher harmonic generated waves due to abrupt changes in the water depth
(the breakwater) along the progressive non-breaking wave path. Accurate examination
of measured water surface deformation along the flume tests (with and without the
submerged breakwater) was carried out to obtain more reliable values of incident and
transmitted wave heights.
Three different methods, the conventional method of Healy (1953), Goda and Suzuki
(1976) and Mansard and Funke (1980) were implemented to calculate incident and
reflected wave height from wave profile data. Comparison of the results obtained from
the different methods revealed that both methods of Mansard and Funke and Goda and
Suzuki are able to calculate incident wave height with reasonable accuracy in
comparison with the conventional method of Healy if wave probes are located at the
appropriate distance from each other. The method of Goda and Suzuki was found more
sensitive to probe spacing and the existence of higher harmonic waves in the flume. The
methods of Healy (1953) and Mansard and Funke (1980) were found more applicable to
calculate incident monochromatic wave height from the recorded water surface data. In
both methods of Healy (1953) and Mansard and Funke (1980), total spectrum (total
energy) of the wave is taken into account which is more realistic with respect to the
amount of energy approaches to the breakwater in the experiments. Finally, data
collected from the traversing probe associated with the conventional method of Healy
were applied in this study to calculate incident wave height due to the availability and
accuracy of the collected data from the traverse probe (this method also being able to
discriminate more reliably the reflected wave heights).
The evolution of waves passing over the submerged breakwater was processed using
data collected from the wave probes over the structure as well as visual analysis of the
recorded images throughout the tests. Analysis of the results showed that the minimum
value of Hi/hs for breaking waves (as a representative of breaker-depth index) increases
slightly with decreasing breakwater crest width. Different developed models of wave
breaking were investigated and the results were compared with the laboratory
experimental observations. The modified breaking criterion of Miche (1944) and Hara
et al. (1992) were found to be appropriate to predict wave breaking over submerged
breakwaters. The proposed breaking model of Hara et al. (1992) showed that breaking
index increases with decreasing the breakwater crest width - the same trend as observed
82
during flume testings. The calculated breaker index will be applied in predicting wave-
induced current over the submerged breakwater in Chapter 5.
The two measurement methods of water surface elevation over and offshore of the
breakwater were also assessed and the results were compared together. The asymmetry
of the wave profile over the breakwater due to high non-linear interaction between
waves and the submerged breakwater indicated that measurement of water surface level
using Franklin water level follower is more accurate and reliable than data collected
from the wave probes.
Finally, wave-induced current over the submerged breakwater were calculated by

integrating the measured velocities over the depth at each measurement points. All
calculated data in this chapter were provided in the corresponding tables and used in
further analysis in Chapters 5 and 7.
83
Table 4.8 Measured maximum set-down and set-up from Franklin Water Level Follower probes.
test No. h(mm) hs(mm) T(sec) B(mm) Lo(mm) Hi(mm) ηmin(mm) ηmax(mm)
t1-1 350 50 0.71 3500 792 48.8 -0.2 2.2
t1-5 350 50 1.00 3500 1470 50.0 -4.1 1.0
t1-9 350 50 1.49 3500 2596 50.3 -2.6 2.1
t1-13 350 50 2.00 3500 3631 55.5 1.1 6.6
t1-14 350 50 2.00 3500 3631 106.3 -12.5 11.0
t1-15 350 50 2.00 3500 3631 146.7 -16.5 14.5
t1-25 400 100 1.49 3500 2729 52.6 -1.2 2.2
t1-26 400 100 1.49 3500 2729 102.4 -8.8 2.2
t1-27 400 100 1.49 3500 2729 145.3 -16.6 9.2
t1-29 400 100 2.00 3500 3859 42.5 -0.7 1.3
t1-30 400 100 2.00 3500 3859 102.5 -2.4 7.7
t1-33 450 150 0.71 3500 786 41.7 0.5 1.7
t1-37 450 150 1.00 3500 1520 54.9 -1.4 0.6
t1-38 450 150 1.00 3500 1520 96.7 -1.5 2.7
t1-41 450 150 1.49 3500 2842 56.7 -0.5 2.7
t1-42 450 150 1.49 3500 2842 99.7 -1.6 10.4
t1-43 450 150 1.49 3500 2842 155.0 -5.7 12.4
t1-45 450 150 2.00 3500 4066 50.5 -0.1 1.3
t2-9 350 50 1.49 1500 2596 43.9 -3.3 0.7
t2-11 350 50 1.49 1500 2596 159.9 -11.5 9.6
t2-14 350 50 2.00 1500 3631 114.7 -4.6 6.3
t2-15 350 50 2.00 1500 3631 145.6 2.2 9.3
t2-22 400 100 1.00 1500 1500 113.3 -5.1 3.7
t2-26 400 100 1.49 1500 2729 107.2 -8.6 0.5
t2-27 400 100 1.49 1500 2729 135.8 -6.7 6.9
t2-30 400 100 2.00 1500 3859 105.9 -6.1 3.6
t2-31 400 100 2.00 1500 3859 136.7 -2.7 8.6
t2-33 450 150 0.71 1500 786 56.2 -0.7 0.5
t2-38 450 150 1.00 1500 1520 86.3 -1.8 1.4
t2-41 450 150 1.49 1500 2842 43.3 -0.5 2.7
t2-42 450 150 1.49 1500 2842 120.7 -0.1 5.0
t2-43 450 150 1.49 1500 2842 152.3 0.9 8.7
t2-46 450 150 2.00 1500 4066 102.0 -5.9 4.5
t2-47 450 150 2.00 1500 4066 155.8 -5.3 2.2
t3-1 350 50 0.71 300 792 52.3 0.0 1.7
t3-5 350 50 1.00 300 1470 69.4 -3.7 1.5
t3-6 350 50 1.00 300 1470 110.4 -5.9 3.5
t3-9 350 50 1.49 300 2596 42.8 -2.5 1.3
t3-10 350 50 1.49 300 2596 96.1 -0.3 4.0
t3-11 350 50 1.49 300 2596 144.7 1.2 5.8
t3-13 350 50 2.00 300 3631 46.7 -0.6 1.5
t3-14 350 50 2.00 300 3631 83.2 -3.5 3.1
t3-15 350 50 2.00 300 3631 98.5 -1.8 4.6
t3-17 400 100 0.71 300 785 51.6 0.1 0.8
t2-37 450 150 1.00 1500 1520 43.7 -0.9 1.3
84
Table 4.8 (continued) Measured maximum set-down and set-up using Franklin Water Level
Follower probe.
test No. h(mm) hs(mm) T(sec) B(mm) Lo(mm) Hi(mm) ηmin(mm) ηmax(mm)
t3-21 400 100 1.00 300 1500 58.4 -0.5 1.1
t3-22 400 100 1.00 300 1500 100.8 -0.6 1.9
t3-25 400 100 1.49 300 2729 39.9 0.3 1.6
t3-26 400 100 1.49 300 2729 103.9 -4.2 3.1
t3-27 400 100 1.49 300 2729 132.4 -5.9 3.6
t3-29 400 100 2.00 300 3859 45.5 0.0 0.7
t3-30 400 100 2.00 300 3859 117.5 -4.0 4.4
t3-31 400 100 2.00 300 3859 170.5 -5.9 7.9
t3-33 450 150 0.71 300 786 43.9 1.0 1.6
t3-37 450 150 1.00 300 1520 37.2 0.5 0.8
t3-38 450 150 1.00 300 1520 89.5 -1.4 0.7
t3-41 450 150 1.49 300 2842 71.9 -0.5 0.7
t3-42 450 150 1.49 300 2842 114.9 -2.1 1.5
t3-43 450 150 1.49 300 2842 133.1 -1.2 3.4
t3-45 450 150 2.00 300 4066 42.8 -0.6 0.2
t3-46 450 150 2.00 300 4066 91.5 -1.2 1.8
t3-47 450 150 2.00 300 4066 147.1 -1.0 5.7
Table 4.9 measured wave-induced discharge over submerged breakwaters (breaking wave
condition)-see Figure 3.19 to 3.21 (Chapter 3) for measurement point locations.
measurement
test No. h(mm) hs(mm) T(sec) B(mm) L(mm) Hi(mm) q(cm2/s)
point
t1-5 350 50 1.00 3500 1470 50.0 27.9 T7
t1-9 350 50 1.49 3500 2596 50.3 212.0 T7
t1-10 350 50 1.49 3500 2596 100.2 160.0 T7
t1-13 350 50 2.00 3500 3631 55.5 144.6 T7
t1-14 350 50 2.00 3500 3631 106.3 300.4 T7
t1-15 350 50 2.00 3500 3631 146.7 297.5 T7
t1-21 400 100 1.00 3500 1500 55.8 163.1 T5
t1-25 400 100 1.49 3500 2729 52.6 176.4 T5
t1-26 400 100 1.49 3500 2729 102.4 352.3 T5
t1-27 400 100 1.49 3500 2729 145.3 461.8 T5
t1-29 400 100 2.00 3500 3859 42.5 108.5 T5
t1-30 400 100 2.00 3500 3859 102.5 254.8 T5
t1-38 450 150 1.00 3500 1520 96.7 184.1 T5
t1-41 450 150 1.49 3500 2842 56.7 0.0 T5
t1-42 450 150 1.49 3500 2842 99.7 315.0 T5
t1-43 450 150 1.49 3500 2842 155.0 474.2 T5
t2-1 350 50 0.71 1500 792 44.7 85.0 T6
t2-5 350 50 1.00 1500 1470 62.2 44.9 T6
t2-9 350 50 1.49 1500 2596 43.9 29.8 T4
t2-11 350 50 1.49 1500 2596 159.9 180.0 T7
t2-14 350 50 2.00 1500 3631 114.7 35.6 T5
t2-15 350 50 2.00 1500 3631 145.6 88.3 T5
t2-21 400 100 1.00 1500 1500 56.9 103.3 T3
t2-22 400 100 1.00 1500 1500 113.3 230.9 T3
85
Table 4.9 (Continued) Measured wave-induced discharge over submerged breakwaters (breaking
wave condition)-see Figure 3.19 to 3.22 (Chapter 3) for measurement point locations.
measurement
test No. h(mm) hs(mm) T(sec) B(mm) L(mm) Hi(mm) q(cm2/s)
point
t2-25 400 100 1.49 1500 2729 49.5 84.0 T3
t2-26 400 100 1.49 1500 2729 107.2 151.8 T3
t2-27 400 100 1.49 1500 2729 135.8 262.1 T3
t2-30 400 100 2.00 1500 3859 105.9 129.5 T3
t2-31 400 100 2.00 1500 3859 136.7 228.7 T3
t2-38 450 150 1.00 1500 1520 86.3 35.5 T3
t2-42 450 150 1.49 1500 2842 120.7 140.0 T3
t2-43 450 150 1.49 1500 2842 152.3 221.0 T2
t2-46 450 150 2.00 1500 4066 102.0 114.7 T3
t2-47 450 150 2.00 1500 4066 155.8 207.7 T3
t3-1 350 50 0.71 300 792 52.3 -9.3 T6
t3-5 350 50 1.00 300 1470 69.4 15.9 T6
t3-6 350 50 1.00 300 1470 110.4 -8.8 T6
t3-9 350 50 1.49 300 2596 42.8 37.3 T6
t3-10 350 50 1.49 300 2596 96.1 47.2 T6
t3-11 350 50 1.49 300 2596 144.7 111.9 T6
t3-13 350 50 2.00 300 3631 46.7 100.4 T6
t3-14 350 50 2.00 300 3631 83.2 64.9 T6
t3-15 350 50 2.00 300 3631 98.5 63.0 T6
t3-21 400 100 1.00 300 1500 58.4 42.7 T2
t3-22 400 100 1.00 300 1500 100.8 70.8 T2
t3-26 400 100 1.49 300 2729 103.9 131.4 T6
t3-27 400 100 1.49 300 2729 132.4 202.3 T6
t3-30 400 100 2.00 300 3859 117.5 146.4 T2
t3-31 400 100 2.00 300 3859 170.5 226.7 T6
t3-38 450 150 1.00 300 1520 89.5 32.6 T2
t3-42 450 150 1.49 300 2842 114.9 27.1 T2
t3-43 450 150 1.49 300 2842 133.1 70.8 T2
t3-46 450 150 2.00 300 4066 91.5 18.0 T2
t3-47 450 150 2.00 300 4066 147.1 69.5 T2
86
120
Probe 3
80
water elvation (mm)

40
0
30 31 32 33 34 35
-40
-80
Time (sec)
120
Probe 4
80
water elvation (mm)
40
0
30 31 32 33 34 35
-40
-80
Time (sec)
80
Probe 5
40
water elvation (mm)
0
30 31 32 33 34 35
-40
-80
Time (sec)
80
Probe 6
40
water elvation (mm)
0
30 31 32 33 34 35
-40
-80
Time (sec)
Figure 4.1. Variation of wave profile at different stations of probes in the empty flume
(without breakwater, h=350 mm, Hi=114 mm & T=1.49 sec)- probe locations are shown
in Figure 3.19, Chapter 3.
Figure 4.2. Spectrum of water level at the position of probe4 (13.0 m onshore the paddle- see Figure 3.19) for the selected tests without
breakwater (for test conditions of the test numbers see Appendix A).
60
(a) no harmonic generation
observed
50
harmonic generation
observed
40
UUR
30
R
20
10
0
0 5 10 15 20 25 30 35
test range
0.6
(b)
0.5
0.4
h/L
0.3
0.2
0.1
0.0
0 5 10 15 20 25 30 35
test range
0.07
(c)
0.06
0.05
0.04
Hi/L
0.03
0.02
0.01
0.00
0 5 10 15 20 25 30 35
test range
Figure 4.3. Variation of dimensionless parameters in the test without breakwater for
harmonic generation and no higher harmonic circumstances.
Figure 4.4. Wave regime around a shelf and a rectangular barrier (Goda et al., 1999)
Figure 4.5. Spectrum for water surface elevation for the selected tests of T1 (B=3500
mm), T2 (B=1500 mm) and T3 (B=300 mm) in front, over and behind the breakwater-
see Figure 3.20 to 3.22 for wave probes locations).
Figure 4.5. (Continued) spectrum for water surface elevation for the selected tests of
T1 (B=3500 mm), T2 (B=1500 mm) and T3 (B=300 mm) in front, over and behind the
breakwater- see Figure 3.20 to 3.22 for wave probes locations).
Figure 4.6. Comparison of zero up-crossing (blue solid line) and down-crossing
methods (red dash line) for calculating wave height from the traverse wave probe.
Figure 4.7. Comparison of smoothed and non-smoothed wave height for zero down-
crossing method.
Figure 4.8. Calculated incident and reflected spectrum using the method of Goda and
Suzuki (1976) for test T0-26 (h=400mm, T=1.49 sec and Hi=87mm; no breakwater).
(a) (b)
160
+10%
B=300 mm 160 No breakwater
H i (mm, Goda & Suzuki method)
H i (mm, Goda & Suzuki method)
120 -10% 120
80 80
40 Tests in the proposed range of 40

Goda
Higher harmonic generation
observed
0 0
0 40 80 120 160 0 40 80 120 160
H i (mm, conventional method) H i (mm, conventional method)
Figure 4.9. Comparison of conventional method of Healy (1953) for calculation of

incident wave height with Goda & Suzuki (1976) method. (a) T3 (B=300 mm) (b) T0
(without breakwater).
Figure 4.10. Calculated incident and reflected spectrum using the method of Mansard
and Funke (1980) for test T3-30 (B=300mm, h=400mm, T=2.0 sec and Hi=117mm).
160
+10%
Hi (mm, Mansard & Funke method)
-10%
120
80
Tests in the proposed range

40
Tests out of the proposed range
Higher harmonic generation

observed
0
0 40 80 120 160
H i (mm, conventional method)
Figure 4.11. Comparison of conventional method for calculation of incident wave

height with Mansard & Funke (1980) method (Test3, B=300 mm).
1
0.8
0.6 0.58
Hi /hs
0.5
0.42
0.4 Breaking,B=3500mm
nonbreaking, B=3500mm
Breaking, B=1500mm
0.2 nonbreaking, B=1500mm
Breaking, B=300mm
nonbreaking, B=300mm
0
0 0.02 0.04 0.06 0.08
H i /L
Figure 4.12. Variation of inverse submergence ratio “Hi/hs” as a function of wave

steepness for breaking and non-breaking waves. The minimum values of Hi/hs for
breaking waves are shown with blue, red and lime dash-dot lines for breakwater crest
width B=3500, 1500 & 300mm, respectively.
breaking,B=3500mm nonbreaking, B=3500mm

Breaking, B=1500mm nonbreaking, B=1500mm
Modified Miche criteria Miche criteria
Bleck & Oumeraci (2002)
0.1
0.08
0.06
Hi /L
H i /L=0.142tanh(0.63kh s )
0.04
0.02
0
0 0.05 0.1 h /L 0.15 0.2
s
Figure 4.13. Variation of wave steepness in relation to the ratio of hs/L for breaking
and non-breaking waves. The modified Miche breaking criteria is shown with solid
line.
breaking, B=3500mm nonbreaking, B=3500mm
1
Breaking
0.8
0.6
Hi/hs
0.4
Non-breaking
0.2
0
0 1 2 3 4 5
B/L o
Figure 4.14. Variation of wave height ratio against relative breakwater crest width for
breaking and non-breaking waves over the breakwater.
0.58
0.56
Breaker-depth index
0.54
0.52
0.50
Hara et al. (1992)
Modified Miche
Nelson (1994)
0.48
0 0.5 1 1.5 2 2.5 3
B/L o
Figure 4.15. Calculated breaker-depth index “γb” as a function of dimensionless

breakwater crest width using different methods.
Figure 4.16. Wave profile shape of different progressive gravity waves (CERC,
2004).
60 1.4
(a) Ur
Sk 1.2
50
1
40
0.8
Sk
30 0.6
Ur
0.4
20
0.2
10
0
0 -0.2
0 5 10 15 20 25 30
Test Number (T0-x)
0.6 1.4
(b) Hi/d
Sk 1.2
0.5
1
0.4
0.8
Hi/d
0.3
Sk
0.6
0.4
0.2
0.2
0.1
0
0.0 -0.2
0 5 10 15 20 25 30
Test Number (T0-x)
Figure 4.17. (a) Variation of Ursell number and skewness of water surface elevation
for the test without breakwater, (b) variation of relative wave height and skewness of
water surface level for the test without breakwater (Test T0).
Test T1-1 Test T1-5 Test T1-9
20 20 20
10 10 10
mean water level (mm)

0 0 0
-10 -10 -10
-20 -20 -20

FWLF
-30 -30 -30
wave probe
-40 -40 -40
breakwater crest
-50 -50
-50
1000 1100 1200 1300 1400 1500 1600 1700 1800 1000 1100 1200 1300 1400 1500 1600 1700 1800
1000 1100 1200 1300 1400 1500 1600 1700 1800
distance from paddle (cm) distance from paddle (cm) distance from paddle (cm)
Test T1-10 Test T1-14-2 Test T1-15-2

20 20 20
10 10 10

mean water level (cm)

0 0 0
-10 -10 -10
-20 -20 -20
-30 -30 -30
-40 -40
-40
-50 -50
-50
1000 1100 1200 1300 1400 1500 1600 1700 1800 1100 1200 1300 1400 1500 1600 1700 1800
1100 1200 1300 1400 1500 1600 1700 1800
distance from paddle distance from paddle (cm) distance from paddle (cm)

40
20 30
20
0
mwan water level (mm)
mean water leve (mm)

0
-20 -30
-20
-40 -60
-40
-60 -90
-60
-80 -120
-80
-100 -150
-100
1100 1200 1300 1400 1500 1600 1700 1800 1100 1200 1300 1400 1500 1600 1700 1800
1100 1200 1300 1400 1500 1600 1700 1800
Figure 4.18. Measured wave-induced set-up over and behind the breakwater using Franklin Water Level Follower and wave probes (B=3500
mm).
20 10 20
10 0 10

0 0
-10
-10 -10
-20
-20 -20
-30
-30 -30
-40
-40 -40
-50 -50 -50

1300 1400 1500 1600 1700 1800 1300 1400 1500 1600 1700 1800 1300 1400 1500 1600 1700 1800
distance from paddle (cm) distance from paddle (cm) distance frpm paddle (cm)
Test T2-14-1 Test T2-15-1 Test T2-25

20 20 20
10 10 0

0 0
-20
-10 -10
-40
-20 -20
-30 -60
-30
-40 -80
-40
-50 -100
-50
1300 1400 1500 1600 1700 1800 1300 1400 1500 1600 1700 1800
1300 1400 1500 1600 1700 1800
distance from paddle (cm) distance from paddle (cm) distance fromm paddle (cm)

20 20 30
mean water elevation (mm)
0 0 0
mean water level

-20 -20 -30
-40 -40 -60
-60 -60 -90
-80 -80 -120
-100 -100 -150

1300 1400 1500 1600 1700 1800 1300 1400 1500 1600 1700 1800 1300 1400 1500 1600 1700 1800
mm).
20 20
20
10 10 0


0 0 -20
-10 -10
-40
-20 -20
-60
-30 -30
-40 -40 -80
-50 -50 -100

1300 1400 1500 1600 1700 1800 1300 1400 1500 1600 1700 1800 1300 1400 1500 1600 1700 1800

20 20 20
0 0 0

-20 -20 -20
-40 -40 -40
-60 -60 -60

-80
-80 -80
-100
-100 -100
1300 1400 1500 1600 1700 1800
1300 1400 1500 1600 1700 1800 1300 1400 1500 1600 1700 1800

20 20 30
0 0 0

-20 -20 -30
-40 -40 -60
-60 -60 -90
-80 -80 -120
-100 -100 -150

1300 1400 1500 1600 1700 1800 1300 1400 1500 1600 1700 1800 1300 1400 1500 1600 1700 1800
distance from paddle (cm) distnace from paddle (cm) distance from paddle (cm)
mm).
Vxi
Vxi-1 Δzi
zi
Figure 4.21. Current velocity profile definition over the breakwater.
3.0 T1-25 (Hi=53 mm)

T1-26 (Hi=102 mm
T1-27 (Hi=145mm)
2.5
Water depth (cm)
2.0
1.5
1.0
0.5
0.0
0.0 10.0 20.0 30.0 40.0 50.0 60.0
Velocity (cm/s)
Figure 4.22. Variation of velocity profile over the breakwater

(B=3500 mm, T=1.49 s and hs=100 mm).
CHAPTER 5
ANALYSIS AND DISCUSSION OF EXPERIMENTAL

RESULTS
Processed data collected from all experiment conditions is analysed in detail and
discussed in this chapter. Wave height transformation over the submerged breakwater,
wave-induced set-up and wave-induced current are analysed and the effectiveness of
several parameters (dimensional and dimensionless) are examined for improved
understanding of hydrodynamic processes. Previously published analytical/experimental
studies for calculating wave transformation, wave-induced set-up and current are
discussed and compared with the present experimental results. Improvements to the
more appropriate previous models are proposed.
5.1 Wave Transmission
5.1.1 Transmitted wave height
The transmitted wave height passing over the experimental submerged breakwaters was
measured using three resistance wire probes located as shown in Figure 3.19 through
3.21, Chapter 3. Transmitted wave onshore of the submerged breakwater includes
higher harmonic free waves, which are generated over the step (breakwater) and
transmitted to the deeper water behind the breakwater as free waves. Hence wave height
behind the submerged breakwater varies spatially and depends on the number of
significant harmonics generated and transmitted over the breakwater. This phenomenon
was clearly observed throughout the laboratory experimental tests and discussed in
detail in Chapter 4.
In order to obtain more accurate measurement results of transmitted wave height behind
the breakwater, different approaches were applied and compared. The average of root
mean squared wave height (Ĥrms) from the three wave probes behind the structure was
calculated with the following three methods:
87
i- Average of Hrms of each probe calculated from the simple statistic method
∑σ i
Hˆ rms t 1 = 2 2 i =1
(5.1)
n
ii- Average of Hrms calculated from zero moment of each spectrum
∑ m0i
Hˆ rmst 2 = 2 2 i =1
(5.2)
n
iii- Square root of mean zero moment of the spectrums
∑m 0i
Hˆ rmst 3 = 2 2 i =1
(5.3)
n
Where σi is the standard deviation of the water surface level enhanced by ith wave probe
behind the breakwater, n is the number of wave probes behind the submerged
breakwater (n=3 in the tests) and m0i is the zero moment of the spectrum of ith wave
probe :
m0 = ∫ S ( f ) df (5.4)
Comparison of the three abovementioned methods for different tests with breaking
waves (Table 5.1) indicates no significant difference between the methods of calculating
transmitted wave heights. In this study, the average of Hrms1 obtained from all three
probes behind the breakwater (Equation 5.1) was adopted as being representative of
transmitted wave height. It should be noted that the calculated transmitted wave height
is an average and includes the fundamental and higher generated harmonic waves since
the total spectrum (wave energy) was taken into account in the calculations. This is
consistent with the total energy of wave approaches to the breakwater that was
considered in the incident wave height calculation.
88
Table 5.1 The average transmitted wave height applying different methods and calculated transmission coefficients for all test conditions with breaking wave
over the submerged breakwater.
Ĥrmst (mm) Ĥrmst (mm)
Test h hs T B Lo Hi Test h hs T B Lo Hi
method method method Kt method method method Kt
No. mm mm sec mm mm mm No. mm mm sec mm mm mm
i ii iii I II III
T1-1 350 50 0.71 3500 792 49 23 23 24 0.47 T2-26 400 100 1.49 1500 2729 107 66 66 66 0.62
T1-5 350 50 1.00 3500 1470 50 24 24 24 0.48 T2-27 400 100 1.49 1500 2729 136 82 82 83 0.60
T1-9 350 50 1.49 3500 2596 50 23 23 24 0.46 T2-30 400 100 2.00 1500 3859 106 64 64 65 0.61
T1-10 350 50 1.49 3500 2596 100 34 34 35 0.34 T2-31 400 100 2.00 1500 3859 137 92 92 92 0.67
T1-13 350 50 2.00 3500 3631 56 47 44 44 0.61 T2-38 450 150 1.00 1500 1520 86 66 66 66 0.77
T1-14 350 50 2.00 3500 3631 106 47 47 47 0.44 T2-42 450 150 1.49 1500 2842 121 75 75 77 0.62
T1-15 350 50 2.00 3500 3631 147 59 59 60 0.40 T2-43 450 150 1.49 1500 2842 152 90 90 91 0.59
T1-21 400 100 1.00 3500 1500 56 28 28 28 0.50 T2-46 450 150 2.00 1500 4066 102 73 73 75 0.70
T1-25 400 100 1.49 3500 2729 53 40 40 40 0.76 T2-47 450 150 2.00 1500 4066 156 101 101 102 0.65
T1-26 400 100 1.49 3500 2729 102 47 47 48 0.46 T3-1 350 50 0.71 300 792 52 26 26 27 0.50
T1-27 400 100 1.49 3500 2729 145 68 68 70 0.47 T3-5 350 50 1.00 300 1470 69 42 42 43 0.60
T1-42 450 150 1.49 3500 2842 100 54 54 54 0.54 T3-6 350 50 1.00 300 1470 110 60 60 60 0.54
T1-43 450 150 1.49 3500 2842 155 69 69 70 0.45 T3-9 350 50 1.49 300 2596 43 36 36 36 0.85
T1-46 450 150 2.00 3500 4066 95 66 66 70 0.69 T3-10 350 50 1.49 300 2596 96 77 77 77 0.80
T2-1 350 50 0.71 1500 792 45 16 16 16 0.35 T3-11 350 50 1.49 300 2596 145 111 111 111 0.76
T2-5 350 50 1.00 1500 1470 62 25 25 25 0.40 T3-13 350 50 2.00 300 3631 47 45 45 45 0.97
T2-6 350 50 1.00 1500 1470 104 41 41 41 0.39 T3-14 350 50 2.00 300 3631 83 69 69 70 0.83
T2-9 350 50 1.49 1500 2596 44 31 31 31 0.70 T3-15 350 50 2.00 300 3631 98 85 85 85 0.86
T2-10 350 50 1.49 1500 2596 99 51 51 51 0.52 T3-21 400 100 1.00 300 1500 58 52 52 52 0.89
T2-11 350 50 1.49 1500 2596 160 80 80 80 0.50 T3-22 400 100 1.00 300 1500 101 58 58 59 0.57
T2-13 350 50 2.00 1500 3631 62 38 38 40 0.62 T3-26 400 100 1.49 300 2729 104 83 83 84 0.80
T2-14 350 50 2.00 1500 3631 115 66 66 67 0.57 T3-27 400 100 1.49 300 2729 132 111 111 112 0.84
T2-15 350 50 2.00 1500 3631 146 69 69 71 0.48 T3-30 400 100 2.00 300 3859 117 101 101 101 0.86
T2-21 400 100 1.00 1500 1500 57 40 40 41 0.71 T3-31 400 100 2.00 300 3859 171 128 128 129 0.75
T2-22 400 100 1.00 1500 1500 113 59 59 60 0.52 T3-42 450 150 1.49 300 2842 115 95 95 95 0.82
T2-25 400 100 1.49 1500 2729 50 42 42 43 0.86 T3-43 450 150 1.49 300 2842 133 94 94 95 0.70
89
5.1.2 Transmission Coefficient
The transmission coefficient of a smooth impermeable submerged breakwater was

calculated using data enhanced from the probes offshore and onshore the breakwater
throughout the tests. The transmission coefficient was calculated as the ratio of the
average transmitted wave height, obtained from the previous section, to the incident
wave height enhanced from the traverse wave probe (Healy’s method) presented in
Chapter 4:
Hˆ rms
Kt = (5.5)
Hi
The values of transmission coefficient for the tests with breaking wave over the
submerged breakwater are provided in Table 5.1.
5.1.3 Effective Parameters for prediction of Kt
Analysis of the transmission coefficient and the effects of different variables

(dimensional and non-dimensional) on Kt were examined both graphically and
statistically. The purpose of this analysis being to define which parameters can most
effectively predict the transmission coefficient. Predicted transmitted wave heights will
also be used in the numerical model for calculating mass flux over the breakwater and
will be discussed in Chapter 7.
The effect of dimensional parameters including wave characteristics, breakwater

geometry and water depth, has been previously investigated by Seelig (1980), Ahrens
(1987), van der Meer (1991) and Seabrook (1997). The variation of transmission
coefficient is plotted in Figures 5.1 through 5.4 as a function of incident wave height
“Hi”, wave period “T ”, breakwater crest width “B ” and submergence depth “hs” (water
depth over the breakwater) for the selected measured data (Table 5.1). The graphical
analysis of data demonstrates transmission coefficient increases with submergence
depth and wave period whilst, breakwater crest width and incident wave height have
inverse effect on Kt which are consistent with the previous studies. However, it is not
possible to confirm, visually, the presence of the trends and statistical analysis of the
variables is needed to determine the significance of each parameter. The variation of
90
correlation coefficient between Kt and other dimensional parameters are plotted in
Figures 5.5 and 5.6. The correlation coefficient can be used as an indicator of
effectiveness of the dimensional parameters on transmission coefficient. It can be
deduced from the figures that the effect of each parameter on transmission coefficient
depends on the other parameters as follows:
• The submergence depth “hs” and incident wave height “Hi ” are more effective
for the wider crest width breakwater while wave period “T ” is more effectual
for narrower breakwater.
• The impact of wave period and submergence depth is stronger for smaller wave
heights whilst for higher waves the breakwater crest width is more effective.
The above results provide valuable guidance for the geometry design of submerged
breakwaters when wave climates are available. However, non-dimensional parameters
can provide a more effective correlation of the above dependent parameters. Based on
dimensional analysis and review of the single parameter analysis above, the following
dimensionless parameters were considered for further analysis:
B hs H H
Kt = f ( , , i , i ) (5.6)
Lo H i Lo B
in which the deepwater wave length Lo is employed instead of wave period “T ”.
Various plots of the variation of transmission coefficient as a function of the above non-
dimensional parameters are presented in Figure 5.7. Graphical analysis of the
dimensionless parameters demonstrates that transmission coefficient increases with
decreasing crest width of breakwater. The reason for this trend can be found in the
length of shallow water over the breakwater to develop wave breaking. For the wider
breakwater, there is sufficient length of shallow water (water depth less than breaking
depth). The breaking process can be developed well and more energy is released as the
wave breaks over the breakwater. It should be noted that the wave transmission analysis
has been made only for the tests with breaking waves. It is assumed that wave breaking
is the main reason for reduction of transmitted wave height behind the breakwater and
friction loss is negligible as the breakwaters with smooth surface were employed
91
through the tests. More discussion is presented in chapter 7 regarding the effects of
bottom friction and other parameters on wave dissipation over the breakwater.
Based on graphical interpretation of Figure 5.7, the effectiveness of the dimensionless

parameters on wave transmission coefficient is described as follows:
• Dimensionless crest width “B/Lo” has an inverse effect on Kt. For the test with
narrower breakwater crest width (B/Lo < 1.0 – 1.0<Lo/B ), transmission
coefficient is more sensitive to B/Lo- effect is less for wider breakwater. It may
be deduced that for breakwater crest width greater than incident wave length the
variation of crest width has small effect on transmission coefficient since the
length of breaking depth area is long enough to encourage wave breaking to be
developed.
• The submergence ratio “hs/Hi ” has direct influence on transmission coefficient.

Increasing submergence ratio allows waves to pass over the breakwater with
less breaking and more energy transfers onshore. Variation of transmission
coefficient with respect to submergence ratio for different breakwater crest
width (Figure 5.7b) reveals that the submergence ratio has more effect in the
tests with wider breakwater.
• Increasing incident wave steepness (Hi/Lo) causes less transmission coefficient

due to stronger wave breaking over the breakwater.
• Transmission coefficient increases with an increasing ratio of incident wave

height to breakwater crest width (wave height ratio). However, the effectiveness
of relative wave height decreases when Hi/B ≥ 0.1. In this case, the incident
wave breaks at the offshore slope of the breakwater. Depth-induced energy
dissipation occurs at the beginning of the breakwater crest and transmission
coefficient (energy dissipation) is no longer dependent on the width of the
breakwater.
Statistical analysis of the selected dimensionless variables was performed for more
investigation of effectiveness of the non-dimensional parameters. The database was
92
subjected to linear correlation test (Pearson) to investigate the relative dependency of
the transmission coefficient on the above independent variables. Although the linear
correlation coefficient is not conclusive for a non-linear relationship, it provides a
general estimation of the effectiveness of the variables. The correlation coefficient
between Kt and other dimensionless variables were calculated for all tests with breaking
wave over the submerged breakwater and the values are provided in Table 5.2. Note that
in some cases the inverse ratio of the dimensionless variables (Lo/B and Lo/Hi) has been
adopted in the table to keep the positive values of correlation coefficients. The
correlation coefficients indicate that the inverse breakwater crest width ratio “Lo/B” , the
ratio of incident wave height to breakwater crest width “Hi/B ”, inverse ratio of wave
steepness “Lo/Hi ” and the submergence ratio “hs/Hi” are respectively the parameters
most effecting transmission coefficient.
Table 5.2 Correlation coefficient between Kt and other dimensionless variables.

