Coastal Dynamics

Coastal Dynamics
Version 1.2 (2023)
Judith Bosboom and Marcel J.F. Stive

Coastal Dynamics
This textbook on Coastal Dynamics focuses on the

interrelation between physical wave, flow and sediment
transport phenomena and the resulting morphodynamics of
a wide variety of coastal systems. The textbook is unique
in that it explicitly connects the dynamics of open coasts
and tidal basins; not only is the interaction between open
coasts and tidal basins of basic importance for the evolution
of most coastal systems, but describing the similarities
between their physical processes is highly instructive
as well. This textbook emphasizes these similarities to
the benefit of understanding shared processes such as
nonlinearities in flow and sediment transport. Some prior
knowledge with respect to the dynamics of flow, waves
and sediment transport is recommended.
Judith Bosboom
TU Delft | Faculty of Civil Engineering
and Geosciences
Dr. Judith Bosboom is a senior lecturer Coastal

Engineering. She developed innovative teaching
methods for the topic of Coastal Dynamics.
Judith has been elected best lecturer for the
MSc Hydraulic Engineering multiple times.
In 2016, she received both the best lecturer
award for the Faculty of Civil Engineering
and Geosciences and the best lecturer
2023 TU Delft Open
award for Delft University of Technology. She ISBN 978-94-6366-370-0
has successfully taught Nanjing University DOI 10.5074/T.2021.001
graduates (China) for several years. Version 1.2
textbooks.open.tudelft.nl
Marcel J.F. Stive Cover image:

TU Delft | Faculty of Civil Engineering Atlantic coast, Angola
and Geosciences (Courtesy Stefanie Ross)
Prof. Marcel Stive is an emeritus professor

Coastal Engineering. He assisted Judith in
developing and teaching the Coastal Dynamics
lectures based on his earlier involvements in
lecturing and teaching. He has successfully
taught Indian, Chinese, Vietnamese, Brazilian
and Iranian graduates and/or professionals for
several years.
Coastal Dynamics
Delft University of Technology
Coastal Dynamics
Judith Bosboom & Marcel J.F. Stive

iv
Bosboom, J. and Stive, M. J. F. (2023). Coastal Dynamics. Delft University of Technology,

Delft, The Netherlands.
Publisher TU Delft Open

Date 11th January 2023
Cover image Atlantic coast, Angola (courtesy Stefanie Ross)
ISBN (softback/paperback) 978-94-6366-370-0
ISBN (e-book) 978-94-6366-371-7
DOI 10.5074/T.2021.001
Version 1.2 (see ‘Errata and improvements’ on page 579)
Subversion (SVN) revision 1489 logged at 2023-01-11 14:53
Corresponding author j.bosboom@tudelft.nl
The latest version of this book is available for online use and free download from the
TU Delft Open Textbook repository at textbooks.open.tudelft.nl. Here, a form is also
available to subscribe to update notifications and provide feedback.
In line with TU Delft Open Science policies, this Open Textbook is licensed under a
Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International Li-
cense (CC BY-NC-SA 4.0), except where otherwise stated. This work can be re-
distributed in unmodified form, or in modified form with proper attribution and
under the same license as the original, for non-commercial uses only.
Every attempt has been made to ascertain the correct source of images and other
potentially copyrighted material and ensure that all materials included in this
book have been attributed and used according to their license. If you believe that
a portion of the material infringes someone else’s copyright, please contact the
author j.bosboom@tudelft.nl.
Copyright © 2010 - 2023 by Judith Bosboom and Marcel J.F. Stive except for some con-
tent and materials that are copyrighted by their respective owners.
The softback is printed by Holland Ridderkerk (HollandRidderkerk.nl) on behalf of the

Faculty of Civil Engineering and Geosciences, Department of Hydraulic Engineering,
Section of Coastal Engineering. It can be ordered at textbooks.open.tudelft.nl.
Graphics produced with [Matlab, Inkscape, Adobe Illustrator, Adobe Photoshop] ©
Typesetting with LATEX
Last change date: 2023-01-11

v
Contents
Preface xiii
1 Overview 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Coastal dynamics for coastal engineers . . . . . . . . . . . . . . . . . . 1
1.2.1 What is coastal engineering? . . . . . . . . . . . . . . . . . . . 1
1.2.2 Position of Coastal Dynamics in the TU Delft curriculum . . . . 3
1.3 Study goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Examples of engineering applications . . . . . . . . . . . . . . . . . . . 5
1.4.1 Overview of coastal area and problems . . . . . . . . . . . . . . 5
1.4.2 Cross-shore profile . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4.3 Morphological development in vicinity of a port . . . . . . . . . 10
1.4.4 Delta near a river mouth . . . . . . . . . . . . . . . . . . . . . . 12
1.4.5 Tidal inlets and basins . . . . . . . . . . . . . . . . . . . . . . . 12
1.4.6 Dune erosion and flooding during a severe storm surge . . . . . 14
1.4.7 Large artificial island in open sea . . . . . . . . . . . . . . . . . 17
1.4.8 Other examples . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.5 Coastal (morpho)dynamics . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.5.1 Definition of the coast . . . . . . . . . . . . . . . . . . . . . . . 19
1.5.2 Coastal morphodynamics . . . . . . . . . . . . . . . . . . . . . 21
1.5.3 Time- and spatial scales . . . . . . . . . . . . . . . . . . . . . . 23
1.5.4 Equilibrium concept . . . . . . . . . . . . . . . . . . . . . . . . 25
1.5.5 Classification of coastal systems . . . . . . . . . . . . . . . . . . 26
1.6 Important parties in the Netherlands . . . . . . . . . . . . . . . . . . . 27
1.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
1.7.1 Lecture notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
1.7.2 Textbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
1.7.3 Internet sources . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
1.7.4 Interesting journals . . . . . . . . . . . . . . . . . . . . . . . . . 31
1.7.5 Conference proceedings . . . . . . . . . . . . . . . . . . . . . . 32
2 Large-scale geographical variation of coasts 35

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.2 Cumulative evolution of coastal systems . . . . . . . . . . . . . . . . . 36
2.2.1 Geological timescale . . . . . . . . . . . . . . . . . . . . . . . . 36
2.2.2 Continental ‘drift’ . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.2.3 Pleistocene inheritance . . . . . . . . . . . . . . . . . . . . . . . 39

vi Contents
2.3 Tectonic control of coasts . . . . . . . . . . . . . . . . . . . . . . . . . . 39

2.3.1 Plate tectonic theory . . . . . . . . . . . . . . . . . . . . . . . . 39
2.3.2 Tectonic plate setting of coasts . . . . . . . . . . . . . . . . . . 41
2.3.3 First-order coastal sedimentary features . . . . . . . . . . . . . 45
2.3.4 Summary of tectonic classification . . . . . . . . . . . . . . . . 49
2.4 Pleistocene inheritance of cliffed coasts . . . . . . . . . . . . . . . . . . 49
2.5 (Holocene) transgression versus regression . . . . . . . . . . . . . . . . 52
2.5.1 Geological sea level changes . . . . . . . . . . . . . . . . . . . . 52
2.5.2 Role of sea level rise in Holocene coastal evolution . . . . . . . 59
2.5.3 More recent coastal development . . . . . . . . . . . . . . . . . 64
2.6 Nature and abundance of coastal material . . . . . . . . . . . . . . . . . 65
2.6.1 Sources of sediments deposits . . . . . . . . . . . . . . . . . . . 65
2.6.2 Sediment sizes . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
2.6.3 Geographical variation . . . . . . . . . . . . . . . . . . . . . . . 69
2.6.4 Muddy coasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
2.6.5 Sandy coasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
2.6.6 Vegetation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
2.7 Process-based classification . . . . . . . . . . . . . . . . . . . . . . . . . 79
2.7.1 Dominance of fluvial, wave or tidal processes . . . . . . . . . . 79
2.7.2 Ternary diagrams for progradation and transgression . . . . . . 81
2.7.3 Classification of deltas . . . . . . . . . . . . . . . . . . . . . . . 84
2.7.4 Overview and examples of coastal forms . . . . . . . . . . . . . 89
2.8 Summary of coastal classification . . . . . . . . . . . . . . . . . . . . . 89
3 Ocean waves 91
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.2 Oscillations of the ocean water surface . . . . . . . . . . . . . . . . . . 92
3.3 Measuring ocean surface elevations . . . . . . . . . . . . . . . . . . . . 98
3.4 Short-term wave statistics . . . . . . . . . . . . . . . . . . . . . . . . . 100
3.4.1 Description of wave characteristics . . . . . . . . . . . . . . . . 100
3.4.2 Analysis of the time series . . . . . . . . . . . . . . . . . . . . . 102
3.4.3 Spectral analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 104
3.4.4 Short-term wave height distribution . . . . . . . . . . . . . . . 107
3.5 Wind wave generation and dispersion . . . . . . . . . . . . . . . . . . . 110
3.5.1 Locally-generated sea . . . . . . . . . . . . . . . . . . . . . . . 110
3.5.2 Wave dispersion . . . . . . . . . . . . . . . . . . . . . . . . . . 111
3.5.3 Wave groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
3.5.4 Sea versus swell waves . . . . . . . . . . . . . . . . . . . . . . . 116
3.6 Long-term statistics and extreme values . . . . . . . . . . . . . . . . . . 119
3.7 Generation of the tide . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
3.7.1 Equilibrium theory of the tide . . . . . . . . . . . . . . . . . . . 121
3.7.2 Gravitational pull . . . . . . . . . . . . . . . . . . . . . . . . . . 122
3.7.3 Differential pull or the tide-generating force . . . . . . . . . . . 123