Dimensionless parameter Lo/Hi Lo/B Hi/B hs/Hi
Correlation coefficient 0.47 0.77 0.55 0.38
5.1.4 Comparison with the Previous Models
The measured transmission coefficients from the experimental tests were compared with
four recent empirical prediction formulas developed by d’Angremond et al. (1996),
Bleck and Oumeraci (2002), Calabrese et al. (2002) and Van der Meer et al. (2004). All
the prediction formulas were fully illustrated in Chapter 2. Graphical comparison of the
calculated transmission coefficient using the proposed formulas with the measured
values derived from the laboratory experiments is plotted in Figure 5.8. The main
reason for the differences between the predictions by the formulae and the experimental
measured values is considered due to the range of test conditions. Table 5.3
demonstrates the test conditions and limitation of the proposed formulae associated with
the present experimental test conditions. As indicated from the table, the present
experimental data covers a wide range of all parameters whilst the mentioned existing
equations are applicable in limited range of the parameters. For instance, high
93
discrepancy between measured and calculated transmission coefficient is observed if the
model of Calabrese et al. (2002) is applied for 8.13<B/ Hi .
The quantitative comparison of the derived transmission coefficient from the

experiments with the existing models was carried out using two statistical indicators,
distortion “β ”, and Root Mean Squared Error “RMSE ” as illustrated in Chapter 4
(Equations 4.18 and 4.19). The values of the two indicators are identified in Figure 5.9.
The closer value of distortion to 1.0 and the lower value of RMSE confirm that formula
of d’Angremond et al. is more reliable than the other methods. Across the extended
range of test results, the developed model by Calabrese et al. (2004) for rubble mound
breakwater also gives sensible results in comparison with the measured data within the
proposed range of B/Hi < 8.13. The values of β and RMSE in Figure 5.9 were calculated
using the data within tested ranges proposed by the indicated authors.
van der Meer et al. (2004) improved the model of d’Angremond et al. (1996) by
adopting different coefficients within the equations. van der Meer et al. proposed two
different expressions (Equations 2.37 and 2.38) for smooth and impermeable
breakwaters considering the breakwater crest width is more effective in the test with
Iribarren number “ξ ” greater than 3.0. Investigation of the trend of transmission
coefficient with respect to “B/Hi ” for the tests with Iribarren number greater and
smaller than 3.0, implies that transmission coefficient is still a function of “B/Hi ” for
the tests with ξ < 3.0. Figure 5.10 shows the variation of experimental Kt against “B/Hi”
for different Iribarren numbers. As indicated in the figure the correlation between
transmission coefficient and breakwater crest width ratio is almost the same for tests
with 3.0 ≤ ξ and ξ < 3.0. Therefore, it seems that the effectiveness of B/Hi should be
taken into account in the transmission coefficient formula for all conditions of wave
steepness.
As indicated before, the present laboratory tests cover a wide range of dimensionless
parameters, which are beyond the limit of the existing methods. Hence, some
modification to the recent proposed expressions is needed to improve the prediction of
transmission coefficient over the wider range of the present data.
94
Table 5.3 Test limitation for the previous proposed transmission coefficient formulas and the
present experimental data.
hs/B B/Hi Hi/h Hi/Lo ξ
Database
min max min max min max min max min Max
d'Angremond et al. (1996) - - - - - 0.54 - 0.60 - -
Bleck and Oumeraci (2001) 0.10 0.60 2.50 12.50 0.11 0.29 0.0014 0.10 - -
Calabrese et al. (2002) -0.30 0.40 1.06 8.13 0.31 0.61 - - 3.00 5.20
Present experiment 0.01 0.50 1.76 71.70 0.12 0.46 0.01 0.08 1.82 4.41
5.1.5 Proposed Empirical Model
Comparison of the existing empirical methods with the measured transmission

coefficient revealed that the models are not applicable in all range of the laboratory
experimental data. Therefore, model modifications are investigated in this section to
develop a more reliable expression for calculating transmission coefficient over
submerged breakwaters with wide range of crest width ratio.
On the base of graphical and statistical analysis, four dimensionless variables were
adopted for investigating a design equation for transmission coefficient:
Ht ⎛h H B L ⎞
Kt = = f ⎜⎜ s , i , , ξ or o ⎟⎟ (5.7)
Hi ⎝ H i B Lo Hi ⎠
The dimensionless parameter, Iribarren numberξ, can be applied as the representative of

wave steepness since it is commonly used in breaking wave analysis in the surf zone.
Three forms of equations were considered for developing an appropriate design
equation of Kt.
The first function was a non-linear equation including submergence ratio, wave height
to breakwater crest with ratio, dimensionless crest width and inverse ratio of wave
steepness:
b2 b4 b6 b8
⎛h ⎞ ⎛ B ⎞ ⎛L ⎞ ⎛B⎞
K t = b1 ⎜⎜ s ⎟⎟ + b3 ⎜⎜ ⎟⎟ + b5 ⎜⎜ o ⎟⎟ + b7 ⎜⎜ ⎟⎟ (5.8)
⎝ Hi ⎠ ⎝ Hi ⎠ ⎝ Hi ⎠ ⎝ Lo ⎠
where the coefficients b1 to b2 were to be defined by non-linear regression analysis.
95
The second form of the equation is similar to Equation 5.8 except that the inverse ratio
of wave steepness was substituted by the Iribarren number:
b2 b4
⎛h ⎞ ⎛ B ⎞
b
⎟⎟ + b5 (ξ )b6 + b7 ⎛⎜ ⎞⎟
B 8
K t = b1 ⎜⎜ s ⎟⎟ + b3 ⎜⎜ (5.9)
⎝ Hi ⎠ ⎝ Hi ⎠ ⎝L⎠
The third form of the equation is similar in form to that proposed by d’Angremond et al.
(1996). As discussed in the previous section, the expression developed by d’Angremond
et al. is beter correlated with the present data:
b3
⎛h ⎞ ⎛ B ⎞
K t = b1 ⎜⎜ s ⎟⎟ + b2 ⎜⎜ ⎟⎟ ⋅ (1 − exp(b4 ⋅ξ )) (5.10)
⎝ Hi ⎠ ⎝ Hi ⎠
in which the coefficients b1 to b4 will be determined by non-linear regression analysis.
Nonlinear regression analysis was carried out using MATLAB software package. Some
scripts were developed to determine the parameters based on nonlinear least-squared
regression. Totally, 10 different design equations for Kt were investigated, applying
Equations 5.8 through 5.10. The calculated values of calibration coefficients b1 to b8
were derived from regression analysis and provided in Table 5.4. The models 1 and 2 in
Table 5.4 correspond to Equation 5.8 and the models 3 and 4 derived from Equation 5.9.
The last term on the right hand side of Equations 5.8 and 5.9 were ignored,
correspondingly, in the models 2 and 4 to assess the sensitivity of Kt to breakwater crest
width ratio B/L- coefficients b7 and b8 were considered zero in regression analysis of the
models 2 and 4. The models 5 through 10 are extracted from the structure of
d’Angremond et al.’s formula (Equation 5.10). In the models 5 through 8, the parameter
b4 was fixed. In the model 9 both coefficients b4 and b3 were considered constant of
prescribed values. The fixed values for b4 and b3 being consistent with values proposed
by van der Meer et al. (2004). The other parameters in the models 5 to 9 were
determined through regression analysis.
Graphical comparison of the measured transmission coefficient with the calculated

values using the proposed equations is displayed in Figure 5.11. All data in these
regression analysis of the models correspond to the experiments with breaking waves
96
over the submerged breakwater. Comparison of the figures for model 1 with model 2
and model 3 with model 4 in Figure 5.11 indicates that the crest width ratio (B/Lo) has
little influence on transmission coefficient. Both dimensional parameters, B and Lo have
been employed in other non-dimensional parameters and the combination of these
values can be ignored from the corresponding equations.
The statistical evaluation of the models was carried out by calculating five statistical
indicators. Bias or distortion (β), Root Mean Squared Error (RMSE) and R2 were
defined in Chapter 4 (see Equations 4.18 through 4.20). Two other indicators, Wilmott
number IW (Wilmott, 1981) and error function ε (Haller et al., 2002), were also
implemented for more discussion of the validity of the various models:
∑ (X )
N
− X mk
2
ck
IW = 1 − k =1
(5.11)
∑( X )
N 2
ck − X m + X mk − X m
k =1
12
⎡ N
(
⎢ ∑ X ck − X mk ) ⎤⎥
2
ε = ⎢ k =1 N ⎥ (5.12)
⎢ ⎥
⎢⎣ ∑
k =1
X mk
2
⎥⎦
where Xc are the calculated values, Xm are the measured (available data) values and the
overbars indicate the average value of the parameter. Perfect agreement is indicated if
the R2, bias β and Wilmott index IW are 1.0 and the error function ε and RMSE are zero.
R2 is considered unsatisfactory as a measure of the goodness of fit for a multivariate
regression relationship (Draper, 1984) and thus the validity of the model will be
assessed mostly by the other parameters. The values of the five statistical parameters are
provided in Table 5.4 and graphically compared within the models in Figure 5.12.
Graphical and statistical comparison of the measured transmission coefficient with the
calculated values from the proposed model reveal that models 1 to 5 and 10 provide
reasonable values of transmission coefficient with respect to the laboratory
measurements. The values of IW and β are greater than 0.96 and the error, RMSE and ε,
97
are less than 0.10. Model 3 and Model 10 present more marginally accurate results in
comparison with the other models:
2.38 0.09
⎛h ⎞ ⎛ B ⎞
0.96
⎛B⎞
K t = 0.06 ⎜⎜ s ⎟⎟ − 1.62 ⎜⎜ ⎟⎟ + 2.08 (ξ ) + 0.09⎜ ⎟
0.20
(5.13)
⎝ Hi ⎠ ⎝ Hi ⎠ ⎝L⎠
−0.26
⎛h ⎞ ⎛ B ⎞
K t = 0.17 ⎜⎜ s ⎟⎟ + 2.84 ⎜⎜ ⎟⎟ ⋅ (1 − exp(− 0.14 ∗ξ )) (5.14)
⎝ Hi ⎠ ⎝ Hi ⎠
The structure of Equation 5.14 (model 10) is less complicated than Equation 5.13
(model 3). Moreover, it is more common in the literature as the predictive form was
reported by d’Angremond et al. (1996) and has been used by other authors. Comparison
of the measured and calculated Kt using Equation 5.14 for different breakwater crest
width is shown in Figure 5.13. As implied from the figure, the proposed equation
provides very sensible results of transmission coefficient over submerged breakwaters
for a range of breaking wave conditions and different crest width ratios. Therefore,
Equation 5.14 is proposed and adopted for calculating transmission coefficient over
smooth submerged breakwaters and will be implemented in the numerical modelling of
submerged breakwater in Chapter 7.
5.2 Wave-induced Flow
Wave-induced discharge flux q (discharge per unit width) over the submerged
breakwater was calculated by integration of the measured current velocity as discussed
in Chapter 4 (Section 4.5). Velocity profiles over the structure were examined in the
previous Chapters and found to be essentially one directional (horizontal) and almost
uniform with depth for fully breaking wave conditions.
The dependency of discharge on dimensional and non-dimensional parameters is

investigated in this section. Moreover, the measured current from the laboratory
experiments will be compared to the existing analytical models and a modified model
will be developed to calculate wave-induced current over submerged breakwaters with
relatively wide crest (reefs).
98
5.2.1 Effectiveness of Dimensional Parameters
Sensitivity of discharge over submerged breakwater to different independent

dimensional parameters was investigated. The variation of depth-induced flux against
incident wave height Hi, wave period T, submergence depth hs and the breakwater crest
width B were examined graphically and statistically.
Figure 5.14 shows the variation of discharge over submerged breakwater with different
breakwater crest width as a function of incident wave height. As implied from the
figure, the amount of wave-induced flow increases with incident wave height due to
higher energy dissipation during the wave breaking over the breakwater. Figure 5.14
also indicates that for constant values of incident wave height, water depth and wave
period, the breakwater with broader crest provides higher value of discharge. For broad
crested breakwater, the breaking zone over the breakwater (shallow water) is long
enough to develop breaking processes and more energy gradient (wave force) exists
over the breakwater to derive the flow.
The values of measured wave-induced discharge as a function of wave period for

constant wave height and given water level and breakwater crest width are plotted in
Figure 5.15. Different trends of discharge variation were observed depending on the
breakwater crest width. For the breakwater with B=3500 mm, discharge initially
increases with wave period up to T=1.5 s and then reduces to the lower value. This trend
is consistent with Gourlay (1996a). He measured discharge over a reef with the crest
width B=15 m. Therefore, a threshold and limit for maximum wave-generated flow over
broad crested breakwaters (reefs) may be derived as a function of wave period or wave
steepness. However, as indicated from Figure 5.15, wave period has inverse effect on
the flow over the relatively broad crested breakwater (B=1500 mm) and flow tends to
reach a constant value with increasing wave period. For the narrow crested breakwater,
the crest width has the least role in wave breaking processes and wave steepness is the
main governing parameter in the breaking wave behaviour (for a given submergence
depth) - as implied from the figure, increasing wave steepness (due to greater wave
period) causes more breaking (wave energy dissipation) and then more discharge over
the breakwater.
99
Table 5.4 Calibration coefficients calculated by nonlinear regression analysis for best fitting of data and the corresponding statistical parameters for evaluation
of the models.
Calibration coefficients
Equation number Model No. criteria R2 β RMSE IW ε
b1 b2 b3 b4 b5 b6 b7 b8
1 0.07 2.30 -0.22 -0.61 0.11 0.17 0.28 -0.39 - 0.88 1.01 0.11 0.97 0.09
(5.8)
2 0.07 2.13 0.76 -0.35 0.01 0.81 - - - 0.85 1.01 0.12 0.96 0.10
3 0.06 2.38 -1.62 0.09 2.08 0.20 0.09 0.96 - 0.89 1.01 0.10 0.97 0.09
(5.9)
4 0.09 1.88 -78.6 0.00 79.11 0.00 - - - 0.86 1.01 0.11 0.96 0.10
*
0.14 1.18 -0.26 0.5 - - - - Ir < 3.0
5 0.85 1.02 0.12 0.96 0.10
*
0.20 1.27 -0.27 0.5 - - - - Ir ≥ 3.0
0.12 0.71 - 0.5* - - - - Ir < 3.0

6 0.38 1.05 0.24 0.73 0.20
*
0.20 1.27 -0.27 0.5 - - - - Ir ≥ 3.0
7 0.17 1.15 -0.24 0.5* - - - - - 0.82 1.03 0.13 0.94 0.11

(5.10) *
0.03 1.26 -0.17 0.5 - - - - B/Hi <10
8 0.85 1.02 0.13 0.95 0.10
*
0.22 1.19 -0.29 0.5 - - - - B/Hi ≥10
0.10 1.35 -0.3* 0.5* - - - - B/Hi <10

9 0.78 1.00 0.14 0.94 0.12
* *
0.30 2.67 -0.65 0.5 - - - - B/Hi ≥10
10 0.17 2.84 -0.26 -0.14 - - - - - 0.87 1.01 0.11 0.96 0.09

*
11 0.52 -0.37 - - - - - - - 0.10 1.07 0.28 0.43 0.24
*
Fixed parameter in the model
100
Figure 5.16 presents the variation of discharge against the submergence depth hs over
the breakwater of crest width 3500 and 1500 mm. Wave energy dissipation and then
wave-induced discharge decreases with increasing water depth over the breakwater.
However, Figure 5.16 reveals that the reduction of wave-induced current is more
significant when submergence ratio is greater than 2 (hs/Hi > 2.0).
Regression analysis between the dependent variable q and the independent variables
Hi, B, T, and hs for all breaking wave data, indicates that Hi with correlation
coefficient 0.54 has the strongest influence on flux variation whilst B, hs and T have
respectively less and reducing effects. The correlation coefficient between discharge
and other dimensional parameters is provided in Table 5.5. The values of correlation
coefficient were calculated for data with different breakwater crest width as well as
the whole data. Although applying correlation coefficient is not appropriate in non-
linear phenomena, it may provide a rough estimate of the trend and effectiveness of
the parameters.
Table 5.5 Correlation coefficient between the measured discharge and other independent
dimensional parameter.
Test range hs T B Hi
B=3500mm 0.18 0.18 - 0.88
B=1500mm 0.20 0.16 - 0.71
B=300mm 0.36 0.53 - 0.72
All data 0.24 0.19 0.53 0.54
As implied from Table 5.5, the values of correlation coefficient for each independent
parameter are different for the different breakwater crest widths. Incident wave height
has the most influence on wave-induced discharge for all three breakwater widths.
After incident wave height, discharge over the narrow-crested breakwater (B=300
mm) is more affected by the incident wave period rather than the submergence depth
which is of second most important for the 1500 and 3500 mm wide breakwaters.
These results are consistent with the graphical analysis of data.
In summation, wave-induced discharge over submerged breakwater is mostly affected

by incident wave height while the effectiveness of submergence depth and wave
period on discharge depends on the breakwater geometry. Wave period is a more
influential parameter when waves approach narrow-crested breakwater (B/Lo < 0.35).
101
5.2.2 Non-dimensional Analysis
The relationship between the independent parameters and dependent parameter, wave-
induced discharge, were also investigated using non-dimensional analysis. Five
dimensionless variables were adopted for evaluating the experimental data:
⎧⎪ q hs B H i H i H ⎫⎪
⎨q =
*
, , , , or i2 ⎬
⎪⎩ gH i3 H i Lo B Lo gT ⎪
⎭
where q is the discharge per unit width, Hi is incident wave height, T is wave period,
Lo is offshore wave length, g is the gravitational acceleration, hs represents the water
depth over the breakwater (submergence depth) and B demonstrates the breakwater
crest width. The first term on the left of the above parameters was proposed by
Gourlay (1996a) as the representative dimensionless flow term over the breakwater
and is called discharge ratio (q*) hereafter. The other dimensionless parameters are
respectively nominated as submergence ratio, crest width ratio, dimensionless wave
height and wave steepness.
The variations of discharge ratio against the four dimensionless parameters (hs/Hi,
B/Lo, Hi/B, Hi/Lo) are plotted in Figure 5.17. Note that Lo /B is used in Figure 5.17b
rather than B/Lo . The maximum value of measured dimensionless discharge ratio is
seen to be about 0.5 in the figures:
⎛ q ⎞
*
qmax =⎜ ⎟ = 0.5 (5.15)
⎜ 3 ⎟
⎝ gHi ⎠ max
The maximum value of dimensionless discharge derived from the present laboratory
measurements corresponds to the broad-crested breakwater (reef type, B/Lo > 1 or
Lo /B < 1) and is consistent with Gourlay (1996a) for wide crested reefs. However,
this maximum value decreases with narrowing breakwater crest width – a value q*
being 0.25 for a moderately wide crested breakwater, 1.0 < Lo/B < 3.0 (0.35 <
B/Lo<1.0) , and q*=0.15 for the narrow-crested breakwater, Lo/B > 3.0 (B/Lo<0.35). In
both the broad and moderately wide crested breakwaters, discharge ratio meets the
102
maximum value at hs/Hi = 2.0 whilst for narrow crested breakwater the maximum
q*=0.15 occurs at lower value of submergence ratio, hs /Hi = 1.0.
The envelopes of dimensionless discharge (for each breakwater crest width) as a

function of submergence ratio are provided in Figure 5.18 and summarised as follows:
hs
*
q max =C⋅ (5.16)
Hi
where the coefficient C is the constant that depends on the breakwater crest width
ratio:
C = 0.35 for 1.0 < B/Lo and hs/Hi ≤ 2.0
C = 0.20 for 0.35 ≤ B/Lo ≤ 1.0 and hs/Hi ≤ 2.0
C = 0.15 for B/Lo < 0.35 and hs/Hi ≤ 1.0
Equation 5.16 can be applied in the preliminary design of submerged breakwater for
prediction of maximum discharge passing over the structure. Therefore,
hydrodynamic and morphodynamic impacts of the breakwaters can be assessed
quickly when wave climate is available.
The correlation coefficients between q* and the four non-dimensional parameters

(hs/Hi, B/Lo, Hi/B, Hi/Lo) were calculated and presented in Table 5.6. Note that the
inverse ratios of B/Hi and Lo/Hi have been presented in the table (instead of Hi/B and
Hi/Lo) to enhance positive values of correlation coefficients. The coefficients are
given for different ranges of breakwater crest width ratio and demonstrate the trend of
discharge ratio variation in terms of other non-dimensional parameters.
Graphical and statistical analysis of the non-dimensional parameters demonstrate the

sensitivity of dimensionless discharge q* to the other non-dimensional parameters as
summarised follows:
103
• In general, inverse wave height ratio, B/Hi , has the strongest influence on flux
variation whilst B/Lo, hs/Lo, and Lo/Hi have respectively less and reducing
effects.
• q* increases with submergence and breakwater crest width ratios and

decreases with increasing wave height ratio Hi/B and wave steepness.
• The effectiveness of the dimensionless parameters depends on the range of

the breakwater crest width ratio so that wave steepness is more dominant in
variation of dimensionless discharge for narrow crested breakwaters. On the
other hand, q* is more affected by Hi/B for Broad and moderately wide
crested breakwaters (B/Lo > 0.35).
Table 5.6 Correlation coefficient between the measured dimensionless discharge and other non-
dimensional parameters.
Range of crest hs/Hi B/Lo B/Hi Lo/Hi
width ratio Correlation coefficient
1.0 < B/Lo 0.70 0.01 0.80 0.59
0.35 ≤ B/L o ≤ 1.0 0.71 0.62 0.70 0.21
B/Lo < 0.35 -0.10 -0.22 0.34 0.49
All tests 0.54 0.73 0.90 0.50
Range of the parameters in the tests
min max min max min Max min max
1.0 < B/Lo 0.52 2.37 1.0 2.33 22.59 82.05 15.56 90.47
0.35 ≤ B/L o ≤ 1.0 0.32 2.02 0.37 1.00 9.56 30.29 13.24 55.11
B/Lo < 0.35 0.35 1.71 0.07 0.20 1.80 7.01 14.87 60.64
All tests 0.32 2.37 0.07 2.33 1.80 82.05 13.24 90.47
5.2.3 Previous Models and Modification
Few analytical models have been reported to predict the amount of wave-induced
discharge over submerged breakwaters and reefs. Drei and Lamberti (1999) provided
a simple empirical equation to approximate discharge over narrow crested submerged
breakwaters. Symonds et al. (1995) provided an analytical model based on the
momentum equation for calculating wave-induced current and set-up across a wide
shallow horizontal reef with a plane inclined seaward slope where wave breaking
occurs. Discharge over a reef was also predicted analytically by Gourlay and Colleter
104
(2005) who developed analytical solutions for calculating discharge over reefs and
compared the results with the measurements of Gourlay (1996a). All the models were
described in Chapter 2 and the measured values of discharge are compared here to the
results from the above-mentioned models. Finally, a modified analytical model is
provided to better predict discharge over relatively broad crested and reef type
breakwaters.
5.2.3.1 Symonds et al. (1995) model
Symonds et al. (1995) provided an analytical model based on the momentum equation
for calculating wave-induced current and set-up across a wide shallow horizontal reef
with a plane inclined seaward slope where wave breaking occurs (see Figure 5.19):
3
U = − gγ r2 tan α
(1 − R1 )(1 − R2 ) × hs (5.17)
2 R1 (1 − R2 ) + R2 (1 − R1 ) r
where
hs
R1 = (5.18)
hb
and
B
R2 = (5.19)
B + (hb − hs ) tan α
in which U is the integrated current velocity over the reef, γr is the wave deformation
index over the reef (γr =Hm/hs), Hm is maximum wave height over the reef, hs is the
water depth over the reef, hb is the wave breaking depth over the slope, B is the reef
length, α is reef face slope and r is the friction term in the momentum equation and
determined by (r=fWr .v) where v is the wave orbital velocity and may be calculated
from linear wave theory. Friction factor fWr is of order 0.01 for sandy bottom
(Longuet-Higgins 1970) and 0.1 to 0.2 for coral reefs (Nelson 1996). In applying the
model to John Brewer reef, Townsville, Australia, Symonds et al. (1995) adopted
wave deformation index γr =0.35 (based on site investigations of Hardy, 1993), an
adjusted reef width B=300 m and friction coefficient r = 0.25 to obtain the best fit
105
between the model and observed current in the field. It was noted by Symonds et al.
that these calibration factors are not unique and other values of B and r may be found
to fit the model to the observed data.
The analytical solution of the model of Symonds et al. (1995) for predicting wave-
induced current over a wide reef has been applied for comparison with the measured
experimental results for breaking waves of this thesis as well as the data reported by
Gourlay (1996a) for a reef width B=15 m. The following variations to the Symonds et
al. (1995) analytical model were made. The friction term “r” was used as calibration
coefficient and the actual tested breakwater crest width was adopted for B without any
adjustment. The deformation index γr was calculated using the equation of Hara et al.
(1992) for solitary waves passing over steps and breakwaters as it was found
consistent with the experimental observations. The details of Hara et al.’s breaking
model and comparison with the laboratory results were defined in Chapter 4.
The nonlinear regression analysis (Seber and Wild, 1989) was applied, using
MATLAB to find the best friction coefficient “r” to fit the model to the measured
data. As indicated before, the experimental data measured over the platform reef by
Gourlay (1996a) were also taken into account in the calibration processes. A different
value of r was calculated for each of the two test breakwater crest widths 1500 and
3500 mm as well as the 15 m reef with open lagoon (Gourlay, 1996a). The values of
calibrated friction term are provided in Table 5.7 and the variation against breakwater
crest width is plotted in Figure 5.20.
As can be seen in Figure 5.20, the calibrated friction term decreases with increasing
crest width that is consistent with the expected influence of reef width on friction
factor (Prandtl, 1952). Comparison of calibrated model and measured data in Figure
5.21 shows that Symonds et al. (1995) method could provide relatively reasonable
prediction for wave-induced flux over wide crested breakwaters if a practical method
of estimating γr and friction coefficient “r ” were available. The discrepancy between
calculated current velocity and measured data is still excessive. The statistical
indicators IW = 0.73, R2 = 0.31 and ε = 0.55 presented in Figure 5.22, also, indicates
relatively poor agreement of current over submerged breakwater between the
106
calibrated Symonds’ et al. model results and the laboratory experimental
measurements. More details of the indicators were provided in Section 5.1.5.
Table 5.7 Calculated current velocity over the breakwater/reef using the calibrated model of
Symonds et al. (1995). The calibration factor r is provided in column 10.
Test B hs Hi T Umeas Ucal
Number (mm) (mm) (mm) (sec)
hs/Hi B/Lo
(cm/s)
γr r
(cm/s)
T1-13 3500 50 56 2.00 0.89 0.96 28.9 0.15 0.20 13.8
T1-14 3500 50 106 2.00 0.47 0.96 60.1 0.15 0.20 30.3
T1-15 3500 50 147 2.00 0.34 0.96 59.5 0.15 0.20 43.6
T1-17 3500 100 50 0.71 1.98 4.46 7.4 0.32 0.20 4.5
T1-21 3500 100 56 1.00 1.79 2.33 16.3 0.32 0.20 6.9
T1-25 3500 100 51 1.49 1.97 1.28 17.6 0.32 0.20 4.7
T1-26 3500 100 102 1.49 0.98 1.28 35.2 0.32 0.20 33.1
T1-27 3500 100 145 1.49 0.69 1.28 46.2 0.32 0.20 61.0
T1-30 3500 100 103 2.00 0.97 0.91 25.5 0.32 0.20 33.4
T1-38 3500 150 98 1.00 1.54 2.30 12.3 0.39 0.20 13.1
T1-42 3500 150 100 1.49 1.50 1.23 21.0 0.39 0.20 14.2
T1-43 3500 150 155 1.49 0.97 1.23 31.6 0.39 0.20 50.2
T2-1 1500 50 45 0.71 1.11 1.89 18.7 0.15 1.03 4.6
T2-11 1500 50 157 1.49 0.32 0.58 36.0 0.15 1.03 21.0
T2-14 1500 50 113 2.00 0.44 0.41 7.1 0.15 1.03 14.6
T2-15 1500 50 146 2.00 0.34 0.41 17.7 0.15 1.03 19.3
T2-22 1500 100 113 1.00 0.88 1.00 23.1 0.32 1.03 18.0
T2-26 1500 100 107 1.49 0.93 0.55 15.2 0.32 1.03 16.3
T2-27 1500 100 136 1.49 0.74 0.55 26.2 0.32 1.03 24.5
T2-30 1500 100 105 2.00 0.95 0.39 12.9 0.32 1.03 15.8
T2-31 1500 100 137 2.00 0.73 0.39 27.6 0.32 1.03 24.8
T2-41 1500 150 121 1.49 1.24 0.53 9.3 0.39 1.03 12.0
T2-43 1500 150 152 1.49 0.99 0.53 14.7 0.39 1.03 21.6
T2-46 1500 150 104 2.00 1.44 0.37 7.6 0.39 1.03 7.5
T2-47 1500 150 156 2.00 0.96 0.37 13.8 0.39 1.03 22.7
15000 50 24 1.48 2.083 5.71 10.5 0.35 0.04 1.4
15000 50 46 1.48 1.087 5.71 22.0 0.35 0.04 13.8
15000 50 80 1.48 0.625 5.71 48.6 0.35 0.04 37.9
15000 50 114 1.48 0.439 5.71 66.5 0.35 0.04 63.3
15000 50 146 1.48 0.343 5.71 86.1 0.35 0.04 87.6
15000 100 72 1.48 1.389 5.45 29.8 0.43 0.04 11.3
Gourlay 15000 100 106 1.48 0.943 5.45 42.1 0.43 0.04 35.7
(1996a) 15000 100 143 1.48 0.699 5.45 56.5 0.43 0.04 66.3
15000 100 161 1.48 0.621 5.45 64.1 0.43 0.04 81.9
15000 100 131 1.10 0.763 8.44 39.9 0.43 0.04 56.1
15000 100 133 2.20 0.752 3.43 42.3 0.43 0.04 57.8
15000 15 118 1.48 0.127 5.93 148.2 0.04 0.04 9.9
15000 30 118 1.48 0.254 5.83 87.8 0.25 0.04 59.3
15000 75 118 1.48 0.636 5.57 54.9 0.40 0.04 58.0
107
The Symonds et al. method can only provide an approximation of wave-induced
current velocity over wide crest breakwaters and even in providing approximation is
reliant upon the development of a practical method of estimating the friction
coefficient “r ”.
5.2.3.2 Drei and Lamberti (1999) model
Laboratory experiments were carried out by Deri and Lamberti (1999) to indicate the
relationship between wave set-up and overtopping discharge over a submerged barrier
with short crest width. The tests were carried out with three different regular wave
conditions, submergence, and constant crest width. The water level was maintained
constant by pumping water from behind the structure to offshore and therefore, there
was no return flow in the flume. Free overtopping discharge over semi-submerged
breakwater was proposed to be approximately evaluated by the following simple
equation:
q
q* = = 0. 4 (5.20)
3
H i .g
Variation of discharge ratio q* against submergence ratio for the measurements of

Drei and Lamberti (1999) and the present laboratory experiment results are plotted in
Figure 5.23. As implied from the figure, the experimental data of Drei and Lamberti
indicates higher possible values of discharge ratio. This discrepancy may be found in
the test conditions. Drei and Lamberti measured discharge over the breakwater in the
condition of no set-up on the lee-side of the breakwater (by pumping the mass flux
from the lee-side to sea-side of shoreward the structure thus keeping the mean water
level the same across the breakwater). Moreover, the present experimental test data
set covers a more extensive range of breakwater crest width and submergence ratio
than the somewhat limited data of Drei and Lamberti (1999).
The average proposed value of discharge ratio by Drei and Lamberti (1999)
overestimates the prediction of average flow over submerged breakwaters. It is noted
that the maximum envelop curves provided in Equation 5.16 (Figure 5.18) can be
used to estimate the maximum value of discharge ratio with respect to the breakwater
108
crest width and that the equations have the same form as proposed by Drei and
Lamberti.
5.2.3.3 Gourlay and Colleter (2005) model
Gourlay (1996a) found that the wave-induced current condition over reefs varies with
submergence ratio S=hs/Hi and he defined two conditions in his experiments (Figure
5.24 and Figure 5.25):
i- S < 0.7, where the downstream reef-rim acts as a broad crested weir control;
ii- 0.7 < S < 2.5, where flow is subcritical and reef-top acts as an open channel.
In condition (i), waves break seaward of the reef-edge and critical flow occurs in the
surf zone. However, in condition (ii) waves break on top of the reef and wave-induced
current is always subcritical. These conditions are in relatively good agreement with
observations from the present laboratory tests with the wider crested breakwaters
( B=3500 and B=1500 mm). The calculated Froude number of wave-induced flow
over the breakwater for experimental tests reported herein is in the range of 0.02 <Fr
< 0.86 for which all flows over the breakwater are subcritical. It is also observed that
for low submergence, the flow at the landward edge of the breakwater appears like
condition (i) with overtopping of a weir as indicated by Gourlay (1996a) and Gourlay
and Colleter (2005).
Comparison of the proposed analytical model of Gourlay and Colleter (2005) with the
measured data was carried out only for reef type control condition (ii) above (S > 0.7)
- the present laboratory measured data is not sufficient to provide confident
comparison sufficient to justify modification to the developed model of Gourlay and
Colleter for low submergence S < 0.7.
Gourlay and Colleter (2005) applied energy equation between the point of the
maximum set-up and the landward edge of the horizontal breakwater (reef) to develop
the model:
q2 q2
η + hs + = h + + hf (5.21)
2 g (η + hs ) 2
s
2 ghs2
109
where η is maximum set-up over the reef, q is discharge per unit width, g is
gravitational acceleration and the friction head loss “ hf ” can be determined using
Darcy’s Equation:
f LVr2
hf = (5.22)
8 g d ave
in which L is effective reef width downstream of the surf zone (= B - xs in Figure

5.25), dave is the average depth over the reef (= hs + η / 2 ) and f is reef surface friction
coefficient.
Gourlay and Colleter (2005) indicated that the kinetic energy term (q2/2gh) in
Equation 5.21 is small and can be neglected from both sides of the equation. The
maximum measured set-up over the breakwater in the laboratory tests showed that
set-up is negligible in comparison to water depth over the breakwater. Moreover, the
effective length of the reef can be approximated with the reef top length in which the
open channel flow regime governs. Hence:
f B Vr2
η = (5.23)
8 g hs
where Vr is depth-averaged current velocity over the reef. By substituting
dimensionless ratio ( S = hs H i ) and dimensionless discharge ( q ∗ = q gH i3 ),
Equation 5.23 simplifies to:
12
∗ ⎛ η ⎞
q = 8 ⎜⎜ ⎟⎟ S 3 2 (5.24)
⎝ f B ⎠
Applying conservation of energy flux over the reef, Gourlay and Colleter (2005)
developed an analytical relationship between set-up “ η ” and discharge ratio “q* ”
which with minor manipulation can be written as:
3K p gH i ⋅ T
η = PT (5.25)
64π ⎛ 3 2 3 ∗⎞
⎜S + Kp ⋅q ⎟
⎝ 2 ⎠
110
where PT is the transmission parameter defined as:
⎡ 2
q∗ ⎛ H i ⎞ ⎤
3 12
2 S
PT = ⎢1 − K R − 4π γ r
2
− 16π 2 ⎜⎜ 2 ⎟⎟ ⎥ (5.26)
⎢⎣ D S ⎝ gT ⎠ ⎥
⎦
in which D is dimensionless reef-top wave length (=g1/2T/hs1/2), Kp is reef profile

shape factor (Gourlay, 1997), KR is the reflection coefficient, γr is reform index
(=Hr/hs) and Hr is the maximum wave height over the reef after breaking. Gourlay
(1994) calculated the reform index by measuring actual water depth and wave height
over the reef at the end of the surf zone. He classified wave breaking over the reef
based on the value of the nonlinearity parameter:
g 1.25 H i0.5T 2.5

Fco = (5.27)
he1.75
where he is the water depth at the reef edge (see figure 5.24). Gourlay (1994) found
that for wave breaking by plunging with nonlinearity parameter Fco ≈ 500 , the
reform index has values of order 0.4 and for 100 < Fco ≤ 140 when waves begin to
spill at the reef-edge the reformed index is very close to the limiting curve for a
horizontal bottom- H/dmax=Fc/(22+1.82Fc) in Figure 5.26. When Fco ≤ 100, waves
pass across the reef-edge and break on the reef-top further the reef-edge. Gourlay and
Colleter (2005) adopted, empirically, γ r = 0.4 for significant dissipation at the reef-
edge with progressive regular wave. This value is the threshold for waves to
propagate across the reef without breaking as found by Gourlay (1994, 1996a).
Substituting Equation 5.25 into Equation 5.24, the dimensionless discharge q* can be
derived:
12
⎡ ⎤
⎢ PT ⎥
q* = A ⋅ ⎢ ⎥ ⋅S3 2 (5.28)
⎢ ⎛ S 3 2 + 3 K q *⎞ ⎥
⎢ ⎜⎝ P ⎟⎥
⎠⎦
⎣ 2
where
111
12 14
⎛ 3 ⎞ ⎛ K P ⎞ ⎛ gT ⎞
12 2
A=⎜ ⎟ ⎜ ⎜
⎟ ⎜ ⎟⎟ (5.29)
⎝ 64π ⎠ ⎝ F ⎠ ⎝ H i ⎠
and
fL
F= (5.30)
8H i
Gourlay (1996b) found the value of dimensionless reef-top wave length D of the order
of 10 to 33 for his idealized horizontal reef experiments. In his analytical model, he
assumed substantial energy dissipation occurs on the reef-rim.
Reef profile factor, KP, increases with increasing reef-rim slope (Gourlay, 1997).
Table 5.8 shows some calculated values of this parameter for different field data sets.
For the idealised horizontal reef, Gourlay (1996b) proposed KP of the order of 0.8 by
fitting the best line to data for some open lagoon reefs. He provided the following
equation from conservation of energy which was applied to determine reef profile
coefficient:
3 g 1 2 H i2T ⎡ 1 ⎛ η + hs ⎞
12
⎤
η = Kp ⋅ ⎢1 − K R2 − 4πK r2 ⎜⎜ ⎟⎟ ⎥ (5.31)
64π (η + hs )3 2 ⎢⎣ T⎝ g ⎠ ⎥⎦
Gourlay (1996b) considered the reflection coefficient and reformation coefficient

equal zero, i.e. KR =0 and Kr=0, and found the slope of the best line fitted to data to
derive Kp:
η = Kp ⋅M (5.32)
in which Kp is the angular coefficient of the fitted line and M reads:
3 g 1 2 H i2T
M= (5.33)
64π (η + hs )3 2
Gourlay’s laboratory experimental data for open lagoon reef has been re-analysed
herein with KP found to be about 0.45 (Figure 5.27) which is different from the
112
proposed value of reef shape factor by Gourlay (1996b). Comparison of calculated KP
for Gourlay’s data in Figure 5.27 verified that ignoring KR and Kr in Equation 5.31
(Figure 5.27b with KR =Kr = 0) does not make a significant difference in calculating
KP (0.43 against 0.45).
Gourlay and Colleter (2005) adopted γr = 0.4 and D = 12.5 reportedly based on
Gourlay (1977). The value of the reflection coefficient in Gourlay’s experiment was
found to be less than 0.3. Hence, they ignored reflection coefficient KR as it at most
decreases the set-up 10%.
Applying Gourlay and Colleter (2005) model to the present set of tests with only
breaking waves passing over the 3500 and 1500 mm breakwater crest widths ensured
the open channel flow assumption is valid. The data set of Gourlay (1996a) for 15 m
long reef was also taken into account in the comparison of the results. Discharge ratio
over the experimental breakwaters and the reef was calculated using Equation 5.28
through 5.30. Comparison of the model with measured laboratory data was made by
keeping the coefficients γ r , Kp and f the same as adopted in Gourlay and Colleter
(2005). The reef top wave length ratio “D” was calculated by adopting corresponding
wave period and submergence depth for each test.
Table 5.8 Calculated reef profile factor Kp for the reefs with different slope (Gourlay, 1996b).
Condition Reef Kp Reef-profile slope
0.018 Inner reef rim
Ala Moana 0.2
0.023 Average reef rim
I
Male 0.36 0.067 Reef rim
Guam 0.58 0.094 Inner reef rim
0.030 Outer reef rim
Hayman Island 0.44
0.222 Reef-face
Ii
0.000 Horizontal reef-top
Idealized horizontal 0.78
1.000 Reef-face
Condition i: wave breaking and surfzone dissipation occur on reef-rim (S< 0.7).
Condition ii: wave breaking occurs at reef-edge and surf zone dissipation occurs on reef-top (S> 0.7).
Comparison of the calculated discharge ratio using the developed model of Gourlay
and Colleter (2005) with the measured values in the laboratory is shown in Figure
5.28. High discrepancy was observed between calculated and measured discharge
with the differences increasing with decreasing breakwater crest width. For narrower
breakwater (B=1500 mm) the width of breakwater crest is not sufficient for open
channel flow to fully develop and high unrealistic friction factor “f ” values would be
113
needed to match the model with measurement. However, as expected, good agreement
is observed between the predictive formulas and the laboratory data of Gourlay
(1996a) for the wide reef.
Gourlay and Colleter (2005) calibrated the model with coefficients Kp and f according
the data from the test with wide reef (B=15 m). The value of the calibration
coefficients are clearly not applicable for the narrower breakwater crest widths
(B=3500 and 1500 mm). The reef shape factor “KP” can only be predicted by knowing
breakwater geometry, incident wave condition and wave transmission coefficient
(Gourlay 1996b).
5.2.4 Modified Model for Discharge
The number of unknown coefficients in Equation 5.28 and unavailability of field or

modelled data for calibrating the model encourages consideration of other methods
for estimating the various model parameters. Variation of calculated to measured
discharge ratio against dimensionless reef top width “B/Lo” is plotted in Figure 5.29.
The figure indicates that for higher submergence (i.e. S > 0.7), Equation 5.28 should
be modified by some correction factor as a function of B/Lo to improve the fit of the
model to measured data. Hence, the following modification is proposed on the
analytical model of Gourlay and Colleter (2005) to be applied for moderately wide
and broad crested breakwaters (B/Lo ≥ 0.35) when submergence ratio is greater than
0.7:
*
q mod
q *
= el
; 0.7 < S and 0.35 ≤ B/Lo (5.34)
∏
amended
where
−1.3
⎛ B ⎞
∏ = 1 + 1.3⎜⎜ ⎟⎟ (5.35)
⎝ Lo ⎠
Figure 5.30 confirms the reliability of Equation 5.34 and 5.35 in amending the method
of Gourlay and Colleter (2005) to give a more reasonable prediction of discharge over
a submerged breakwater/reef. Comparison of Figure 5.30 with 5.28 clearly shows the
114
improvement. The statistical comparison of the model and amended model was
carried out by calculating Wilmott index “IW ” and error function “ε ” in Figure 5.31.
The experimental data of the breakwater with both crest width 3.5 m and 1.5 m and
the reported data of Gourlay (1996a) for the reef with B=15 m, were applied
separately in computing the indicators. As indicated in the figure, the amended model
can successfully predict discharge ratio over breakwater with B/Lo > 0.35 in
comparison with the model provided by Gourlay and Colleter (2005) while no field or
laboratory data is needed. In addition, the value of calculated discharge ratio for the
long reef (B=15 m) is improved by applying the amended model in comparison with
the values reported by Gourlay and Colleter (2005).
The amendment to Gourlay and Colleter (2005) model for relatively narrow reefs was
made only for submergence ratio (S=hs/Hi) greater than 0.7, crest width ratio greater
than 0.35 and steep reef/breakwater slope (tan β ≥ 0.5).
5.3 Wave-induced Set-up
Wave-induced set-up in the surfzone on a plane beach is mainly governed by the

amount of wave energy dissipation due to water depth changes. In 2D depth-averaged
application, when no wave-induced currents exist, the wave-induced set-up is based
on the vertically integrated momentum balance equation which is a balance between
the wave force (gradient of the wave radiation stress) and the hydrodynamic pressure
gradient:
∂η
Fx = g h (5.36)
∂x
where Fx is the wave-induced forces and h is the total water depth.
However, Equation 5.36 is not applicable for submerged breakwaters/reefs due to

existence of considerable wave-induced current over the structure. In the presence of
current, wave-induced set-up varies with flow conditions and other parameters such as
bottom friction and boundary conditions may affect the value of set-up. For instance,
115
set-up may be different in the case of 2D laboratory test with closed lagoon (behind
the structure) and 2D or 3D test with an open lagoon for which a longshore current
exists. Therefore, adopting the concept of set-up in the surfzone (over a plane beach)
is not appropriate in the presence of submerged breakwaters and studying wave-
induced set-up should be integrated with current over submerged breakwaters. The
variation of set-up as function of other parameters and comparison of measured data
with existing models for set-up prediction behind submerged breakwaters are
investigated.
5.3.1 Sensitivity of Set-up to Other Parameters
Variation of maximum wave-induced set-up over and onshore of the breakwater

against incident wave height is plotted in Figure 5.32. The analysed data presented in
Table 4.9 (Chapter 4) are used in this section for detailed discussion. The maximum
value of wave-induced set-up increases with incident wave height for all ranges of
breakwater crest width that is consistent with the measurement conducted by Gourlay
(1996a). Moreover, set-up increases with breakwater crest width for a given wave
height and submergence depth.
Figure 5.33 reveals that the submergence depth over the breakwater has inverse effect
on wave-induced set-up as it decreases with increasing water depth for a given
incident wave height. Increasing submergence depth causes less energy dissipation
(less wave forces) over the breakwater to create set-up.
Wave breaking dissipation over submerged breakwater increases with wave period
(wave steepness) that leads to higher set-up over and behind the structure. The
influence of wave period on wave-induce set-up is shown in Figure 5.34. As indicated
in the figure, set-up increases with wave period for a given incident wave height and
water depth.
All the above results are in agreement with the laboratory observations reported by
Gourlay (1996a). As mentioned in the previous section, Gourlay’s measurements
were carried out for a wide reef with B=15 m. Nevertheless, the same trends were
observed for variation of set-up against incident wave height, wave period and water
depth in the present laboratory test with different narrower breakwaters.
116
5.3.2 Comparison with Existing Models
Wave-induced set-up, measured from the present laboratory experiments, were

compared with two developed models for prediction of set-up behind the submerged
breakwaters:
Diskin et al. (1970)
A series of two-dimensional experimental tests were carried out by Diskin et al.