Contents vii
3.7.4 Spring and neap tide . . . . . . . . . . . . . . . . . . . . . . . . 127

3.7.5 Daily inequality . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
3.7.6 Tidal constituents . . . . . . . . . . . . . . . . . . . . . . . . . . 130
3.8 Propagation of the tide . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
3.8.1 Dynamic theory of tides . . . . . . . . . . . . . . . . . . . . . . 133
3.8.2 Amphidromic systems . . . . . . . . . . . . . . . . . . . . . . . 137
3.8.3 Kelvin waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
3.9 Tidal analysis and prediction . . . . . . . . . . . . . . . . . . . . . . . . 144
4 Global wave and tidal environments 149

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
4.2 Zonal wind systems and ocean circulation . . . . . . . . . . . . . . . . 150
4.2.1 Solar radiation and temperature distribution . . . . . . . . . . . 150
4.2.2 Atmospheric circulation and wind patterns . . . . . . . . . . . 153
4.2.3 Oceanic circulation . . . . . . . . . . . . . . . . . . . . . . . . . 157
4.3 Large-scale variation in wave environments . . . . . . . . . . . . . . . 158
4.3.1 Wave height variation . . . . . . . . . . . . . . . . . . . . . . . 158
4.3.2 Wave environments . . . . . . . . . . . . . . . . . . . . . . . . 160
4.3.3 Coastal impact of different wave conditions . . . . . . . . . . . 162
4.4 Large-scale variation in tidal characteristics . . . . . . . . . . . . . . . 163
4.4.1 Global tidal environments . . . . . . . . . . . . . . . . . . . . . 163
4.4.2 Coastal impact of tide and classification . . . . . . . . . . . . . 166
5 Coastal hydrodynamics 169

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
5.2 Wave transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
5.2.1 Energy balance . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
5.2.2 Shoaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
5.2.3 Refraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
5.2.4 Diffraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
5.2.5 Wave-breaking . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
5.3 Wave asymmetry and skewness . . . . . . . . . . . . . . . . . . . . . . 184
5.4 Wave orbital velocity, pressure and bed shear stress . . . . . . . . . . . 191
5.4.1 Wave orbital velocities . . . . . . . . . . . . . . . . . . . . . . . 191
5.4.2 Dynamic pressure . . . . . . . . . . . . . . . . . . . . . . . . . 193
5.4.3 Wave boundary layer . . . . . . . . . . . . . . . . . . . . . . . . 194
5.5 Wave-induced set-up and currents . . . . . . . . . . . . . . . . . . . . . 200
5.5.1 Wave-induced mass flux or momentum . . . . . . . . . . . . . . 200
5.5.2 Radiation stress . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
5.5.3 Wave-induced forces . . . . . . . . . . . . . . . . . . . . . . . . 210
5.5.4 Cross-shore balance: wave set-up and set-down . . . . . . . . . 212
5.5.5 Alongshore balance: longshore current . . . . . . . . . . . . . . 220
5.5.6 Vertical structure of the wave-induced currents . . . . . . . . . 230

viii Contents
5.5.7 3D effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231

5.6 Wind-induced set-up and currents . . . . . . . . . . . . . . . . . . . . . 235
5.7 Tidal propagation in coastal waters . . . . . . . . . . . . . . . . . . . . 238
5.7.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
5.7.2 Tidal propagation along the shore . . . . . . . . . . . . . . . . . 238
5.7.3 Tidal propagation into basins . . . . . . . . . . . . . . . . . . . 248
5.7.4 Tidal asymmetry . . . . . . . . . . . . . . . . . . . . . . . . . . 256
5.7.5 Overtides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
5.7.6 Residual currents . . . . . . . . . . . . . . . . . . . . . . . . . . 264
5.8 Long-wave phenomena in coastal waters . . . . . . . . . . . . . . . . . 269
5.8.1 Seiches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
5.8.2 Bound long waves and surf beat . . . . . . . . . . . . . . . . . . 270
6 Sediment transport 273

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
6.2 Sediment properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
6.2.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
6.2.2 Grain size, density and bulk properties . . . . . . . . . . . . . . 274
6.2.3 Fall velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
6.3 Initiation of motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
6.3.1 Forces on a single grain . . . . . . . . . . . . . . . . . . . . . . 279
6.3.2 Shields curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
6.4 Basic principles of transport modelling . . . . . . . . . . . . . . . . . . 283
6.4.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
6.4.2 Practical modelling of sediment transport . . . . . . . . . . . . 289
6.5 Bed load based on the Shields parameter . . . . . . . . . . . . . . . . . 290
6.5.1 Importance of the Shields parameter . . . . . . . . . . . . . . . 290
6.5.2 Including waves . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
6.5.3 Instantaneous bed load transport . . . . . . . . . . . . . . . . . 293
6.5.4 Bed load transport based on time-averaged shear stress . . . . . 296
6.5.5 Summary and concluding remarks . . . . . . . . . . . . . . . . 299
6.6 Diffusion approach for suspended transport . . . . . . . . . . . . . . . 301
6.6.1 General formulation . . . . . . . . . . . . . . . . . . . . . . . . 301
6.6.2 Sediment continuity . . . . . . . . . . . . . . . . . . . . . . . . 304
6.6.3 Time-averaged concentration distribution . . . . . . . . . . . . 307
6.7 Energetics approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
6.7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
6.7.2 Energetics approach for combination of waves and currents . . 310
6.8 Some aspects of (very) fine sediment transport . . . . . . . . . . . . . . 313
6.8.1 Memory effects . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
6.8.2 Critical shear stress and settling velocity . . . . . . . . . . . . . 315
6.8.3 Environmental issues . . . . . . . . . . . . . . . . . . . . . . . . 317
6.9 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318

Contents ix
6.9.1 Choice of models . . . . . . . . . . . . . . . . . . . . . . . . . . 318

6.9.2 Specific situations . . . . . . . . . . . . . . . . . . . . . . . . . . 319
7 Cross-shore transport and profile development 321

7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
7.2 Equilibrium shoreface profile . . . . . . . . . . . . . . . . . . . . . . . . 325
7.2.1 The concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
7.2.2 (Semi-)empirical derivations . . . . . . . . . . . . . . . . . . . . 328
7.2.3 Engineering applications . . . . . . . . . . . . . . . . . . . . . . 330
7.3 Morphodynamics of the upper shoreface . . . . . . . . . . . . . . . . . 333
7.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
7.3.2 Beach states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
7.3.3 Storm and seasonal changes . . . . . . . . . . . . . . . . . . . . 339
7.3.4 Bar cycles over years . . . . . . . . . . . . . . . . . . . . . . . . 341
7.3.5 Episodic changes (dune erosion) . . . . . . . . . . . . . . . . . . 344
7.4 Structural losses or gains . . . . . . . . . . . . . . . . . . . . . . . . . . 348
7.5 Cross-shore sediment transport . . . . . . . . . . . . . . . . . . . . . . 350
7.5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
7.5.2 Decomposition of the transport rate . . . . . . . . . . . . . . . 350
7.5.3 Analytical solutions for the middle and lower shoreface . . . . 352
8 Longshore transport and coastline changes 361

8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361
8.2 Longshore transport formulas . . . . . . . . . . . . . . . . . . . . . . . 362
8.2.1 General transport formulas . . . . . . . . . . . . . . . . . . . . 362
8.2.2 Cross-shore distribution of longshore transport . . . . . . . . . 364
8.2.3 Bulk longshore transport formulas . . . . . . . . . . . . . . . . 365
8.2.4 The (S,𝜑)-curve . . . . . . . . . . . . . . . . . . . . . . . . . . . 373
8.2.5 Yearly-averaged sediment transport . . . . . . . . . . . . . . . . 376
8.3 Calculation of coastline position . . . . . . . . . . . . . . . . . . . . . . 379
8.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
8.3.2 Single-line theory . . . . . . . . . . . . . . . . . . . . . . . . . . 380
8.3.3 Analytical solution for accretion near breakwater or jetty . . . 385
8.3.4 Multiple-line theory . . . . . . . . . . . . . . . . . . . . . . . . 390
8.4 Coastal features and coastal change due to longshore transport . . . . . 392
8.4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 392
8.4.2 Blockage of longshore transport by shore-normal structures . . 393
8.4.3 Shadow effects due to obstacles away from the shoreline . . . . 396
8.4.4 Shoreline perturbation . . . . . . . . . . . . . . . . . . . . . . . 400
8.4.5 Interrupted coasts: spits . . . . . . . . . . . . . . . . . . . . . . 402
8.4.6 Deltaic coastlines . . . . . . . . . . . . . . . . . . . . . . . . . . 405

x Contents
9 Coastal inlets and tidal basins 409

9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409
9.2 Basin and inlet types . . . . . . . . . . . . . . . . . . . . . . . . . . . . 410
9.2.1 Bays, lagoons and estuaries . . . . . . . . . . . . . . . . . . . . 410
9.2.2 Hydrodynamical classification . . . . . . . . . . . . . . . . . . . 413
9.2.3 Hydraulic boundary conditions and geometric controls . . . . . 416
9.3 The main morphological elements . . . . . . . . . . . . . . . . . . . . . 418
9.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418
9.3.2 Tidal deltas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419
9.3.3 Basin characteristics . . . . . . . . . . . . . . . . . . . . . . . . 422
9.4 The ebb-tidal delta or outer delta . . . . . . . . . . . . . . . . . . . . . . 424
9.4.1 Waves and currents at the outer delta . . . . . . . . . . . . . . . 424
9.4.2 Sediment transport patterns . . . . . . . . . . . . . . . . . . . . 430
9.4.3 Empirical relationships: volume of the ebb-tidal delta . . . . . . 435
9.5 Stability of the inlet cross-sectional area . . . . . . . . . . . . . . . . . . 438
9.5.1 Escoffier’s model . . . . . . . . . . . . . . . . . . . . . . . . . . 438
9.5.2 Empirical equilibrium cross-sectional area . . . . . . . . . . . . 441
9.6 The inner basin geometry . . . . . . . . . . . . . . . . . . . . . . . . . . 444
9.6.1 Complex geometry of tidal basins . . . . . . . . . . . . . . . . . 444
9.6.2 Equilibrium relations for tidal channels and flats . . . . . . . . 446
9.7 Net sediment import or export . . . . . . . . . . . . . . . . . . . . . . . 449
9.7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
9.7.2 Tide-induced residual transport of (medium to) coarse sediment 450
9.7.3 Fine sediment transport and siltation . . . . . . . . . . . . . . . 456
9.7.4 Overview of the relation between morphology and sediment
transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 460
9.7.5 Large-scale morphodynamics . . . . . . . . . . . . . . . . . . . 461
9.8 Changes in dynamic equilibrium . . . . . . . . . . . . . . . . . . . . . . 462
9.8.1 Closure of a part of a tidal basin . . . . . . . . . . . . . . . . . . 462
9.8.2 Accretion of new land . . . . . . . . . . . . . . . . . . . . . . . 464
9.8.3 Relative sea level rise . . . . . . . . . . . . . . . . . . . . . . . . 464
9.8.4 Adaptation time . . . . . . . . . . . . . . . . . . . . . . . . . . . 465
10 Coastal protection 467

10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467
10.2 Coastal protection strategies and methods . . . . . . . . . . . . . . . . 468
10.2.1 Management strategies . . . . . . . . . . . . . . . . . . . . . . . 468
10.2.2 Selection of protection method . . . . . . . . . . . . . . . . . . 469
10.2.3 From problem definition to realisation . . . . . . . . . . . . . . 471
10.3 Coastal erosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471
10.3.1 Structural erosion of coasts . . . . . . . . . . . . . . . . . . . . 471
10.3.2 Beach and dune erosion during severe storm surges . . . . . . . 474
10.3.3 Dynamic behaviour of tidal inlets . . . . . . . . . . . . . . . . . 475

Contents xi
10.4 Modification of longshore transport processes . . . . . . . . . . . . . . 476

10.5 Structures influencing longshore transport rates . . . . . . . . . . . . . 478
10.5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 478
10.5.2 Jetties or shore-normal breakwaters . . . . . . . . . . . . . . . 478
10.5.3 Groynes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483
10.5.4 Detached shore-parallel offshore breakwaters . . . . . . . . . . 487
10.5.5 Piers and trestles . . . . . . . . . . . . . . . . . . . . . . . . . . 490
10.5.6 Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . 491
10.6 Structures protecting against storm-induced erosion . . . . . . . . . . . 492
10.6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 492
10.6.2 Seawalls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493
10.6.3 Revetments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497
10.6.4 Sea dikes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 498
10.7 Nourishments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500
10.7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500
10.7.2 Design aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . 501
10.7.3 Counteracting structural erosion of coasts . . . . . . . . . . . . 504
10.7.4 Dune reinforcement . . . . . . . . . . . . . . . . . . . . . . . . 508
10.7.5 Beach widening and creation . . . . . . . . . . . . . . . . . . . 509
10.7.6 A new nourishment strategy: the Sand Engine . . . . . . . . . 512
A Linear wave theory 515
B Waves breaking on a beach 521

B.1 Scale comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 521
B.2 Periodic wave results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 521
B.3 Random wave results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 524
C Hydrographic charts 525

C.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 525
C.2 Units and their background . . . . . . . . . . . . . . . . . . . . . . . . . 526
C.3 Explanatory notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 526
C.4 The map itself . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 530
C.5 Interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 532
C.6 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 532
D Stability of structures 533

D.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 533
D.2 Initiation of transport and damage . . . . . . . . . . . . . . . . . . . . . 533
D.3 Other protections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 536
E Responses to the closures of Dutch tidal basins 537

E.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 537
E.2 Closures in the Wadden Sea . . . . . . . . . . . . . . . . . . . . . . . . 539

xii Contents
E.2.1 Closure of Zuiderzee . . . . . . . . . . . . . . . . . . . . . . . . 539

E.2.2 Closure of Lauwerszee . . . . . . . . . . . . . . . . . . . . . . . 541
E.2.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 543
E.3 Closures in the delta area . . . . . . . . . . . . . . . . . . . . . . . . . . 543
E.3.1 Overview of the closures . . . . . . . . . . . . . . . . . . . . . . 543
E.3.2 Developments outside area . . . . . . . . . . . . . . . . . . . . 544
E.3.3 Impact on the (semi-)closed basins . . . . . . . . . . . . . . . . 546
E.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 546
Acronyms and abbreviations 549
Bibliography 551
Subject index 565
Credits 575
Errata and improvements 579

xiii
Preface
This Open Textbook is the Open Access version of lecture notes that have been de-
veloped to support and supplement the Delft University of Technology (TU Delft) first-
year Master of Science (MSc) course on Coastal Dynamics. Since its introduction in the
academic year 2009-2010, this first-year MSc course has been taught to MSc graduate
students following the two-year Hydraulic Engineering MSc curriculum of the Faculty
of Civil Engineering and Geosciences. It focuses on the interrelation between physical
wave, flow and sediment transport phenomena and the resulting morphodynamics of
a wide variety of coastal systems. The objective is to provide hydraulic and coastal
engineering MSc students with insights into the phenomenological and theoretical, as
well as applied aspects of these phenomena. The course builds upon Bachelor (BSc)
and MSc courses treating the dynamics of flow, waves and sediment transport, and
may serve as a basis for a course focusing on coastal modelling, as in the TU Delft cur-
riculum. Although several valuable course monographs and books exist on the topic
of coastal dynamics and coastal engineering (see Sect. 1.7), we felt that no standard
teaching books existed for the purposes that we had in mind. It has been our explicit
intention to connect coastal and estuarine dynamics. Not only is the interaction of
coasts and estuaries of basic importance for the evolution of many coastal systems,
but describing the similarities between their physical processes is highly instructive
as well.
This textbook emphasizes these similarities for the benefit of understanding shared
processes such as nonlinearities in flow and sediment transport. Furthermore, our
textbook differs from other coastal textbooks in that we connect engineering scales
with larger Holocene scales, emphasizing among other aspects the impact of leading
and trailing edge coasts on coastal evolution. While primarily developed for the pur-
pose of teaching, we hope that these notes are also useful as a reference book for MSc
and PhD students.
The present form and format of this book has been developed to follow the Open
Access rules and is subject to change, both in language and contents. We feel that this
is a matter of evolution, strongly steered by our continuous learning through and our
experience with both teaching and examination. In the years to come we will provide
regular updates of this Open Textbook, also in response to comments we expect to
receive from our readers.
Our list of acknowledgements is long, but our view is that the art of teaching is to
accumulate and digest the efforts of many of our colleagues nationally and globally.