(1970) to measure piling up over rouble mound submerged breakwaters, using regular
waves. They empirically developed a relationship between set-up “ η ”, incident wave
height “Hi” and submergence depth “hs”:
η ⎡ ⎛ hs ⎞ ⎤
2
= 0.60 ⋅ exp ⎢− ⎜⎜ 0.7 + ⎟ ⎥ (5.37)

Hi ⎢⎣ ⎝ H i ⎟⎠ ⎥
⎦
The equation was based on experimental points falling in the following range of
submergence ration:
hs
− 2 .0 < < +1 .5 (5.38)
Hi
Comparison of the measured data with Equation 5.37 in the range of the model
application is shown in Figure 5.35. High discrepancy between the model and
measured data is observed in the figure. Equation 5.37 overestimates wave-induced
set-up behind submerged breakwaters, especially for higher incident wave heights.
The main reason for the differences is that the Diskin et al.’s model was derived from
2D experimental tests in which water was allowed to pill up behind the breakwater
and no longshore current existed in the flume to reduce wave-induced set-up behind
the structure. In contrast, in the present experimental tests, water passing across the
breakwater is free to return through the side channels with no return flow over the
breakwater (see Section 2 for more details of the experiment).
117
Gourlay and Colleter (2005)
Wave-induced set-up was calculated applying Equation 5.25 for the data set
corresponding to the breakwaters with 3500 and 1500 mm crest width. Applying the
parameters, Kp=0.8 and Kr=0.4 (the same as proposed by Gourlay and Colleter) the
model greatly overestimated set-up over the breakwater with 1500 mm crest width
(Figure 5.36). The value of reflection coefficient KR was neglected in all test
conditions (KR 2 ~ 0). Figure 5.37 indicates that the calibrated model (adopting the
reef shape coefficient Kp = 0.2 for the narrower breakwater, B = 1500mm provides
more reasonable prediction of set-up over the breakwater. However, the model of
Gourlay and Colleter (2005) still shows considerable scatter for estimating set-up over
either width breakwater.
5.4 Summary and Discussion
The laboratory measured data of transmitted wave height, wave-induced current and
wave-induced set-up were analysed in detail in this chapter. All data corresponds to
the tests with breaking wave over the submerged breakwater.
The transmission coefficient was calculated as the ratio of the average transmitted
wave behind the breakwater to the incident wave height calculated from the envelope
of water surface offshore the structure. Sensitivity of transmission coefficient “Kt ” to
other dimensional parameters, e.g. incident wave height “Hi ” “, wave period “T ”,
breakwater crest width “B ” and submergence depth “hs ”, were examined. Graphical
and statistical analysis of the parameters revealed that transmission coefficient
increases with submergence depth and wave period whilst, breakwater crest width and
incident wave height have inverse effect on Kt. However, the effectiveness of each
parameter on transmission coefficient depends on the other parameters. The
submergence depth “hs” and incident wave height “Hi ” are more effective for the
wider crest width breakwater while wave period “T ” is more effectual for narrower
breakwater. Moreover, the impact of wave period and submergence depth is stronger
with smaller wave heights whilst for higher waves the breakwater crest width is more
effective.
118
Analysing the variation of transmission coefficient against dimensionless parameters
showed that the inverse breakwater crest width ratio “Lo/B ” , inverse ratio of wave
steepness “Lo/Hi ”, the ratio of incident wave height to breakwater crest width “Hi/B ”
and the submergence ratio “hs/Hi ” are respectively more influential parameters on
transmission coefficient. Variation of transmission coefficient with respect to
submergence ratio for different breakwater crest width revealed that the submergence
ratio is more influential in the tests with wider breakwater.
The measured transmission coefficient was compared with the existing models for
predicting Kt over submerged breakwaters. The selected models predicted a
transmission coefficient with some scatter in comparison with the measured results.
The main reason for the discrepancy in the results was found in the tested range of the
breakwater crest width ratio (B/Lo). The present laboratory tests cover a wide range of
B/Lo including broad crested and reef type breakwaters while the existing models
were mostly calibrated within a range of narrow crested breakwater.
The modified model (Equation 5.14) was developed for a wide range of crest width
and submergence ratio to provide more consistent results of transmission coefficient
over submerged breakwaters. Comparison of measured and calculated transmission
coefficient using the proposed equation provides good accuracy of the model. The
proposed model is a significant advance in that it can be used to predict Kt
consistently over submerged breakwaters/reefs in the range of crest width ratio 0.08
< B/Lo < 4.4.
The measured wave-induced discharge flux q (discharge per unit width) over the
submerged breakwater was examined to investigate the dependency of discharge on
dimensional and non-dimensional parameters. Graphical and analytical analysis of
data revealed that the amount of wave-induced flow over the breakwater increases
with incident wave height due to higher energy dissipation during the wave breaking
and for constant values of incident wave height, water depth and wave period, the
breakwater with broader crest width provides higher value of discharge. The
effectiveness of wave period depends on the breakwater crest width so that for the
breakwater with B=3500 mm, discharge initially increases with wave period up to
T=1.5 s and then reduces to the lower value. On the other hand, increasing wave
119
period (steepness) with a given wave height causes more breaking and hence more
discharge over submerged breakwater with narrow crest width (B = 300 mm).
The effect of non-dimensional parameters (as used in study of transmission

coefficient) on wave-induced current was investigated by introducing discharge ratio
“q*=q/(gHi3 )0.5 ”. Graphical and statistical analysis of the values of dimensionless
parameters revealed that the ratio of B/Hi has the strongest influence on flux variation
whilst B/Lo, hs/Lo, and Hi/Lo have respectively less and reducing effects. q* increases
with submergence and breakwater crest width ratios and decreases with increasing
wave height ratio “Hi/B ” and wave steepness. However, the effectiveness of the
dimensionless parameters depends on the range of the breakwater crest width ratio so
that wave steepness is more dominant in variation of dimensionless discharge for
narrow crested breakwaters. On the other hand, q* is more affected by Hi/B for
relatively broad-crested and reef type breakwaters (B/Lo > 0.35).
Variation of discharge ratio against other dimensionless parameters showed the

maximum limit of dimensionless discharge ratio q* = 0.5. The maximum discharge
ratio increases with breakwater crest width and the envelopes of maximum
dimensionless discharge are provided in Equation 5.16 as a function of submergence
ratio in different range of crest width ratio (Figure 5.18). The proposed equation may
be applied in the preliminary design of submerged breakwater for prediction of
maximum discharge passing over the structure.
Two models were assessed for predicting current velocity and discharge over wide
and relatively broad crested submerged breakwaters (0.35 < B/Lo) and predictive
capabilities of the models are improved with adjustment to the calibration factors and
modification of models. The Symonds et al. (1995) method provides qualitatively
reasonable results but for broader application needs a practical method for estimating
the friction coefficient.
The model provided by Gourlay and Colleter (2005) was found to overestimate
discharge over submerged breakwater/reef if the unknown coefficients in Equation
5.24 to 5.26 were adopted the same as Gourlay and Colleter. The number of unknown
coefficients in the corresponding equations and unavailability of field or modelled
data for calibrating the model encouraged consideration of other methods for
120
estimating the various model parameters. The model of Gourlay and Colleter (2005)
provides reasonable calculation of dimensionless discharge for high submergence (0.7
< S) if amended by the inclusion of a correction factor (Equation 5.34) being a
function of crest width ratio (B/Lo). No further field or laboratory data are needed for
finding unknown coefficients incorporated in the amended Gourlay and Colleter
method.
Wave set-up over and behind the breakwater was also investigated. It was found that
the maximum value of wave-induced set-up increases with incident wave height for
all ranges of breakwater crest width. In addition, for a given wave height and
submergence depth, set-up increases with breakwater crest width. Submergence depth
over the breakwater has inverse effect on wave-induced set-up as it decreases with
increasing water depth for a given incident wave height. Increasing submergence
depth causes less energy dissipation (less wave forces) over the breakwater to create
set-up.
The measured maximum set-up over (behind) the submerged breakwater in the
laboratory was compared to calculated values using the models of Diskin et al. (1970)
and Gourlay and Colleter (2005). Comparison of the results showed that the model of
Diskin et al. (1970) overvalues maximum set-up behind the breakwater since it has
been derived from 2D laboratory data with piled up water behind the structure (no
longshore current).
The model of Gourlay and Colleter (2005) provided more reasonable prediction of
set-up over reef type breakwater (B=3500 mm). However, the maximum set-up was
overestimated for the breakwater with B=1500 mm. The model can be improved by
adopting a revised value of reef profile factor (K P =0.2) that gives better results for
narrower breakwater. Nevertheless, the model still shows significant scatter for
estimating set-up over different crest width submerged breakwaters.
121
1.0
hs=50 m m
0.8 hs=100 m m
0.6
Kt
0.4
0.2
0.0
0 50 100 150 200
H i (mm)
Figure 5.1. Variation of transmission coefficient against incident wave height

(T=1.49 sec and B=1500 mm).
1.0
Hi~50 m m
0.8 Hi~100 mm
0.6
Kt
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5
T (sec)
Figure 5.2. Variation of transmission coefficient against wave period

(B=1500 mm and hs=50 mm).
1.0
Hi~100 m m
0.8 Hi~150 m m
0.6
Kt
0.4
0.2
0.0
0 1000 2000 3000 4000
B (mm)
Figure 5.3. Variation of transmission coefficient against breakwater crest width

(hs=150 mm and T=1.49 sec).
1.0
0.8
0.6
Kt
0.4
B= 3500 m m
0.2
B=1500 m m
B=300 m m
0.0
0 50 100 150
h s (mm)
Figure 5.4. Variation of transmission coefficient against submergence depth with

different breakwater crest width. (Hi~100 mm and T=2.0 sec).
1.00
"hs"
"T"
0.80
correlation coefficient
"Hi"
0.60
0.40
0.20
0.00
0 1000 2000 3000 4000
B (mm)
Figure 5.5. Variation of correlation coefficient between transmission coefficient "Kt"

and submergence depth "hs" (solid blue line), wave period "T" (dash red line) and
incident wave height "Hi" (dash-dot purple line) against breakwater crest width "B".
1.00
"hs"
"T"
correlation coefficient
0.80 "B"
0.60
0.40
0.20
0.00
0 20 40 60 80 100 120 140 160
H i (mm)
Figure 5.6. Variation of correlation coefficient between transmission coefficient " Kt

" and submergence depth " hs " (solid blue line), wave period "T" (dash red line) and
Breakwater crest width" B " (dash-dot purple line) against incident wave height " Hi "
1.0 1.0
(a) (b)
0.8 0.8
0.6 0.6
Kt
Kt
0.4 0.4
Test T1 (B=3500mm)
0.2 Test T2 (B=1500mm) 0.2
Test T3 (B=300mm)
0.0 0.0
0.00 1.00 2.00 3.00 4.00 5.00 0 0.5 1 1.5 2 2.5
B /L o h s /H i
1.0 1.0
(c) (d)
0.8 0.8
0.6 0.6
Kt
Kt
0.4 0.4
0.2 0.2
0.0 0.0
0.00 0.02 0.04 0.06 0.08 0.0 0.1 0.2 0.3 0.4 0.5 0.6
H i /L o Hi/B
Figure 5.7. Variation of transmission coefficient against different dimensionless variables for the whole tests
with breaking wave over the breakwater. a) inverse crest width ratio, b) submergence ratio, c) wave steepness, d) wave height to crest width
ratio.
1.00
+10%
0.80
-10%
0.60
calculated
0.40
d'Angremond et al. (1996)

0.20
Van der Meer et al. (2004)
Bleck & Oumeraci (2002)
Calabrese et al. (2002)
0.00
0.00 0.20 0.40 0.60 0.80 1.00
measured
Figure 5.8. Comparison of measured transmission coefficient with previous

experimental models.
1.4
Distortion
1.2 RMSE
1.0
0.8
0.6
0.4
0.2
0.0
d'Angremond Van der Meer Calabrese Bleck
Figure 5.9. The values of distortion and Root Mean Squared Error for previous
empirical models.
1.2
1
R2 = 0.40
0.8
0.6 R2 = 0.48
Kt
0.4
0.2 Ir<3
3<Ir
0
0 0.1 0.2 0.3 0.4 0.5 0.6
B /H i
Figure 5.10. Variation of transmission coefficient against the ratio of breakwater crest
width to incident wave height for the tests with Iribarren numbers greater and smaller
than 3.
1 1 1
Model 1 Model 2 Model 3
0.8 0.8 0.8
Kt calculated
Kt calculated
Kt calculated
0.6 0.6 0.6
0.4 0.4 0.4
0.2 0.2 0.2
0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
K t measured K t measured K t measured
1 1 1
0.8 0.8 0.8
Kt calculated
Kt calculated
Kt calculated
0.6 0.6 0.6
0.4 0.4 0.4
0.2 0.2 0.2
0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Figure 5.11. Comparison of the measured and calculated transmission coefficient using the proposed models.
1 1 1
0.8 0.8 0.8
Kt calculated
Kt calculated
Kt calculated
0.6 0.6 0.6
0.4 0.4 0.4
0.2 0.2 0.2
0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
1.00 1
Model 10 Model 11
0.80 0.8
Kt calculated
Kt calculated
0.60 0.6
0.40 0.4
0.20 0.2
0.00 0
0.00 0.20 0.40 0.60 0.80 1.00 0 0.2 0.4 0.6 0.8 1
K t measured K t measured
Figure 5.11. (Continued) Comparison of the measured and calculated transmission coefficient using the proposed models.
1.08
(a)
1.06
1.04
Bias
1.02
1.00
0.98
0.96
model1 model 2 model 3 model 4 model 5 model 6 model 7 model 8 model 9 model 10 model 11
m ode l No.
1.20
(b)
1.00
0.80
0.60
IW
0.40
0.20
0.00
m ode l No.
1.00
(c)
0.90
0.80
0.70
0.60
0.50
R2
0.40
0.30
0.20
0.10
0.00
model No.
0.30
(d)
0.25
0.20
RMSE
0.15
0.10
0.05
0.00
m ode l No.
0.30
(e)
0.25
0.20
0.15
e
0.10
0.05
0.00
m ode l No.
Figure 5.12. Statistics comparison of measured transmission coefficient with the

calculated values from the proposed models, a) bias, b) Wilmott index, c) squared
correlation coefficient, d) root mean square error, e) error function.
1.0
Test1, B=3500 mm
Test2, B=1500 mm +10%
Test3, B=300 mm
0.8
-10%
Kt calculated
0.6
0.4
0.2
0.0
0.0 0.2 0.4 0.6 0.8 1.0
K t measured
Figure 5.13. Comparison of the measured and calculated transmission coefficient

using the proposed model (Equation 5.14).
500
B=3500mm
B=1500mm
400 B=350mm
q (cm3/s/cm)
300
200
100 d=450 mm
T=1.49 s
0
0 50 100 150 200
H i (mm)
500
400
q (cm 3/s/cm)
300
200
100 d=400 mm
T=1.49 s
0
0 50 100 150 200
H i (mm)
200
160
q (cm 3/s/cm)
120
80
40 d=350mm
T=1.49sec
0
0 50 100 150 200
H i (mm)
300
250
q (cm 3/s/cm)
200
150
100
d=400 mm
50
T=2.0 s
0
0 50 100 150 200
H i (mm)
Figure 5.14. Variation of discharge over the breakwater as a function of incident

wave height for different breakwater crest width, water depth and wave period.
250
Hi~50 mm
200
q (cm 3/s/cm)
150
100
50 d= 350 mm
B=3500 mm
0
0.5 1.0 1.5 2.0 2.5
T (sec)
400
Hi~50 mm
Hi~100 mm
300
q (cm 3/s/cm)
200
100
d= 400 mm
B=3500 mm
0
0.5 1.0 1.5 2.0 2.5
T (sec)
400
Hi~50 mm
Hi~100 mm
300 Hi~135 mm
q (cm 3/s/cm)
200
100
d= 400 mm
B=1500 mm
0
0.5 1.0 1.5 2.0 2.5
T (sec)
125
Hi~50 mm
100 Hi~100 mm
q (cm 3/s/cm)
75
50
25 d= 350 mm
B=300 mm
0
1.0 1.3 1.5 1.8 2.0 2.3
T (sec)
Figure 5.15. Wave-generated flow as a function of wave period for constant wave
height and different water level and breakwater crest width.
350 Hi ~50mm,T=2.0sec,B=3500mm
300 Hi~50mm,T=1.49,B=3500mm
Hi~100 mm,T=2.0sec,B=1500mm
250
q (cm 2/s)
200
150
100
50
0
0 50 100 150 200
hs (mm)
Figure 5.16. Variation of wave-induced discharge as functions of submergence depth

for given wave height and wave period.
0.6 0.6
(a) (b) B=3500 mm
B=1500 mm
0.5 0.5 B=300 mm
0.4 0.4
q/(gHi 3)0.5
q/(gHi 3)0.5
0.3 0.3
0.2 0.2
B=3500 mm
0.1 0.1
B=1500 mm
B=300 mm
0 0
0 0.5 1 1.5 2 2.5 0 3 6 9 12 15
hs/Hi Lo /B
0.6 0.6
(c) B=3500 mm (d) B=3500 mm
0.5 B=1500 mm B=1500 mm
B=300 mm 0.5
B=300 mm
0.4 0.4
q/(gHi 3)0.5
q/(gHi 3)0.5
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0 0.02 0.04 0.06 0.08
Hi /B Hi /Lo
Figure 5.17. Variation of discharge ratio against: a) submergence ratio, b) inverse of breakwater crest width ratio, c) dimensionless wave
height, d)wave steepness.
0.6
B=3500 mm
B=1500 mm
B=300 mm q*=0.35 (hs/Hi) 0.5
0.5
Envelop (B/Lo>1.0)
Envelop (0.35<B/Lo<1.0)
0.4 Envelop (B/Lo<0.35)
q/(gHi 3)0.5
0.3 q*=0.20 (hs/Hi) 0.5
0.2
q*=0.15 (hs/Hi) 0.5
0.1
0
0 0.5 1 1.5 2 2.5
hs/Hi
Figure 5.18. Envelop of maximum dimensionless discharge as a function of

submergence ratio for different ranges of breakwater width ratio.
Wave
propagation
l
B
x
hs
hb
Figure 5.19. Schematic definition of parameters used in Symonds’ et al. (1995)

model.
1.2
(a) Test1 (B=3500mm)
1 Test2 (B=1500mm)
Gourlay (1996a)
friction term "r "
0.8
0.6
0.4
0.2
0
0 2 4 B /L 6 8
o
1.2
(b)
1
friction term "r "
0.8
0.6
0.4
0.2
0
0 5 10 15 20
B (m)
Figure 5.20. Variation of calibrated friction term in the model of Symonds et al.
(1995) as functions of : a) Breakwater crest width ratio, b) Breakwater crest width.
100.0
Test1 (B=3500mm), r=0.2
Test2 (B=1500mm), r=1.03 +10%
Gourlay (1996a), r=0.04
80.0
-10%
60.0
Ucalculated (cm/s)
40.0
20.0
0.0
0.0 20.0 40.0 60.0 80.0 100.0
Umeasured (cm/s)
Figure 5.21. Comparison of measured and calculated current velocity over the
submerged breakwaters/reefs using the calibrated method of Symonds et al. (1995).
1.00
0.80 0.73
0.60 0.55
0.40 0.31
0.20
0.00
R2 Iw e
statistics indicators
Figure 5.22. Calculated statistics indicators for comparison the results of calibrated
method of Symonds et al. (1995) with the measured data.
0.8
measured data
Drei and Lamberti (1999)
0.6 Average of data
Average of Drei and Lamberti
q/(gHi3)0.5
0.4 0.4
0.2
0.17
0
0 0.5 1 1.5 2 2.5 3 3.5
hs/Hi
Figure 5.23. Comparison of discharge ratio between measured by Drei and Lamberti
(1999) and the present experimental test. The red and blue dash line demonstrates the
average value of measured discharge ratio by Drei and Lamberti and the present
study, respectively.
hs hs
h d
Figure 5.24. Coral reef definitions (after Gourlay and Colleter, 2005).
(i)
hs
(ii)
hs
Figure 5.25. Wave-induced flow over a horizontal ideal reef: (i) reef-rim control (ii)
Reef-top control (after Gourlay and Colleter, 2005).
Figure 5.26. Wave breaking conditions and reformation over the reef as a function of
nonlinearity parameter Fco (after Gourlay, 1994).
70
(a)
60
y = 0.4461x
50
Set-up
40
30
20
10
0
0 40 80 120 160
M
70
(b)
60
y = 0.4246x
50
40
set-up
30
20
10
0
0 40 80 120 160
M
Figure 5.27. variation of wave set-up as a function of M=3g0.5Hi2T/(64п (ή+hs)1.5). (a)

Calculated Kr and KR, (b) Kr=KR=0.
+10%
0.8
-10%
0.6
q* model
0.4
0.2
Test 1 (B=3.5 m)
Test 2 (B=1.5 m)
Gourlay, 1996a (B=15 m)
0
0 0.2 0.4 0.6 0.8 1
q* measured
Figure 5.28. Comparison of Gourlay and Colleter (2005) model with measured
laboratory data.
10
Test 1, B=3.5 m
9 Test 2, B=1.5 m
8 Gourlay (1996a), B=15m
fitted curve
q *model/q *measured 7
6
5
4 y =1+1.3x -1.3
3
2
1
0
0 2 4 B/L o 6 8 10
Figure 5.29. Variation of overestimating of model of Gourlay and Colleter (2005) as

a function of crest with ratio for S>0.7.
0.5
+10%
0.4
-10%
0.3
q* amended
0.2
0.1
Test 1 (B=3.5 m), S>0.7
Test 2 (B=1.5 mm), S>0.7
Gourlay, 1996a (B=15 m), S>0.7
0
0 0.1 0.2 0.3 0.4 0.5
q* measured
Figure 5.30. Comparison of amended model of Gourlay and Colleter (2005) with
measured laboratory data for dimensionless discharge.
2.0
Gourlay and Colleter (a)
(2005) model 1.53
1.6 modified model
1.2
0.93
0.8
0.34
0.4 0.21
0.0
Iw e
2.0
(b)
1.6
1.2
0.84 0.89
0.8
0.4
0.15 0.11
0.0
Iw e
Figure 5.31. Statistics indicator, Wilmott index (IW) and error function (ε) for
evaluating the goodness of Gourlay and Colleter (2005) model and the modified
model. (a) Data series of B=3.5 m and 1.5m, (b) data series of B=15 m (Gourlay,
1996a).
h =350 mm & T =2.0 sec
16
B=3500 mm
B=1500 mm
12 B=300 mm
set-up (mm)
0
0 50 H i (mm) 100 150
h =400 mm & T =2.0 sec

10
8
set-up (mm)
0
0 50 100 150 200
H i (mm)
h=450 mm & T=1.49 sec

14
12
10
set-up (mm)
0
0 50 100 150 200
H i (mm)
Figure 5.32. Variations of maximum wave set-up over (behind) the breakwater as a
function of incident wave height for different breakwater crest width.
3.5 12
B=1500 mm
3 10
B=300 mm
2.5 8
set-up (mm)
set-up (mm)
2
6
1.5
4
1
0.5 2
Hi~50 m m , T=0.71 s Hi~150 m m , T=1.49 s
0 0
0.0 100.0 200.0 300.0 400.0 500.0 0.0 100.0 200.0 300.0 400.0 500.0
h (mm) h (mm)
Figure 5.33. Maximum wave set-up over (behind) the submerged breakwater with different crest width as functions of total water depth for
given wave heights and period.
9.0 12.0
B=3500
8.0
B=1500 mm 10.0
7.0
B=300 mm
6.0 8.0
set-up (mm)
set-up (mm)
5.0
6.0
4.0 h=450 m m , Hi~100 m m
3.0 4.0
2.0
2.0
1.0 h=400 m m , Hi~100 m m
0.0 0.0
0.0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5
T (sec) T (sec)
Figure 5.34. Variation of maximum wave set-up over (behind) submerged breakwater with different crest width and water depth for a given
wave height.
40
35
30
25
calculated
20
15
10
B=3500 mm
5 B=1500 mm
B=300 mm
0
0 5 10 15 20 25 30 35 40
measured
Figure 5.35. Comparison of measured maximum set-up with calculated values (in
mm) using the proposed equation of Diskin et al. (1970).
30
25
20
Calculated
15
10
5
B=3500 mm
B=1500 mm
0
0 5 10 15 20 25 30
Measured
Figure 5.36. Comparison of measured and calculated maximum set-up over

submerged breakwater with two different breakwater crest width using the proposed
model of Gourlay and Colleter (2005).
20
+10%
16
-10%
Calculated
12
4
B=3500 mm
B=1500 mm
0
0 4 8 12 16 20
Measured
Figure 5.37. Comparison of measured maximum set-up over submerged breakwater

with two different breakwater crest width with the calibrated model of Gourlay and
Colleter (2005). The reef shape factor KP was adopted 0.2 for the narrower breakwater
(B=1500 mm).
CHAPTER 6
CONCEPTUAL DESCRIPTION OF WAVE AND FLOW

MODULES WITHIN DELFT3D NUMERICAL MODEL
SYSTEM
Numerical hydrodynamic modelling of flow over a submerged breakwater was

investigated using Delft3D. The Delft3D computer package, developed by WL | Delft
Hydraulics, is a numerical modelling system that consists of a number of integrated
modules which together allow the simulation of short wave generation and propagation,
hydrodynamic flow, computation of the transport of salinity and heat, sediment
transport and morphological changes, and the modelling of ecological processes and
water quality parameters.
Two modules of Delft3D, WAVE (SWAN) and FLOW, were applied in the numerical
modelling part of this study to simulate wave propagation and flow over/behind
submerged breakwaters. The conceptual parts of the two modules that were
implemented in the numerical simulation are described in this Chapter. Only the 2D
depth averaged aspects of the model modules are presented, this being consistent with
the application of the model described in detail in the following Chapter 7. For more
explanation and description of Delft3D, the reader is referred to the user manual of
Delf3D (Delft Hydraulics, 2005a and 2005b).
6.1 Conceptual Description of Wave Model (SWAN)
SWAN model (Simulating Waves Nearshore) can be applied to simulate the evolution
of random, short-crested wind-generated waves in coastal waters, estuaries, tidal inlets
and lakes. In SWAN the waves are described using the two-dimensional wave action
density spectrum, even when non-linear phenomena dominate (e.g., in the surf zone).
The spectrum that is considered in SWAN is the action density spectrum N (σ, θ) rather
than the energy density spectrum E (σ, θ) since in the presence of currents, action
122
density is conserved whereas energy density is not (Whitham, 1974). σ and θ are
respectively the independent variables relative frequency and wave direction. The action
density is equal to the energy density divided by the relative frequency and may vary in
time and space in SWAN:
E (σ , θ )
Nˆ (σ ,θ ) = (6.1)
σ
In SWAN the evolution of the wave spectrum is described by the spectral action balance
equation which is provided by Hasselmann et al. (1973). For Cartesian co-ordinates it
may be written:
∂ ˆ ∂ ∂ ∂ ∂ S
N + c x Nˆ + c y Nˆ + cσ Nˆ + cθ Nˆ = o (6.2)
∂t ∂x ∂y ∂σ ∂θ σ
where So is the source term in terms of energy density representing the effects of
generation, dissipation and non-linear wave-wave interactions. The first term in the left-
hand side of the above equation corresponds to the local rate of change of action density
in time and the second and third term represent propagation of action in geographical
space (with propagation velocities cx and cy in x- and y- direction, respectively). The
fourth term indicates shifting of the relative frequency due to variations in depths and
currents (with propagation velocity cσ in σ -space) and the fifth term represents depth-
induced and current-induced refraction (with propagation velocity cθ in θ -space). For
definitions of the propagation speeds see Whitham (1974); Mei (1983) and Dingemans
(1997).
The following processes are accounted for in SWAN wave modelling:
• wind generation by wind,
• wave dissipation by depth-induced breaking, whitecapping and bottom friction,
• non-linear wave-wave interaction,
• wave propagation through obstacles,
123
• computation of wave-induced set-up (when the effects of set-up are accounted in
wave propagation).
Among the above processes, wave dissipation by depth-induced breaking,

whitecapping, bottom friction as well as propagation through obstacles will be
described in the following sections.
6.1.1 Wave dissipation in SWAN
6.1.1.1 Depth-induced Breaking dissipation
The process of wave-induced breaking dissipation in SWAN is based on the model

developed by Battjes and Janssen (1978), and Eldeberky and Battjes (1995) for random
waves. Battjes and Janssen (1978) developed the following expression for the average
power dissipated in the breaking process, per unit area:
α
Db = Qb f ρ g H m2 (6.3)
4
in which α is a constant of order 1 (if the model is good), f represents the mean
frequency of the progressive wave, ρ is the water density, g is gravity acceleration and
Hm is the maximum wave height which the given water depth can sustain. Qb is the
probability that at a given point the wave height is equal or greater than the breaking
point which is a function of the ratio of root mean squared wave height to Hm:
2
1 − Qb ⎛H ⎞
= −⎜⎜ rms ⎟⎟ (6.4)
lnQb ⎝ Hm ⎠
Variation of Qb as a function of Hrms/Hm in Figure 6.1 indicates that the value of Qb may
be disregarded (Q b <0.001) for the value of Hrms/Hm less than 0.14. Therefore,
according to Battjes and Jansenn’s theory, there is almost no wave-induced breaking
dissipation for the range of 0 < Hrms/Hm< 0.14.
Battjes and Jansen (1978) used the Miche (1944) criteria for wave breaking of periodic
waves and provided a model for maximum wave height:
124
⎛ 2πh ⎞ −1
H m ≅ 0.14 Lo tanh ⎜ ⎟ ≅ 0.88 k tanh kh (6.5)
⎝ L ⎠
where k =2π/Lo is wave number which can also be defined in terms of the dispersion
equation:
(2π f )2 = gk tanh kh (6.6)
in which f is wave frequency , g is gravity acceleration and h represents water depth.
In shallow water, the expression for Hm reduces to Hm= 0.88h and Battjes and Janssen
used a similar functional relationship for random waves with some freedom adjustment
to consider the effects of beach slope on transformation of random waves. They adopted
the reduced form of H m = γ h in which γ was introduced as adjustable (calibration)
parameter. This expression is applied in SWAN to calculate maximum wave height.
Battjes and Stive (1985) re-analysed wave data of a number of laboratory and field
experiments and found the values of breaker parameter, γ, vary within 0.60 to 0.83 for
different types of bathymetry. From a large number of experiments Kaminsky and
Kraus (1993) found breaker parameter in the range of 0.60 to 1.59.
In SWAN wave energy dissipation due to depth-induced breaking “Sds,br ( f , θ )” is

calculated using the extended formulation of Eldeberky and Battjes (1995):
E ( f ,θ )
S ds ,br ( f , θ ) = Dtot (6.7)
Etot
where Dtot is power dissipated in the breaking processes (Equation 6.3), E ( f , θ ) is the
wave energy density at frequency f and in θ direction and Etot is the total wave energy of
the spectrum.
6.1.1.2 Dissipation Due to Whitecapping
The effect of whitecapping on the spectral balance of surface waves in SWAN is

calculated according to the pulse-based model of Hasselmann (1974). Extending the
model to shallow water and assuming that the whitecapping dissipation is controlled by
125
the wave slope, WAMDI group (1988) reformulated the model of Hasselmann (1974) to
the following equation:
k
S ds ,w (σ ,θ ) = −Γσ~ ~ E (σ ,θ ) (6.8)
k
where Sds,w ( f , θ ) is wave energy dissipation due to whitecapping, k is the wave