xiv Preface
First, we wish to acknowledge the efforts of all our colleagues in the Netherlands who
developed earlier lecture notes that we used as inspiration for these notes: Professor
Eco Bijker and Associate Professors Walt Massie and Jan van de Graaff, Professor Kees
d’Angremond and MSc Liesbeth van der Velden, Professors Huib de Vriend and Zheng
Bing Wang and PhDs Anneke Hibma and Edwin Elias, and Professor Job Dronkers (see
‘Credits’ on page 575).
Second, we wish to acknowledge some key international references that we used while
developing new material that we found missing. Chapter 2, especially Sect. 2.3, is
heavily inspired by Davis Jr. (1994), whose well-illustrated textbook offers a good in-
sight into larger-scale coastal behaviour and classification. The energetics approach
for cross-shore sediment transport in Sect. 6.1 and Ch. 7 is largely based on Bowen
(1980), whose elegant work we still find very instructive. In a more general sense we
have been inspired by a number of books we reference in Sect. 1.7, notably the books
by Fredsøe and Deigaard (1992), Kamphuis (2000) and Masselink and Hughes (2003).
Further, we need to acknowledge many Delft colleagues who have contributed to
our knowledge by their work and comments and/or provided us with course and lec-
ture materials. Regarding knowledge contributions, we mention Jurjen Battjes, Leo
Holthuijsen, Jacobus van de Kreeke and Han Winterwerp. Stefan Aarninkhof, Jan van
Overeem, Roshanka Ranasinghe and Zheng Bing Wang gave valuable contributions
to the content of our lectures. Revisions of sections of early versions of our notes
by Ad Reniers, Dano Roelvink, Ad van der Spek and Zheng Bing Wang are highly
appreciated. Special thanks go to Howard Southgate, who has read through our com-
plete lecture notes in detail making relevant comments and suggestions. PhDs Stuart
Pearson, Yorick Broekema and Alejandra Gijón Mancheño greatly improved the con-
nection between the notes and the lecturing of the course. We thank our former Delft
Hydraulics colleague Hans de Vroeg for providing cross-shore distributions of wave
heights, longshore currents and sediment transports calculated with Unibest-CL+. Fi-
nally, we are grateful for the many suggestions we received from our students over
the years.
We have invested heavily in making the illustrations and graphs “our own”. For this we
were assisted by many of our talented students Liang Li, Dáire Stive, Marcio Boechat
Albernaz, Janbert Aarnink, Ascha Simons, Viktorija Usevičiūtė and Tim van Dam,
whose creative contributions we gladly acknowledge. We thank Marcel Mol for the
subject index.
TU Delft library convinced, stimulated and sponsored us to publish these notes Open
Access. Their specialists Michiel de Jong, Jacqueline Michielen-van den Riet and Mo-
nique de Bont guided us to understand and comply with copyright licenses. Besides
their creative qualities, Ascha Simons and Tim van Dam used their ample technical
competences to successfully finish this ambitious project. The management qualities

Preface xv
of Carolina Piccoli to streamline this process are highly appreciated. The broad sup-
port by Stefan Aarninkhof in this venture has been essential. We thank Saskia Roselaar
for proofreading the book.
As our knowledge and experience is continuously developing, lecture notes are al-
ways work-in-progress. We would be very grateful for feedback from our readers on
contents and didactics. Also, we would appreciate to be informed of any copyright
infringements in the textbook. We will gladly adjust and/or correct.
Delft, The Netherlands
January 2023
Note to our readers: We would like to track the reach and use of this Open Text-
book. Therefore, we would be very pleased if you could provide information on
your intended use of this book using the form on the TU Delft Open Textbook
repository. The form also provides options to subscribe to update notifications
and give feedback.
Note to our students: No one reads a textbook more thoroughly than a student
studying the material for a course. Therefore, if you find any smaller or larger
mistakes in this book – from typographical errors and incorrect cross-references
to unclarities and inconsistencies – we would greatly appreciate to be informed
of these. In order to suggest a correction or improvement, please fill out the form
accessible via the TU Delft Open Textbook repository.

1
1
Overview
1.1. Introduction
Throughout history, humans have extensively used the coastal zone for many pur-
poses, such as fishing, tourism, transport of goods, water treatment and housing. Ag-
riculture has benefited from the very fertile grounds created by marine and riverine
deposits. Approximately three billion people – half the world’s population – live and
work within a few hundred kilometres of a coastline, notwithstanding the vulnerab-
ility of coastal areas to flooding. Due to the high population densities and extensive
infrastructure and property development in coastal areas, disasters will have major
consequences. Coastal engineers play an important role in both developing the coastal
zone and protecting the coast and the hinterland.
In Sect. 1.2, the course contents and the position in the curriculum are explained.
Sect. 1.3 lists the study goals. Section 1.4 gives some examples of the problems coastal
engineers may be faced with. In doing so coastal engineers need a thorough know-
ledge of the natural dynamics of the coastal system. An introduction on that topic
is given in Sect. 1.5. Important players in the (Dutch) coastal engineering sector are
summarised in Sect. 1.6. The chapter concludes with a list of handbooks, journals,
conference proceedings and internet sources for further reading (Sect. 1.7).
1.2. Coastal dynamics for coastal engineers

1.2.1. What is coastal engineering?
Coastal Engineering is the branch of civil engineering concerned with the planning,
design, construction and maintenance of works in the coastal zone. Coastal engineer-
ing usually involves either 1) the transport and stabilisation of sand and other coastal
sediments or 2) the construction of structures.

2 1. Overview
Measures in the first category are called ‘soft’ measures since they make use of natural
(soft) coastal material. Examples are beach nourishments, maintenance dredging and
land reclamation.
The second category of structures or ‘hard’ measures can be divided into various func-
tional groups:
• Seawalls and revetments are built parallel or nearly parallel to the shoreline at
the land-sea interface with the objective of preventing further shoreline reces-
sion. Seawalls are usually massive and rigid, while a revetment is an armour-
ing of the beachface with rock armour or artificial units. Although often used,
the efficacy of seawalls and revetments is debatable, as we will discuss later on
(Sect. 10.6);
• Groynes are built perpendicular to the shore and usually extend out through
the surf zone under normal wave and water level conditions (the surf zone or
breaker zone is the zone close to the shore where waves are breaking). They
help widen and protect a beach by trapping sand from the alongshore transport
system (see Ch. 8) or by retaining artificially placed sand;
• Jetties are structures built at the entrance to a river or tidal basin to stabilise the
entrance as well as to protect vessels navigating the entrance channel;
• Breakwaters primarily protect a shoreline or harbour anchorage area from wave
attack. Breakwaters may be located completely offshore and oriented parallel
to the shore (detached breakwaters), or they may be oblique and connected to
the shore. Traditionally, detached breakwaters have been designed as emerged
structures, but submerged breakwaters have now also become a popular option.
The latter are not easily noticeable because of their low crests;
• Other structures such as submerged pipelines.
The purposes of these ‘soft’ and ‘hard’ works are diverse:
• Control of shoreline erosion;

• Defence against flooding caused by storms and/or tides;
• Development of coastal functions, for instance coastal recreation;
• Development of navigation channels and harbours.
Coastal engineering works are carried out in a highly dynamic and energetic environ-
ment. The various sources of coastal energy are:
• Marine forces (waves, tides, currents and other oceanographic phenomena);

• Terrestrial forces (river outflow);
• Atmospheric forces (coastal winds and climate).
These forces not only directly impact the planned ‘soft’ or ‘hard’ measures, but also
permanently change the physical shape and structure of a coastal system. This shape
of the coast is called morphology. Coastal morphology for a sandy stretch of coast
thus is the topography of the sandy dunes and beach and the underwater topography

1.2. Coastal dynamics for coastal engineers 3
of the seabed. When changes occur in the external forcing, the coastal morphology
will change accordingly. Changes in the forcing can have a natural cause or can be
human-induced. Examples of changes in the forcing conditions are:
• High waves and piling up of water against the coast (surge) due to the occurrence
of a storm;
• Long-term sea level rise;
• Changes in the wave impact on the adjacent coast due to the construction of a
harbour;
• The deprivation of a coastal system of sediment supply due to the construction
of a dam in a river.
Although coastal changes take place on a variety of timescales, coastal engineers and
managers are mostly interested in timescales ranging from 1 year to (a few) hundred
years and in large-impact events like storms causing dune erosion and flooding.
1.2.2. Position of Coastal Dynamics in the TU Delft curriculum

Coastal engineering is a very broad profession. As an illustration, the program of the
two-yearly International Conference on Coastal Engineering (ICCE) includes discip-
lines such as hydrodynamics, coastal morphology, coastal protection, structures and
ports and waterways. At TU Delft these disciplines are covered in a number of courses
within the specialisations Coastal Engineering, Rivers, Ports and Waterways, Environ-
mental Fluid Mechanics and Hydraulic Structures and Flood Risk1 .
Coastal Dynamics I and II

For an appropriate choice and design of measures, the dynamics of the forcing and re-
sponse of the coastal system should be taken into account. For that reason the courses
Coastal Dynamics I and II focus on the dynamics of the coastal system (waves, cur-
rents, sediment transport and morphology). Coastal Dynamics I (CIE4305), which is
served by this textbook, treats the most important coastal and estuarine phenomena.
Furthermore, attention is paid to functional design and impacts of engineering works.
Coastal Dynamics II (CIE4309) goes into more detail regarding coastal processes and
pays special attention to coastal modelling. Coastal Dynamics I is compulsory for
Coastal Dynamics II.
In both Coastal Dynamics I and II the focus is on the dynamics of coastal systems built
up by loose and relatively fine material (we consider mainly sand and to a lesser extent
mud). This material has been delivered to the coast in recent geological history, mostly
through Holocene marine and fluvial sediment supply (see Ch. 2).
1
The Faculty of Civil Engineering is implementing a new MSc curriculum. During the transition period,
we will continue to refer to the old course names and codes and update these in the near future.

4 1. Overview
Relevant prior knowledge

Fluid mechanics and wave theories are indispensable topics in coastal dynamics. Coastal
Dynamics I builds upon the Bachelor courses treating flow, wave and transport dynam-
ics. Of the MSc courses, preferably Ocean Waves (CIE4325) should be followed prior to
Coastal Dynamics I. A summary of the (offshore) aspects of wind waves and tides that
are relevant to Coastal Dynamics I are given in Ch. 3 and will be built on in subsequent
chapters.
Related topics
The technical design of structures is not treated in Coastal Dynamics I and II. Other
courses like Bed, Bank and Shoreline Protection (CIE4310) deal with this.
1.3. Study goals

At TU Delft, this textbook accompanies the course Coastal Dynamics I, which will be
renamed Coastal Systems in the new curriculum. On completion of either course, the
student will be able to:
1. Discuss the nature and complexity of typical coastal engineering problems (Chs. 1
and 10, but also Chs. 7 to 9);
2. Explain the imprint on present-day coastal systems – considering uninterrupted
coastlines as well as tidal inlets and basins – of a) geological processes, sea level,
and climate (Ch. 2); and b) fluvial, wave and tidal processes and their relative
influence (Ch. 2);
3. Understand the global variation of a) coastal systems (Chs. 2 and 4); and b) wave
and tidal climate (Ch. 2);
4. Analyse the hydrodynamics of waves and tides in a) the oceans (Ch. 3); b) open
coasts (Ch. 5); and c) tidal basins (Ch. 5);
5. Evaluate sediment transport processes under waves, tides, and currents (Ch. 6),
distinguishing between: a) cross-shore transport (Ch. 7); b) longshore transport
(Ch. 8); and c) transport in tidal inlets and basins (Ch. 9);
6. Apply system knowledge to determine the morphological response to environ-
mental conditions as well as interventions (groynes, breakwaters, nourishments,
land reclamations, sand mining) of a) open coasts (Chs. 7 and 8); and b) tidal in-
lets and basins (Ch. 9);
7. Assess the merits and disadvantages of various coastal interventions for protec-
tion against flooding and erosion (Ch. 10).

1.4. Examples of engineering applications 5
1.4. Examples of engineering applications

1.4.1. Overview of coastal area and problems
Figure 1.1 shows a schematic plan view of a coastal area with most of the natural
forcing conditions, natural features and some examples of man-made interventions in
a natural coastal system. Most of the items are dealt with in these lecture notes.
The coastline in Fig. 1.1 receives sediment from rivers and is interrupted by openings
called tidal or coastal inlets (see Intermezzo 9.2 for the terminology). The name tidal
inlet refers to the fact that the tide is important in maintaining the inlet, viz. keeping
the inlet from closing naturally. Tidal inlets are either found along barrier island coasts
or along coasts interrupted by estuaries or lagoons. Tidal inlets and their associated
basins are common features of lowland coasts all around the world.
Seen from the sea side, an estuary is an arm of the ocean that is thrust into the mouth
and lower course of a river as far as the tide reaches. Estuaries receive fresh water from
rivers, and salt water from the sea. Lagoons do not have a major point source of fresh
water input, such as a river. These tidal systems play a crucial role in the sediment
budget of the coastal zone and thus influence the long-term coastal evolution.
At the uninterrupted stretches of coast, waves are the dominant forcing agent. Wind
waves and tides are treated in detail in Chs. 3 and 5 for oceanic and coastal waters
respectively. The global variation in wave and tidal climate is discussed in Ch. 4.
Some examples of practical cases are given in this section by briefly discussing the
following items from Fig. 1.1:
• Cross-shore profile (section A–A in Fig. 1.1);
• Morphological development in vicinity of a port;
• Delta near a river mouth;
• Tidal inlet;
• Dune erosion during a severe storm surge;
• Large artificial island in open sea.
1.4.2. Cross-shore profile

The upper panel of Fig. 1.2 shows the shape of a cross-section (called cross-shore pro-
file) as measured perpendicular to a sandy coast at two moments in time. The lower
panel shows the variation in time around an average profile. Please note that the ver-
tical and horizontal scales of the plot are quite different. Dunes, beach and a part of
the so-called shoreface can be discerned. The actual slope of the dune face is 1:3 to 1:4.
The slope of the beach is decreasing from the upper part of the beach near the foot of
the dunes (1:20) towards the sea; near the waterline the slope is approximately 1:50.
At the shoreface some (breaker) bars are present. The average bottom slope becomes

6 1. Overview
jettie
s
primary waves
our
harb
secondary waves
delta
rge
cha
dis
er
riv
accretion
High Water (HW)
t
breakwaters
our
harb l Low Water (LW)
anne
entrance ch
e r o sio
win
artif
icia
df
l is
n
arm
la
nd
gr
oyn
es
y
ar
stu
e
sto barr
rm ier
su
rge
e
s bl
en and ca
d
gi
ne e an
lin
pe
pi A
be
ac
du
al
h
-tid
ne
ent ebb lta

s
evetm de
ry A r
stua
e t
el
nn
le
rmer c h a
in
fo
al
tid
milit
in
as
a l b on
ary z
d
ti lag o
or
one
al
r tid
e a barrier
int are island
in let
al
ti d
Figure 1.1: Plan view of a coastal area.

flatter with longer distance from the waterline. At the seaward end of the plot (water
depth: MSL −10 m) the slope is approximately 1:125.
20
2004
2011
10
-10
20
Range 1970-2015
Time-averaged
10
-10
-200 0 200 400 600 800 1000
cross-shore distance [m]
Figure 1.2: Cross-shore profiles at Egmond aan Zee (RSP 7004125) from the long-running Jarkus
dataset (JARKUS, n.d.), which contains profiles of the entire Dutch coast measured in sub-
sequent summers. The upper panel shows two instantaneous profiles, whereas the lower panel
displays the temporal variation from 1970 to 2015. The cross-shore coordinate is relative to a
local beach pole or Rijksstrandpaal (RSP) in Dutch. The elevation is relative to the Dutch refer-
ence level NAP.
The water level as indicated in Fig. 1.2 reflects the Mean Sea Level (MSL). This is the sea
level averaged over a period of time such as a month or a year, such that periodic sea
level changes e.g. due to waves and tides are averaged out. In the Netherlands, MSL
is approximately equal to the Dutch reference level called NAP (Normaal Amsterdams
Peil, in Dutch)2 . The seabed consists of sandy material that generally fines when going
further offshore. A typical grain size for a sandy coast is 𝐷50 = 200 µm; 150 million of
these particles fit into a volume of 1 litre!
Under the influence of waves, the position of the coastline (represented by the intersec-
tion of MSL and the profile) will continuously change. Variations will take place on the
timescale of storms and seasons. During storms, high and long waves cause erosion
of the beach. This sediment is deposited in the surf zone (the zone where the waves
are breaking). A typical storm profile therefore has a narrow beach and a relatively
2
Recently, MSL has been around NAP + 0.06 m in the Netherlands.

8 1. Overview
flat slope. Seasonality is especially evident in the Northern Hemisphere which exper-
iences a large number of storms in winter (see Sect. 4.3.1). Hence, a storm profile is
also often called a winter profile. During storms high water levels can also cause dune
erosion and flooding (see Sect. 1.4.6). In summer the sand is moved back towards the
beach and dunes by lower and shorter waves. This cross-shore transport of sediment
causes oscillations of the coastline (Fig. 1.3), but in principle the mean position of the
coastline does not change. The mean position of the coastline will only change in the
case of a structural loss or gain of sediment; structural erosion may occur when sand
disappears in offshore canyons or in the alongshore direction (Sect. 1.4.3). Seasonal
variations are relevant to tourism (beach width) and the safety of property close to the
brink (highest point) of the dune.
storm/winter MSL
summer MSL
rm
be
storm/winter MSL
summer MSL
bar
Figure 1.3: Summer and winter profiles showing the annual changes in beach profile. In summer
the beach is rich in sediment (see the berm in the upper panel). In winter the beach is lower and
sediment is found further offshore (see the bar in the lower panel).
Suppose that irregular waves (also called random waves) approach the coast perpen-
dicularly with a significant deep-water wave height of 𝐻𝑠,0 = 2 m and a peak period
of 𝑇𝑝 = 10 s (for the definitions see Sects. 3.4.2 and 3.4.3). The wave height distribu-
tion along the profile (Fig. 1.4a) can be calculated with numerical models based on a
spectral energy or action balance (Sects. 3.5.3 and 5.2.1). Note that the wave height
in Fig. 1.4a reduces gradually when the waves approach the waterline. This is due to
energy dissipation due to (partial) wave-breaking and bottom friction.
In Fig. 1.4b the maximum horizontal components of the orbital velocities near the bed
are plotted as a function of the position in the cross-shore profile. Note the rather large
magnitudes (>0.6 m/s to 0.8 m/s at many positions). Realizing that the critical velocity
to initiate motion in uniform flow for particles with 𝐷50 = 200 µm is approximately
0.2 m/s (see Sect. 6.3), one may understand that the waves of Fig. 1.4a are able to stir