number, σ is wave frequency, θ is wave direction, Γ is a steepness dependent
~
coefficient, σ~ and k denote a mean frequency and a mean wave number, respectively:
2π −1
σ~ = ⎛⎜ Etot−1 ∫ σ −1 E (σ ,θ ) dσ dθ ⎞⎟
∞
⎝ 0 ∫0 ⎠
2π ∞ −2
~
k = ⎛⎜ Etot
−1
∫0 ∫0 k −1 2
E (σ , θ ) d σ d θ ⎞⎟ (6.9)
⎝ ⎠
2π ∞
Etot = ∫ ∫ E (σ ,θ ) dσ dθ
0 0
Komen et al. (1984) estimated the value of Γ by closing the energy balance of the
waves in fully developed conditions for which Γ depends of overall wave steepness.
Günther et al. (1992) developed the following equation for Γ:
p
⎛ k ⎞⎛ ~ s ⎞
Γ = Cds ⎜ (1 − δ ) + δ ~ ⎟⎜⎜ ~ ⎟⎟ (6.10)
⎝ k ⎠⎝ s PM ⎠
The coefficients Cds, δ and exponent p are tuneable coefficients, ~ s is the overall wave
~ 0.5
steepness ( ~
s = k .Etot ) and ~
s PM is the value of ~
s for the Pierson-Moskowitz spectrum
(~
s PM =(3.02×10-3)1/2 ). The values of the tuneable coefficients Cds, δ and exponent p in
Equation (6.10) have been obtained by Komen et al., (1984) by closing the energy
balance of the waves in idealised wave growth conditions for deep water. The tuneable
coefficients in Equation (6.10) depend on the wind input formulation that is used. For
the wind input of Komen et al. (1984) Cds = 2.36×10-5, δ = 0 and p = 4.0.
6.1.1.3 Bottom Friction Dissipation
Three optional models are available in SWAN to calculate wave dissipation due to
bottom friction - the empirical model of JONSWAP (Hasselmann et al., 1973), the drag
126
law model of Collins (1972), and the eddy viscosity model of Madsen et al. (1988). All
three models are investigated in various numerical simulations reported in Chapter 7
and are discussed here. All the three models can be expressed in the same formulation
as provided by Bouws and Komen (1983):
σ 2
S ds ,b (σ , θ ) = −Cbottom E ( σ ,θ ) (6.11)
g 2 sinh 2 (kh )
where Sds,b ( f , θ ) is wave energy dissipation due to bottom friction, Cbottom is a bottom
friction coefficient that generally depends on the bottom roughness and horizontal
orbital motion Ub rms:
2π ∞
σ2
= ∫∫ sinh (kh) E (σ ,θ )dσ dθ
2
U b rms 2
(6.12)
0 0
The JONSWAP experiment (Hasselmann et al. 1973) suggested the mean value of
Cbottom= 0.038 m 2 s - 3 for swell condition. Bouws and Komen (1983) adopted the value
of 0.067 m 2 s - 3 for Cbottom when waves are fully developed in shallow water.
Hasselmann and Collins (1968) applied the quadratic friction law to compute the local
dissipation of the wave spectrum. The drag-law expression is proportional to a friction
factor Cf that has to be determined empirically:
Cbottom = C f gU rms (6.13)
Hasselmann and Collins (1968) found C f = 0.015 according to field and experimental
data.
The eddy-viscosity model for bottom friction dissipation was derived by Madsen et al.
(1988). Their model was based on the formulation of the turbulent flow in a bottom
boundary layer. The bottom friction coefficient Cbottom may be derived from the model
of Madsen et al. (1988) as:
g
Cbottom = f wr U rms (6.14)
2
127
in which fwr is the wave friction factor similar to that proposed by Jonsson (1966):
1 1 A
+ log10 = log10 br − 0.17 (6.15)
4 f wr 4 f wr kb
where kb denotes the equivalent bottom roughness (= 30 z o ) and Abr is a representative

near-bottom orbital excursion amplitude:
2 2π ∞
⎛U ⎞ 1
A = ⎜⎜ b rms
2
⎟⎟ = 2 ∫ ∫ E (σ ,θ ) dσ dθ (6.16)
br
⎝ ωr ⎠ 0 0
sinh 2
(kh )
where Ubrms is bottom orbital velocity (Equation 6.12) and ωr is radian frequency
(Madsen et al., 1988).
6.1.2 Wave Propagation Through Obstacles
SWAN can provide estimates of wave transmission over/through a structure (obstacle)

such as a breakwater. An obstacle is modelled as a line in SWAN since obstacles
usually have a transversal area that is too small to be resolved by the bathymetry grid. If
the crest of the breakwater is at a level where (at least part of the) waves can pass over,
the transmission coefficient Kt is a function of wave height and the difference between
crest level and water level. The method of Goda et al. (1967) is applied in the SWAN
model:
⎡ ⎛ π ⎛ hs ⎞ ⎞⎤ hs
K t = 0.5 ⎢1 + sin ⎜⎜ ⎜⎜ − β ⎟⎟ ⎟⎟⎥ for β −α < <α + β (6.17)
⎣⎢ ⎝ 2α ⎝ H i ⎠ ⎠⎦⎥ Hi
where hs is the submergence depth over the breakwater and Hi is the incident
(significant) wave height at the offshore side of the obstacle. The coefficients α, β
depend on the shape of the obstacle as provided by (Seelig, 1979):
Case α β
Vertical thin wall 1.8 0.1
Caisson 2.2 0.4
Dam with slope 1:1.5 2.6 0.15
128
Transmission coefficient in SWAN can also be specified by the user regardless of
Equation 6.17. In this case, the obstacle is defined as a sheet with a constant defined
transmission coefficient.
It should be noted that in the SWAN module it is assumed that the frequencies of the
incident spectrum remain unhampered over the breakwater (no spectral shape changes)
and only the energy scale of the spectrum is affected by the obstacle.
6.1.3 Wave-induced forces
Depth averaged wave-induced forcing “Fx” (in x- direction) can be defined in the
SWAN module using two different approaches: (i) radiation stress tensor or (ii) wave
energy dissipation after Dingemans et al. (1987).
i- The radiation stress tensor in SWAN can be approximated using the energy
spectrum and wave celerity ratio n:
⎡ 1⎤
S xx = ρ g ∫ ⎢n cos 2 θ + n − ⎥ E dσ dθ
⎣ 2⎦
⎡ 1⎤
S yy = ρ g ∫ ⎢n sin 2 θ + n − ⎥ E dσ dθ (6.18)
⎣ 2⎦
S xy = S yx = ρ g ∫ n sin θ cosθ E dσ dθ
where Sxx, Syy, Sxy and Syx are radiation stress components in x, y, xy and yx directions.
Depth average wave-induced forces per unit surface area in x- and y- directions can then
be defined in terms of the radiation stress tensor as follows:
∂S xx ∂S yx
Fx = − −
∂x ∂y
(6.19)
∂S yx ∂S yy
Fy = − −
∂x ∂y
ii- Dingemans et al. (1987) showed that under the commonly occuring conditions
of slowly varying wave fields, the driving force per unit mass is closely proportional
to the wave energy dissipation per unit area and formulation of the driving forces in
129
terms of the wave dissipation yields more reliable results than those obtained by
numerical differentiation of the radiation stress tensor. They provided the following
equation in which wave forces are determined only by the wave dissipation:
kx
Fx = Dtot
ω
(6.20)
ky
Fy = Dtot
ω
in which kx and ky are wave numbers in x and y direction, respectively, ω is wave

frequency and the total dissipation rate Dtot is computed from the calculated wave
dissipation due to bottom friction, depth-induced breaking, and whitecapping that are
defined in Section 6.1.1 (i.e. Equation 6.3, 6.8 and 6.11).
6.2 Conceptual Description of Flow Model
Hydrodynamic numerical modelling of the flow over submerged breakwaters was

carried out using Delft3D-FLOW. FLOW module of Delft3D is a multi-dimensional
(2D or 3D) hydrodynamic (and transport) simulation program which calculates non-
steady flow and transport phenomena that result from tidal and meteorological forcing
on a rectilinear or a curvilinear, boundary fitted grid. The area of application of
Delft3D-FLOW can be summarised as follows (for more description see Delft
Hydraulic 2005b):
• Tide and wind-driven flows (i.e. storm surges).
• Stratified and density driven flows.
• River flow simulations.
• Simulations in deep lakes and reservoirs.
• Simulation of tsunamis, hydraulic jumps, bores and flood waves.
• Fresh-water river discharges in bays.
130
• Salt intrusion.
• Thermal stratification in lakes, seas and reservoirs.
• Cooling water intakes and waste water outlets.
• Transport of dissolved material and pollutants.
• Online sediment transport and morphology.
• Wave-driven currents.
• Non-hydrostatic flows.
Among the areas of application of Delft3D-FLOW, the wave-driven current simulation

is of interest in the present research. Two approaches for wave-current interaction are
available in the FLOW module. The first one is the application of conventional depth-
averaged momentum equations for which the wave climate at each grid point is
imported from the WAVE (SWAN) model. The second approach that is available in the
latest version of Deft3D-FLOW (3.49.02, June 2005) includes a “roller model” for wave
breaking dissipation. The roller model will be discussed in Section 6.3.
The three dimensional Navier Stokes equations are applied and solved in Delft3D-
FLOW under the shallow water and Boussinesq assumptions. The system of equations
consists of the horizontal equations of motion, the continuity equation and the transport
equations for the conservative constituents.
6.2.1 Continuity Equation
The depth-averaged continuity equation in FLOW module is given by:
∂ζ
+
1 [
∂ (h + ζ )U Gηη
+
1 ]∂ (h + ζ )V Gξξ[=Q
] (6.21)
∂t Gξξ Gηη ∂ξ Gξξ Gηη ∂η
where ζ is water level above the datum, Gξξ and Gηη are the transformation coefficient
of curvilinear co-ordinate to rectangular, U and V are depth-averaged velocity in ξ and η
131
direction, h is water depth and Q is the global source or sink per unit area. For further
description of the parameters, see Delft Hydraulics (2005b).
6.2.2 Momentum Equation
The momentum equation in ξ-direction, averaged over the wave motion is given by:
∂U U ∂U V ∂U UV ∂ Gξξ V2 ∂ Gηη
+ + + − − fV =
∂t Gξξ ∂ξ Gηη ∂η Gξξ Gηη ∂η Gξξ Gηη ∂ξ
(6.22)
1 gU U + V
2 2
− Pξ − + Fξ + M ξ
ρ o Gξξ C 22D (d + ζ )
where f is Coriolis coefficient, ρo is reference density of water, Pζ is gradient

hydrostatic pressure in ξ direction, Fξ is turbulent momentum flux in ξ direction, C2D is
the 2D-Chézy coefficient (friction coefficient) and Fξ is the depth averaged wave-
induced forcing in ξ direction. Mξ represents the contributions due to external sources or
sinks of momentum (external forces by hydraulic structures, discharge or withdrawal of
water, etc.).
The momentum equation for η direction is of similar form to that for ξ direction given
in Equation (6.22).
Wave induced forces can be approximated from either (i) radiation stress gradient or
(ii) wave energy dissipation in SWAN (see Section 6.1.3).
6.2.3 Bottom Friction in FLOW
Bed shear stress due to flow alone is assumed to be given by a quadratic friction law:
r r
ρ 0 gU U
τc = (6.23)
C 22D
r
where U is the depth average horizontal velocity. The 2D-Chézy coefficient C2D can be
determined directly from the corresponding tables or calculated using Manning (1890)
or Colebrook (1939) equations in Delf3D-FLOW.
132
Manning (1890) provided the following empirical relationship for Chézy coefficient:
6
h
C2D = (6.24)
nˆ
where h is total water depth and n̂ is Manning coefficient that can be determined
depends on surface condition of the channel.
Colebrook (1939) developed an empirical equation to calculate Chézy coefficient using

water depth h and Nikuradse roughness length ks:
⎛ 12h ⎞
C 2 D = 1810 log⎜⎜ ⎟⎟ (6.25)
k
⎝ s ⎠
6.2.4 Wave-induced enhancement of bed shear stress
Wave-current interaction has the effect of enhancing both the mean and oscillatory bed
r
shear stresses. The bed shear stress due to the combination of waves and current “ τ m ”
is enhanced beyond the value which would result from a linear addition of the bed shear
r r
stress due to waves “ τ w ”and the bed shear-stress due to current “τ c ”.
r
For 2D depth average current, τ c can be determined from Equation (6.23). The wave-
induced shear stress component is computed using wave orbital velocity near the
bottom “u2orb” and the friction coefficient fwr :
r 1
τ w = ρ o f wr u orb
2
(6.26)
2
where uorb may be determined using linear wave theory:
1 π ω ⋅ H rms
u orb = (6.27)
2 2 sinh (kh )
Hrms and wave frequency are computed in SWAN and read from the communication
file. For more definition of wave friction coefficient fwr see Delft Hydraulic (2005b).
133
The maximum bed shear stress (wave and current combined) is calculated in FLOW
using the proposed formula by Soulsby et al. (1993):
r r r
τ m = y (τ c + τ w ) (6.28)
where the shear stress enhancement factor is given by
[
y = x 1 + bx p (1 − x )
q
] (6.29)
More discussion of the parameters in Equation 6.0 and 6.0 are provided in Delft
Hydraulic (2005b).
6.3 Roller Model
The surf beat or “roller model”, a recent extension to the Delft3D-FLOW was
developed to include the effect of short waves on long waves. The spatial variations of
radiation stresses cause long waves to travel with short waves as bound waves. The
roller model is able to include the forcing due to the short waves. The dominant wave
frequency and direction from narrow-banded wave spectrum are used to determine the
direction of propagation of the short waves as carrier waves.
The mean direction of the wave field is imported from the SWAN model using the usual
coupling provided by the WAVE module whilst the short-wave (roller) energy is
modelled in Delf3D-FLOW (it being directly determined from short-wave components
provided at the boundaries).
The propagation of the obliquely incident short wave group on a variable bathymetry is
obtained from the short wave energy balance, Ew (Reniers et al. 2004):
∂E w ∂E w C g cos θ ∂E w C g sin θ
+ + = − Dw (6.30)
∂t ∂x ∂y
where Cg is the group velocity corresponding to the peak period of the waves, θ is the
wave direction and Dw is the wave energy dissipation. x and y are the distance in the
cross-shore and alongshore direction, respectively. Roelvink (1993) provided the
134
following equation for calculating the expected dissipation rate for a given wave energy
and water depth using the dissipation model for random waves developed by Battjes and
Janssen (1978):
⎡ ⎡ n
⎤⎤
⎢ ⎢ ⎛⎜ E w ⎞ 2 ⎥⎥
Dw = 2α f p E w ⎟
⎢1 − exp⎢− ⎜ γ 2 E ⎟ ⎥⎥
(6.31)
⎢ ⎢⎣ ⎝ ref ⎠ ⎥⎥
⎣ ⎦⎦
where
1
E ref = ρ g h 2 (6.32)
8
in which α is a constant of order 1 (see Battjes and Janssen, 1978 for more details), fp is
the peak frequency, γ is the wave breaking parameter, ρ is the water density and h is
total water depth including set-up. n is a variable to be determined from observation and
acts as a dissipation parameter corresponding to the randomness of the incident waves.
Roelvink (1993) calibrated the probabilistic model of Thornton and Guza (1983) using
several experimental data and found that the values of α=1.0, n=10 and γ=0.55 give the
best results for the probabilistic model of Thornton and Guza (1983).
Through the process of wave breaking, the wave energy is transformed into roller
energy which is located in the down-wave area after breaking and is rapidly dissipated
in shallow region:
∂Er ∂ ∂
+ 2 Er c cosθ + 2 Er c sin θ = Dw − Dr (6.33)
∂t ∂x ∂y
Er is roller energy, c is the phase velocity, Dw is wave energy dissipation (Equation

6.31) and Dr represents the roller energy dissipation contributed by Nairn et al. (1990)
to be:
βρgA 2 gβ E r
Dr = = (6.34)
T C
135
in which β is a dissipation coefficient given by the angle of inclination of the roller
(sinθ. cosθ) (Figure 6.2) and A is the roller area. Dally and Brown (1995) found that the
value of β=0.10 provides a very good comparison of the model to data of Hansen and
Svendsen (1984). However, it can be used as a calibration parameter of the model.
Svendsen (1984a) applied Duncan (1981) measurement of breaker surface behind a

hydrofoil and provide the following equation for calculating A:
A = 0.9 H 2 (6.35)
The variation of wave and roller energy produce variation in the radiation stress
equations. Reniers et al. (2002) presented the radiation stress equations including the
roller contribution:
⎡⎛ ⎤
( ) 1⎞
S xx = ⎢⎜ n 1 + cos 2 θ − ⎟ E w + 2 Er cos 2 θ ⎥
⎣⎝ 2⎠ ⎦
⎡⎛ ⎤
( ) 1⎞
S yy = ⎢⎜ n 1 + sin 2 θ − ⎟ E w + 2 E r sin 2 θ ⎥ (6.36)
⎣⎝ 2⎠ ⎦
S xy = [(n cosθ sin θ )E w + 2 Er cos θ sin θ ]
where n is the ratio of group velocity to phase velocity (Cg/C) and the wave energy Ew
and the roller energy Er are calculated from a wave and roller energy balance,
respectively.
Therefore, in the application of the “roller model”, the Delft3D-FLOW module only
reads wave direction and period (wave celerity) from the WAVE module and then
calculates the wave energy using wave height at the up-wave boundary. Values of roller
and wave energy for the next grid are calculated using wave energy dissipation
equations, iteratively. It is assumed that the amount of roller energy Er at the boundary
is zero.
136
1
0.8
0.6
Qb
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
H rms /H m
Figure 6.1. Variation of breaking probability Qb in Battjes and Janssen (1978) theory
against non-dimensional ratio of Hrms/Hm
Figure 6.2. Definition sketch for wave roller (after Dally and Brown, 1995).
CHAPTER 7
NUMERICAL MODELLING RESULTS
Numerical modelling of the hydrodynamic effects of wave breaking and flow over
submerged breakwater was undertaken using Delft3D - the results were then compared
to the laboratory experimental data. Two modules of Delft3D, WAVE (SWAN) and
FLOW, were applied in the numerical model. Wave climate deformation along the
flume without and in the presence of a submerged breakwater was computed in SWAN
and the calculated wave characteristics were imported in the communication file to be
used by the FLOW module to calculate wave-induced current in the flume. This
chapter reports on the capability of numerical modelling to simulate wave height
transformation and wave-induced current over submerged breakwaters.
Wave-current interaction was included in all models to enhance simulation results.

Activating online wave-current interaction in FLOW and WAVE modules, the effects of
waves on current (forcing, enhanced turbulence and bed shear stresses) and the effect of
flow on waves (set-up, current refraction and enhanced bottom friction) are accounted
for in the model. Every time the communication file is written by FLOW module,
subsequently a WAVE simulation is performed using the updated flow and water level
data. Then FLOW resumes, using the latest WAVE results.
From the extensive laboratory data sets discussed in Chapter 3, the tests with breaking
wave conditions over the breakwater and the submergence ratio greater than 0.5 (hs / Hi
> 0.5) were examined - in these cases all breaking waves pass over the breakwater
without overtopping (Hi < 2hs). 26 tests were examined in wave transformation
simulations and 18 tests were adopted for investigating wave-induced current over the
breakwater. The list and conditions of the tests applied in the numerical modelling are
provided in Table 7.1. As will be discussed in the following section, the experimental
tests were scaled up for implementation within the numerical modelling.
137
7.1 Model Setup
7.1.1 Scale of the Model
The stability of the model and accuracy of results are determined by the scale (time and
space) and grid size of the model. In Delft3D-FLOW, the time step is chosen to satisfy a
criteria based on Courant number (Cr) for wave propagation;
⎛ 1 1 ⎞
C r = 2Δt ⋅ gh ⎜⎜ 2 + 2 ⎟⎟ (7.1)
⎝ Δx Δy ⎠
where Δt is the time step (in seconds), g is the gravitational acceleration, h is the water
depth and Δx and Δy are the smallest grid sizes in x- and y- direction. Delft3D is
inaccurate for Courant numbers larger than 4 2 ( Delft Hydraulics, 2005b) and the
upper bound for the Courant number occurs in the most critical situation, namely in case
of a narrow channel (width of few grid sizes) that makes an angle of 45 degrees with the
computational grid. In practical situations, the Courant number should not exceed a
value of 1.0. However, it is advised to carry out sensitivity tests in order to determine
the largest time step for which the model still yields accurate results.
A grid size of 1 m (at scale of 10 represents 10 cm in the laboratory tests) combined

with a 0.01 minute time step (0.20 second in the laboratory tests) achieved stable and
suitably accurate Delft3D numerical model results (Cr ~ 0 . 1 8 ) . Preliminary sensitivity
tests at smaller grid sizes and smaller time steps (small Cr) indicate that solution
accuracy was not significantly improved for grid sizes and time steps less than the 1 m
and 0.01 seconds. Figure 7.1 indicates that applying finer grid size (0.2 m) in the
WAVE model produces only 4% higher transmitted wave height than the model with
1.0 m grid size (test T1-27; Hi= 1.01 m, T=4.71 sec, B=35 m and h= 4.0 m) – more
CPU time is needed to run the model with finer grid size. Therefore, 1m grid size and
0.01 minute time step were judged suitable for ongoing comparative testing. Suitable
accuracy for assessing the relative performance of the various Delft3D options being all
that required for the thesis evaluation.
138
Using dimensionless analysis and keeping the values of gravitational acceleration and
fluid properties the same as the prototype (experiment), the scale, N, of the fluid flow
and wave characteristics was applied in the numerical model as follows (ξ and η are
used in the numerical model for x- and y- dimensions/measurements from the
experiments):
Ng = Nρ = Nμ = 1.0
Nξ =Nη = Nz = NL = N=10
Nu = Nv = NC = N ½ =3.3 (7.2)
NQ = N 5/2 = 3 1 6 . 2
NM = NP = NF = N 3 =103
Nn = N 1/6 =1.47
where the subscripts indicate the properties of fluid flow (For definition see the List of
Symbols).
Table 7.1 Scaled test conditions used in the numerical model.

Test No. h (m) B (m) hs(m) Hi (m) T (sec) Lo (m)
T1-21 4.0 35.0 1.0 0.56 3.2 15.0
T1-25 4.0 35.0 1.0 0.51 4.7 27.3
T1-26 4.0 35.0 1.0 1.02 4.7 27.3
T1-27 4.0 35.0 1.0 1.45 4.7 27.3
T1-30 4.0 35.0 1.0 1.03 6.3 38.6
T1-42 4.5 35.0 1.5 1.00 4.7 28.4
T1-43 4.5 35.0 1.5 1.55 4.7 28.4
T2-21 4.0 15.0 1.0 1.58 3.2 15.0
T2-22 4.0 15.0 1.0 1.13 3.2 15.0
T2-25 4.0 15.0 1.0 0.50 4.7 27.3
T2-26 4.0 15.0 1.0 1.07 4.7 27.3
T2-30 4.0 15.0 1.0 1.06 6.3 38.6
T2-43 4.5 15.0 1.5 1.52 4.7 28.4
T2-46 4.5 15.0 1.5 1.18 6.3 40.7
T2-47 4.5 15.0 1.5 1.56 6.3 40.7
T3-21 4.0 3.0 1.0 0.58 3.2 15.0
T3-22* 4.0 3.0 1.0 1.01 3.2 15.0
T3-26 4.0 3.0 1.0 1.02 4.7 27.3
T3-27 4.0 3.0 1.0 1.32 4.7 27.3
*
T3-30 4.0 3.0 1.0 1.17 6.32 38.6
T3-43* 4.5 3.0 1.5 1.33 4.7 28.4
* only used in the wave model.
139
The width of grid size in the wave model is one cell greater that the actual width of the
flume at both sides to eliminate the effect of lateral boundaries on wave fields in the
main calculation grids. A sample of the computational grid associated with the modelled
breakwater is shown in Figure 7.2. The model grid included the central channel that
contained the submerged breakwater/reef and the two full depth side channel the same
as the laboratory experimental model (see Chapter 3).
7.1.2 Boundary Conditions
In the WAVE (SWAN) model, all boundaries at both longitudinal sides and the
downstream dissipating beach were assumed as non-reflective open boundary in which
waves can pass, perpendicularly, through the boundaries with no reflection. The
offshore boundary conditions at the west side of the model (paddle) were defined so as
to eliminate the differences between target wave conditions and the generated waves
measured in front of the breakwater (in laboratory). The experimental unidirectional
monochromatic waves were defined in SWAN as a narrow width spectrum. The
spectrum was adopted such that the total energy of the monochromatic wave and the
spectrum are equal:
1 1
∫S df = ρ g H i2 = ρ g H rms
2
f (7.3)
8 8
The spectrum was specified in SWAN as an open boundary by editing the spectrum.bnd
file. A sample of defined spectrum file is presented in Appendix C.
In the FLOW model, the offshore boundary was defined as an open boundary with the
fixed defined water level (still water level) since wave conditions are defined offshore
using the “roller” model. All other lateral boundaries are represented as impermeable
boundaries in the flow model. The main central channel was simulated by defining 2
thin dams along the flume in the flow model with no lateral friction.
7.1.3 Physical Parameters
The physical parameters used in the numerical simulation for all test conditions were
defined as follows:
140
• Gravity acceleration (g) – g =9.81 m/s2 in experiment and Delft3D.
• Water density (ρ)
The flume in the laboratory was filled with fresh water transferred from Manly dam
where located at the top of Water Research Laboratory. During the physical and
chemical measurements of the water properties, the density was found close to 1000
Kg/m3 which are used in the Delft3D model.
• Bottom roughness
The Manning formula was applied to determine the 2D-Chézy coefficient. The value of
Manning coefficient n for prototype was adopted 0.012 over the breakwater (planed
wood) and 0.014 over the finished concrete bottom of the flume (Streeter et al., 1998).
In the scaled model, Manning coefficients were calculated from Equation 1 with respect
to the scale factor of 10. The spatially varied bottom roughness was defined in the flow
model by creating the t.rgh file. In the *.rgh file the values of Manning coefficients were
identified at each grid point (see Appendix C).
7.2 Wave Model
Wave propagation along the flume and over the breakwater was simulated using the
wave module (SWAN) of Delft3D. To simulate the evolution of random, short-crested
wind-generated waves in estuaries, tidal inlets, lakes etc., the SWAN model can be
applied. The scaled model of the flume was simulated in SWAN to calculate the wave
field along the flume. In order to prevent the effects of side boundaries on wave field in
the model the width of the wave computational grid was defined one grid (1.0 m) bigger
than the actual scaled model on both sides.
The breakwater was defined with the same scaled length, height and location as the
experiment in the flume. The width of the modelled breakwater was considered one grid
size (1.0 m) wider than the scaled width in order to prevent the effect of abrupt water
141
depth changes across the flume on the wave field through the main channel in which all
laboratory measurements have been made.
7.2.1 Sensitivity Analysis
In order to better understand the effects of various physical parameters on wave

propagation, the experimental flume was simulated in SWAN with and without (empty
flume) the breakwater.
7.2.1.1 Empty flume
In the so called empty flume (without the submerged breakwater) the effects of bottom
friction, wave-induced breaking and whitecapping on wave climate along the flume is
investigated.
Bottom friction
As indicated in Chapter 6, three processes are optionally available in SWAN to

calculate wave energy dissipation due to bottom friction:
• The empirical model of JONSWAP (Hasselmann et al., 1973),
• The drag law model of Collins (1972) and
• The eddy viscosity model of Madsen et al. (1988).
The effects of the bottom friction on wave propagation along the empty flume with the
above three approaches was examined for the test condition (T1-27) which has the
highest value of wave steepness (Hi/L) thus highest importance of bottom friction. The
scaled test conditions for test T1-27 are redefined in Table 7.2.
The bottom friction Cbottom for JONSWAP and Collins’ method in Equations 6.11
(Chapter 6) was defined 0.067 and 0.015 according to the values proposed by Bouws
and Komen (1983) and Collins (1972), respectively. In the eddy viscosity model of
Madsen et al. (1988), the bottom friction is as a function of non-dimensional friction
factor, which is calculated in SWAN.
142
The bottom friction models of Hasselmann et al. (1973) and Madsen et al. (1988) are
based on depth-induced friction energy loss of irregular waves. The Collins’ (1972)
method is based on a conventional formulation for periodic waves. The variations of
Table 7.2 Wave and water depth condition for scaled model of tests T1-26 and T1-27.
Water
Test Breakwater crest Breakwater crest Incident wave Incident wave
depth “h”
Number hight “d” (m) width “d” (m) height “Hrms” (m) period “T” (sec)
(m)
T1-26 4.0 3.0 35.0 1.02 4.71
T1-27 4.0 3.0 35.0 1.45 4.71
wave height along the flume for the three approaches are shown in Figures 7.3, 7.4 and
7.5. As indicated in the figures the attenuation of wave height along the empty flume is
very small with respect to the incident wave energy. The attenuation of wave height
along the whole flume in the numerical model (226 m length) was calculated to be 2.8%
for Hasselmann et al., 4.1% for Collins and 8.6% for Madsen et al. method. The wave
attenuation along the flume from the location of wave generation (paddle) up to the
offshore toe of the breakwater (where waves travel unhampered in the experiment) were
calculated 1.4%, 2.1% and 4.5% for Hasselmann et al., Collins and Madsen et al.’s
methods, respectively. The method of Collins (1972) is considered to be more
appropriate herein since the method is based on a conventional formulation for periodic
waves which applied during the experimental test (See Chapter 6 for more declaration).
The effect of bottom friction on wave propagation from the paddle up to the breakwater,
where waves break, can be neglected in the ongoing numerical modelling. The
attenuation of wave height due to bottom friction was found negligible during the
laboratory experimental tests without breakwater.
Whitecapping
The effect of whitecapping on wave energy dissipation and gradient of wave-induced

forces were investigated for the flume without breakwater with the same wave and
water depth conditions of test T1-27. As defined in Chapter 6, the processes of
whitecapping in the SWAN model are represented according to the pulse-based model
of Hasselmann (1974) and the values of the “tuneable” coefficients in the model have
been adopted from Komen et al. (1984). The value of wave dissipation due to
143
whitecapping depends on wave steepness and can be identified as a steepness-limited
breaking process (Johnson, 2006). The variation of wave height and energy dissipation
along the flume for the test condition of T1-27 (Figure 7.6) reveals that there is no
significant steepness-induced wave energy dissipation along the flume without
breakwater, which is consistent with the laboratory experimental observations.
Therefore, the effect of whitecapping on wave deformation between the paddle wave
generator and the offshore toe of the breakwater can be neglected in the numerical
simulation.
Depth-induced wave breaking
The other process that was investigated through the model without breakwater is wave
energy dissipation due to depth-induced breaking. SWAN applied the proposed method
of Battjes and Janssen (1978) and Eldeberky and Battjes (1995) for calculating depth-
induced wave breaking dissipation. The theory has been fully illustrated in Chapter 6.
The sensitivity of the model to breaker parameter γ is examined here. The model was
set-up for the flume without breakwater with the scaled wave and water level conditions
of test T1-26 (Hi=1.02m, T=4.7sec and h=4.0m). Running the SWAN model with
different values of breaker parameter γ, for the flume without breakwater and wave
conditions of T1-26, represents the trend of wave dissipation along the flume. Figure
7.7 indicates the rate of significant wave height (Hs) attenuation and total wave energy
dissipation (Dtot) along the flume predicted by SWAN for different values of adopted
breaker index “γ ” . As indicated in the figure wave height and wave energy dissipation
along the flume vary between 0.4% to 25% along the flume for the range of assumed
breaker index 0.5≤ γ ≤1.0.
The measurements of incident wave heights in the laboratory were mainly carried out
using the traverse probe travelling along the main flume. The probe travelled in the
distance of 4m to 1m (in laboratory) offshore the breakwater in the main channel (2.5m
up-wave from the breakwater in average). Therefore, it is desired that the numerical
model provides almost the same incident wave height at the middle of the scaled
incident wave height measurement area from the laboratory. The SWAN model accepts
the wave conditions only at the boundaries and there are no facilities to introduce wave
144
conditions at a specific position along the model grids. Moreover, the breaking index γ
is considered constant for the entire grid of the model. Investigation of wave dissipation
trend along the flume with different values of γ (Figure 7.7) revealed that for the breaker
index greater than 0.7 (corresponds to Hrms/Hm < 0.39) wave energy dissipation between
the paddle (open boundary) and the measurement point of incident wave height (in the
laboratory) is less than 5%. Therefore, the amount of energy that approaches the
breakwater (incident wave height at offshore toe of the breakwater) is almost the same
in the experiment and numerical model.
7.2.1.2 Flume with Breakwater
The sensitivity of the SWAN model to bottom friction and whitecapping were
investigated for the flume in the presence of a submerged breakwater. The influence of
depth-induced wave breaking on wave deformation will be inclusively demonstrated in
Section 7.2.2.
Bottom friction
The influence of bottom friction will be more important in shallow water (over the
breakwater) for non-breaking waves since the horizontal vector of wave orbital velocity
near the bottom is higher in shallow water condition. However, it is expected to be less
important (possibly negligible) in the presence of depth-induced wave breaking. The
scaled model of test T1-27 with 35 m wide breakwater was run in SWAN. The method
of Collins (1972) within SWAN was used for the bottom friction case. To match (within
±2%) the SWAN calculated transmission coefficient to the test T1-27 experimental
result (Kt = 0.47) a breaker index value of 1.85 was found to be needed within SWAN.
Further detailed discussion of the breaking parameter in SWAN is presented in Section
7.2.3. Figure 7.8 and Figure 7.9 show the variation of wave height and wave energy
dissipation along the flume, respectively, for the condition of depth-induced wave
breaking over the breakwater with and without bottom friction. The solid red and dash
green lines in Figure 7.8 indicate wave height variation when only depth-induced
breaking and both bottom friction and depth-induced breaking are considered in the
model. As implied from Figure 7.8, wave height behind the breakwater has a slight
difference when depth-induced breaking is accompanied with bottom friction in the
145
model (less than 4%). Therefore, wave energy dissipation due to bottom friction has
been neglected in the ongoing numerical modelling with respect to depth-induced
breaking dissipation.
Whitecapping
The comparison of wave height attenuation and wave energy dissipation along the
flume in the presence of a breakwater with and without applying steepness-induced
breaking (whitecapping) in SWAN model associated with depth-induced breaking are
shown in Figure 7.10 and Figure 7.11. Test conditions are the same as the previous
section (T1-27) with breaking parameter γ =1.85. As indicated in Figure 7.10, there is
only small wave height difference (2.2%) behind the breakwater between the models
with and without whitecapping effects.
7.2.2 Breaking Dissipation in SWAN
The previous section confirmed that the dominant process of wave height attenuation in
the numerical model (SWAN) is depth-induced breaking in the presence of a submerged
breakwater. The only breaking dissipation process which is available in Delft3D-
WAVE is the method proposed by Battjes and Janssen (1978) and Eldeberky and
Battjes (1995) which are comprehensively illustrated in Chapter 6. The dissipation for a
spectral component per unit time is calculated in SWAN with:
E ( f ,θ )
S ds ,br ( f , θ ) = Dtot (7.4)
Etot
where E(f, θ) is the wave energy density at frequency f and in θ direction, Etot is the total
wave energy of the spectrum and Db indicates the average power dissipated in the
breaking process, per unit area:
α
Db = Qb f ρ g H m2 (7.5)
4
in which α is a constant of order one, f represents the mean frequency of the

progressive wave, ρ is the water density, g is gravity acceleration and Hm is the
maximum wave height which the given water depth can sustain. Qb is the probability
146
that at a given point the wave height is equal or greater than the breaking point that is a
function of the ratio of root mean-squared wave height to Hm:
2
1 − Qb ⎛H ⎞
= −⎜⎜ rms ⎟⎟
lnQb ⎝ Hm ⎠ (7.6)
The maximum wave height Hm in SWAN is considered with Hm=γ h, in which the
breaker parameter γ is defined as an adjustable parameter (Battjes and Janssen, 1972)
and h is the total water depth (including the wave-induced set-up if computed by
SWAN). In SWAN, the breaker parameter is constant along the entire wave grid. In the
literature, the breaker parameter γ is often a constant or spatially variable as a function
of bottom slope or incident wave steepness (McCowan, 1891; Miche, 1944; Galvin,
1968; Collins and Wier, 1969; Smith and Kraus, 1991; Hara et al., 1992; Nelson, 1997)
However, the dependency of breaker parameter on incident wave steepness is not
available in SWAN (see Chapter 4, Section 4.4, for more discussion on wave breaker
parameter).
The theory of Battjes and Janssen (1978) has been applied successfully for modelling
the breaking processes on a mild slope or barred beaches. For a wide range of wave
conditions over plane and barred beaches, Battjes and Stive (1985) used both laboratory
and field data to calibrate the dissipation model of Battjes and Janssen (1978). They
adopted the value of α=1.0 in the model and calibrated the model by estimating optimal
values of γ to enhance the best fit of the model to the experimental data. The value of
breaker index (adjustment parameter) was found in the range of 0.6 to 0.83 with the
average of 0.73 for plane and barred beaches (Battjes and Stive, 1985).
As indicated in Chapter 6, the value of breaker parameters γ in SWAN was considered

the average value of 0.73 (Battjes and Stive, 1985) to investigate the applicability of the
breaking dissipation model of Battjes and Janssen (1978) for submerged breakwaters.
The same wave and water depth conditions as Tests T1-26, T2-26, and T3-26 were
modelled in SWAN for the breakwater crest width 35 m, 15m, and 3m, respectively.
The calculated transmission coefficients were compared with the measured transmission
coefficient from laboratory experiments to investigate the validity of the dissipation
model for submerged breakwaters/reefs. The transmitted wave height in the model was
147
calculated at the same scaled locations where the equivalent measurements were carried
out in the laboratory experiments for each test. Figure 7.12 shows the attenuation of
wave height due to depth-induced wave breaking over the breakwaters calculated by
SWAN (Battjes and Janssen, 1978). With the constant value of γ = 0.73 for all test
conditions the condition of incident wave height and water depth for the three tests are
almost the same (Hrms ~ 1.0 m, T=4.7 sec and h=4.0 m). As indicated from the figures,
wave dissipation is higher for the breakwater with wider crest width, which is,
qualitatively, consistent with experimental results. However, SWAN calculated wave
height dissipation considerably higher than that measured in the experiment with
resultant SWAN model predicted transmission coefficient (Kt) notably lower than those
measured in the laboratory flume experiments. The transmission coefficient was
calculated from the computed wave height. The transmitted wave height was calculated
from the average of computed wave heights at the same location as the probes 8 to 10 in
the experiment Figure 7.13 shows the comparison of measured and calculated (model)
transmission coefficient for three tests.
The calculated values of transmission coefficient (from the SWAN model) are plotted
against the measured values in Figure 7.14 for all test conditions in Table 7.1. As shown
in the figure, the value of transmission coefficients calculated by SWAN with the
breaker parameter γ =0.73 are lower than the measurements.
Therefore, applying the energy dissipation model of Battjes and Janssen (1978) leads to
higher energy dissipation over submerged breakwaters, although, it has been shown to
be appropriate for plane or barred beaches (Battjes and Stive, 1985).
This finding is consistent with the investigations of Zanuttigh et al. (2003) and Johnson
(2006) which used the experimental wave basin results of the tests on low-crested
breakwaters carried out by Zanuttigh and Lamberti (2003) at Aalborg University,
Denmark. The wave basin tests consist of regular and irregular waves for overtopping of
30 cm crest height submerged breakwaters with 20 cm and 60 cm crest width. Johnson
(2006) calculated the transmission coefficient using the dissipation method of Battjes
and Janssen (1978) method applied in the MIKE 21 PMS numerical wave model. He
adopted the 8 test results of Zanuttigh and Lamberti (2003) with submerged breakwater
148
and random directional waves in his modelling of wave breaking over submerged
breakwater. Table 7.3 shows the test conditions used by Johnson (2006).
Table 7.3 Test conditions applied in the numerical analysis by Johnson (2006).
Test B h hs Tp Hi
Lo (mm)
series (mm) (mm) (mm) (sec) (mm)
17 200 340 70 1.97 147 6075