up many particles in the cross-shore profile. The turbulence generated by breaking

waves is very effective in keeping those particles away from the bed. The particles
can subsequently be transported by for instance wave-generated and tidal currents
(see Ch. 5). Asymmetric waves can also give a net sediment transport. Chapter 6 dis-
cusses sediment transport in general, whereas Ch. 7 focuses on cross-shore transport.
Both the magnitude and direction (onshore-offshore) of the cross-shore transport may
change depending on the local hydrodynamic conditions.
2
elevation [m + MSL]
-2
-4
-6
0 100 200 300 400 500
1
(a)
1
orbital velocity [m/s]
0.8
0.6
0.4
0.2
0
0 100 200 300 400 500
(b)
Figure 1.4: Wave height and orbital velocity along a cross-shore profile computed with XBeach
(Roelvink et al., 2009) based on a JONSWAP spectrum (Sect. 3.5) with peak period 𝑇𝑝 = 7 sec
and significant wave height 𝐻𝑠 = 1.9 m. (a) 𝐻𝑠 (solid line) and 1 % exceedance wave height 𝐻1 %
(dashed line) calculated from a Rayleigh distribution (Sects. 3.4.2 and 3.4.4); (b) Horizontal or-
bital velocity near the bed (root-mean-square velocity 𝑢𝑟𝑚𝑠 ).
Often the effect of cross-shore oscillations is assumed to average out in the longer
term. In those cases structural trends in coastline position are due to sediment trans-
port along the coast (or more correctly due to gradients in longshore transport, as we
will see in Sect. 1.4.3). Nevertheless, some important practical problems are related to
changes in the shape of the profile with time (e.g. dune erosion and the behaviour of
beach and shoreface nourishments).

10 1. Overview
1.4.3. Morphological development in vicinity of a port

Figure 1.5 shows in plan view a part of a uniform sandy coast. Uniform means that the
depth contours are assumed to be straight and parallel. Waves obliquely approach the
coast, viz. there is a non-zero angle 𝜑 between the wave crests and the depth contours
(or equivalently between a wave ray and the shore normal). As the waves approach
the shore, the angle becomes smaller due to refraction (Sect. 5.2.3).
Inside (and a little bit outside) the surf zone, sediment is transported along the coast,
the so-called longshore transport (Ch. 8). Waves continuously stir up material from
the bed. This sediment is then transported by the longshore current. This current is
generated by the breaking of obliquely incident waves in the surf zone, see Sect. 5.5.5.
Due to the wave action in stirring up the material, a longshore current of 1 m/s is much
more effective in transporting sand than a river flow with the same magnitude.
If the coast is uniform, the sediment transport 𝑆 is constant along the coast and the
coast remains stable. The coastal section under consideration will only change when
the amount of sediment transported into the section is different than the sediment
leaving the section; or in other words: when there is a gradient in longshore transport
rates. Erosion will occur in case of a positive gradient in the transport direction (the
sediment transport is increasing along the shore, and hence more sediment is leaving
than entering the section). Accretion occurs in case of a negative gradient in the drift
direction. A uniform sediment transport along the coast (no gradient) does not change
the coast. This leads to one of the most important notions of these lecture notes:
Coastal changes occur in case of transport gradients. A positive gradient (an in-
crease in sediment transport in the transport direction) leads to erosion. A negat-
ive gradient (a decrease in sediment transport in the transport direction) creates
accretion. If the gradient is zero, there are no changes in morphology.
If along the uniform sandy coast a port is built with the help of two rather long break-
waters, much longer than the width of the surf zone (see Fig. 1.6), the longshore sedi-
ment transport will be interrupted. Seaward of the breakwaters no sediment transport
waves
S [m3/yr] longshore current surf

zone
Figure 1.5: Plan view of a uniform sandy coast. In the zone where waves are breaking (called
breaker or surf zone), a wave-driven sediment transport takes place along the coast.

is assumed to occur. On the updrift side of the port, accumulation of sand will occur
(negative transport gradient); at the downdrift side (lee side) erosion will take place
(due to a positive transport gradient). In Fig. 1.6 some coastline positions have been
sketched as a function of time.
waves
S=0
S [m3/yr] longshore current
Figure 1.6: Plan view of a uniform sandy coast with port breakwaters. Note the updrift accretion
and lee-side erosion. The latter is an oversimplified sketch of the real shape of lee-side erosion,
see Ch. 8.
After having studied these lecture notes, you should be able to understand and describe
the shapes of the coastline at updrift and downdrift sides as a function of time. It is
to be expected that sooner or later the accreting coastline on the updrift side reaches
the end of the updrift breakwater; you should be able to say at what time after the
construction of the port this will happen. For the time being it will be clear that as long
as no sediment will pass the breakwaters (𝑆 = 0 m3 /yr seaward of the breakwaters) the
total accumulation of sand in 𝑡 years after completion of the port on the updrift side will
be 𝑡 × 𝑆 m3 (with 𝑆 the undisturbed transport rate). That is, when we assume that 𝑆 in
m3 /yr is expressed including pores between the grains, so that the volumes represent
deposited volumes. (In Sect. 6.4 other units of sediment transport are discussed.) The
total erosion on the lee side will be 𝑡 × 𝑆 m3 as well.
In most cases accumulation will not be considered problematic by the coastal zone
manager involved; valuable new land has been gained (unless sediment is deposited
in navigation channels). However, the erosion on the lee side of the port will sooner
or later cause serious problems. How to resolve such types of erosion problems will
be discussed in Ch. 10. An obvious solution will be to artificially transfer volumes of
sand from the one side of the port to the other (𝑆 m3 /yr on average; a so-called sand
bypass system). Summarizing:
Engineering problems are often related to longshore transport gradients. Struc-

tural coastal problems arise when the longshore transport is changing alongshore,
for instance when the longshore current is interrupted by harbour breakwaters.
In the example of Fig. 1.6, the longshore sediment transport was due to obliquely ap-
proaching waves only. If tidal currents also occur along the coast, the morphological

12 1. Overview
behaviour becomes more complicated. The combination of a wave-induced and tidal

longshore current is discussed in Sect. 5.7.2.
1.4.4. Delta near a river mouth

Figure 1.7 shows in plan view a part of a sandy coast with a river outfall. Waves are
assumed to approach perpendicular to the (initial straight) coastline. The river dis-
charges a volume of water 𝑄𝑟 m3 /s to the sea; at the same time sediment is transported
by the river; say 𝑆𝑟 m3 /yr. 𝑆𝑟 is expressed in m3 /yr since we are looking at large mor-
phological timescales.
Like many rivers all over the world, the river of Fig. 1.7 acts as source of sediments for
the coastal system. It is interesting to understand the interaction processes between
river and sea and to be able to predict the morphological changes with time, as shown
in Fig. 1.7. Such a clear delta coastline develops when the supply of riverine sedi-
ments to the coast is faster than they can be dispersed along the coast by tidal and
wave-generated currents. (See Chs. 2 and 8 for a more detailed discussion of deltaic
coastlines.) The stability of a deltaic coastline is very dependent on the river sediment
supply. If the sediment supply is cut off or reduced by, for instance, the construction of
dams, sand mining or irrigation schemes, the system is deprived of its regular supply
of sediments. This leads to erosion of the coastline on either side of the river mouth
and is quite common in present-day deltaic coastlines.
1.4.5. Tidal inlets and basins

Tidal inlets are openings in the shoreline, for example between two barrier islands
that connect bays or lagoons to the open ocean. They are maintained (viz. kept from
closing naturally) by tidal currents. Figure 1.8 shows various tidal inlets in the Wadden
Sea, the Netherlands. Essential for a tidal inlet is the tidal variation in the open sea;
the tide is the engine that determines most of the features of the inlet and the basin it
connects to.
original coastline
Qr [m3/yr] Qr [m3/yr]
Sr [m3/yr] Sr [m3/yr]
(a) (b)
Figure 1.7: Plan view of a sandy coast with a river outfall in initial situation (a) and with a devel-
oping delta (b).

The tidal range – the difference between High Water Level (HWL) and Low Water
Level (LWL) – and the surface area of the tidal basin together determine, in principle,
the volumes of water that have to flow in and out through the inlet during a tide.
This tidal prism is in some cases in the order of magnitude of one billion (1 × 109 )
m3 . For instance, during one tidal cycle about 109 m3 of water enters the Texel Inlet or
Marsdiep (between Texel and Den Helder, see Fig. 1.8) and leaves the Texel Inlet again.
Figure 1.8: Tidal inlets in the Wadden Sea (the Netherlands). The elevations are taken from
the Vaklodingen dataset (https://publicwiki.deltares.nl/display/OET/Dataset+documentation+
Vaklodingen) and are w.r.t. to NAP.
A typical tidal inlet system consists of several morphological units (see Ch. 9 for more
details):
• The actual entrance or tidal inlet, often dominated by a main channel;
• A shallow ebb-tidal delta seaward of the inlet that often folds around a deep
channel.
• The flood basin, with possibly a distinct flood-tidal delta just landward of the
inlet and an inner tidal basin consisting of the channels that are followed by the
tidal currents and of the lower and higher tidal flats that alternately inundated
and exposed by the tides and possibly covered by salt marshes or mangroves.
The position of the different elements of the tidal inlet system (e.g. ebb-tidal delta,
flood-tidal delta, flood channels, ebb channels, shoals, tidal flats and gorge) changes
over time. Figure 1.9 shows a schematic plan view of a tidal inlet with some typical
notions.

14 1. Overview
η
HW
t
gorge LW
ebb
water flood water

shed tidal basin shed
Figure 1.9: Schematic plan view of a tidal inlet system with the tidal water level variation, outflow-
ing ebb and inflowing flood currents and watersheds. At watersheds, tidal flow through different
inlets meets and tidal currents are zero.
From a morphological point of view, tidal inlets form highly dynamical systems, which
are interlinked with the adjacent coast and the tidal basin or backbarrier area to which
they give access. Often, unnatural constraints (e.g. coastal defence works) and the
effects of human utilisation (e.g. sand mining) interfere with the natural morphody-
namic behaviour. Sometimes measures are taken to restrict the dynamic behaviour of
the morphological system, for instance when it complicates navigation.
1.4.6. Dune erosion and flooding during a severe storm surge

Figure 1.10 shows an instantaneous cross-shore profile (cf. Fig. 1.2). Under normal
conditions the water level changes due to vertical tidal variations. On top of that,
shorter variations due to waves are present. Due to changing water levels and chan-
ging wave characteristics (wave height, wave period and wave direction), cross-shore
sediment transport rates (and cross-shore transport gradients) continuously change in
magnitude and direction. This normal variation was already discussed in Sect. 1.4.2.
20
elevation [m + NAP]
10
MHW = NAP + 0.82 m

0 MSL ≈ NAP + 0.06 m
MLW = NAP – 0.79 m
-10
0 200 400 600 800 1000
cross-shore distance w.r.t. RSP [m]
Figure 1.10: Instantaneous (summer 2011) cross-shore profile at Egmond aan Zee (RSP
7004125). Tidal elevations (see App. C) are calculated for 2018 and assumed static for the
entire JARKUS period of 1965-present (JARKUS, n.d.).
During a severe storm the waves generated in the open sea will be much higher than
normal. Depending on the direction and strength of the wind during the storm and

on the shape of the sea bordering the coast under consideration, water can pile up at
the coastline, raising the Still Water Level (SWL) (without the effect of waves). This
piling up of water is called surge.
During the severe storm surge of January 31st and February 1st 1953 in the Nether-
lands, the water level along the Dutch coast was about 2.5 to 3.0 m higher than nor-
mal. The wind associated with this storm came from north-westerly direction, blew
over the funnel-shaped North Sea and forced the water to pile up against the coasts
of the southern part of the North Sea (see Fig. 1.11). The rather small gap between
United Kingdom and France (The Channel) prevented the raised water from escaping
the North Sea basin. During the 1953 storm surge, many dikes in the southwestern
part of the Netherlands broke. Nearly 1850 lives were lost and there were large eco-
nomical consequences as well. This disaster triggered the execution of the so-called
Delta plan in the Netherlands.
NW
depth [m]
0 150 1000
Figure 1.11: Funnel-shaped North Sea with the wind blowing from the northwest (NW). The darker
the shade the deeper the water. Note the distinction between the shallower shelf region and
the deeper oceanic waters. Bathymetric data from GEBCO (https://www.gebco.net/data_and_
products/gridded_bathymetry_data/).
Figure 1.12 shows some characteristics of the measured and predicted water levels in
Flushing (Vlissingen in Dutch). Note that the storm effect (the surge) is the differ-
ence between the actual measured and the predicted astronomical (tidal) water level
variation. Note also that the surge lasted in fact a rather small period of time; within
two days the storm effect rose from zero to approximately 2.8 m and fell down to zero
again.

16 1. Overview
4
recorded level
elevation [m + NAP]
3 predicted level
-1
-2
2
meteorological
1
effect [m]
-1
/20
/20
/20
/20
2/2
2/2
/0 2
/02
/0 2
/0 2
/0
/0
09
10
11
12
13
14
5
recorded level
4 predicted level
elevation [m + NAP]
-1
-2
Vlissingen
storm surge [m]
0
3
/53
3
1/5
2/5
2/5
/02
/0
/0
/0
31
01
02
03
Figure 1.12: Measured and predicted water levels at Vlissingen (the Netherlands). The upper
panel displays the predicted astronomical tide (grey line) and the actual measured water levels
(black line) during a storm in March 2020 (https://waterinfo.rws.nl/). The difference between
these two is the meteorological effect (second panel) and consists of both the storm surge and
the interaction between the tide and the storm surge (see (Bijlsma et al., 1989)). The third panel
shows the astronomical tide and recorded water levels during the disastrous storm of 1953.
The fourth panel shows the corresponding storm surge only, which is not exactly equal to the
difference between predicted and measured water levels, because of the effect of interaction
between the tide and the storm surge.

In Fig. 1.13 the same cross-section as in Fig. 1.10 has been sketched, but now under
maximum storm surge conditions. Not only is the SWL much higher than in Fig. 1.10,
but also much higher waves are present. The SWL even exceeds the level of the dune
foot; all the beaches have disappeared under water and the waves hit the dunes. It can
be argued that the shape of the cross-shore profile does not correspond to the storm
surge conditions (the profile shape is far out of equilibrium for these conditions). Large
offshore-directed sediment transports ensure that the profile shape is transformed to-
wards the equilibrium shape associated with the storm conditions. The dunes erode
and the eroded dune material settles in deeper water, where the cross-shore profile
gradually flattens (see Fig. 1.13; cf. the winter profile of Fig. 1.3).
20
pre-storm
15 post-storm
elevation [m +NAP]
erosion
10 deposition
SSL = MSL + 5 m
5
MSL ≈ NAP + 0.06 m

0
-5
-100 0 100 200 300 400 500 600
Figure 1.13: Illustration of a storm impact on the cross-shore profile. The pre-storm profile
(dashed line) is a typical summer profile (summer 2015) for the Dutch coast, averaged over
a 3 km long stretch near Zandvoort (data from JARKUS, n.d.). The post-storm profile (solid
black line) as a consequence of a storm surge with Storm Surge Level (SSL) and storm wave
conditions is estimated after (Vellinga, 1986), see Fig. 7.18 for details.
In Ch. 7 a more detailed description is given of the associated processes. For the time
being it is sufficient to understand that a coastal zone manager likes to know what the
loss of dune area will be under a given set of storm surge conditions, e.g. to assess the
safety of properties built close to the brink of the dunes. In the Netherlands, further-
more, the safety of a large part of the population, living well below MSL behind the
dunes, is at stake if a break-through of the dunes occurs.
1.4.7. Large artificial island in open sea

In densely populated areas (e.g. Japan, Taiwan or the Netherlands) it is becoming
increasingly complicated to find large open areas on the mainland to start new devel-
opments (e.g. for a new airport, hosting new industries or even housing the growing
population).
The open sea in front of the existing coast may be used to build an artificial island. In
the Netherlands, for instance, there has been ongoing discussion whether it is useful
to build an artificial island off the coast near IJmuiden for hosting an extension of
Schiphol Airport. A specific airport-island requires rather large dimensions (amongst

18 1. Overview
others due to the length of runways). A typical size for such an island is 5 km. With
respect to the distance of an island from the existing coast, an optimum has to be found
taking into account transportation requirements, noise limitations and morphological
implications.
An artificial island in open sea has large morphological implications for the existing
coasts and affects the stability of coasts in a wide area. In the shadow area behind
the island, the wave characteristics will fundamentally change. Another aspect of an
airport-island is that the (tidal) current patterns will be affected in the vicinity of the
island. The adaptation time of the existing coastal area is generally relatively long.
The water depth off the coast of IJmuiden, where such an island would probably be
built, is approximately 15 m. Because of this depth, the required dimensions of an
island and the required level above MSL (approximately 5 m) huge volumes of sand
are required to construct such an island (order of magnitude 600 million m3 of sand.)
These volumes are easily available from the bed material of the North Sea, but call for
large borrow pits which also impact the morphology.
During dredging operations fine-grained silt (see Sect. 2.6.2) will also be mobilised,
since within the sand deposits at the seabed often small volumes of silt occur (say 2 %).
With a required volume of 600 million m3 , with 2 % silt, this results in 12 million m3 of
silt that is mobilised in the North Sea environment. The associated turbidity can have
large ecological effects.
It is clear that in the final decision whether to build an artificial island or not, coastal
morphology topics have to play a role.
1.4.8. Other examples

A few examples relevant to coastal engineering practice have been discussed in this
section. Other examples are:
• Siltation of (dredged) navigation channels;
• Erosion near the toe of breakwaters, seawalls or offshore structures;
• Structural or gradual erosion of coasts;
• Impact of coastal protection tools, like groynes, offshore breakwaters and sea-
walls and revetments;
• Behaviour of artificial beach and shoreface nourishments;
• Blockage of sediment supply by a river.
In all cases sediment transports due to waves and currents are important. Compared
to sediment transports due to currents alone (like in rivers), the waves enhance the
sediment transport and make the calculation of the transport rates more complicated.
In Ch. 6, sediment transport due to waves and currents is discussed in detail.