18 200 340 70 1.4 122 3038
21 200 340 70 1.32 66 2700
22 200 340 70 0.93 48 1350
33 600 340 70 1.97 142 6075
34 600 340 70 1.4 126 3038
35 600 340 70 1.32 65 2700
36 600 340 70 0.93 48 1350
In summation, comparison of the measured transmission coefficient with the calculated

value from SWAN (Figure 7.14) revealed that the wave dissipation model of Battjes
and Janssen (1978) in SWAN with the proposed breaker index for plain (barred)
beaches is not applicable to modelling of submerged breakwaters - the model if used
with γ =0.73 will overestimate energy dissipation over the submerged breakwater. This
is consistent with the finding of Zanuttigh and Lamberti (2003) and Johnson (2006).
7.2.3 Modified Dissipation Model in SWAN
The significant discrepancies noted above between measured and calculated

transmission coefficients, using the dissipation model of Battjes and Janssen (1978),
supports the necessity for modification of the dissipation method in WAVE modelling
of submerged breakwaters.
In the present study, SWAN was run for different incident wave conditions, water
depths and breakwater crest widths over the range of laboratory experimental test
conditions. As indicated in Chapter 3, the present experimental testing covers a
comprehensive range of crest width and submergence ratio. The selected tests (Table
7.1) for numerical modelling, varied over 0.11 ≤ B/L ≤ 2.33 and 0.63 ≤ hs/Hi ≤ 2.0. The
value of breaker index in SWAN was adjusted for each test so that the calculated
149
transmission coefficient in the model agrees with the measured value.
Variability/dependency on γ with various dimensionless parameters was then examined.
In analysing the variation of γ, the dependency on the following introduced

dimensionless parameters were considered: Ursell number “UR”, the ratio of wave
height to breakwater crest height “Hi/d”, Iribarren number “ξ ”, inverse ratio of wave
steepness “Lo/Hi”, the ratio of breakwater crest height to wave height “ds/Hi”, the
inverse ratio of breakwater submergence over the breakwater “Hi/hs”, breakwater crest
width to wavelength ratio “B/Lo” and the ratio of incident wave height to deepwater
depth “Hi/h”.
The values of calculated breaker index associated with test conditions and
dimensionless parameters for the selected tests are provided in Tables 7.4. Figure 7.15
indicates the variation and effectiveness (correlation) of calculated breaker parameter γ
against the value of the dimensionless parameters. The correlation coefficient is a
simple normalized measure of linear relationship strength between the variables. As
indicated in Figure 7.15, strong linear relationship exists between the calibrated derived
breaker parameter in SWAN and the submergence ratio - this was also reported by
Johnson (2006). The effects of other (independent) dimensionless parameters are not
significant over the range of conditions tested/modelled.
As mentioned previously, only tests with submergence ratio greater than 0.5 were
included in the numerical simulations. At present the SWAN wave model cannot
correctly include wave energy dissipation for submergence ratios hs/Hi < 0.5 irrespective
of breakwater crest width ratio B/Lo.
Attempts of numerical simulation of the test conditions with submergence ratio less
than 0.5 in SWAN leads to unrealistic high values of breaking parameter being required
to fit the calculated transmission coefficient to experimental test results. In this case, the
water surface in the wave trough is lower than the offshore breakwater crest edge. In the
laboratory tests the breakwater experienced wave overtopping with no plunging
breaking over the breakwater. The length of the sea-side slope of the breakwater was
not sufficient to develop wave breaking over the slope. The depth-induced dissipation
model of Battjes and Janssen (1978) in SWAN provides very high wave energy
150
dissipation over the breakwater due to very shallow water over the breakwater (Figure
6.1, Chapter 6) – this was not observed in the laboratory since waves transmitted across
the breakwater with overtopping rather than plunging breaking. Therefore, eliminating
the tests with low submergence values (hs/Hi ≤ 0.5) in the numerical model is reasonable
to ignore the effect of overtopping which is not included in SWAN.
Variation of derived calibrated breaker parameter against submergence ratio (Figure

7.16) reveals that for submergence ratio 1.0 < hs/Hi, the calculated breaker parameter is
almost constant with the average of 0.92 and standard deviation of 0.13 with no longer
dependency on submergence ratio. This value of breaker parameter is close to the
maximum value of breaker index proposed for plane beaches by Battjes and Stive
(1985). For the values of submergence ratio 0.5 < hs/Hi < 1.0, Figure 7.17 indicates that a
simple linear regression provides the best fit to the derived γ. The following equations
have been derived for calculating the breaker index as a function of submergence ratio
to be used in the dissipation model of Battjes and Janssen (1978) within Delft3D-
SWAN over submerged breakwater:
hs hs
γ = −2.62 + 3.75 ≤ 2.0 ; 0.5 < ≤ 1.0
Hi Hi
(7.7)
hs
γ = 0.90 ; > 1.0
Hi
Equation 7.7 is plotted and compared with the derived calibrated values in Figure 7.17
as solid lines. Three different statistics indicators, the squared multiple correlation
coefficient R2, Wilmott index IW (Wilmott, 1981) and error function ε (Haller et al.,
2002) were employed for investigation of the validity of the model:
2
⎛ N ⎞
⎜ ∑ ( X mk − X mk )( X ck − X ck )⎟
R 2 = N⎝ k =1 ⎠ (7.8)
∑ (X mk − X mk ) ∑ (X ck − X ck )
N
2 2
k =1 k =1
151
Table 7.4 Values of derived calibrated breaker parameter γ for different selected test conditions (scaled).
Test h hs T B Lo Hi (Hrmso)
Kt UR B/Lo Hi/d Hi/h ξ (Hi/L)-1 ds/Hi hs/Hi γBattjes γRoelvink
series (m) (m) (sec) (m) (m) (m)
T1-21 4.00 1.00 1.00 35 15.00 0.56 0.50 1.96 2.33 0.14 0.14 2.59 26.86 0.19 1.79 0.73 0.51
T1-25 4.00 1.00 1.49 35 27.29 0.53 0.76 6.12 1.28 0.13 0.13 3.60 51.85 0.18 1.89 1.15 0.79
T1-26 4.00 1.00 1.49 35 27.29 1.02 0.46 11.91 1.28 0.26 0.26 2.58 26.65 0.34 0.98 1.30 0.91
T1-27 4.00 1.00 1.49 35 27.29 1.45 0.47 16.90 1.28 0.36 0.36 2.17 18.78 0.48 0.69 1.85 1.32
T1-30 4.00 1.00 2.00 35 38.59 1.02 - 23.84 0.91 0.26 0.26 3.07 37.66 2.93 0.98 - -
T1-42 4.50 1.50 1.49 35 28.42 1.00 0.54 8.84 1.23 0.22 0.22 2.67 28.51 0.33 1.52 0.90 0.64
T1-43 4.50 1.50 1.49 35 28.42 1.55 0.45 13.74 1.23 0.34 0.34 2.14 18.34 0.52 0.97 1.15 0.82
T1-46 4.50 1.50 2.00 35 40.66 0.95 0.69 17.31 0.86 0.21 0.21 3.26 42.63 0.32 1.56 1.10 0.79
T2-21 4.00 1.00 1.00 15 15.00 0.57 0.71 2.00 1.00 0.14 0.14 2.57 26.35 0.19 1.75 0.93 0.68
T2-22 4.00 1.00 1.00 15 15.00 1.13 0.52 3.98 1.00 0.28 0.28 1.82 13.24 0.38 0.88 1.30 0.98
T2-25 4.00 1.00 1.49 15 27.29 0.50 0.86 5.76 0.55 0.12 0.12 3.71 55.11 0.17 2.00 1.10 0.79
T2-26 4.00 1.00 1.49 15 27.29 1.07 0.62 12.48 0.55 0.27 0.27 2.52 25.44 0.36 0.93 1.55 1.18
T2-30 4.00 1.00 2.00 15 38.59 1.06 0.61 24.64 0.39 0.26 0.26 3.02 36.45 0.35 0.94 1.40 1.18
T2-43 4.50 1.50 1.49 15 28.42 1.52 0.59 13.50 0.53 0.34 0.34 2.16 18.67 0.51 0.98 1.20 0.97
T2-46 4.50 1.50 2.00 15 40.66 1.02 0.70 18.51 0.37 0.23 0.23 3.16 39.87 0.34 1.47 0.95 0.79
T2-47 4.50 1.50 2.00 15 40.66 1.56 0.65 28.28 0.37 0.35 0.35 2.55 26.09 0.52 0.96 1.30 1.12
T3-21 4.00 1.00 1.00 3 15.00 0.58 0.89 2.05 0.20 0.15 0.15 2.53 25.66 0.19 1.72 0.85 0.77
T3-26 4.00 1.00 1.49 3 27.29 1.04 0.80 12.08 0.11 0.26 0.26 2.56 26.27 0.35 0.96 1.30 0.67
T3-27 4.00 1.00 1.49 3 27.29 1.32 0.84 15.41 0.11 0.33 0.33 2.27 20.60 0.44 0.76 1.80 1.29
152
∑ (X )
N
− X mk
2
ck
IW = 1 − k =1
(7.9)
∑( X )
N 2
ck − X m + X mk − X m
k =1
12
⎡ N ⎤
(
⎢ ∑ X ck − X mk )
2
⎥
ε = ⎢ k =1 N ⎥ (7.10)
⎢ ⎥
⎢⎣ ∑
k =1
X mk
2
⎥⎦
where Xc is the calculated values (Delft3D), Xm is the measured (available data) values
and the barred parameters demonstrate the average values of the parameters. Perfect
agreement is achieved if the R2 and Wilmott index IW are 1.0 and the error function ε is
zero. R2 alone is unsatisfactory as a measure of the goodness of a regression relationship
(Draper, 1984). As shown in Figure 7.17 the fit is good.
The transmission coefficient over the breakwater was re-calculated using the calibrated
value of the breaker parameter γ (Equation 7.7) in rerunning the SWAN and then
compared with the measured Kt in the laboratory. Figure 7.18 now shows very good
agreement between the calculated and measured transmission coefficient regardless of
breakwater crest width. Therefore, Equation 7.7 provides a significant improvement to
the application of the dissipation model of Battjes and Janssen (1978) in predicting
wave climate over and behind submerged breakwaters when hs/Hi ≥ 0.5.
Johnson (2006) proposed a similar procedure to calibrate the dissipation model of

Battjes and Janssen for submerged breakwaters (Figure 7.19). He split wave breaking
into steepness and depth-induced breaking dissipation components. For steepness-
induced dissipation, he used the whitecapping energy dissipation model of Hasselmann
(1974) which has been extended by Komen et al. (1984). In MIKE21 PMS, Johnson
considered the value of breaker index γ in Battjes and Janssen theory as a calibration
parameter for the depth induced breaking dissipation. He calculated the calibration
parameter γ so that the measured and the calculated transmission coefficients agree and
finally provided the following empirical equation for γ over submerged breakwater.
153
1.55 hs H i ≤ 0.5
h
γ = 1.91 − 0.72 s 0.5 < hs H i <1.5 (7.11)
Hi
0.8 hs H i ≥1.5
Equation 7.11 by Johnson (2006) is similar but significantly different to Equation 7.7
developed within this research. The proposed Equation (7.11) by Johnson (2006) for
breaker parameter calculation were limited by the range of the experimental results
(only 8 tests as per Table 7.3) in comparison to the 21 measurement points in the
present laboratory tests (Table 7.4).
7.3 Flow Module
Delft3D-FLOW is a multi-dimensional (2D or 3D) hydrodynamic simulation program

that calculates non-steady flow and transport phenomena that result from tidal and
meteorological forcing on a rectilinear or a curvilinear, boundary fitted grid. Delft3D-
FLOW can be applied in calculation of tidal and Coriolis forces, vertical turbulent
viscosity, wave induced stresses and mass fluxes, influence of waves on the bed shear
stress and many other hydrodynamic features. For more description of Defl3D-FLOW
see the manual (Delft Hydraulics, 2005a).
Delft3D-FLOW has been applied in this study to simulate the wave-induced current
over submerged breakwaters. As described in Chapter 6, the FLOW module reads the
wave-induced forces from the communication file (imported from WAVE module) and
applies it in the momentum equation. The gradient of wave forces drives flow in the
flume. Three different approaches were examined for the calculation of wave-induced
forces and FLOW simulation of hydrodynamics at submerged breakwaters:
• Gradient of radiation stresses due to wave height changes passing over a

non-reflective obstacle (as a submerged barrier) which is called FLOW-
OBSTACLE hereafter.
• Gradient of radiation stresses or energy dissipation due to depth-induced

wave breaking dissipation over the submerged breakwater that are
nominated as FLOW-WAVE model.
154
• Gradient of radiation stresses due to wave-roller energy dissipation which is
called FLOW-ROLLER hereafter.
The conceptual description and details of all above three methods have been illustrated
in Chapter 6 and only the simulation processes and results are presented for discussion
in this chapter.
The whole flume including the main central channel and the breakwater was modelled
in Delft3D-FLOW with the same scale and grid size as the WAVE module. Time step
was adopted as 0.01 minutes according to the Courant number (Equation 7.1) to
enhance the best accuracy and convergence in the model. The duration of run times in
the FLOW model was sufficient for the resulting hydrodynamic parameters to become
stable. Figure 7.20 shows the variation of discharge over the time in the model with Test
T1-27 conditions and 1 hour time duration. As can be observed in the figure the value of
calculated discharge over the breakwater is sustained after 20 minutes. Therefore,
adopting 30 minutes time duration for FLOW modelling gives reasonable results of
hydrodynamic parameters for the conditions of this test/model case.
7.3.1 FLOW-OBSTACLE model
In this section, the breakwater was modeled as an obstacle in SWAN with defined
transmission coefficient. The bathymetry used in the wave model does not contain the
submerged breakwaters as these are modelled as ‘obstacles’.
Defining obstacles in SWAN along the grids, the propagation of waves are interrupted
by the barrier from one grid point, where the obstacle line is located, to the
neighbouring grid point in direction of wave propagation. Two options are available to
identify transmissible barriers in SWAN:
i- SWAN can estimate wave transmission through a (line-) structure such as a

breakwater. The method of Goda et al. (1967) is applied in this case, in which the
transmission coefficient depends on the incident wave height, freeboard and the
offshore slope of the structure.
ii- Transmission coefficient can also be defined by the user for each run test.
155
The definition of transmission coefficient over submerged breakwaters has been
comprehensively discussed in Chapter 2 and Chapter 5. As discussed through the
chapters, the method of Goda et al. (1967) does not provide accurate results of
transmission coefficient for submerged breakwaters. An empirical expression was
developed (modified) in Chapter 5 to calculate transmission coefficient as function of
incident wave height “Hi ”, submergence depth “hs”, breakwater crest width “B ” and
Iribarren number “ξ ”:
−0.26
⎛h ⎞ ⎛ B ⎞
K t = 0.17 ⎜⎜ s ⎟⎟ + 2.84 ⎜⎜ ⎟⎟ ⋅ (1 − exp(− 0.13∗ξ )) (7.12)
⎝ Hi ⎠ ⎝ Hi ⎠
The above expression was shown to provide good agreement with experimental
measured transmission coefficient for smooth and impermeable submerged breakwaters
and in comparison with other reported methods in the literature is superior (see Chapter
4, Section 4.1). Therefore, the transmission coefficient is best defined in SWAN using
Equation (7.12) for each test conditions (the Goda et al.’s (1967) alternative method for
defining Kt is within SWAN will not be used). Table 7.5 provides the test conditions
associated with the calculated transmission coefficients that are applied in the numerical
model for application of FLOW-OBSTACLE.
Wave-induced forces are calculated in SWAN to be used in the FLOW model to drive a
current over a submerged breakwater. There are two options available in SWAN to
calculate wave-induced forces using radiation stresses gradients or wave energy
dissipation, which were comprehensively discussed in Chapter 6. As discussed
previously, in FLOW-OBSTACLE model wave grids have no changes in water depth
and no depth-induced breaking occurs along the flume. Therefore, no currents would be
driven over the obstacle if wave energy dissipation method was applied in SWAN.
The radiation stress gradients are used in SWAN to create the driving forces. Radiation
stress tensors are calculated in SWAN using wave height gradients at the location of the
transmissive obstacle. A sample of variation of wave height, water level changes and
wave-induced forces along the flume are shown in Figure 7.21 through 7.23,
respectively. The figures correspond to test T1-26 with 3500mm laboratory breakwater
156
crest width, 400 mm water depth, 1.49 sec wave period and almost 100 mm incident
wave height (scaled up values were used in the numerical model).
Table 7.5 Calculated discharge “qc” and transmission coefficient “Kt” over submerged breakwaters
for the test series applied in FLOW-OBSTACLE model (scaled model).
qm qc
Test No. h (m) B (m) hs(m) Hi (m) T (sec) Kt B/L
(m3/s/m) (m3/s/m)
T1-21 4.0 35.0 1.0 0.56 3.2 0.60 2.33 0.52 0.31
T1-25 4.0 35.0 1.0 0.51 4.7 0.69 1.28 0.56 0.29
T1-26 4.0 35.0 1.0 1.02 4.7 0.50 1.28 1.11 0.38
T1-27 4.0 35.0 1.0 1.45 4.7 0.44 1.28 1.46 0.27
T1-30 4.0 35.0 1.0 1.03 6.3 0.55 0.91 0.81 0.20
T1-42 4.5 35.0 1.5 1.00 4.7 0.60 1.23 1.00 0.50
T1-43 4.5 35.0 1.5 1.55 4.7 0.49 1.23 1.50 0.44
T2-21 4.0 15.0 1.0 1.58 3.2 0.66 1.00 0.33 0.30
T2-22 4.0 15.0 1.0 1.13 3.2 0.47 1.00 0.73 0.60
T2-25 4.0 15.0 1.0 0.50 4.7 0.81 0.55 0.27 0.19
T2-26 4.0 15.0 1.0 1.07 4.7 0.58 0.55 0.48 0.40
T2-30 4.0 15.0 1.0 1.06 6.3 0.64 0.39 0.41 0.21
T2-43 4.5 15.0 1.5 1.52 4.7 0.57 0.53 0.70 0.48
T2-46 4.5 15.0 1.5 1.18 6.3 0.75 0.37 0.36 0.37
T2-47 4.5 15.0 1.5 1.56 6.3 0.63 0.37 0.66 0.18
T3-21 4.0 3.0 1.0 0.58 3.2 0.84 0.20 0.14 0.17
T3-26 4.0 3.0 1.0 1.02 4.7 0.81 0.11 0.42 0.40
T3-27 4.0 3.0 1.0 1.32 4.7 0.75 0.11 0.38 0.41
As indicated in Figure 7.23, the gradient of wave-induced forces occurs only at the
location of the obstacle (line-grid) defined in SWAN. The wave forces are created from
the radiation stress gradients. Therefore, the effect of surfzone length is ignored in the
model and only the difference of wave forces between two adjacent grids at the location
of the obstacle drives water current in the flume.
Using estimated wave forces in the communication file (created by SWAN), FLOW
module then calculates depth-averaged velocity and discharge in the flume. In the
FLOW module, the breakwater has to be represented by the grid, where the crest width
can be defined by the water level points (water depth changes). Figure 7.24 represents
the pattern of calculated velocity and flow in the flume and around the breakwater.
Qualitatively, the pattern of flow in the model is in good agreement with the dye tracing
tests in the laboratory (Chapter 3). The values of calculated wave-induced flow over the
submerged breakwater at the same position of measurement in the laboratory are
presented in Table 7.5. Comparison of the calculated and measured discharge over the
157
breakwater per unit width is shown in Figure 7.25. The results show that modelling of
submerged breakwaters, as a transmissive barrier would give satisfactory results for the
tests series T3 in which the breakwater crest width ratio B/Lo is less than 0.3. The figure
also shows that the discrepancy between the calculated discharge and the measured
values from the laboratory tests increases with B/Lo since the effect of surfzone length
(breakwater crest width) dominates the wave breaking processes when B/Lo is greater
than 0.3 (wide crested breakwaters).
The quantitative comparison of the calculated discharge with the measurements has
been carried out using the statistics indicators R2, IW and ε for two ranges of breakwater
crest width ratio B/L o ≤ 0.3 and B/L o > 0.3. The values of the statistics parameters are
given in Table 7.9 and compared in Figure 7.26. As indicated in the figure, the values of
Willmot index and error function for B/Lo ≤ 0.3 are 0.99 and 0.08, respectively that
indicates a high “goodness of fit”. The fit is however poor for B/L o > 0.3. The values of
statistics indicators confirms that FLOW-OBSTACLE model can predict, reasonably,
2D depth-averaged discharge passing over narrow crested (B/Lo< 0.3) submerged
breakwaters due to breaking waves- this outcome being dependent upon the adoption of
the proposed Equation 7.12 for the definition of the transmission coefficients Kt.
7.3.2 FLOW-WAVE Model
Wave deformation in presence of a submerged breakwater can also be simulated in

WAVE module (SWAN) by defining the breakwater in the wave grid with water depth
changes at the location of the breakwater in the flume. The dominant process in wave
evolution over the breakwater was found to be depth-induced breaking dissipation
whilst the effects of bottom friction and whitecapping were found to be small in the
numerical simulation (see Section 7.2.1). As illustrated previously, depth-induced wave
dissipation in SWAN is calculated by using the dissipation theory of Battjes and Janssen
(1978).
As detailed in Chapter 6, depth-averaged wave induced forces are calculated in SWAN

by either (i) the gradients of the radiation stresses tensor or (ii) approximated by wave
energy dissipation (after Dingemans et al. 1987). Application of both methods of
calculating wave forces in SWAN was investigated in the simulation. The definitions
158
FLOW-WAVE-1 and FLOW-WAVE-2 are nominated for the models that use the
radiation stress gradients and wave energy dissipation for calculation of wave forces,
respectively.
The calibrated values of breaker parameter as a function of submergence ratio (Equation

7.7) were employed for both models of FLOW-WAVE whilst other physical parameters
in SWAN and FLOW were adopted the same as defined in Section 7.1.3. The variation
of wave-induced forces per unit area in x- direction using both models FLOW-WAVE-1
(radiation stress gradients) and FLOW-WAVE-2 (wave dissipation) are shown in Figure
7.27 for some targeted tests. The solid green and red lines correspond to (i) gradient
radiation stress tensor and (ii) wave energy dissipation methods, respectively. As can be
observed in the figure, radiation stress gradient method leads to high gradient of wave
forces around abrupt water depth changes. Theoretically, varying water depth causes
changes in n (the ratio of group velocity over phase velocity) which affects directly the
radiation stress tensor:
⎡ 1⎤
S xx = ρ g ∫ ⎢n cos 2 θ + n − ⎥ E dσ dθ
⎣ 2⎦
⎡ 1⎤
S yy = ρ g ∫ ⎢n sin 2 θ + n − ⎥ E dσ dθ (7.13)
⎣ 2⎦
S xy = S yx = ρ g ∫ n sin θ cosθ E dσ dθ
In reality, wave celerity ratio n cannot be much hampered by the water depth changes
along the offshore slope of the breakwater due to the short length of the slope. The
sudden (unrealistic) changes of wave forces shown in Figure 7.27 may introduce errors
in the numerical computation of submerged breakwaters. Dingemans et al. (1987)
showed that the large irrotational part of the driving force per unit volume is not able to
drive 2D depth-averaged currents and using the spatially rapidly varying radiation
stresses (as seen in Figure 7.27) can result in spurious discharges being indicated by
numerical models. They showed that wave-induced current is closely proportional to the
wave energy dissipation per unit area (see Chapter 6). This matter will be investigated
in the following paragraphs by comparing the calculated and measured discharge over
the submerged breakwater.
159
The spatial variation of calculated depth-averaged velocity and discharge for Test
conditions of T1-26, T2-26 and T3-26 are shown in Figures 7.28 through 7.30,
respectively. Wave climate and water depth are almost the same for different
breakwater crest width in the above tests. As indicated from the figures, calculated
discharge over the breakwater decreases with breakwater crest width which confirms
the dependency of the discharge to the breakwater crest width as observed in the
laboratory experiments (see Chapter 5). The computed discharges in FLOW module
over the breakwater at the location of measurement probes in the laboratory are
provided in Table 7.6. The value of discharge was calculated using both radiation stress
gradient (FLOW-WAVE-1) and wave energy dissipation (FLOW-WAVE-2) methods.
Comparison of calculated discharge over the breakwaters using both approaches of

calculating wave forces (Figures 7.31 and 7.32) shows that in 50% of test conditions,
applying wave energy dissipation for calculating wave-induced forces provides higher
values of discharge than radiation stress gradient method. Deviation between calculated
and measured discharge as a function of breakwater crest width “B/Lo” for both
radiation stress gradients and energy dissipation methods is shown in Figure 7.33. As
indicated from the figure, the discrepancy of measured and calculated discharge
increases with decreasing B/Lo. The statistics indicators for “goodness of fit ”
assessment of the models in Figure 7.34 also confirms that the discrepancy between
measured and calculated discharges are higher for both methods when B/L o < 0.55.
In more detail, a quantitative comparison of the calculated and measured results has
been carried out using the statistics indicators R2, IW and ε for three ranges of
breakwater crest width ratio B/oL < 0.35 , 0.35≤ B/L o ≤ 1.0 and 1.0 < B/Lo. Comparison
of the statistics parameters of the two adopted approaches in Figure 7.35 shows
negligible difference between the models FLOW-WAVE-1 and FLOW-WAVE-2 for
the breakwater crest width ratio “B/Lo” greater than 1.0 - in which both models are in
very good conformity with the measured discharge from laboratory experiments. Both
methods do not provide reasonable results of computed discharge over submerged
breakwaters when 0.35 ≤ B/L o ≤ 1.0. The statistics parameters indicate that the method
of wave energy dissipation (FLOW-WAVE-2) is more appropriate to be employed for
160
wave force and then discharge calculation in narrow crested breakwater when the
breakwater crest width is smaller than 0.35 (B/Lo < 0.35).
breakwaters for the test series applied in FLOW-WAVE model (scaled model). qc1 and qc2
correspond to the calculated discharge using radiation stress gradient (FLOW-WAVE-1) and wave
energy dissipation (FLOW-WAVE-2), respectively.
Crest
Water Crest Wave Wave Calibrated Measured Calculated
Submergence width
Test No. depth width height period Gamma discharge discharge
ratio
qm qc1 qc2
h (m) B (m) hs(m) Hi (m) T (sec) γcal B/L
(m3/s/m) (m3/s/m) (m3/s/m)
T1-21 4.0 35.0 1.0 0.56 3.2 0.60 2.33 0.52 0.26 0.35
T1-25 4.0 35.0 1.0 0.51 4.7 0.69 1.28 0.56 0.35 0.35
T1-26 4.0 35.0 1.0 1.02 4.7 0.50 1.28 1.11 0.97 1.28
T1-27 4.0 35.0 1.0 1.45 4.7 0.44 1.28 1.46 1.3 1.59
T1-30 4.0 35.0 1.0 1.03 6.3 0.55 0.91 0.81 1.04 0.98
T1-42 4.5 35.0 1.5 1.00 4.7 0.60 1.23 1.00 0.78 0.86
T1-43 4.5 35.0 1.5 1.55 4.7 0.49 1.23 1.50 1.48 1.5
T2-21 4.0 15.0 1.0 1.58 3.2 0.66 1.00 0.33 0.24 0.27
T2-22 4.0 15.0 1.0 1.13 3.2 0.47 1.00 0.73 0.78 0.92
T2-25 4.0 15.0 1.0 0.50 4.7 0.81 0.55 0.27 0.27 0.26
T2-26 4.0 15.0 1.0 1.07 4.7 0.58 0.55 0.48 0.86 0.89
T2-30 4.0 15.0 1.0 1.06 6.3 0.64 0.39 0.41 0.77 0.84
T2-43 4.5 15.0 1.5 1.52 4.7 0.57 0.53 0.70 1.45 1.95
T2-46 4.5 15.0 1.5 1.18 6.3 0.75 0.37 0.36 0.79 0.85
T2-47 4.5 15.0 1.5 1.56 6.3 0.63 0.37 0.66 0.94 1.71
T3-21 4.0 3.0 1.0 0.58 3.2 0.84 0.20 0.14 0.13 0.14
T3-26 4.0 3.0 1.0 1.02 4.7 0.81 0.11 0.42 0.16 0.54
T3-27 4.0 3.0 1.0 1.32 4.7 0.75 0.11 0.38 0.16 0.44
7.3.3 FLOW-ROLLER Model
In the recent research version of Deft3D-FLOW (3.49.02, June 2005) new wave and
roller modules are provided. Reniers et al. (2004) described the ROLLER model for
wave propagation and breaking coupled to a depth averaged nonlinear flow model to
predict the time-dependent infragravity flow field. The mean direction of wave field is
imported from SWAN model using usual coupling provided by the WAVE module. The
FLOW module only reads wave direction, period (wave celerity) from the WAVE
module and then calculates the wave energy along the grids using the wave climate
introduced at the open boundary. Then, the values of roller and wave energy in the grids
are calculated using the wave energy dissipation model of Roelvink (1993) and energy
161
balance equations of Reniers et al. (2004) in computing radiation stress. Comprehensive
demonstration of the roller model is presented in Chapter 6.
7.3.3.1 Breaking Parameter in ROLLER Model
As indicated in Chapter 6, the depth-induced wave breaking dissipation theory of

Roelvink (1993) was applied for calculating wave dissipation in the short wave energy
balance for roller model. The breaking parameter γ and other parameters in dissipation
formula of Roelvink were determined empirically as a calibration coefficient. He used
three data bases containing a total of 11 tests to calibrate the probabilistic and
parametric models. The experimental calibration sets include 10 laboratory tests with
plane and barred beach and one field data with barred beach. Roelvink (1993) found
n=10 and γ =0.55 to fit the model to those data set. n in the Roelvink’s equation is a
variable to be determined from the observation which acts as a dissipation parameter
corresponding to the randomness of the incident waves.
To investigate the applicability of Roelvink’s model for calculating depth-induced

breaking dissipation over submerged breakwaters a spreadsheet was developed in
MATLAB (Appendix B). The spreadsheet calculates wave dissipation and transmission
coefficient over submerged breakwater using the dissipation model of Roelvink (1993).
The adjustment coefficient n was retained at the value of 10 in the model and
transmission coefficient was calculated based on the proposed value of γ =0.55.
Comparison of calculated and measured (in the experiments conducted during this
thesis; Chapter 5) transmission coefficient (Figure 7.36) reveals that applying breaking
parameter γ =0.55 in the dissipation model of Roelvink (1993) predicts higher energy
dissipation (lower transmission coefficient) than the measured values in the laboratory
experiments for submerged breakwaters. This result was also observed when comparing
SWAN model results for Battjes and Janssen (1978) dissipation model with the
proposed value of breaking parameter (Section 7.2.4).
Therefore, some modification to the Roelvink’s (1993) breaking parameter is needed

prior to application to adjust the data to the calculated transmission coefficient. During
the calibration processes, the data of tests with submergence ratio hs/Hi greater than 0.5
was employed to ensure wave breaking without overtopping occurs over the submerged
162
breakwater (see Section 7.2.4 for more explanation). The value of γ in Roelvink’s
formulae was calculated for each test so that the computed transmission coefficient
agrees with the measured transmission coefficient in the laboratory experiments.
The same calibration approach as that applied to evaluation of the model of Battjes and
Janssen (1978) was implemented in Roelvink’s model (see section 7.2.4). The variation
and effectiveness of the relevant dimensionless parameters are shown in Figure 7.37.
The correlation coefficient between the inverse value of submergence ratio (Hi/hs) and
calculated breaker parameter was found to be the most significant. Breakwater crest
width ratio (B/L) was also found to have significant effect on γ in Roelvink’s
dissipation method- unlike in application of Battjes and Janssen (1978).
Two different approaches considering (i) only submergence ratio and (ii) both
breakwater crest width and submergence ratio were adopted and compared to
investigate the most appropriate model for calibrated breaking parameter. Firstly, the
calibrated breaker parameter was considered only as a function of the submergence
ratio. Variation of calculated breaking parameter against submergence ratio or inverse
value of submergence (Figure 7.38) reveals that for submergence ratio 1.0 ≤ hs/Hi, the
calculated breaker parameter is almost constant with an average of 0.75 and standard
deviation of 0.10. Therefore, in relatively high submergence ratio circumstances,
applying the dissipation model of Roelvink (1993) for calculation of depth-induced
wave dissipation over submerged breakwaters would yield reasonable results if a
breaker index γ = 0.75 was adopted - this value is 35% greater than the proposed breaker
parameter by Roelvink (1993) for the surfzone and plane beach cases. In Figure 7.38, a
simple linear regression model was found the best fit to the derived breaker parameter
for submergence ration hs/Hi <1.0. The maximum value of γ was adopted 1.50 that
gives the best statistics results for fitting the data to the calculated value of the
calibrated breaking parameter. This maximum value of breaker parameter is consistent
with the maximum value reported by Kaminsky and Kraus (1993). They collected and
examined 416 points of different data sets that include very steep beach slopes.
Kaminsky and Kraus (1993) found the maximum value of breaker parameter 1.56 for
censored and 1.49 for uncensored data points. Combining the fitted line with the two
upper and lower limits leads to the following expression for calculating calibrated
163
breaking parameter as a function of submergence ratio to be used in the dissipation
model of Roelvink (1993):
hs hs
γ = −1.93 + 2.89 ≤ 1.5 ; ≤ 1.0
Hi Hi
(7.14)
hs
γ = 0.75 ; > 1.0
Hi
Figure 7.38 represents the fitted lines to the scattered plot of derived breaking
parameters from calibration processes. Graphical and statistical comparison of measured
transmission coefficient with calculated values using Equation 7.14 is presented in
Figure 39. As implied by the figure, the proposed calibrated breaker parameter γ in
Equation 7.14 provides relatively good agreement between the calculated results and
measurements. However, the model shows more scatter for narrow crested breakwaters
(B/L<0.35). Statistical indicators for comparison of calculated and measured
transmission coefficient (Table 7.7) confirms that the proposed calibrated breaker index
in Equation 7.14 is more applicable in relatively broad crested submerged breakwater
(B/L≥ 0.35).
Table 7.7 Statistics comparison of calculated and measured transmission coefficient applying the
proposed calibrated breaker parameter γ (Equation 7.14) in the dissipation model of Roelvink
(1993).
Statistics whole
≥0.35 B/L<0.35
B/L≥
index data
2
R 0.55 0.50 0.27
IW 0.82 0.81 0.68
ε 0.14 0.17 0.14
As mentioned in the previous paragraph, the derived calibrated breaking parameter was
found to be more affected by breakwater crest width in the dissipation model of
Roelvink (1993) than the model of Battjes and Janssen (1978). Therefore, more
investigation was carried out so that the effect of breakwater crest width and wavelength
was also taken into consideration in the calibration processes. Statistical investigation of
the variables and predictors revealed that the multiple linear relationship between the
variables provides sensible fit of the breaker index between the model and the
experimental parameters:
164
⎛ hs ⎞ ⎛B⎞
γ = α 1 ⎜⎜ ⎟⎟ + α 2 ⎜⎜ ⎟⎟ + β (7.15)
⎝ Hi ⎠ ⎝ Lo ⎠
The coefficients in Equation 7.15 were defined using the data for 1.0<hs/Hi and
0.5<hs/Hi≤1.0, separately, since for submergence ratio greater than 1.0 the derived
calibrated γ is no longer dependent on hs/Hi. Multiple regression analysis in MATLAB
was applied to define the coefficients α1, α2 and β as the best fit of the equation to data
that reads:
⎛ hs ⎞ ⎛B⎞ hs
γ = −1.97⎜⎜ ⎟⎟ − 0.27⎜⎜ ⎟⎟ + 3.07 ≤ 1.5 ; 0.5 < ≤ 1.0
⎝ Hi ⎠ ⎝ Lo ⎠ Hi
(7.16)
⎛B⎞ h
γ = −0.14⎜⎜ ⎟⎟ + 0.84 ; 1.0 < s
⎝ Lo ⎠ Hi
The statistics indicators (R2=0.85, IW=0.96 and ε=0.12) signify that Equation 7.16 leads
to more reliable results of calibrated breaking parameter than Equation 7.14. The
transmission coefficient was calculated applying Equation 7.16 in the dissipation model
of Roelvink (1993). The predicted values of transmission coefficient are compared
graphically and statistically in Figure 7.40. Very good agreement was found when both
submergence and breakwater crest width ratio were employed to determine the
calibrated breaking parameter and hence, Equation 7.16 was implemented to calculate
calibrated breaking parameter in the roller model.
Depth-induced flow was calculated using FLOW-ROLLER module in Delft3D. The

calibrated value of breaking parameter γ was calculated from Equation 7.16 for each test
and applied in the model. All physical parameters (e.g. water density, gravity
acceleration, bottom roughness) remained the same as FLOW-WAVE and FLOW-
OBSTACLE models for each test. Wave-current interaction was also taken into account
in the model to enhance more realistic results from the numerical simulation. The tests
with submergence ratio hs/Hi >0.5 were also implemented in the model to ensure no
overtopping occurs over the submerged breakwater.
165
As illustrated in Chapter 6, wave energy in the roller model consists of short-wave and
roller energy. The variation of calculated short-wave energy and roller energy along the
flume for tests T1-26, T2-26 and T3-26 are shown in Figure 7.41 through Figure 7.43.
The tests are at the same conditions with varying breakwater crest width. As indicated
in the figures, roller energy is maximum over the submerged breakwater where depth-
induced breaking occurred. The total roller energy is higher in the test with the broadest
crest width structure (T1-26, B=35 m) which is consistent with the breaking processes
over the breakwater. For the wide crested breakwater, the length of the breakwater is
long enough to develop wave breaking while the wave passes over the breakwater. On
the other hand, in the narrower submerged breakwaters the length of breaking is not
sufficient to develop wave breaking and roller over the structure and hence, short-wave
energy gradient is more dominant in the computing current over narrow crested
breakwaters.
The computed depth-averaged current flow direction along the flume and over the
submerged breakwater agree well, qualitatively with the laboratory observations with
dye plumes (Chapter 3). Figure 7.44 and Figure 7.45 demonstrate the calculated vectors
and values of velocity and discharge along the flume for test T1-26 conditions. The
values of calculated discharge (qc) over the breakwater, using the calibrated breaking
parameters γcal, for the desired tests are provided in Table 7.8. The quantities were
computed at the same locations as measured in the laboratory. Graphical and statistical
comparisons of the calculated and measured discharge were carried out to examine the
capability of calibrated FLOW-ROLLER model in hydrodynamic modelling of
The measured wave-induced discharge per unit width “qm” over the breakwater is
compared with those obtained from numerical model (FLOW-ROLLER) in Figure 7.46.
The figure implies that in general the roller model can predict discharge over submerged
breakwaters with some confidence. The overall agreement between estimated and
measured discharge is quantified by calculating the statistics indicators R-squared,
Wilmott index IW and error function ε that are defined in Section 7.2.4. The values of
statistics parameters for all tests were calculated R2=0.90, IW=0.95 and ε=0.21 that
indicate good agreement between the numerical simulation and laboratory experiment.
166
However, as observed in Figure 7.46, the accuracy of the predicted discharge decreases
for the narrow crested breakwater (experiment crest width B=300 mm) and the roller
model underestimates the discharge over the narrow crested breakwater.
breakwaters for the test series applied in FLOW-ROLLER model (scaled model).
Crest
Water Crest Wave Wave Calibrated Measured Calculated
Submergence width
Test No. depth width height period Gamma discharge discharge
ratio
qm qc
h (m) B (m) hs(m) Hi (m) T (sec) γcal B/Lo
(m3/s/m) (m3/s/m)
T1-21 4.0 35.0 1.0 0.56 3.2 0.75 2.33 0.52 0.52
T1-25 4.0 35.0 1.0 0.51 4.7 0.75 1.28 0.56 0.58
T1-26 4.0 35.0 1.0 1.02 4.7 1.01 1.28 1.11 0.80
T1-30 4.0 35.0 1.0 1.03 6.3 1.01 0.91 0.81 0.63
T1-42 4.5 35.0 1.5 1.00 4.7 0.75 1.23 1.00 0.96
T1-43 4.5 35.0 1.5 1.55 4.7 1.02 1.23 1.50 1.45
T2-21 4.0 15.0 1.0 1.58 3.2 0.75 1.00 0.33 0.47
T2-22 4.0 15.0 1.0 1.13 3.2 1.19 1.00 0.73 0.52
T2-25 4.0 15.0 1.0 0.50 4.7 0.75 0.55 0.27 0.36
T2-26 4.0 15.0 1.0 1.07 4.7 1.09 0.55 0.48 0.56
T2-30 4.0 15.0 1.0 1.06 6.3 1.07 0.39 0.41 0.45
T2-43 4.5 15.0 1.5 1.52 4.7 0.99 0.53 0.70 0.86
T2-46 4.5 15.0 1.5 1.18 6.3 0.75 0.37 0.36 0.53
T2-47 4.5 15.0 1.5 1.56 6.3 1.03 0.37 0.66 0.63
T3-21 4.0 3.0 1.0 0.58 3.2 0.75 0.20 0.14 0.15
T3-26 4.0 3.0 1.0 1.02 4.7 1.03 0.11 0.42 0.22
T3-27 4.0 3.0 1.0 1.32 4.7 1.43 0.11 0.38 0.14
For more investigation of accuracy of the model comparison of measured and calculated
discharge were made in three different groups of data (results) distinguished by the
breakwater crest width ratio B/Lo:
Broad crested breakwater 1.0 < B/Lo

moderately wide crest width 0.35 ≤ B/Lo ≤ 1.0
Narrow crested breakwater B/Lo< 0.35
The values of calculated statistics parameters in Table 7.9 and Figure 7.47 indicate that
the roller model is most applicable to predict discharge over wide submerged
breakwaters with crest width ratio greater than 1.0 - in this case the width of the
breakwater is wide enough to let roller become fully developed in the breaking process
over the breakwater. On the other hand, poor agreement is observed for the narrow
crested breakwater (B/Lo< 0.35) since breaking length over the breakwater is not
167
sufficient to develop roller in breaking processes. For breakwaters with intermediate
crest width (0.35 ≤ B/Lo ≤ 1.0) the roller model (FLOW-ROLLER) predicts reasonable
values of discharge over the structure however, the accuracy is less than the broad
crested submerged breakwaters.
Table 7.9 Statistical parameters for the four models

0.1≤B/Lo≤2.4
1.0<B/Lo 0.35≤B/Lo≤1.0 B/Lo<0.35
(whole data)
FLOW-OBSTACLE
R2 0.98 0.17 0.95 0.65
Iw 0.55 0.53 0.96 0.57
ε 0.60 0.48 0.15 0.55
FLOW-WAVE-1
R2 0.99 0.76 0.99 0.77
Iw 0.95 0.65 0.53 0.87
ε 0.17 0.65 0.58 0.38
FLOW-WAVE-2
R2 0.98 0.69 0.99 0.69
Iw 0.97 0.47 0.93 0.75
ε 0.14 1.09 0.24 0.58
FLOW-ROLLER
R2 0.94 0.71 0.50 0.90
Iw 0.96 0.80 0.54 0.95
ε 0.14 0.24 0.53 0.21
7.4 Summary and Discussion
The 2D depth averaged numerical modelling of wave-induced discharge over

submerged breakwaters was investigated in this chapter using the recent research
version of Delft3D. The entire experimental 3m wide wave flume was simulated in the
model with the scale of 10:1 (model dimensions 10 times those of the flume in the
laboratory) and time duration of 15 minutes or more due to the stability condition of the
numerical model. Four different approaches using combinations/options within the two
main modules of Delft3D (SWAN and FLOW) were tested in the numerical
simulations. The discharges over the modelled submerged breakwater were calculated at
the equivalent measurement points in the laboratory for 18 selected tests. The calculated
discharges per unit width were compared graphically and statistically with the measured
laboratory experimental results and appropriate models were proposed for
hydrodynamic modelling of submerged breakwaters. Wave-current interaction was
168
taken into account in all models to give more realistic results with respect to the
laboratory experimental conditions.
The first model was called FLOW-OBSTACLE in which the breakwater was simulated
as a linear transmissive obstacle in the wave grid lines. The developed formula for
transmission coefficient in Chapter 4 (Equation 7.12) was applied in SWAN to calculate
Kt for each test condition. Wave forces were calculated using the gradient radiation
stresses due to wave height changes at the location of the obstacle and used in the
FLOW module to calculate wave-induced flow over the breakwater. The breakwater
was modelled in the FLOW grid cells with defining water depth changes. Poor
agreement was found in comparison of calculated and measured discharge over the
submerged breakwater for crest width ratio (B/Lo) greater than 0.35. Simulation of the
breakwater as an obstacle in SWAN was found appropriate for predicting discharge
over narrow crested submerged breakwater with B/Lo < 0.35. In this case, the
breakwater crest width is so small with respect to the incident wavelength and the
length of breaking over the structure is so short to affect wave-breaking processes.
Hence, modelling the breakwater as a linear transmissive obstacle would be reasonable
for narrow-crested breakwaters. On the other hand, wave breaking is more affected by
the breakwater crest width and simulation of the structure as a line causes some errors in
hydrodynamic computations.
The experimental flume was also simulated with the online interaction between FLOW
and WAVE module in which the breakwater was modelled in both flow and wave grids.
The model of Battjes and Janssen (1978) was found to overestimate the wave energy
dissipation due to breaking. A revised equation (Equation 7.7) for prediction the value
of breaking parameter was proposed to be used in the theory of Battjes and Janssen to
calculate transmitted wave height over and behind submerged breakwaters,
appropriately. The calibrated function of breaking parameter γ was investigated and
found to be sensitive to submergence ratio rather than other dimensionless parameters.
The FLOW module reads the calculated wave forces from the calibrated model of
SWAN to predict flow over the breakwater.
Two different methods, radiation gradient stresses and wave energy dissipation, were
assessed for the calculation of wave-induced forces in SWAN. Comparison of the
169
results from the two methods showed that both models provide almost the same
outcomes for wide crested submerged breakwaters. However, the discrepancy between
the results increases with decreasing breakwater crest width ratio B/Lo. In theory, when
waves pass over the submerged breakwater the wave length and thus the radiation stress
changes rapidly due to abrupt changes in water depth whilst, in reality, the wave length
changes gradually as water depth decreases in the surf zone and no sudden changes in
wave length happens due to the abrupt water changes. Therefore, the calculated
radiation stress gradients due to abrupt water depth changes provide more discrepancy
between the calculated and measured current over the breakwater - applying wave
energy dissipation is more suitable to calculate wave forces for waves passing over
submerged narrow crested breakwaters.
The influence of roller in the energy balance equation was taken into account in the
other approach to predict flow field around the submerged breakwater. The roller model
available in the research version of Delft3D-FLOW (3.52.06.00) was applied in the
simulation. The wave dissipation model of Roelvink (1993) is implemented in the
model to calculate short-wave energy dissipation. The dissipation model of Roelvink
(1993) was also found to overestimate the amount of energy dissipated by depth-
induced wave breaking. The calibrated values of breaking parameters in Roelvink’s
model were defined so that the measured and computed transmission coefficients agree.
The variations of derived calibrated breaking index were found to be sensitive to
breakwater crest width as well as submergence ratio and the effects of B/Lo was also
taken into account in the calibrated γ equations. Comparison of measured and calculated
transmission coefficient revealed that respecting both breakwater crest width and
submergence ratio provides more sensible results when the dissipation model of
Roelvink (1993) is applied in the model.
The proposed calibrated value of the breaking parameter (Equation 7.16) was
implemented in the roller model to compute the wave-induced current over the
breakwater. Comparison of measured discharge over the breakwater with the calculated
values using roller model and the calibrated breaking parameter showed that the roller
model produces reasonable results for wide and relatively wide submerged breakwaters
(B/L ≥ 0.35) however, poor agreement was found for narrow crested breakwater.
170
Overall results of goodness of fit of the four model approaches are given in Table 7.9
and Figure 7.48.
In summary, the following guidance is given as to most appropriate modelling to predict

discharge over submerged breakwaters. The primary division being governed by
relative breakwater crest width:
• B/Lo< 0.35;
Applying WAVE and FLOW module in simulating the breakwater as a transmissive

obstacle in grid lines (FLOW-OBSTACLE) and using the relevant proposed
transmission coefficient of Equation 7.12. The calibrated dissipation model of
Battjes and Janssen (1978) in SWAN is also applicable if wave energy dissipation is
applied in the wave force calculation (FLOW-WAVE-2) and the proposed value of
breaker parameter (Equation 7.7) is used in SWAN.
• 0.35 ≤ B/Lo ≤ 1.0;
Applying the roller model (FLOW-ROLLER) using the proposed calibrated breaker
parameter (including B/Lo and hs/Hi, Equation 7.16) in the wave dissipation model
of Roelvink (1993).
• 1.0 < B/Lo;
Both roller (FLOW-ROLLER) and wave model (FLOW-WAVE) associated with

the relevant proposed calibrated breaking parameter (Equation 7.7 and 7.16,
respectively) in dissipation formula are applicable.
171
2.5
Normal grid size
refined grid
2.0
1.5
Hs (m)
1.0
0.5
0.0
0 50 100 150 200 250
Distance (m)
Figure 7.1. Comparison of calculated wave transformation over submerged breakwater in the model with 1m and 0.2 m grid size
(test T1-27; Hrms= 1.45 m, T=4.71 sec, B=35 m and h= 4.0 m).
Paddle Beach
Figure 7.2. A sample of computational grid used in the numerical model.

breakwater due to bottom friction (Hasselmann et al., 1973) for test T1-27.
breakwater due to bottom friction (Collins, 1972) for test T1-27.
breakwater due to bottom friction (Madsen et al., 1988) for test T1-27.
Figure 7.6 Wave height and wave energy dissipation along the flume without
breakwater for test T1-27 due to steepness-induced breaking (whitecapping).
Total energy dissipation
2400
Gamma=1.02
2200 Gamma=0.91
Gamma=0.80
2000 Gamma=0.75
Gamma=0.71
1800 Gamma=0.65
Gamma=0.61
1600 Gamma=0.55
Dissipation (J/m /sec)
Gamma=0.50
1400
2
1200
1000
800
600
400
200
0
0 25 50 75 100 125 150 175 200 225
Distance (m)
Energy dissipation%
25
Gama=1.02
Gamma=0.91
Gamma=0.80
Gamma=0.75
20 Gamma=0.71
Gamma=0.65
Gamma=0.61
Gamma=0.55
Dissipation%
15 Gamma=0.50
10
0
0 25 50 75 100 125 150 175 200 225
Distance (m)
Figure 7.7. Wave height attenuation and wave energy dissipation along the flume
without breakwater (h=4.0 m, Hi=1.02m, T=4.71 sec).
3 10
2.5
8
Water depth (m)

2
6
Hs (m)
1.5
4
1
Hs (breaking+friction)
Hs (breaking) 2
0.5 SWL
breakw ater
0 0
0 50 100 150 200
Distance (m)
wave breaking was considered in SWAN with and without bottom friction effects
(Hrms=1.45 m, T=4.71 sec, h=4.0 m and B=35 m).
400 10
Hs (breaking+friction)
Hs (breaking)
SWL
breakw ater
8
300
Dissipation (N/ms)
Water depth (m)

6
200
4
100
2
0 0
0 50 100 150 200
Distance (m)
wave breaking was considered in SWAN with and without bottom friction effects
3 10
2.5
8
Water depth (m)

2
6
Hs (m)
1.5
4
1
Hs (breaking+w hitecapping)
Hs (breaking)
2
0.5 SWL
breakw ater
0 0
0 50 100 150 200
Distance (m)
wave breaking was considered in SWAN with and without whitecapping effects
400 10
Hs (breaking+w hitecapping)
350 Hs (breaking)
SWL 8
300
Dissipation (N/ms)
breakw ater
Water depth (m)
250 6
200
150 4
100
2
50
0 0
0 50 100 150 200
Distance (m)
Figure 7.11. Calculated wave energy dissipation along the flume when depth-
induced wave breaking was considered in SWAN with and without whitecapping
effects (Hrms=1.45 m, T=4.71 sec, h=4.0 m and B=35 m).
1.2 10
(a) Hs
1 SWL
breakwater 8
Water depth (m)

0.8
6
Hrms (m)
0.6
4
0.4
2
0.2
0 0
100 110 120 130 140 150 160 170 180 190 200
Distance (m)
1.2 10
(b)
1
8
Water depth (m)

0.8
6
Hrms (m)
0.6
4
0.4
2
0.2
0 0
100 110 120 130 140 150 160 170 180 190 200
Distance (m)
1.2 10
(c)
1
8
Water depth (m)
0.8
6
Hrms (m)
0.6
4
0.4
2
0.2
0 0
100 110 120 130 140 150 160 170 180 190 200
Distance (m)
Figure 7.12. Wave height attenuation around submerged breakwater due to

depth-induced breaking [Hrms~1.0 m, T=4.71 s, h= 4.0 m, γ=0.73]. a) B=35 m,
b) B=15 m, c) B=3 m.
1
measured
0.8 Calculated
0.6
Kt
0.4
0.2
0
T1-26 T2-26 T3-26
Test series
Figure 7.13. Comparison of calculated transmission coefficient by

SWAN model with the measured one in laboratory experiments
(Hrms~1.0 m, T=4.71, h= 4.0 m, γ=0.73).
1.0
0.8
Kt (measured)
0.6 Experiment
B=3.5 m & h=0.4 m
B=3.5 m & h=0.45 m
B=1.5 m & h= 0.4 m
0.4 B=1.5 m & h=0.45 m
B= 0.3 m & h= 0.4 m
B=0.3 m & h=0.45 m
0.2
0.2 0.4 0.6 0.8 1.0
(clculated)
Kt (calculated)
Figure 7.14. Comparison of measured transmission coefficients with calculated

using SWAN (dissipation model of Battjes and Janssen, 1978 with γ=0.73).
2.0 2.0
correl = 0.49 correl = 0.70
1.5 1.5
Gamma
Gamma
1.0 1.0
0.5 B=3500 mm 0.5

B=1500 mm
B=300 mm
0.0 0.0
0.0 10.0 20.0 30.0 0.0 0.1 0.2 0.3 0.4
UR Hi/d
2.0 2.0
correl = -0.30
correl = -0.32
1.5 1.5
Gamma
1.0 Gamma 1.0
0.5 0.5
0.0 0.0
1.0 2.0 3.0 4.0 0.0 20.0 40.0 60.0
Ir L/Hi
2.0 2.0
correl = 0.88 correl =-0.31
1.5 1.5
Gamma
Gamma
1.0 1.0
0.5 0.5
0.0 0.0
0.0 0.5 1.0 H /h 1.5 2.0 0.0 0.5 1.0 1.5 2.0 2.5
i s B/Lo
2.0
correl = 0.70
1.5
Gamma
1.0
0.5
0.0
0.0 0.1 0.2 0.3 0.4
Hi/h

calculated breaker parameter (Battjes and Janssen, 1978 dissipation model).
5.0
B=3500 mm
4.0 B=1500 mm
B=300 mm
3.0
Gamma
2.0
1.0
0.0
0.0 0.5 1.0 1.5 2.0
hs/Hi
Figure 7.16. Variation of derived calibrated breaking index against submergence

ratio hs/Hi for all experimental data (dissipation model of Battjes and Janssen, 1978).
2.5
y =-2 .62x+ 3.75
2.0 R 2 =0.85
I W =0.95
1.5 e=0.12
Gamma
1.0
0.5
0.0
0.5 1.0 1.5 2.0 2.5
hs/Hi
Figure 7.17. Proposed model for calculating breaker parameter in SWAN (selected
experimental data, 0.5<hs/Hi ≤2.0).
1.0
R2=0.72
+10%
IW =0.92
0.8 e=0.11
-10%
Kt (measured)
0.6
B=3.5 m & h=0.4 m

B=3.5 m & h=0.45 m
0.4
B=1.5 m & h= 0.4 m
B=1.5 m & h=0.45 m
B= 0.3 m & h= 0.4 m
B=0.3 m & h=0.45 m
0.2
0.2 0.4 0.6 0.8 1.0
Kt (clculated)
Figure 7.18. Comparison of measured and calculated transmission coefficient Kt

using the proposed calibrated values of breaker index in dissipation equation of
Battjes and Janssen (1978).
hs/Hi
Figure 7.19. Proposed linear model for calculating breaker index γ in wave
dissipation model (after Johnson, 2006).
Figure 7.20. The variation of calculated discharge over the time in FLOW module.
Figure 7.21. Calculated incident wave height variation along the flume for FLOW-
OBSTACLE model (Test T1-26, B=35m, h=4.0m, T=4.70 sec, Hi=0.51 m and
Kt=0.50).
Figure 7.22. Calculated water level changes along the flume for FLOW-OBSTACLE
model (Test T1-26, B=35m, h=4.0m, T=4.70 sec, Hi=0.51 m and Kt=0.50).
Figure 7.23. Calculated wave-induced forces along the flume for FLOW-
OBSTACLE model (Test T1-26, B=35m, h=4.0m, T=4.70 sec, Hi=0.51 m and
Kt=0.50).
Figure 7.24. Calculated depth averaged velocity in the flume and around the
submerged breakwater for FLOW OBSTACLE model (Test T1-26, B=35m, h=4.0m,
T=4.70 sec, Hi=0.51 m and Kt=0.50).
2.0
Test1, h=400mm
Test2, h=400mm
Test1, h=450mm +10%
Test2, h=450mm
1.5 Test3, h=400mm
Claculated q (m 3/s/m)
-10%
1.0
0.5
0.0
0 0.5 1 1.5 2
3
Measured q (m /s/m)
Figure 7.25. Comparison of calculated and measured discharge over submerged

breakwater when FLOW-OBSTACLE model applied in the simulation.
1.20
FLOW-OBSTACLE model B/L<0.3
1.00 B/L>0.3
0.80
0.60
0.40
0.20
0.00
R2 IW ε
Figure 7.26. Statistical comparison of measured values of discharge with the

calculated values from FLOW-OBSTACLE model for different ranges of breakwater
crest width (B/L). R2, IW and ε are squared correlation coefficient, Willmot index and
error function, respectively.
200 Radiation stresses method 200 Test T1-26 200 Test T2-21
Energy dissipation method
Wave forces (N/m 2)

Wave forces (N/m 2)
Wave forces (N/m 2)

Breakw ater Layout
100 100 100
0 0 0
-100 -100 -100

Test T1-21
-200 -200 -200
100 125 150 175 200 100 125 150 175 200 100 125 150 175 200
DIstance (m ) DIstance (m ) DIstance (m )
350
200 Test T2-26 200 Test T3-21 Test T3-26
250
Wave forces (N/m 2)
Wave forces (N/m 2)
Wave forces (N/m 2)

150
100 100
50
0 0 -50
-150
-100 -100
-250
-200 -200 -350

100 125 150 175 200 100 125 150 175 200 100 125 150 175 200
DIstance (m ) DIstance (m ) DIstance (m )
Figure 7.27. Calculated wave-induced forces by SWAN in x- direction around the breakwater using radiation stress gradients method (green
solid line) and wave energy dissipation method (red solid line).
FLOW-WAVE-2 model (Test T1-26, B=35 m, h= 4 m, T=4.7 sec and Hi=1.02 m).
2
FLOW-WAVE-1
FLOW-WAVE-2
1.6
q (m 3/s/m) 1.2
0.8
0.4
0
T1-21
T1-25
T1-26
T1-27
T1-30
T1-42
T1-43
T2-21
T2-22
T2-25
T2-26
T2-30
T2-43
T2-46
T2-47
T3-21
T3-26
T3-27
Test series
Figure 7.31. Computed discharge using radiation stress gradients (FLOW-WAVE-1)

and wave energy dissipation (FLOW-WAVE-2) methods.
2
+10%
WAVE-FLOW-1
-10%
1.5
Test1, h=400mm
0.5 Test2, h=400mm
Test1, h=450mm
Test2, h=450mm
Test3, h=400mm
0
0 0.5 1 1.5 2
Measured q (m3/s/m)
2 +10%
WAVE-FLOW-2
-10%
1.5
0.5
0
0 0.5 1 1.5 2
Measured q (m3/s/m)
Figure 7.32. Comparison of calculated and measured discharge using radiation stress
gradients (FLOW-WAVE-1) and wave energy dissipation (FLOW-WAVE-2)
methods.
200
FLOW-WAVE-1
FLOW-WAVE-2
150
%Difference
100
50
0
2.33
1.28
1.28
1.28
1.23
1.23
1.00
1.00
0.91
0.55
0.55
0.53
0.39
0.37
0.37
0.20
0.11
0.11
B/L
Figure 7.33 Deviation between calculated and measured discharge as a function of

breakwater crest width for radiation stress gradients and energy dissipation methods.
1.50
(a) FLOW-WAVE-1
1.20 B/L>0.55 FLOW-WAVE-2
0.90
0.60
0.30
0.00
R2 IW ε
Statistics parameters
1.50
(b)
1.20 B/L<0.55
0.90
0.60
0.30
0.00
R2 IW ε

calculated values from FLOW-WAVE-1 and FLOW-WAVE-2 models. a) Breakwater
crest width B/L≥0.55, b) Breakwater crest width B/L<0.55.
1.20
(a)
1.00 B/L<0.35
0.80
0.60
0.40
0.20
0.00
R2 Iw e
Statistics param eters
1.20
(b)
1.00 0.35<B/L<1.0
0.80
0.60
0.40
0.20
0.00
R2 Iw e
1.20
(c)
FLOW-WAVE-1
1.0<B/L
1.00
FLOW-WAVE-2
0.80
0.60
0.40
0.20
0.00
R2 Iw e
Figure 7.35 Statistical comparison of measured values of discharge with the

calculated values from FLOW-WAVE-1 and FLOW-WAVE-2 models for different
range of breakwater crest width B/L.
1.0
0.8
Kt (measured)
0.6
0.4 B=3.5 m & h=0.4 m

B=3.5 m & h=0.45 m
B=1.5 m & h= 0.4 m
0.2 B=1.5 m & h=0.45 m
B= 0.3 m & h= 0.4 m
B=0.3 m & h=0.45 m
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Kt (calculated)
Figure 7.36. Comparison of measured transmission coefficient with the calculated Kt

using Roelvink’s (1993) dissipation model and the proposed value of breaking
parameter γ=0.55.
2.0 2.0
correl = 0.69 correl = 0.61
1.5 1.5
Gamma
Gamma
1.0 1.0
0.5 B=3500 mm 0.5

B=1500 mm
B=300 mm
0.0 0.0
0.0 10.0 20.0 30.0 0.0 0.1 0.2 0.3 0.4
UR Hi/d
2.0 2.0
correl = 0.61 correl ~ 0.00
1.5 1.5
Gamma
Gamma
1.0 1.0
0.5 0.5
0.0 0.0
0.0 0.1 0.2 0.3 0.4 0.0 1.0 2.0 3.0 4.0
Hi/h Ir
2.0 2.0
correl ~ 0.00 correl = 0.51
1.5 1.5
Gamma
Gamma
1.0 1.0
0.5 0.5
0.0 0.0
0.0 20.0 40.0 60.0 0.0 0.2 0.4 0.6
L/Hi ds/Hi
2.0 2.0
correl = 0.78 correl =0.51
1.5 1.5
Gamma
Gamma
1.0 1.0
0.5 0.5
0.0
0.0
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0
0.0 1.0 2.0 3.0
Hi/hs Lo/B

calculated breaker parameter (Roelvink, 1993 depth-induced dissipation model).
2.00
1.60 y = -1.93x +2.89

R2=0.75
IW =0.92
1.20
Gamma
e=0.15
0.80
0.40
0.00
0.50 0.75 1.00 1.25 1.50 1.75 2.00
hs/Hi
Figure 7.38. Proposed model for calculating the calibrated breaker parameter in the
dissipation model of Roelvink (1993). Selected experimental data with 0.5<hs/Hi ≤2.0
was adopted during calibration processes.
1.0
R2=0.48
+10%
I W =0.81
0.8 e=0.16
-10%
Kt (measured)
0.6
B=3.5 m & h=0.4 m

0.4 B=3.5 m & h=0.45 m
B=1.5 m & h= 0.4 m
B=1.5 m & h=0.45 m
B= 0.3 m & h= 0.4 m
B=0.3 m & h=0.45 m
0.2
0.2 0.4 0.6 0.8 1.0
Kt (calculated)
Figure 7.39. Comparison of measured transmission coefficients Kt with the

calculated values using the calibrated breaker parameters (Equation 7.14).
1.0
R2=0.82
I W =0.95 +10%
0.8 e=0.09
-10%
Kt (measured)
0.6
B=3.5 m & h=0.4 m

0.4 B=3.5 m & h=0.45 m
B=1.5 m & h= 0.4 m
B=1.5 m & h=0.45 m
B= 0.3 m & h= 0.4 m
B=0.3 m & h=0.45 m
0.2
0.2 0.4 0.6 0.8 1.0
Kt (calculated)
Figure 7.40. Comparison of measured and calculated transmission coefficients Kt

with the using the calibrated breaker parameters (Equation 7.16) in the Roelvink’s
dissipation model.
Figure 7.41. Variation of calculated short-wave and roller energy by FLOW-
ROLLER model along the flume and over the submerged breakwater for test T1-26
(B=35m, h=4.0m, hs=1.0m, Hi=1.02m and T=4.71 sec).
(B=15.0m, h=4.0m, hs=1.0m, Hi=1.07m and T=4.71 sec).
(B=3.0m, h=4.0m, hs=1.0m, Hi=1.04m and T=4.71 sec).
Figure 7.44. Computed depth-averaged current velocity along the flume and over the
submerged breakwater using FLOW-ROLLER model for Test T1-26 (B=35m,
h=4.0m, hs=1.0m, Hi=1.02m and T=4.71 sec).
Figure 7.45. Computed depth-averaged discharge per unit width along the flume and
over the submerged breakwater using FLOW-ROLLER model for Test T1-26
(B=35m, h=4.0m, hs=1.0m, Hi=1.02m and T=4.71 sec).
Test1, h=400mm, B=3500mm
Test2, h=400mm, B=1500mm +10%
1.5 Test1, h=450mm, B=3500mm
-10%
0.5
0
0 0.5 1 1.5
Measured q (m 3/s/m)
Figure 7.46. Calculated discharge over the submerged breakwater using roller model
(FLOW-ROLLER) against the measured value from laboratory experiments.
1.20
FLOW-ROLLER B/L>1.0
1.00
0.35<B/L<1.0
0.80 B/L<0.35
0.60
0.40
0.20
0.00
R2 Iw e
Statistics param eters

calculated values from FLOW-ROLLER models for different groups of crest width
ratio.
1.20
1.0<B/L FLOW-WAVE-1
1.00 FLOW-WAVE-2
FLOW-ROLLER
0.80 FLOW-OBSTACLE
0.60
0.40
0.20
0.00
R2 Iw e
1.20
0.35<B/L<1.0
1.00
0.80
0.60
0.40
0.20
0.00
R2 Iw e
1.20
B/L<0.35
1.00
0.80
0.60
0.40
0.20
0.00
R2 Iw e
Figure 7.48. Comparison of all applicable models in Deflt3D in predicting discharge

over submerged breakwaters for different relative breakwater crest width groups.
CHAPTER 8
CONCLUSIONS AND RECOMENDATIONS
8.1 Summary and Conclusions
Wave transformation and wave-induced mass flux over impermeable smooth

submerged breakwaters/reefs has been investigated through this thesis. An extended set
of laboratory experimental tests was conducted to examine the effect of different
parameters on wave transmission and current over submerged breakwaters. Numerical
modelling of wave deformation and current has also been carried out and calibrated to
provide functional design guidelines to predict hydrodynamic effects of flow over
submerged breakwaters/reefs. A wide ranging review of relevant literature is presented
in the thesis.
Experimental testing in a 1m wide and 12m long internal channel at the centre of the 3m
wide, 1.5m deep and 30m long flume ensures wave-induced current returns through the
channels on both sides of the central channel. Therefore, no return flow due to
shoreward set-up is observed above the central test submerged breakwater section and
set-up over and behind the structure remains a function of breakwater top friction.
Moreover, measurements of shoreward directed free flow overtopping the structure are
more reliable.
Analysis of the results showed that the observed breaker-depth index increases slightly
with decreasing breakwater crest width. A modified breaking criterion following Miche
(1944) was provided for more appropriate prediction of wave breaking over submerged
breakwaters.
The laboratory measured data of transmitted breaking wave height showed that the
transmission coefficient “Kt ” increases with submergence depth and wave period
whilst, breakwater crest width and incident wave height have inverse effect on Kt.
172
Analysing the variation of transmission coefficient against dimensionless parameters
showed that the inverse breakwater crest width ratio “Lo/B ”, the ratio of incident wave
height to breakwater crest width “Hi/B ”, inverse ratio of wave steepness “Lo/Hi ” and
the submergence ratio “hs/Hi” are in order the more influential parameters on
transmission coefficient. Variation of transmission coefficient with respect to
submergence ratio for different breakwater crest width revealed that the submergence
ratio is more influential in the tests with wider breakwater. For a wide range of B/Lo
including broad crested and reef type breakwaters, an improved predictive model of Kt
has been provided (Equation 5.14). The proposed model is a significant advance in that
it can be used to predict Kt consistently over submerged breakwaters/reefs in the range
of crest width ratio 0.08 < B/Lo < 4.4 whereas existing models in the literature are
restricted to a limited range of narrow crested breakwater.
The measured wave-induced discharge flux q (discharge per unit width) over the
submerged breakwater was examined to investigate the dependency of discharge on
dimensional and non-dimensional parameters. Graphical and analytical analysis of data
revealed that the amount of wave-induced flow over the breakwater increases with
incident wave height due to higher energy dissipation during the wave breaking and for
constant values of incident wave height, water depth and wave period, the breakwater
with broader crest width provides higher value of discharge.
Analysis of the values of dimensionless parameters revealed that water depth ratio
“B/Hi ” has the strongest influence on discharge flux variation whilst B/Lo, hs/Lo, and
Hi/Lo have respectively less and reducing effects. The dimensionless discharge flux
ratio, q *=q/(gH3)0.5 increases with submergence and breakwater crest width ratios and
decreases with increasing water depth ratio “Hi/B ” and wave steepness. However, the
effectiveness of the dimensionless parameters depends on the range of the breakwater
crest width ratio so that wave steepness is more dominant in variation of dimensionless
discharge for narrow crested breakwaters (B/Lo <0.35). On the other hand, q* is more
affected by Hi/B for relatively wide and broad-crested breakwaters (B/Lo > 0.35).
Variation of discharge ratio against other dimensionless parameters showed that the
maximum discharge ratio increases with breakwater crest width and the envelopes of
maximum dimensionless discharge are provided in Equation 5.16 as a function of
173
submergence ratio in different range of crest width ratio (Figure 5.18). The proposed
equation may be applied in the preliminary design of a submerged breakwater for
prediction of maximum discharge passing over the structure as a function of
submergence ratio.
Two analytical models of Symonds et al. (1995) and Gourlay and Colleter (2005) were
assessed for predicting current velocity and discharge over relatively wide and broad-
crested submerged breakwaters (B/Lo > 0.35). Comparison of the models with
experimental results revealed that Symonds et al.’s (1995) method provides
qualitatively reasonable results but for broader application needs a practical method for
estimating the friction coefficient as a calibration factor in the model.
The model provided by Gourlay and Colleter (2005) was found to overestimate
discharge over submerged breakwaters/reefs if the unknown coefficients in Equation
5.24 to 5.26 were adopted as the same as Gourlay and Colleter. The number of
unknown coefficients in the corresponding equations and unavailability of field or
modelled data for calibrating the model encouraged consideration of other methods for
estimating the various model parameters. The model of Gourlay and Colleter (2005)
provides a reasonable calculation of dimensionless discharge for high submergence
(0.7 < S) if amended by the inclusion of a correction factor (Equation 5.34) being a
function of crest width ratio (B/Lo). No further field or laboratory data are needed for
finding the unknown coefficients incorporated in the amended Gourlay and Colleter
method.
Wave set-up over and behind the breakwater was also investigated. It was found that the
maximum value of wave-induced set-up increases with incident wave height for all
ranges of breakwater crest width. In addition, for a given wave height and submergence
depth, set-up increases with breakwater crest width.
The 2D depth averaged numerical modelling of wave-induced discharge and wave

transmission over submerged breakwaters was investigated using the recent research
version of Delft3D. Both models of Battjes and Janssen (1978) and Roelvink (1993)
were found to overestimate wave-induced breaking dissipation over submerged
breakwaters. Two alternative equations (Equation 7.7 and 7.16) for wave breaker
174
parameter were developed for incorporation in the models of Battjes and Janssen
(1978) and Roelvink (1993) to improve the calculation of transmitted wave height over
and behind submerged breakwaters.
Applying the corresponding proposed calibrated breaker parameter (Equation 7.7 and
7.16) in Delft3D-WAVE and Delft3D-FLOW (ROLLER), the following guidance was
given as to the most appropriate Delft3D modelling to predict discharge flux over
submerged breakwaters. The primary division being governed by relative breakwater
crest width:
• B/Lo< 0.35 – narrow crested submerged breakwaters;
(1) Apply WAVE and FLOW modules in simulating the breakwater as a

transmissive obstacle in grid lines (FLOW-OBSTACLE) – use the relevant
proposed transmission coefficient of Equation 5.14.
(2) The calibrated dissipation model of Battjes and Janssen (1978) in SWAN is also
applicable if wave energy dissipation is applied in the wave force calculation
(FLOW-WAVE-2) and the proposed value of breaker parameter (Equation 7.7) is
used in SWAN.
• 0.35 ≤ B/Lo ≤ 1.0 moderately wide crested submerged breakwaters;
Apply the roller model (FLOW-ROLLER) using the proposed calibrated breaker
parameter (including effects of B/Lo and hs/Hi, Equation 7.16) in the wave
dissipation model of Roelvink (1993).
• 1.0 < B/Lo broad crested submerged breakwaters ;
Both roller (FLOW-ROLLER) and wave model (FLOW-WAVE) associated with

the relevant proposed calibrated breaking parameter (Equation 7.7 and 7.16,
respectively) in appropriate dissipation formulae are applicable.
175
8.2 Discussion and Recommendations for Future Research
As discussed in Chapter 3, the present laboratory tests were restricted to smooth

impermeable submerged breakwaters subjected to monochromatic waves. Although
Van der Meer et al. (2005) indicated that transmission coefficient is affected by the
surface roughness of low crested breakwaters, other research (Davies and Kriebel , 1992
and Seabrook, 1997) indicates that breakwater roughness has little effect when the
breakwater is fully submerged (crest height below still water level), particularly for low
steepness waves. It is anticipated that the value of transmission coefficient over fully
submerged and smooth breakwaters does not vary significantly in comparison with
rubble mound submerged breakwaters. However, a few detailed tests with the same
geometric, water level and wave conditions but different type of breakwater materials is
recommended to be carried out to better investigate the effect of breakwater materials
on wave transmission coefficient.
Wave-induced discharge over submerged breakwaters is however influenced by the

surface roughness of the breakwater. The effect of submerged crest roughness increases
with breakwater crest width. For narrow (B/Lo< 0.35) and moderately wide
(0.35 ≤ B/Lo ≤ 1.0) crested submerged breakwaters negligible changes in wave-induced
discharge over the breakwater is anticipated since 100% increase in top surface
roughness (rubble mound) leads to 6% decrease in calculated discharge over the
breakwater using Delft3D-FLOW. Applying appropriate values of roughness coefficient
in the momentum equation of flow can provide reasonable prediction of wave-induced
discharge over broad-crested submerged breakwaters. However, more experimental
tests are needed to investigate the effect of roughness on current and verify the
application of conventional methods (e.g. Chézy and Manning) in estimating top surface
roughness of submerged breakwaters.
More experimental tests with irregular incident waves on smooth submerged

breakwaters with the same geometry, water depth, wave height and wave period are
needed to investigate the applicability of the proposed approaches in this thesis to
predict transmitted wave height and wave-induced current over fully submerged
breakwaters under irregular wave conditions.
176
Application of water quality and sediment transport morphological models (e.g.
Delft3D-MOR) to predict shoreline response in presence of submerged breakwaters
(being conditional upon the reliability of the hydrodynamic modelling of wave
breaking, transmission and discharge flux over the submerged breakwaters) can now be
undertaken with increased confidence. The thesis has provided valuable guidance as to
the best combination of WAVE and FLOW (ROLLER) modules within Delft3D for the
correct modelling of the hydrodynamic processes. The effect of different parameters
(beyond submergence, crest width, wave height and period) such as the distance of the
breakwater from the beach, breakwater length alongshore and wave directionality can
now be investigated more confidently using the morphological numerical models.
177
REFERENCES
Abdul Khader M.H. and Rai, S.P. (1980). “A study of submerged breakwaters.” Journal of
Hydraulic Research, Vol. 18, No. 2, 113-121.
Adams, C.B. and Choule J.S. (1986). “Wave transmission across submerged near-surface
breakwater.” Proc. 20th Int. Conf. on Coastal Engineering, ASCE, 1730-1738.
Ahrens, J.P. and McCartney. B.L. (1975). “Wave period effect on the stability of riprap.”
Proc. of Civil Engineering in the Ocean/ІІІ, ASCE, 1019-1034.
Ahrens, J.P. (1987). “Characteristics of reef breakwaters.” Technical Report, CERC-87-17,

45 p.
Ahrens, J.P. (1989). “Stability of reef breakwaters.” Journal of the Waterway, Port, Coastal
and Ocean Engineering, Vol. 115, No. 2, 221-234.
Allsop, N.W.H. (1983). ”Low-crested breakwaters, studies in random waves.” Proc.

Coastal Structure’83, ASCE, 94-107.
Amnity, P. Lamberti, A. and Liberatore, G. (1983). “Experimental studies on submerged

barriers as shore protection structures.” Proc. Int. Conf. on Coastal and Port Eng. in
Developing Countries, Colombo, 201-218.
Battjes, J.A. and Janssen, J.P.F.M. (1978). “Energy loss and set-up due to breaking of
random waves.” Proc. 16th Int. Conf. on Coastal Engineering, ASCE, 569-587.
Battjes, J.A. and Stive, M.J.F. (1985). “Calibration and verification of a dissipation model
for random breaking waves.” Journal of Geophysical Research, Vol. 90, No. C5, 9159-
9167.
Beach Erosion Board (1940). “A model study of the effect of submerged breakwaters on
wave action.” Technical memorandum No. 1, Chief of Engineers, U.S. War Dept.
Biesel, F. and Suquet, F. (1953). “Laboratory wave generating apparatus.” Project report
39, St. Anthony Falls Hydraulic Laboratory, Minnesota University, Minneapolis.
Bleck, M. and Oumeraci, H. (2001). “Wave damping and spectral evolution at artificial
reefs.” Proc. 4th Int. Symp. on Ocean Wave Measurement and analysis, ASCE, 1062-
1072.
Bleck, M. and Oumeraci, H. (2002). “Hydraulic performance of artificial reefs: Global and
178
local description.” Proc. 28th Int. Conf. on Coastal Engineering, ASCE, 1778-1790.
Bleck, M. and Oumeraci, H. (2004).“Analytical model for wave transmission at artificial

reefs.” Proc. 29th Int. Conf. on Coastal Engineering, ASCE, 269-281.
Browder, A. and Dean, R.G. (1996). “Performance of a submerged breakwater for shore
protection.” Proc. 25th Int. Conf. on Coastal Engineering, ASCE, 2312-2323.
Buhr Hansen, J. and Svendsen, A. (1974). “Laboratory generation of waves of constant

form.” Proc. 14th Int. Conf. on Coastal Engineering, ASCE, 321-339.
Bouws, E. and Komen, G.J. (1983). “On the balance between growth and dissipation in an
extreme, depth-limited wind-sea in the southern North Sea.” Journal of Physical
Oceanography., Vol. 13, 1653-1658.
Calabrese, M., Vicinanza, D. and Buccino M. (2002). “Large-scale experiments on the

behaviour of low crested and submerged breakwaters in presence of broken waves.”
Proc. 28th Int. Conf. on Coastal Engineering, ASCE, 1900-1912.
Calabrese, M., Vicinanza, D. and Buccino M. (2003). “Low-crested and submerged

breakwater in presence of broken waves.” Proc. 13th ISOPE, 831-836.
Chasten, A., Rosati, J., McCormick, J. (1993). “Engineering Design Guidance for Detached
Breakwaters as Shoreline Stabilization Structures.” Coastal Engineering Research
Center, U.S. Army Corps of Engineers, Waterways Experiment Station, CERC-93-19.
Chatham, C.E. (1972). “Movable-bed model studies of perched beach concept.” Proc. 13th
Int. Conf. on Coastal Engineering, ASCE, 1191-1216.
Chiaia, G., Damiani, L. and Petrillo, A. (1992). “Evolution of a beach with and without a
submerged breakwater: Experimental investigation.” Proc. 23th Int. Conf. on Coastal
Engineering, ASCE, 1959-1972.
Chiang et al. (1969). “Scattering of surface waves by rectangular obstacles in waters of

finite depth.” Journal of Fluid Mechanics, Vol. 38, Part 3, 499-511.
CERC (2004).Coastal Engineering Manual, Co. Eng. Res. Centre, U.S Corps of Eng.,
Vicksburg. http://www.usace.army.mil/publications/eng_manuals/cecw.htm
Colebrook, C.F. (1939). “Turbulent flow in pipes, with particular reference to the transition
between the smooth and rough pipe lows.” Journal of Inst. Civil Engr. London, Vol. 1l,
133-156.
Collins, J.I., (1972). “Prediction of shallow water spectra.” Journal of Geophysical

Research, Vol. 77, No. 15, 2693-2707.
179
Collins, J. I. And Wier, W. (1969), “Properties of wave characteristics in the surf zone.”
Tetra Tech Report No. TC-149.
Daemen, I.F.R. (1991). “Wave transmission at low crested structures.” MSc. Thesis, Delft
University of Technology, Faculty of Civil Engineering, Delft Hydraulics Report H
462.
Daemrich, K. and Kahle, W. (1985). “Schutzwirkung von Unterwasser Wellen brechern

unter dem Einfluss unregelmassiger seegangswellen.” Technical Report, Franzius
Instituts fur Wasserbau und Kusteningenieurswesen, Report Heft 61 (in Germany).
Dally, W.R. and Brown, C.A. (1995). “A modeling investigation of the breaking wave
roller with application to cross-shore current.” Journal of Geophysical Research, Vol.
100, No. C12, 24873-24883.
d’Angremond, K., Van der Meer, J.W. and De Jong, R.J. (1996). “Wave transmission at
low-crested structures.” Proc. 25th Int. Conf. on Coastal Engineering, ASCE, 2418-
2427.
Davies, B. and Kriebel, D.L. (1992). “Model testing on wave transmission past low crested
breakwaters.” Proc. 23th Int. Conf. on Coastal Engineering, ASCE, 1115-1128.
Dean, R.G. (1977). “Equilibrium beach profiles: U.S Atlantic and Gulf coasts.”
Department of Civil Engineering, Ocean Engineering Report No. 12, University of
Delaware, Newark, DE.
Dean, W.R. (1945). “On the reflection of surface waves by a submerged plane barrier.”
Proc. Cambridge Philosophy Society, 41, Part 3, 231-236.
Debski, D. and Loveless, J.H (1997). “The design and performance of submerged
breakwaters.” Research Report, University of Bristol.
De Later, J. (1996). “Effect of submerged breakwaters on a beach profile exposed to

regular waves in a wave basin.” Msc. Thesis, Delft University of Technology,
Netherlands.
Delft Hydraulics. (2005a). “Delft3D-WAVE, Simulation of short-crested waves with

SWAN.” User manual, WL/Delft Hydraulics, The Netherlands.
Delft Hydraulics. (2005b). “Delft3D-FLOW, simulation of multi-dimensional

hydrodynamic flow and transport phenomena, including sediment.” User manual,
WL/Delft Hydraulics, The Netherlands.
Dick, T.M. and Brebner, A. (1968). “Solid and permeable submerged breakwaters.” Proc.
11th Int. Conf. on Coastal Engineering, ASCE, 1141-1158.
180
Dingemans, M.W., Radder, A.C. and de Vriend, H.J. (1987). “Computation of the driving
forces of wave-induced currents.” Journal of Coastal Engineering, 11, 539-563.
Dingemans, M.W. (1997). “Water wave propagation over uneven bottoms. Part 1 -linear
wave propagation.” Advanced Series on Ocean Engineering, 13, World Scientific, 471
p.
Diskin, M.H., Vajda, M.L. and Amir, I. (1970). “Piling-up behind low and submerged
permeable breakwaters.” Journal, Waterways and Harbours Division, WW2, ASCE,
359-372.
Draper, N.R. (1984). “The Box-Wetz criterion versus R2.” Journal of Statistical Society,
147, 100-103.
Drei, E. and Lamberti, A. (1999). ”Wave pumping effect of a submerged barrier.” Coastal
Structure’99, ASCE, 667-673.
Driscoll, A.M., Darlymple, R.A. and Grilli, S.T. (1992). “Harmonic generation and
transmission past a submerged rectangular obstacle.” Proc. 23th Int. Conf. on Coastal
Duncan, J.H. (1981). “An experimental investigation of breaking waves produced by a

towed hydrofoil.” Proc. The Royal Society of London, Ser. A, 377, 331-348.
Eldeberky, Y. and Battjes, J.A. (1996). “Spectral modelling of wave breaking: Application
to Boussinesq equations” . Journal of Geophysical Research, Vol. 101, No. C1, 1253-
1264.
Flik, R.E. and Guza, R.T. (1980). “Paddle generated waves in laboratory.” Journal of the
Waterway, Port, Coastal and Ocean Division, Vol. 106, No. WW1, 79-97.
Galvin, C.J. (1968). “Shapes of unbroken periodic gravity water waves.” Transaction of the
American geophysical Union, Vol. 49, No. 1, p 206.
Galvin, C.J. (1969). “Breaker travel and choice of deign wave height.” J. Waterw.
Harbours Div. Am. Soc. Civ. Eng., Vol. 95 (WW2), 175-200.
Garcia, N., Lara, J.L. and Losada, I.J. (2004). “2-D numerical analysis of near-field flow at
low-crested permeable breakwaters.” Journal of Coastal Engineering, 51, 991-1020.
Gilbert, G., Thompson, D.M. and Brewer, A.J. (1970). “Design curves for regular and
random wave generators.” Journal of Hydraulic Research, Vol. 9, No. 2, 161-196.
Gironella, X., Sanchez-Arcilla, A. (1999). “Hydrodynamic behaviour of submerged

breakwaters. Some remarks based on experimental results.” Proc. Coastal
Structures’99, ASCE, 891-896.
181
Goda, Y. (1969). “Re-analysis of laboratory data on wave transmission over breakwaters.”
Report of the port and Harbour Research Institute, Vol. 18, No. 3, 3-18.
Goda, Y., Takeda, H. and Moriya, Y. (1967). “Laboratory investigation on wave

transmission over breakwaters.” Report of the port and Harbour Research Institute, No.
13, 38 pp.
Goda, Y. and Suzuki, Y. (1976). “Estimation of incident and reflected waves in random
wave experiments.” Proc. 15th Int. Conf. on Coastal Engineering, ASCE, 828-845.
Goda, Y., Okazaki, K. and Kagawa, M. (1999). ”Generation and evolution of harmonic
wave components by abrupt depth changes.” Proc. Coastal Structure’99, ASCE, 649-
658.
Gómez Pina, G. and Valdés, J.M. (1990). “Experiments on coastal protection submerged
breakwaters: A way to look at the results.” Proc. 22th Int. Conf. on Coastal
González, M. Medina, R. and Losada, M.A. (1999). “Equilibrium beach profile for perched
beaches.” Journal of Coastal Engineering, 36, 343-357.
Gourlay, M.R. (1993). “Wave setup and wave generated currents on coral reefs.” Proc.
11th Australian Conf. on Coastal and Ocean Engineering, Townsville, 479-484.
Gourlay, M.R. (1994). “Wave transformation on a corral reef.” Journal of Coastal

Engineering, 23, 17-42.
Gourlay, M.R. (1996a). “Wave set-up on coral reefs. 1. Set-up and wave-generated flow on
an idealized two dimensional horizontal reef.” Journal of Coastal Engineering, 27, 161-
193.
Gourlay, M.R. (1996b). “Wave set-up on coral reefs. 2. Set-up on reefs with various
profiles.” Journal of Coastal Engineering, 27, 161-193.
Gourlay, M.R. (1997). “Wave set-up on coral reefs: Some practical applications.”
Combined Australian Coastal Engineering and Ports Conf, 959-964.
Gourlay, M.R. and Colleter, G. (2005). “Wave-generated flow on coral reefs-an analysis for
two-dimensional horizontal reef-tops with steep faces.” Journal of Coastal
Gunbak, A. (1979). “Rubble mound breakwaters.” Report no. 1, Division of Port and
Ocean Engineering, University of Trondheim, Trondheim, Norway.
Günther, H., Hasselmann, S. and Janssen P.A.E.M. (1992). “The WAM model Cycle 4
(revised version)”, Deutsch. Klim. Rechenzentrum, Techn. Rep. No. 4, Hamburg,
182
Germany.
Groenewoud, M.D., van der Graaff, J., Claessen, E.W.M. and van der Beizen, S. (1996).
“Effect of submerged breakwater on profile development.” 25th Int. Conf. on Coastal
Engineering, ASCE.2428-2441.
Haller, M.C., Darlymple, R.A. and Svendsen, I.A. (2002). “Experimental study of
nearshore dynamics on a barred beach with rip channels” Journal of Geophysical
Research, Vol. 107, C6, 14-1 – 14-21.
Hansen, J.B. and Svendsen, I.A. (1994). “A theoretical and experimental study of
undertow.” Proc. 24th Int. Conf. on Coastal Engineering, ASCE, 2246-2262.
Hanson, H. and Kraus, N.C. (1990). “Shoreline response to a single transmissive

breakwater.” Proc. 22th Int. Conf. on Coastal Engineering, ASCE, 2034-2046.
Hara, M., Yasuda, T. and Sakakibara, Y. (1992). “Characteristics of a solitary wave

breaking caused by a submerged obstacle.” Proc. 23th Int. Conf. on Coastal
Engineering, ASCE. 253-266.
Hardy, T.A. (1993). “The attenuation and spectral transformation of wind waves on a coral
reef.” PhD. Thesis, 336 pp., James Cook University, Townsville, Queensland.
Hasselmann, K. (1974). “On the spectral dissipation of ocean waves due to whitecapping.”
Boundary-Layer Meteorology, Vol. 6., No. 1-2, 107-127.
Hasselmann, K. and Collins, J.I. (1968). “Spectral dissipation of finite-depth gravity waves
due to turbulent bottom friction.” Journal of Marine Research, Vol. 26, No. 1, 1-12.
Hasselmann, K., Barnett, T.P., Bouws, E., Carlson, H., Cartwright, D.E., Enke, K., Ewing,
J.A., Gienapp, H., Hasselmann, D.E., Kruseman, P., Meerburg, A., Müller, P., Olbers,
D.J., Richter, K., Sell, W. and Walden, H. (1973). “Measurements of wind-wave
growth and swell decay during the Joint North Sea Wave Project (JONSWAP).” Dtsch.
Hydrogr. Z. Suppl., 12, A8.
Hattori, M. and Sakai, H. (1994). “Wave breaking over permeable submerged

breakwaters.” Proc. 24th Int. Conf. on Coastal Engineering, ASCE. 1101-1113.
Healy, J.J. (1953). “Wave damping effect of beaches.” Proc. Minnesota Int. Hydraulics
Convention, 213-220.
Hearn, C.J. (1999). “Wave-breaking hydrodynamics within coral reef systems and the
effect of changing relative sea level.” Journal of Geophysical Research, Vol. 104, No.
C12, 30007-30019.
Iwagaki, Y. and Saki, T. (1970). “Horizontal water particle velocity of finite amplitude
183
waves.” Proc. 12th Int. Conf. on Coastal Engineering, ASCE, 309-325.
Iwata, K., Kawasaki, K. and Kim, D. (1996). “Breaking limit, breaking and post-breaking
wave deformation due to submerged structures.” Proc. 25th Int. Conf. on Coastal
Johnson, H.K. (2006). “Wave modeling in the vicinity of submerged breakwaters.” Journal
of Coastal Engineering, 53, 39-48.
Johnson, H.K., Karambas, T.V., Avgeris, I. Zanuttigh, B., Gonzalez-Marco, D. and

Caceres, I. (2005). “Modelling of waves and currents around submerged breakwaters.”
Journal of Coastal Engineering, 52, 949-969.
Johnson, J.W., Fuchs, R.A. and Morison, J.R. (1951). “The damping action of submerged
breakwaters.” Transactions, American Geophysical Union, Vol. 32, No. 5, 704-718.
Jonson, I.G. (1966). “Wave boundary layers and friction factors.” Proc. 10th Int. Conf. on
Coastal Engineering, ASCE, 127-148.
Kamphuis, J.W. (2000). Introduction to coastal engineering and management, World

Scientific, 437 p.
Kaminsky, G.M. and Kraus, N.C. (1993). “Evaluation of depth-limited wave breaking
criteria.” Proc. of 2nd Int. Symposium on Ocean Wave Measurement and Analysis,
New Orleans, 180-193.
Kobayashi, N., Otta, A.K. and Roy, I. (1987). “Wave reflection and run-up on rough
slopes”. Journal of the Waterway, Port, Coastal and Ocean Engineering, Vol. 113, No.
3, 282-298.
Kobayashi, N. and Wurjanto, A. (1988). “Wave transmission over submerged

breakwaters”. Journal of the Waterway, Port, Coastal and Ocean Engineering, Vol.
115, No. 5, 662-680.
Komen, G.J., Hasselmann, S. and Hasselmann, K. (1984). “On the existence of a fully
developed wind-sea spectrum”. Journal of Geophysical Oceanography, Vol. 14, 1271-
1285.
Lamberti, A. (2005). “Editorial.” Journal of Coastal Engineering, 52, 815-818.
Lamberti, A. and Mancinelli, A. (1996). “Italian experience on submerged barrier as beach

defence structures.” Proc. 25th Int. Conf. on Coastal Engineering, ASCE, 2352-2365.
Lamberti, A., Archetti, R., Kramer, M., Paphitis, D., Mosso, C. and Risio, M.Di. (2005).
“European experience of low crested structures for coastal management.” Journal of
184
Coastal Engineering, 52, 841-866.
Lara, J.L., Garcia, N. and Losada, I.J. (2006). “RANS modelling applied to random wave
interaction with submerged permeable structures .” Journal of Coastal Engineering, 53,
395-417.
Le Mehaute, B., Divoky, D. and Lin, A. (1968). “Shallow water waves: a comparison of
theories and experiments.” Proc. 11th Int. Conf. on Coastal Engineering, ASCE, 86-
106.
Lesser, G.L., Vroeg, de J.H., Roelvink, J.A., Gerloni, de M. and Ardone, V. (2003).
“Modeling the morphological impact of submerged offshore breakwaters.” Coastal
Sediment’2003, Annual Symposium on Coastal Engineering and Science of Coastal
Sediment Processes, 5th, Clearwater Beach, Fla., May 18-23, 2003, Proceedings, 14p.
[CD-ROM].
Lin, P. and Liu, P.L.-F. (1998). “A numerical study of breaking waves in the surf zone.”
Journal of Fluid mechanics, Vol. 359, 239-264.
Longuet-Higgins, M.S. (1964). “On the wave induced difference in mean sea level between
two sides of a submerged breakwater.” Journal of Marine Research, Vol. 25, No. 2,
148-1153.
Longuet-Higgins, M. S. (1970). “Longshore currents generated by obliquely sea waves, 1.”

Journal of Geophysical Research, Vol. 75, No. 33, 6778-6789.
Longuet-Higgins, M.S. (1977). “The mean forces exerted by waves on floating or

submerged bodies with applications to sand bars and wave power machines.” Proc. The
Royal Society of London, A 352, 463-480.
Losada, I.J., Silva, R. and Losada, M.A. (1996), “3-D non-breaking regular wave
interaction with submerged breakwaters”. Journal of Coastal Engineering, 28, 229-248.
Losada, I.J., Patterson, M.D. and Losada, M.A. (1997). “Harmonic generation pas a
submerged porous step.” Journal of Coastal Engineering, 31, 281-304.
Loveless, J.H. and Debski, D. (1997). “Wave transmission and set-up at detached
breakwater” Coastal Dynamics’97, ASCE.
Loveless, J.H., Debski, D. and MacLeod, A.B. (1998). “Sea level set-up behind detached
Madsen, O.S. (1971). “On the generation of long waves.” Journal of Geophysical
Research, Vol. 76, No. 36, 8672-8683.
Madsen, O.S., Poon, Y.-K. and Graber, H.C. (1988). “Spectral wave attenuation by
185
bottom friction: Theory.” Proc. 21th Int. Conf. on Coastal Engineering, ASCE, 492-
504.
Mansard, E.P.D. and Funke, E.R. (1980). “The measurement of incident and reflected
spectra using a least square method.” Proc. 17th Int. Conf. on Coastal Engineering,
ASCE, 154-172.
Massel, S.R. (1983). “Harmonic generation by waves propagating over a submerged step.”
Massel, S.R and Brinkman, R.M. (2001). “Wave-induced set-up and flow over shoals and
coral reefs. Part 1. A simplified bottom geometry case.” Oceanologia, Vol. 43, No. 4,
373-388.
McCowan, J. (1891), “On the solitary wave.” Philosophical Magazine, Journal of Science,
London, Vol. 32.
Mei, C.C. (1983). The applied dynamics of ocean surface waves, Wiley, New York, 740 p.
Melito, I. and Melby, J.A. (2002). “Wave runup, transmission, and reflection for structures
armored with CORE-LOC.” Journal of Coastal Engineering, 42, 33-52.
Miche, R. (1944). “Mouvements ondulatoires de la mer en profoundeur constante ou

decroissante (Breaking wave motion in water of constant water depth).” Ann. Ponts et
Chaussees, 121, 285-318, (in French).
Michell, J. H. (1893), “On the highest waves in water.” Philosophical Magazinem, Vol.
36, 5th series, 430-437.
Morison J.R. (1949). “Model study of wave action on underwater barriers.” Report HE-
116-304, Inst. Eng. Res., Univ. Calif., Berkley.
Morison J.R. and Crooke, R.C. (1953). “The mechanics of deep water, shallow water and
breaking waves.” Tech. Memo No. 40, Beach Erosion Board.
Multer, R.H. (1970). “A nonlinear numerical-hydrodynamic model of a mechanical water

generator.” PhD. Thesis, Mech. And Hydraulics, University of Iowa.
Muñóz-Pérez, J.J., Tejedor, L. and Medina, R. (1999). “Equilibrium beach model for reef-
protected beaches.” Journal of Coastal Research, Vol. 15, No. 4, 950-957.
Nairn, R.B., Roelvink, J.A. and Southgate, H.N. (1990). “Transition zone width and
implication for modeling surfzone hydrodynamics.” Proc. 22th Int. Conf. on Coastal
Nelson, R. C. (1985), “Wave heights in depth limited conditions.” Civil Eng. Trans. Inst.
186
Eng. Aust., Vol. 27, 210-215.
Nelson, R. C. (1987), “Design wave heights on very mild slopes – An experimental

study.” Civil Eng. Trans. Inst. Eng. Aust., Vol. 29, 157-161.
Nelson, R. C. (1994), “Depth limited design wave heights in very flat regions.” Journal of
Coastal Engineering, 23, 43-59.
Nelson, R. C. (1996). “Hydraulic roughness of corral reef platforms.” Appl. Ocean

Research, 18, 265-274.
Ohyama, T., Kioka, W. and Tada, A. (1995). “Applicability of numerical models to

nonlinear dispersive waves.” Journal of Coastal Engineering, 24, 297-313.
Orfanidis, S.J. (1996). Introduction to Signal Processing, Prentice-Hall, Englewood Cliffs,

NJ.
Petti, M. and Ruol, P. (1992). “Laboratory tests on the interaction between nonlinear long
waves and submerged breakwaters.” Proc. 23th Int. Conf. on Coastal Engineering,
ASCE, 792-803.
Pilarczyk, W.K. and Zeidler, R.B. (1996). Offshore breakwaters and shore evolution
control, A.A. Balkema, Rotterdam, Netherlands, 560 p.
Powell, K.W. and Allsop, N.W.H. (1985). “Low-crested breakwaters, hydraulic

performance and stability.” Technical report SR 57, HR Wallingford.
Prandtl, L. (1952). Essential of Fluid Dynamics, 452 pp., Haffner, New York.
Ranasinghe, R., Turner, I.L. and Symonds, G. (2006). “Shoreline response to multi-
functional artificial surfing reefs: A numerical and physical modelling study.” Journal
Reniers, A.J.H.M. and Battjes, J.A. (1997). “A laboratory study of longshore currents over
barred and non-barred beaches.” Journal of Coastal Engineering, 30, 1-22.
Reniers, A.J.H.M., van Dongeren, A.R., Battjes, J.A. and Thornton, E.B. (2002). “Linear
modeling of infragravity waves during Delilah.” Journal of Geophysical Research, Vol.
107, C10, 1-18.
Reniers, A.J.H.M., Roelvink, J.A. and Thornton, E.B. (2004). “Morphodynamic modeling
of an embayed beach under wave group forcing.” Journal of Geophysical Research,
Vol. 109, C01030, 1-22.
Rivero, F.J., Sánchez-Arcilla, A. Gironella, X. and Corrons, A. (1997). “Large-scale

hydrodynamic experiments on submerged breakwaters.” Proc. Coastal Dynamics’97,
187
ASCE, 754-763.
Roelvink, J.A. (1993). “Dissipation in random wave groups incident on a beach.” Journal
Rojanakamthorn, S., Isobe, M. and Watanabe, A. (1990). “Modeling of wave

transformation on submerged breakwater.” Proc. 22th Int. Conf. on Coastal
Roul, P. and Faedo, A. (2002). “Physical model study on low-crested structures under
breaking wave conditions.” Proc. Int. MEDCOAST Workshop on Beaches of the
Mediterranean & the Black Sea, Kusadasi, Turkey, 83-96.
Roul, P., Faedo, A. and Paris A. (2004). “Physical model study of water piling-up behind
low-crested structures.” Proc. 29th Int. Conf. on Coastal Engineering, ASCE, 4165-
4177.
Schlurmann, T., Bleck, M. and Oumeraci, H. (2002). “Wave transformation at artificial

reefs described by the Hilbert-Huang transformation (HHT).” Proc. 28th Int. Conf. on
Seabrook, S.R. (1997). “Investigation of the performance of submerged rubblemound

breakwaters.” MSc. Thesis, 199 pp., Queen’s University, Kingston, Ontario, Canada.
Seabrook, S.R. and Hall, K.R. (1997). “Effect of crest width and geometry on submerged
breakwater performance.” Proc. Canadian Coastal Conference, CCSEA, 58-72.
Seabrook, S.R. and Hall, K.R. (1998). “Wave transmission at submerged rubblemound
Seber, G.A.F, and Wild, C. J. (1989). Nonlinear Regression, Wiley. 768 p.
Seelig, W.N. (1979).”Effect of breakwaters on waves: laboratory tests of wave transmission

by overtopping.” Proc. Coastal Structures’79, ASCE, 941-961.
Seelig, W.N. (1980). “Two-dimensional tests of wave transmission and reflection

characteristics of laboratory breakwaters.” Technical report, CERC, Fort Belvoir,
Report No. 80-1.
Shen, Y.M, Ng, C.O. and Zheng, Y.H. (2004). “Simulation of wave propagation over a
submerged bar using the VOF method with a two-equation k-ε turbulence modeling.”
Smith, E.R. and Kraus, N.C. (1991). “Laboratory study of wave-breaking over bars and
artificial reefs.” Journal of the Waterway, Port, Coastal and Ocean Engineering, Vol.
117, No. 4, 307-325.
188
SonTek (1997). “Acoustic Doppler Velocimeter (ADV) principles of operation.” SonTek
Technical Notes, San Diego, USA.
Soper, A. (1940). “Beach erosion studies (a discussion).” Trans. Amer. Soc. Civil
Engineers. Vol. 105, 893-895.
Sorensen, R.M. and Beil, N.J. (1988). “Perched beach profile response to wave action.”
Proc. 21st Int. Conf. on Coastal Engineering, ASCE, 1482-1491.
Soulsby, R.L., Hamm, L., Klopman, G., Myrhaug, D., Simons, R.R. and Thomas, G.P.
(1993). “Wave-current interaction within and outside the bottom boundary layer.”
Stauble, D.K. and Tabar, J.R. (2003). “The use of submerged narrow-crested breakwaters
for shoreline erosion control.” Journal of Coastal Research, Vol. 19, No. 3, 684-722.
Stokes, G.G. (1847). “On the theory of oscillatory waves.” Trans. Cambridge Philos. Soc.,
8, 441-455.
Stokes, G.G. (1880). “Considerations relative to the greatest height of oscillatory waves
which can be propagated without change of form..” Mathematical and Physical papers,
1, 225-228.
Streeter, V.L., Wylie, E.B. and Bedford, K.W. (1998). Fluid Mechanics, 9th ed., McGrow
Hill, 740 p.
Stucky, A. and Bonnard, D. (1937). “Contribution to the experimental study of marine rock
fill dikes.” Bull. Technique de Suizze Romande, (also see summary in Tech. Memo. 1,
Beach Erosion Board, 1940).
Svendsen, I.A. (1984a). “Wave heights and set-up in a surf zone.” Journal of Coastal
Svendsen, I.A. (1984b). “Mass flux and undertow in a surf zone.” Journal of Coastal
Symonds, G. (1994). ”Theory and observation of currents and setup over a shallow reef.”
Proc. Coastal Dynamics’98, ASCE, 1-13.
Symonds, G., Black, K.P. and Young, I.R. (1995). “Wave-driven flow over shallow reefs.”
Journal of Geophysical Research, Vol. 100, No. C2, 2639-2648.
Tanaka, N. (1976). “Effects of submerged rubble-mound breakwater on wave attenuation

and shoreline stabilization.” Proc. 23th Japanese Coastal Engineering Conference,
ASCE, 152-157.
189
Thornton, E.B. and Guza, R.T. (1983). “Transformation of wave height distribution.”
Journal of Geophysical Research, Vol. 89, No. C10, 5925-5938.
Tomasicchio, U. (1996). “Submerged breakwaters for the defence of the shoreline at Osita
field experiences, comparison.” Proc. 25th Int. Conf. on Coastal Engineering, ASCE,
2404-2417.
Turner, I.L. (2006). “Discriminating Modes of Shoreline Response to Offshore-Detached

Structures.” Journal of the Waterway, Port, Coastal and Ocean Engineering, Vol. 132,
No. 3, 180-191.
Ursell, F. (1953). “The Long Wave Paradox in the Theory of Gravity Waves,” Proc.
Cambridge Philos. Soc., Vol 49, 685-694.
Ursel, F., Dean, R.G. and Yu, Y.S. (1960). “Forced small-amplitude water waves: a
comparison of theory and experiment.” Journal of Fluid mechanics, Vol. 7, Part 3, 33-
52.
van der Biezen, S.C., Van der Graaff, J. and de Later. (1996). “3D model tests of the
influence of submerged breakwaters on a beach profile exposed to regular waves.”
Proc. Conf. on Physics of Estuaries and Coastal Seas.
van der Biezen, S.C., Van der Graaff, J., Schaap, J. and Torrini, L. (1997). “Small scale
tests and numerical modeling of the hydrodynamic and morphological effects of
submerged breakwaters.” Proc. Combined Australian Coastal Engineering and Ports
Conf., 219-224.
van der Biezen, S.C., Roelvik, J.A., Graaff van der, J., Schaap, J. and Torrini, L. (1998).
“Wave transmission at low-crested structures.” Proc. 26th Int. Conf. on Coastal
van der Meer, J.W. (1988). “Rock slopes and gravel beaches under wave attack.” PhD.
Thesis, Delft University of Technology, Delft Hydraulics Report No. 396.
van der Meer, J.W. (1990). “Data on wave transmission due to overtopping.” Technical
report, Delft Hydraulic, Report No. H 989.
van der Meer, J.W. (1991). “Stability and transmission at low-crested structures.” Technical
report, Delft Hydraulic, Report No. H 453. 33p.
van der Meer, J.W. and Daemen, I.F.R. (1994). “Stability and wave transmission at low-
crested rubble-mound structures.” Journal of the Waterway, Port, Coastal and Ocean
Engineering, Vol. 120, No. 1, 1-19.
van der Meer, J.W., Briganti, R., Wang, B. and Zanuttigh, B. (2004). “Wave transmission
190
at low-crested structures, including oblique wave attack.” Proc. 29th Int. Conf. on
van der Meer, J.W., Briganti, R., Zanuttigh, B. and Wang, B. (2005). “Wave transmission
at low-crested structures, including oblique wave attack.” Journal of Coastal
Vidal, C., Lomónaco, P., Migoya, L., Archetti, R., Turchetti, M. and Sorci, M., Sassi, G.
(2001). “Laboratory experiments on flow around and inside LCS structures. Description
of tests and data base.” DELOS European Project, Technical Report.
WAMDI group. (1988). “The WAM model - a third generation ocean wave prediction
model.” Journal of Physical Oceanography, Vol. 18, 1775-1810.
Wamsley, T., Hanson, H., and Kraus, N. C. (2002). “Wave transmission at detached
breakwaters for shoreline response modeling.” ERDC/CHL CHETN-II-45, U.S. Army
Engineer Research and Development Center, Vicksburg, MS.
http://chl.wes.army.mil/library/publications/chetn/
Whitham, G.B. (1974). Linear and nonlinear waves, Wiley, New York, 636 p.
Wilmott, C.J. (1981). “On the validation of models.” Progress in Physical Geography, 2,
184-194.
Zanuttigh, B. and Lamberti, A. (2003). “3D hydrodynamic tests at Aalborg University,

DK.” In: Kramer et al. (Eds.), Internal report, DELOS deliverable D31.
Zanuttigh, B., Guerrero, M. and Lamberti, A. (2003). “3D experimental analysis and
numerical simulations of hydrodynamics around low crested structures”. Proc. IAHR
Conference, Thessaloniki, Greece, 369-376.
Zanuttigh, B. and Lamberti, A. (2006). “Experimental analysis and numerical simulations

of waves and current flows around low-crested rubble-mound structures”. Journal of
the Waterway, Port, Coastal and Ocean Engineering, Vol. 132, No. 1, 10-27.
191
APPENDIX A
LIST OF TEST CONDITIONS AND VELOCITY

MEASUREMENT POINTS IN 2D LABORATORY WAVE
FLUME MODEL
Table A.1 Test conditions in wave flume without breakwater.

Breakwater submergence wave wave wave
water depth
test No. crest width period length height
h (mm)
B (mm) hs (mm) T (sec) L (mm) Hi (mm)
T0-1 350 - - 0.71 792 47.2
T0-5 350 - - 1.00 1470 56.9
T0-6 350 - - 1.00 1470 95.1
T0-9 350 - - 1.49 2596 41.4
T0-10 350 - - 1.49 2596 81.7
T0-11 350 - - 1.49 2596 128.2
T0-13 350 - - 2.00 3631 42.7
T0-14-1 350 - - 2.00 3631 93.4
T0-14-2 350 - - 2.00 3631 121.3
T0-15-1 350 - - 2.00 3631 121.4
T0-15-2 350 - - 2.00 3631 173.2
T0-17 400 - - 0.71 785 47.4
T0-21 400 - - 1.00 1500 62.5
T0-22 400 - - 1.00 1500 96.8
T0-25 400 - - 1.49 2729 42.5
T0-26 400 - - 1.49 2729 87.3
T0-27 400 - - 1.49 2729 113.6
T0-29 400 - - 2.00 3859 24.1
T0-30 400 - - 2.00 3859 98.2
T0-31 400 - - 2.00 3859 153.9
T0-33 450 - - 0.71 786 45.0
T0-37 450 - - 1.00 1520 43.1
T0-38 450 - - 1.00 1520 97.6
T0-41 450 - - 1.49 2842 52.4
T0-42 450 - - 1.49 2842 117.5
T0-43 450 - - 1.49 2842 147.6
T0-45 450 - - 2.00 4066 42.6
T0-46 450 - - 2.00 4066 101.8
T0-47 450 - - 2.00 4066 155.1
A.1
Table A.2 Test conditions in wave flume with 3500mm crest with breakwater.
Breakwater submergence wave wave wave Velocity
water depth
test No. crest width period length height sampling
h (mm)
B (mm) hs (mm) T (sec) L (mm) Hi (mm) location
T1-1 350 3500 50 0.71 792 48.8 T3,T7
T1-5 350 3500 50 1.00 1470 50.0 T7
T1-9 350 3500 50 1.49 2596 50.3 T7
T1-10 350 3500 50 1.49 2596 100.2 T7
T1-13 350 3500 50 2.00 3631 55.5 T3,T7
T1-14 350 3500 50 2.00 3631 106.3 T7
T1-15 350 3500 50 2.00 3631 146.7 T7
T1-17 400 3500 100 0.71 785 50.4 T5
T1-21 400 3500 100 1.00 1500 55.8 T5
T1-25 400 3500 100 1.49 2729 52.6 T5
T1-26 400 3500 100 1.49 2729 102.4 T3,T5
T1-27 400 3500 100 1.49 2729 145.3 T3,T5
T1-29 400 3500 100 2.00 3859 42.5 T3,T5
T1-30 400 3500 100 2.00 3859 102.5 T3,T5
T1-33 450 3500 150 0.71 786 41.7 T4
T1-37 450 3500 150 1.00 1520 54.9 T3,T5
T1-38 450 3500 150 1.00 1520 96.7 T3,T5
T1-41 450 3500 150 1.49 2842 56.7 T3,T5
T1-42 450 3500 150 1.49 2842 99.7 T3,T5
T1-43 450 3500 150 1.49 2842 155.0 T3,T5
T1-45 450 3500 150 2.00 4066 50.5 T3
T1-46 450 3500 150 2.00 4066 95.4 T5
A.2
Breakwater submergence wave wave wave Velocity
test No. water depth crest width period length height sampling
h (mm) B (mm) hs (mm) T (sec) L (mm) Hi (mm) location
T2-1 350 1500 50 0.71 792 44.7 T6
T2-5 350 1500 50 1.00 1470 62.2 T6
T2-6 350 1500 50 1.00 1470 103.8 T7
T2-9 350 1500 50 1.49 2596 43.9 T4
T2-10 350 1500 50 1.49 2596 99.3 T7
T2-11 350 1500 50 1.49 2596 159.9 T7
T2-13 350 1500 50 2.00 3631 62.4 T4
T2-14 350 1500 50 2.00 3631 114.7 T5
T2-15 350 1500 50 2.00 3631 145.6 T5
T2-17 400 1500 100 0.71 785 57.9 T3
T2-21 400 1500 100 1.00 1500 56.9 T3
T2-22 400 1500 100 1.00 1500 113.3 T3
T2-25 400 1500 100 1.49 2729 49.5 T3
T2-26 400 1500 100 1.49 2729 107.2 T3
T2-27 400 1500 100 1.49 2729 135.8 T3
T2-29 400 1500 100 2.00 3859 40.3 T3
T2-30 400 1500 100 2.00 3859 105.9 T3
T2-31 400 1500 100 2.00 3859 136.7 T3
T2-33 450 1500 150 0.71 786 56.2 T3
T2-37 450 1500 150 1.00 1520 43.7 T3
T2-38 450 1500 150 1.00 1520 86.3 T3
T2-41 450 1500 150 1.49 2842 43.3 T3
T2-42 450 1500 150 1.49 2842 120.7 T3
T2-43 450 1500 150 1.49 2842 152.3 T2,T3
T2-45 450 1500 150 2.00 4066 57.2 T3
T2-46 450 1500 150 2.00 4066 102.0 T3
T2-47 450 1500 150 2.00 4066 155.8 T3,T3
A.3
water Breakwater submergence wave wave wave Velocity
test No. depth crest width period length height sampling
h (mm) B (mm) hs (mm) T (sec) L (mm) Hi (mm) location
T3-1 350 300 50 0.71 792 52.3 T6
T3-5 350 300 50 1.00 1470 69.4 T6
T3-6 350 300 50 1.00 1470 110.4 T6
T3-9 350 300 50 1.49 2596 42.8 T7
T3-10 350 300 50 1.49 2596 96.1 T7
T3-11 350 300 50 1.49 2596 144.7 T7
T3-13 350 300 50 2.00 3631 46.7 T6
T3-14 350 300 50 2.00 3631 83.2 T6
T3-15 350 300 50 2.00 3631 98.5 T6
T3-17 400 300 100 0.71 785 51.6 T2
T3-21 400 300 100 1.00 1500 58.4 T2,T3
T3-22 400 300 100 1.00 1500 100.8 T2,T6
T3-25 400 300 100 1.49 2729 39.9 T2
T3-26 400 300 100 1.49 2729 103.9 T2,T6
T3-27 400 300 100 1.49 2729 132.4 T6
T3-29 400 300 100 2.00 3859 45.5 T2
T3-30 400 300 100 2.00 3859 117.5 T2,T6
T3-31 400 300 100 2.00 3859 170.5 T6
T3-33 450 300 150 0.71 786 43.9 T2
T3-37 450 300 150 1.00 1520 37.2 T2
T3-38 450 300 150 1.00 1520 89.5 T2
T3-41 450 300 150 1.49 2842 71.9 T2
T3-42 450 300 150 1.49 2842 114.9 T2
T3-43 450 300 150 1.49 2842 133.1 T2
T3-45 450 300 150 2.00 4066 42.8 T2
T3-46 450 300 150 2.00 4066 91.5 T2
T3-47 450 300 150 2.00 4066 147.1 T2
A.4
APPENDIX B
SAMPLES OF MATLAB PROGRAMS DEVELOPED IN

THE THESIS
B-1 inciwav.m
This spreadsheet reads the measured water surface level of the traverse probe (P2) from
the relevant EXCEL file and calculates incident and reflected wave height using the
conventional method of Healy (1953). The results of filtered data can also be compared
with non-filtered data. Both zero down-crossing and up-crossing methods are applied in
the calculation of wave period and wave height. See Chapter 4 for more details.
% Analysis of traverse probe 22 for calculating incident wave height,

% reflection. analyasis is consistent of two parts: 1- nonfiltered
data
% and 2- filtered data.
clear all;
srate=input('sample rate(Hz)=')
close all;
%xf=1.4;
%yf=1.8;
xf=input('pass-band factor=')
yf=input('stop-band factor=')
%N=140;
N=input('filter order=')
n=0;
data1=0;
data1m=0;
data1=xlsread('D:\usr\mojtaba\Test\3m Wave
flume\data\Wave\results\ct0-6.xls','spectrum','c3:c2050');
data1m=data1-mean(data1)-20;
n=length(data1);
time=[(1/srate):(1/srate):(n/srate)];
c=0;
w=0;
for p=1:n-1;
if (data1m(p)>=0)...
& (data1m(p+1)<=0);
c=c+1;
w(c)=p;
ttd(p)=p/srate;
end
end
B.1
d=0;
v=0;
for q=1:n-1 ;
if (data1m(q)<0)...
& (data1m(q+1)>=0);
d=d+1;
v(d)=q;
ttu(q)=q/srate;
end
end
% Determine the crest, trough, height, and period of each wave
k=0;
wvd=0;
crestd=0;
troughd=0;
htd=0;
for k=1:c-1;
wvd=data1m((w(k)+1):w(k+1));
crestd(k)=max(wvd);
troughd(k)=min(wvd);
htd(k)=crestd(k)-troughd(k);
end
l=0;
wvu=0;
crestu=0;
troughu=0;
htu=0;
for l=1:d-1;
wvu=data1m((v(l)+1):v(l+1));
crestu(l)=max(wvu);
troughu(l)=min(wvu);
htu(l)=crestu(l)-troughu(l);
end
close all;
A1=figure;
plot(1:c-1,htd(1:c-1),'r-.',1:d-1,htu(1:d-1),'b-');grid;;
axis([0,c,0,200])
legend('wave height (zero down-crossing)','wave height (zero up-
crossing)',2)
xlabel('Number of zero crossing')
ylabel('wave height (mm)')
title('comparison of zero up & down crossing method (Test0-6)')
htdav=sgolayfilt(htd,0,5); % smoothing the signals
htuav=sgolayfilt(htu,0,5); % smoothing the signals
A2=figure;
subplot(2,1,1), plot(1:c-1,htdav(1:c-1),'r-.',1:c-1,htd(1:c-1),'b-');
axis([0,c,0,200])
legend('smoothed ','non-smoothed',2)
title('zero down-crossing method (Test0-6)')
subplot(2,1,2), plot(1:d-1,htuav(1:d-1),'r-.',1:d-1,htu(1:d-1),'b-');
axis([0,c,0,200])
legend('smoothed','non-smoothed',2)
title('zero up-crossing method (Test0-6)')
Hmaxup=max(htuav(1:l));
Hminup=min(htuav(1:l));
Hmaxd=max(htdav(1:k));
B.2
Hmind=min(htdav(1:k));
Hmax=(Hmaxup+Hmaxd)/2;
Hmin=(Hminup+Hmind)/2;
Hi=(Hmax+Hmin)/2;
Kr=(Hmax-Hmin)/(Hmax+Hmin);
%figure
% plot(1:c-1,htdav(1:c-1),'b',1:d-1,htuav(1:d-1),'r');hold on
%axis([0,c,0,200])
%legend('down crossing moving average wave height','up crossing moving
average wave height',2)
%xlabel('time (sec)')
%ylabel('wave height (mm)')
%title('results of analysis');hold off
% *********************** Filtering **************************
A3=figure;
B=0;
B=fft(data1);
spec=plot(abs(B));
title('spectrum (Test0-6)')
figure
plot(unwrap(angle(B)));grid;
title('unwrap (Test0-6)')
Y=0;
I=0;
[y,I]=max(abs(B(2:size(data1,1)))); %This line is used to find the
fundamental frequency component
%which has the maximum
magnitude after the DC component
%To calculate the optimum pass band and stop band

frequencies (omega_p and omega_s)
%to suppress the harmonics.The passband extends
from 0 to omega_p=2*pi*(X*I)/size(sample1,1)
%and the stopband is from
omega_s=2*pi*(Y*I)/size(sample1,1)to pi.
omega_p=2*pi*(xf*I)/size(data1,1);
omega_s=2*pi*(yf*I)/size(data1,1);
% Filter program
delta_p=0.1;
delta_s=0.01;
pi=3.1416;
F=0;
A=0;
W=0;
F=[0 omega_p/pi omega_s/pi 1];
A=[1 1 0 0];
W=[1/delta_p 1/delta_s];
h_equiripple=remez(N,F,A,W);
wp=0;
np=0;
wp=(0:0.001*pi:pi);
np=0:N;
j=sqrt(-1);
mag=abs(h_equiripple*exp(-j*np'*wp));
A4=figure; % plot of the
desired and designed filters
semilogy(wp,max(mag,0.001),'r');hold on;
semilogy([0 omega_p],[1+delta_p 1+delta_p],'b');
semilogy([0 omega_p],[1-delta_p 1-delta_p],'b');
B.3
semilogy([omega_s pi],[delta_s delta_s],'b');hold off
A5=figure;
fdata1=0;
fdata1m=0;
% To compare the initial samples and filtered ones
fdata1=filter(h_equiripple,1,data1);
fdata1m=fdata1-mean(fdata1);
plot(1:n,data1m(1:n),'b',1:n,fdata1m(1:n),'r');grid;
legend('non-filtered data','filtered data',2)
xlabel('Number of sample')
ylabel('water level (mm)')
title('comparison of filtered and non-filtered data (Test0-6)')
% Analyzing the filtered data1
n=length(fdata1)
r=0;
w=0;
for p=1:n-1;
if (fdata1m(p)>=0)...
& (fdata1m(p+1)<=0);
r=r+1;
w(r)=p;
end
end
s=0;
v=0;
for q=1:n-1 ;
if (fdata1m(q)<0)...
& (fdata1m(q+1)>=0);
s=s+1;
v(s)=q;
end
end
% Determine the crest, trough and height of wave
k=0;
fwvd=0;
fcrestd=0;
ftroughd=0;
fhtd=0;
for k=1:r-1;
fwvd=fdata1m((w(k)+1):w(k+1));
fcrestd(k)=max(fwvd);
ftroughd(k)=min(fwvd);
fhtd(k)=fcrestd(k)-ftroughd(k);
end
l=10;
fwvu=0;
fcrestu=0;
ftroughu=0;
fhtu=0;
for l=1:s-1;
fwvu=fdata1m((v(l)+1):v(l+1));
fcrestu(l)=max(fwvu);
ftroughu(l)=min(fwvu);
fhtu(l)=fcrestu(l)-ftroughu(l);
end
A6=figure;
plot(10:r-1,fhtd(10:r-1),'r-.',10:s-1,fhtu(10:s-1),'b-')
axis([0,c,0,200])
legend('zero down-crossing','zero up-crossing',2)
B.4
title('comparison of up & down crossing method for filtered data
(Test0-6)')
fhtdav=sgolayfilt(fhtd,0,5);
fhtuav=sgolayfilt(fhtu,0,5);
A7=figure;
subplot(2,1,1), plot(10:r-1,fhtdav(10:r-1),'r-.',10:r-1,fhtd(10:r-
1),'b-')
axis([0,c,0,200])
legend('zero down-crossing smoothed','zero down-crossing non-
smoothed',2)
title('zero down-crossing method for filtered data (Test0-6)')
subplot(2,1,2), plot(10:s-1,fhtuav(10:s-1),'r-.',10:s-1,fhtu(10:s-
1),'b-')
axis([0,c,0,200])
legend('zero up-crossing smoothed','zero up-crossing non-smoothed',2)
title('zero up-crossing method for filtered data (Test0-6)')
fHmaxup=max(fhtuav(10:l));
fHminup=min(fhtuav(10:l));
fHmaxd=max(fhtdav(10:k));
fHmind=min(fhtdav(10:k));
fHmax=(fHmaxup+fHmaxd)/2;
fHmin=(fHminup+fHmind)/2;
fHi=(fHmax+fHmin)/2;
fKr=(fHmax-fHmin)/(fHmax+fHmin);
fHiup=(fHmaxup+fHminup)/2;
fKrup=(fHmaxup-fHminup)/(fHmaxup+fHminup);
fHid=(fHmaxd+fHmind)/2;
fKrd=(fHmaxd-fHmind)/(fHmaxd+fHmind);
results={'time' 'data1m' 'fdata1m' 'Hi' 'Kr' 'fHi' 'fKr' 'pass-band'
'stop-band' 'filter order' ...
;time' data1m fdata1m Hi Kr fHi fKr xf yf N }
xlswrite('D:\usr\mojtaba\Test\3m Wave flume\data\Wave\results\ct0-6'
,results','reflection analysis','A2')
xlswrite('D:\usr\mojtaba\Test\3m Wave flume\data\Wave\results\ct0-
6',[time' data1m fdata1m],'reflection analysis','A2')
6',data1m,'reflection analysis','B2')
6',fdata1m,'reflection analysis','C2')
6',[Hi Kr fHi fKr xf yf N],'reflection analysis','D2')
6',Kr,'reflection analysis','E2')
6',fHi,'reflection analysis','F2')
6',fKr,'reflection analysis','G2')
6',xf,'reflection analysis','H2')
6',yf,'reflection analysis','I2')
6',N,'reflection analysis','J2')
A8=figure;
B.5
subplot(2,1,1); plot(1:c-1,htdav(1:c-1),'r-.',1:d-1,htuav(1:d-1),'b-
');hold on
axis([0,c,0,200])
legend('zero down-crossing smoothed','zero up-crossing smoothed',2)
title('Wave height analysis for non-filtered data (Test0-6)');hold off
subplot(2,1,2); plot(10:r-1,fhtdav(10:r-1),'r-.',10:s-1,fhtuav(10:s-
1),'b-')
axis([0,r,0,200])
legend('zero down-crossing smoothed','zero up-crossing smoothed',2)
title('Wave height analysis for filtered data (Test0-6)');
saveas(A1,'D:\usr\mojtaba\Test\3m Wave flume\MATLAB\work\Test\incident
wave height analysis\Test0-6-1.fig');
B-2 spectral.m
This spreadsheet calculates incident and reflected wave height using the methods of
Goda and Suzuki (1976) and Mansard and Funke (1980). See Chapter 4 for more
details.
% calculate transmission coefficient from energy transmission over

submerged breakwater.(averaging the energy measured from probes in
front and behind
% the structure)
% **************************** Test 3
*********************************
% **************************** Goda and Suzuki (1976) , Mansard and
Funke (1980)
close all;
clear all;
srate=input('(Test3-46);sample rate(Hz)=')
Tpeak=input('(Test3-46)designed period (sec) T=')
d=input('(Test3-46)water depth in the flume (mm) d=')
x12=input('(Test3-46)distance of probe 1&2 (mm) =')
x13=input('(Test3-46)distance of probe 1&3 (mm)=')
deltaL=x12;
B.6
data=xlsread('D:\usr\mojtaba\Test\3m Wave flume\data\Wave\results\ct3-
46.xls','spectrum','B3:K2050');
A=xlsread('D:\usr\mojtaba\Test\3m Wave flume\data\Wave\results\ct3-
46.xls','reflection analysis','D2:E2');
Hi=A(1);
KR=A(2);
N=length(data(:,1));
FFTdata=fft(data);
Sf=2*(abs(FFTdata)).^2; % wave spectrum
a=2*abs(FFTdata)./N; % harmonic amplitude
f=srate*(1:N)/N;
E1=trapz(Sf);
E=E1./(N/srate);
H=2*sqrt(2)*std(data);
% --------- computing incident and reflected wave height (Goda and
Suzuki, 1976) ------------------------
A1=((real(FFTdata(1:N,3)))')./(N);
B1=((-imag(FFTdata(1:N,3)))')./(N);
A2=((real(FFTdata(1:N,4)))')./(N);
B2=((-imag(FFTdata(1:N,4)))')./(N);
T=f.^(-1);
L=(9806*T.^2/(2*pi)).*(tanh(4*pi^2*d./(T.^2*9806))).^(0.5);
k=2*pi*L.^(-1);
aI=(0.5*(abs(sin(k*deltaL)).^(-1))).*((A2-A1.*cos(k*deltaL)-
B1.*sin(k*deltaL)).^2+(B2+A1.*sin(k*deltaL)-
B1.*cos(k*deltaL)).^2).^0.5;
aR=(0.5*(abs(sin(k*deltaL)).^(-1))).*((A2-
A1.*cos(k*deltaL)+B1.*sin(k*deltaL)).^2+(B2-A1.*sin(k*deltaL)-
B1.*cos(k*deltaL)).^2).^0.5;
SI=2*(N/2*aI).^2;
SR=2*(N/2*aR).^2;
n=N/(Tpeak*srate);
EI=trapz(SI(round(n-20):round(n+20)));
ER=trapz(SR(round(n-20):round(n+20)));
aIGoda=max(aI(round(n-20):round(n+20)))
aRGoda=max(aR(round(n-20):round(n+20)))
KRGoda=sqrt(ER/EI)
figure
plot(f,Sf(:,3),':o',f,SI,':*',f,SR,':d')
legend('Sf','SI','SR')
xmin=round(n-5);
xmax=round(n+5);
axis([f(xmin) f(xmax) 1 max(Sf(xmin:xmax,3))])
% ------------------------------------------------------------------
----------------------------
% ********** Mansard & Funke (1980) for wave reflection ***********
beta=2*pi*x12./L; %probe3&4&5
gama=2*pi*x13./L;
D=2*((sin(beta)).^2+(sin(gama)).^2+(sin(gama-beta)).^2);
R1=(sin(beta)).^2+(sin(gama)).^2;
Q1=sin(beta).*cos(beta)+sin(gama).*cos(gama);
R2=sin(gama).*sin(gama-beta);
Q2=sin(gama).*cos(gama-beta)-2*sin(beta);
R3=-sin(beta).*sin(gama-beta);
Q3=sin(beta).*cos(gama-beta)-2*sin(gama);
ZI=((FFTdata(:,3))'.*(complex(R1,Q1))+(FFTdata(:,4))'.*(complex(R2,Q2)
)+(FFTdata(:,5))'.*(complex(R3,Q3)))./D;
B.7
ZR=((FFTdata(:,3))'.*(complex(R1,-
Q1))+(FFTdata(:,4))'.*(complex(R2,-Q2))+(FFTdata(:,5))'.*(complex(R3,-
Q3)))./D;
SZI=2*(abs(ZI)).^2;
SZR=2*(abs(ZR)).^2;
EI2=trapz((abs(SZI(round(n-20):round(n+20)))));
ER2=trapz((abs(SZR(round(n-20):round(n+20)))));
aIMansard=2*max(abs(SZI(round(n-20):round(n+20)))/N);
aRMansard=2*max(abs(SZR(round(n-20):round(n+20)))/N);
KRMansard=sqrt(ER2/EI2)
%---------------------------------------------------------------------
---
figure
plot(f,Sf(:,3),'--o',f,SI,':*',f,SR,':d',f,SZI,'-.*',f,SZR,'-.d')
legend('Sf','SI','SR','SZI','SZR')
xmin=round(n-5);
xmax=round(n+5);
axis([f(xmin) f(xmax) 1 max(Sf(xmin:xmax,3))])
result=[Hi/2 KR*Hi KR;aIGoda aRGoda KRGoda;aIMansard aRMansard
KRMansard]
Kt1=sqrt(mean(E(:,8:10))/mean(E(:,3:4)));
Kt2=mean(H(:,8:10))/Hi;
KtGoda=(sqrt(EI/mean(E(:,8:10))))^-1;
[var(data);E./1000000];
figure
plotyy(1:10,E,1:10,H.^2)
result2=[Kt1 Kt2 KtGoda]
B-3 Symonds.m
This script calculates the best friction factor "r" for fitting measured data (velocity over
the breakwater) to Symonds theory. The definition of Symonds et al. (1995) model is
presented in Chapter 5.
% This script calculates the best "r" for fitting measured data
% to Symonds theory.with calculated gama by Hara et al. 1992'
clear Hi;
clear U;
clear d;
clear T;
clear H;
clear hb;
clear R2;
clear beta;
clear Fc;
clear x(:,:);
Hi=xlsread('D:\usr\mojtaba\Test\3m Wave flume\final
analysis\symonds.xls','test3','E3:E29')/1000;
B.8
U=xlsread('D:\usr\mojtaba\Test\3m Wave flume\final
analysis\symonds.xls','test3','G3:G29')/10000;
d=xlsread('D:\usr\mojtaba\Test\3m Wave flume\final
analysis\symonds.xls','test3','B3:B29')/1000;
T=xlsread('D:\usr\mojtaba\Test\3m Wave flume\final
analysis\symonds.xls','test3','C3:C29');
Xr=input('crest width(m) Xr=')
beta=input('initial value for r0=')
n=length(U);
H=(d-0.3);
%Fc=((9.81^1.25)*(Hi.^0.5).*(T.^2.5))./H.^1.75 %Nelson 1994
%GAMA=Fc./(22+1.82*Fc) %Nelson 1994
exi=(((Xr*d.^(-1))+(0.3*d.^(-1))/(3.5*0.5)).^0.2).*((0.3*d.^(-
1))./(Hi./d).^0.4); %Hara et al. 1992
GAMA=-0.463*exi.^0.133+1.039; %Hara et al. 1992
hb=Hi./GAMA;
R2=Xr./(Xr+2*(hb-H));
% x1=H , x2=R2
x(:,:)=0;
x(1:n,1)=H;
x(1:n,2)=R2;
x(1:n,3)=Hi;
x(1:n,4)=GAMA;
betafit=nlinfit(x,U,@formula,beta)
current=(1.5*9.81*0.5*(GAMA.^2).*(H.*((1-((GAMA.*H)./Hi)).*(1-
R2)))./(betafit*((1-R2).*((GAMA.*H)./Hi)+(1-
((GAMA.*H)./Hi)).*R2))).*10000;
x(1:n,5)=current;
Formula:
function velocity=formula(beta,x)
b1=beta;
% b2=r
d=xlsread('D:\usr\mojtaba\Test\3m Wave flume\final
analysis\symonds.xls','test3','B3:B29')/1000;
n=length(d);
x1=x(1:n,1);
x2=x(1:n,2);
x3=x(1:n,3);
x4=x(1:n,4);
velocity=(1.5*9.81*0.5*(x4.^2).*(x1.*((1-((x4.*x1)./x3)).*(1-
x2)))./(b1*((1-x2).*((x4.*x1)./x3)+(1-((x4.*x1)./x3)).*x2)));
B-4 Gourlay.m
This script calculates the dimensionless discharge over the breakwater using the
proposed analytical model of Gourlay and Colleter (2005). See Chapter 5 more details.
% ********* %This script calculates non-dimensionless discharge based

on ******
% ********* Gourlay and Colleter (2005) analytical method and
GOurlay (1996a) data series. *****
% Formulation is presented in Chapter 5.
% The script reads all wave specification from "D:\usr\mojtaba\Test\3m
Wave
B.9
% %flume\final analysis\Gourlay\summary (breaking waves)" and asks for
the
% following parameter:
% gamar= wave deformation factor
% KR=wave reflection coefficient
% f=reef-top friction factor
% Kp=reef profile shape factor
% CD=weir discharge coefficient
% The script calculate dimensionless discharge (qstar) using the model
for
% reef-top control and reef-rim control by submergence ratio "S"
% determinator.
clear All
clear f1
clear KR1
clear KP1
clear Ls
clear CD1
clear gamar1
clear Hi
clear d
clear T
clear hs
clear S
clear qstarmeas
clear qstarcal
clear l
clear F
clear D
clear L
clear A1
clear A2
clear A4
clear p
clear R2
clear resid
gamar=input('deformation factor gamar=')
KP=input('reef shape factor=')
KR=input('wave reflection coefficient KR=')
f=input('reef-top friction factor f=')
Ls=input('reef-top crest width(mm) Ls=')
CD=input('weir discharge coefficient CD=')
Hi=xlsread('D:\usr\mojtaba\Test\3m Wave flume\final
analysis\Gourlay\Gourlay.xls','data','E2:E25');
T=xlsread('D:\usr\mojtaba\Test\3m Wave flume\final
analysis\Gourlay\Gourlay.xls','data','D2:D25');
hr=xlsread('D:\usr\mojtaba\Test\3m Wave flume\final
analysis\Gourlay\Gourlay.xls','data','K2:K25');
S=xlsread('D:\usr\mojtaba\Test\3m Wave flume\final
analysis\Gourlay\Gourlay.xls','data','L2:L25');
qstarmeas=xlsread('D:\usr\mojtaba\Test\3m Wave flume\final
analysis\Gourlay\Gourlay.xls','data','M2:M25');
%KR=xlsread('D:\usr\mojtaba\Test\3m Wave flume\final
analysis\Gourlay\reflection.xls','Test1','F2:F25')
l=length(Hi);
for n=1:l
d(n)=320;
F(n)=f*Ls/(8*Hi(n));
KP1(n)=KP;
KR1(n)=KR; %calculated KR
B.10
f1(n)=f;
CD1(n)=CD;
gamar1(n)=gamar;
D(n)=9810^0.5*T(n)/hr(n)^0.5;
% D(n)=12.5;
L(n)=(9806*T(n)^2/(2*pi))*sqrt(tanh(4*pi^2*d(n)/(T(n)^2*9806)));
if S(n)>0.7
B=(3/(64*pi))*(KP/F(n))*(9810*T(n)^2/Hi(n))^0.5;
A1(n)=(3*KP/(2*B*S(n)^3))+(16*pi/S(n)^2)*(Hi(n)/(9810*T(n)^2))^0.5;
A2(n)=1/(B*S(n)^1.5);
A4(n)=KR1(n)^2-1+4*pi*gamar^2*S(n)^2/D(n);
p=[A1(n) A2(n) 0 A4(n)];
q=roots(p);
for m=1:3
if q(m)>0
qstarcal(n)=(q(m)./(1+1.3*(Ls/L(n))^-1.3))'; % dividing
calculated discharge by
% qstarcal(n)=(q(m))';
end
end
else
C=0.5443*CD;
% x=(C*S(n)^2/(S(n)+C^2*F(n))^0.5)/(1.9*(Ls/L(n))^-0.4);
x=(C*S(n)^2/(S(n)+C^2*F(n))^0.5);
qstarcal(n)=x;
end
end
resid=qstarcal-qstarmeas';
SSE=sum(resid.^2);
SSyy=sum((qstarmeas'-mean(qstarmeas')).^2);
%R2(1:l)=1-SSE/SSyy;
R2(1:l)=(sum((qstarmeas'-mean(qstarmeas')).*(qstarcal-
mean(qstarcal))))^2/(sum((qstarmeas'-mean(qstarmeas')).^2)...
*sum((qstarcal-mean(qstarcal)).^2));
correlation2=(corr(qstarcal',qstarmeas))^2
Test1results=[qstarcal' qstarmeas R2' KP1' KR1' f1' CD1' gamar1'
S qstarcal'./qstarmeas (Ls./L)' ]
B-5 Roelvink.m
The program calculate wave energy dissipation and wave transformation (transmission
coefficient) over submerged breakwater for a given breaker parameter “γ ” using the
model of Roelvink (1993).
clear all
clear Hrms
clear h
clear l
clear C
Hrms(1,1)=input('Incident wave height (mm)=')
%Hrms(1,1)=53;
T=input('Period (sec)=')
%T=1.49;
h(1,1)=input('offshore water depth (mm)=')
B.11
%h(1,1)=400;
%hs(1,1)=input('submergence (mm)=')
hs(1,1)=h-300;
%Alfa=input('Alfa=')
Alfa=1;
%N=input('N=')
N=10;
Gamma=input('Gama=')
%Deltax=input('Deltax (mm)=')
Deltax=100;
%Nmesh=input('Number of meshes=')
Nmesh=47;
n=0;
close all;
for n=1:Nmesh;
if n<7;
h(n,1)=h(1,1)-(n-1)*Deltax*0.5;
h(n+1,1)=h(1,1)-(n)*Deltax*0.5;
elseif n>=7 & n<=(Nmesh-7);
h(n,1)=hs(1,1);
h(n+1,1)=hs(1,1);
else n>Nmesh-7;
h(n+1,1)=h(1,1)-(Nmesh-n)*Deltax*0.5;
end
L(n,1)=(9806*T^2/(2*3.14))*(tanh(4*3.14^2*h(n,1)/(9806*T^2)))^0.5;
C(n,1)=(9806*L(n,1)/(2*3.14)*tanh(2*3.14*h(n,1)./L(n,1)))^0.5;
L(n+1,1)=(9806*T^2/(2*3.14))*(tanh(4*3.14^2*h(n+1,1)./(9806*T^2)))^0.5
;
C(n+1,1)=((9806*L(n+1,1)./(2*3.14)).*tanh(2*3.14*h(n+1,1)./L(n+1,1)))^
0.5;
Gama(n+1,1)=0.76*(h(n+1,1).*2*3.14/L(n+1,1))+0.29; % Ruessink
et al. (2003)
%Hrms(n+1,1)=(((Hrms(n,1).^2).*C(n,1)-Deltax*(1-exp(-
((Hrms(n,1))./(Gama(n+1,1)*h(n+1,1))).^N))...
% .*(2*Alfa*(Hrms(n,1)).^2))./C(n+1,1)).^0.5;
Hrms(n+1,1)=(((Hrms(n,1).^2).*C(n,1)-Deltax*(1-exp(-
((Hrms(n,1))./(Gamma*h(n+1,1))).^N))...
.*(2*Alfa*(Hrms(n,1)).^2))./C(n+1,1)).^0.5;
end
Hrms
Kt=Hrms(48,1)/Hrms(1,1)
B.12
APPENDIX C
SAMPLE OF ATTRIBUTE FILES USED IN THE

NUMERICAL MODEL
spectrum.bnd
The generated monochromatic wave was simulated as a narrow band spectrum at the
open boundary in SWAN. the amplitude of the spectrum was defined so that the total
energy of wave spectrum is the same as monochromatic wave energy. A sample of the
spectrum defined at open boundary is as follows:
SWAN 1 Swan standard spectral file, version

$ Data produced by SWAN version 40.41A
$ Project: Test1 ; run number: 26
LOCATIONS locations in x-y-space
1 number of locations
1.0000 1.0000
AFREQ absolute frequencies in Hz
25 number of frequencies
0.0400
0.0800
0.1200
0.1600
0.2000
0.2400
0.2800
0.3162
0.3200
0.3600
0.4000
0.4400
0.4800
0.5200
0.5600
0.6000
0.6400
0.6800
0.7200
0.7600
0.8000
0.8400
0.8800
0.9200
C.1
1.0000
QUANT
3 number of quantities in table
EnDens energy densities in J/m2/Hz
J/m2/Hz unit
-0.9900E+02 exception value
NDIR average nautical direction in degr
degr unit
DSPRDEGR directional spreading
degr unit
LOCATION 1
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.9545E+04 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
0.0 270.0 0.0
C.2
t.rgh
The spatially varying bottom roughness along the flume and over the breakwater is
defined for FLOW module by editing *.rgh file. The values of Manning coefficients
were identified at each grid point along the flume. A sample of defined manning
coefficient in t.rgh file along the flume (one grid line) is presented as follows:
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761
0.01761 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761
0.01761 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761
0.01761 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761
0.01761 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761 0.01761
0.01761 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
0.02055 0.02055 0.02055 0.02055 0.02055 0.02055 0.02055
The value of Manning coefficient is varying as a function of water depth over the
submerged breakwater.
C.3

Wave Script Example

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Wave Script Example

Uploaded by

Copyright:

Available Formats

EXPERIMENTAL AND NUMERICAL

MODELLING OF WAVE-INDUCED CURRENT

A thesis submitted in fulfillment of the

School of Civil and Environmental Engineering

Two dimensional experimental and numerical modeling of wave transmission and

Two preliminary 3D and a comprehensive series of 2D laboratory experiments have been

Different approaches to experimental data processing are examined in producing

I would like to express my sincere thanks to my supervisor, Associate Professor Ron

I would like to extend my appreciation to staff members at Water Research Laboratory

Figure 1.1 Salient and Tombolo formation behind offshore breakwaters.

Figure 2.3. Influence of breaking conditions on wave transmission coefficient.

Figure 2.4. Proposed wave transmission coefficient criteria.

Figure 2.6. The effect of armour rock size on transmission coefficient.

Figure 2.7. Variation of transmission coefficient against breakwater crest width.

Figure 2.9. Prediction of transmission coefficient as a function of relative crest height.

Figure 2.11. The influence of crest width on set-up.

Figure 2.13. Piling-up levels as function of net transmitted discharge.

Figure 2.14. Dimensionless set-up and current as a function of submergence ratio.

Figure 2.15. Comparison of theoretical and experimental values of discharge ratio.

Figure 2.16. Empirical relationship for preliminary prediction of shoreline response to

Figure 2.18. Intercomparison of computed longshore current velocities.

Figure 3.1. 24.5x16.0x0.7m wave basin used for preliminary 3D tests.

Figure 3.3 Layout of 2D laboratory test in 1.5x3.0x35m wave flume.

Figure 3.4. 1m wide internal channel located in the wave flume.

Figure 3.6 Combination of gravel beach and wave absorber panel.

Figure 3.9. Rectangular breakwater used in 3D tests.

Figure 3.10. Trapezoidal breakwaters used in the wave flume tests.

Figure 3.15. Long continuous rectangular breakwater in 3D preliminary test "TP-1".

Figure 4.2. Spectrum of water level at the position of probe4.

Figure 4.6. Comparison of zero up-crossing and down-crossing methods for

Figure 4.9. Comparison of conventional method of Healy (1953) for calculation of

Figure 4.11. Comparison of conventional method for calculation of incident wave

Figure 4.12. Variation of inverse submergence ratio “Hi/hs” as a function of wave

Figure 4.15. Calculated breaker-depth index “γb” as a function of dimensionless

Figure 4.21. Current velocity profile definition over the breakwater.

Figure 4.22. Variation of velocity profile over the breakwater.

Figure 5.1 Variation of transmission coefficient against incident wave height.

Figure 5.2 Variation of transmission coefficient against wave period.

Figure 5.3 Variation of transmission coefficient against breakwater crest width.

Figure 5.4 Variation of transmission coefficient against submergence depth.

Figure 5.5 Variation of correlation coefficient between transmission coefficient "Kt"

Figure 5.7 Variation of transmission coefficient against different dimensionless

Figure 5.8 Comparison of measured transmission coefficient with previous

Figure 5.12 Statistics comparison of measured transmission coefficient with the

Figure 5.16 Variation of wave-induced discharge as functions of submergence depth

Figure 5.17 Variation of discharge ratio against: a) submergence ratio, b) inverse of

Figure 5.18 Envelop of maximum dimensionless discharge as a function of

Figure 5.24 Coral reef definitions.

Figure 5.27 variation of wave set-up as a function of M=3g0.5Hi2T/(64п (ή+hs)1.5).

Figure 5.29 Variation of overestimating of model of Gourlay and Colleter (2005) as a

Figure 5.37 Comparison of measured maximum set-up over submerged breakwater

Figure 7.1. Comparison of calculated wave transformation over submerged breakwater

Figure 7.13. Comparison of calculated transmission coefficient by SWAN model with

Figure 7.14. Comparison of measured transmission coefficients with calculated using

Figure 7.15. Variation and effectiveness of different dimensionless parameters on

Figure 7.25. Comparison of calculated and measured discharge over submerged

Figure 7.27. Calculated wave-induced forces by SWAN in x- direction around the

Figure 7.31. Computed discharge using radiation stress gradients (FLOW-WAVE-1)

Figure 7.33 Deviation between calculated and measured discharge as a function of

Figure 7.36. Comparison of measured transmission coefficient with the calculated Kt

Figure 7.37. Variation and effectiveness of different dimensionless parameters on

Figure 7.39. Comparison of measured transmission coefficients Kt with the calculated

Figure 7.40. Comparison of measured and calculated transmission coefficients Kt with

Figure 7.41. Variation of calculated short-wave and roller energy by FLOW-ROLLER