1.5. Coastal (morpho)dynamics 19
1.5. Coastal (morpho)dynamics

1.5.1. Definition of the coast
So far we have assumed that it was clear what we meant by ‘coast’. The definition of
coast however depends on the objective and the timescale under consideration. Coasts
are the transition zones between oceans and continents. The coastal zone is made up
of:
• the part of the land that is affected by being near to the ocean (coastlands); and
• the part of the oceans that is affected by being near to the land (coastal waters).
The coastlands encompass all terrain features that are influenced by coastal processes,
like dunes, cliffs and low-lying areas (coastal plains). In a practical situation, the in-
land extent depends on the timescale under consideration. A coastal engineer, who is
mostly concerned about timescales of years, would say that the coast extends inland as
far as the influence of the tides and storm waves reaches. Although for a large estuary
the limit of inland tidal propagation can be hundreds of kilometres, this definition of
the landward extent is relatively narrow in the eyes of a geologist. A geologist would
be aware of the fact that in former times the sea has reached higher levels than the
present levels and might find evidence of that in the form of coastal deposits far from
the present influence of the sea. Therefore, a geologist would include all these areas in
a definition of the coast. Half of the Netherlands consisting of polders below sea level
would then belong to the coast.
Similarly, the seaward limit of the coast is dependent on the timescale under consid-
eration. Engineers have introduced the so-called depth of closure or closure depth as
the most seaward point of interest. The depth of the beach profile closure is the depth
beyond which repeated field observations over a certain period of time show no signi-
ficant changes in bed height. It can be empirically determined by examining a series
of profile measurements over a period of months to years so that both calm periods
and storm conditions are included (Fig. 1.14).
The beach profile envelope in Fig. 1.14 includes typical accretional profiles built up
during months of moderate wave conditions and typical erosional profiles under the
influence of large storm waves. It shows that the profile is divided into an inactive
offshore and a more active nearshore zone called shoreface. The shoreface is the part
of the sandy profile affected by wave action and typically extends to water depths
of 10 m to 20 m. The shoreface can further be separated into the lower shoreface (or
shoal zone) where waves gain amplitude up to the point of breaking, and the most
active zone where waves are breaking and the majority of the changes takes place
(upper shoreface or littoral zone).
When considering larger timescales, we have to take into account that coastal pro-
cesses have reached up to the continental shelf (Fig. 1.15) when sea levels were lower

20 1. Overview
SSL
MHWS
MSL
MLWS
seasonal range
of sand level
shelf lower shoreface upper shoreface (littoral zone)
Figure 1.14: Envelope of beach profiles measured at different times, over a period of for instance
a year. The two depth limits 𝑑𝑖 and 𝑑𝑙 correspond to the closure depth definition of Hallermeier
(1978, 1981), see also Sect. 7.2.3. The profile is dynamic landward of the outer depth limit 𝑑𝑖 .
The majority of the bed dynamics takes place at depths smaller than the inner depth limit or
annual closure depth 𝑑𝑙 , where 𝑑𝑙 is the maximum water depth for nearshore erosion by extreme
conditions exceeded for twelve hours per year. The tidal levels Mean High Water Spring (MHWS)
and Mean Low Water Spring (MLWS) are explained in App. C.
than present. The continental shelf is the edge of a continent and is covered with
relatively shallow seas up to 100 m to 200 m water depth (up to the shelf break).
Quaternary coastal morphology

continental
slope continental shelf shoreface coastal plain
Quaternary sediments
shelf break barrier estuary
waves start ‘feeling’ bottom
Figure 1.15: Spatial boundaries of the coastal zone. The Quaternary coastal morphology reaches
up to the shelf break. The transition between the shelf and the shoreface is located where waves
propagating towards the shore start feeling the bottom.
In the broadest sense of the word, we can now characterise coastal systems as relative
shallow areas (i.e. depths less than order 100 m) bordered by or partly enclosed by
land that are influenced by the sea and connected to the oceans, and in which ocean
disturbances propagate. Due to the intensive interaction between land and water at
the interfaces of the two, a large variety of coastal ocean systems has developed. The
landward side includes the partially enclosed basins (such as the estuaries of Fig. 1.16b).

1.5.2. Coastal morphodynamics

In Sect. 1.4.3 we already concluded from continuity considerations that coastal changes
occur in the case of gradients in sediment transport rates. This of course not only holds
for the coastal stretch considered there, but for coastal systems in general. Consider
for instance the sediment budgets along the estuarine and deltaic coasts of Fig. 1.16.
Changes in the morphology of these systems depend on the spatial and temporal fluc-
tuations in the sediment transport rates. In terms of a continuity equation or mass
balance:
𝜕𝑧𝑏 𝜕𝑆𝑥 𝜕𝑆𝑦

+ + =𝑉 (1.1)
𝜕𝑡 𝜕𝑥 𝜕𝑦
where:
𝑧𝑏 (𝑥, 𝑦, 𝑡) bed level above a certain horizontal datum m

𝑆𝑥 (𝑥, 𝑦, 𝑡), 𝑆𝑦 (𝑥, 𝑦, 𝑡) sediment transport rates per m width of flow m3 /m/s
in the horizontal 𝑥- and 𝑦- direction,
including the effect of porosity
𝑉 (𝑥, 𝑦, 𝑡) sink or source term per unit area m3 /m2 /s
representing local sediment gains and losses,
often taken as zero
If the net sediment flux into a certain area is negative, meaning that the outgoing
sediment flux is larger than the incoming one, the bottom will supply the sediment
deficit (we assume that the bottom is erodible). In that case a lowering of the bot-
tom occurs (erosion). However, while the morphology changes, the waves and tides
– being dependent on the water depth – respond to the adjusted bed level. As a res-
ult the sediment transport rates change and this again affects the development of the
morphology. Apparently, a feedback (named morphodynamics) exists between hy-
drodynamic processes and morphology. The coupling between the two is provided by
sediment transport:
Coastal morphodynamics is the mutual adjustment of morphology and hydro-

dynamic processes involving sediment transport.
The morphodynamic feedback can be positive or negative. In the case of negative

feedback the adjustment process continues until a new situation is reached where no
changes occur. An example is the response of a straight coast with normally incident
waves to a beach nourishment (see also Example 8.4). Waves will disperse the sediment
until eventually the coastline orientation is such that the transport gradients are zero
again (hence, the end result is a straight coast again). Negative feedback thus is a
stabilizing process that makes sure that after a disturbance a new equilibrium develops.

22 1. Overview
inputs outputs
t
pu
out
output
ut
inp
rwash
ove
ut
in p ut
inp
ach
ss ive be
re
cliff prog ge plain
erosion rid
(a)
outputs inputs
input output
output
storm
cut
rier
bar ash
rw
ove
sive
rocky s g res plain
headland tran ridge
ch
bea
t
u
inp
(b)
Figure 1.16: Sediment budgets along (a) deltaic and (b) estuarine coasts. In the example, the
deltaic coast gains sediment; the sum of the inputs is larger than the sum of the outputs. This
is reversed for the estuarine coast that loses sediment.
Note that in practice, an equilibrium is never static, since the external conditions are
also changing during the adaptation process.
Positive feedback is exactly the opposite from negative feedback in that the system is
pushed away from equilibrium. An example is a small disturbance of the bed, which
itself generates the sediment transport convergence that makes sure that the disturb-
ance grows larger and larger, forming a larger-scale feature, such as a shoal.

1.5.3. Time- and spatial scales

As we have seen in the previous two sections, the behaviour of a natural coastal system
is dynamic on a variety of time- and spatial scales. The spatial scale is generally de-
termined by the dimensions (in m) of a particular morphological element; it indicates
the extent of the element. Examples of spatial scales (see also Fig. 1.17) are:
• A whole tidal inlet system, comprising a flood basin, an inlet gorge and an ebb
or outer delta. Dimensions vary between 50 to 700 km2 . Morphological changes
on such a large spatial scale generally take decades to centuries (e.g. the gradual
migration of the entire inlet system);
• Large tidal channels and sand banks. Typical surface dimensions are 5 to 20 km2 ,
while significant changes generally take place within years to decades (such as
the landing of the sand bank ‘De Onrust’ on the south coast of Texel around
1910);
• Smaller tidal channels and bars, with typical surface dimensions of no more than
a few km2 . Significant morphological changes occur within years (such as the
development and landing of the ‘Bornrif’ sand bar on the west coast of Ameland
in the nineties of the last century);
• Accretion and erosion patterns at both sides of a port entrance, built with the
help of long breakwaters. The spatial scale is several kilometres;
• Smaller bed forms like ripples and (bottom) dunes. The related spatial scales are
a few metres or less.
10 8
oceans
geology
continents
10 6 seas
catchment basins
coastal deltas
sandbanks
spatial scale [m]
estuaries
10 4 channels
tidal flats
morphology
bars beach
10 2 rips
dunes
cusps
sedimentology
10 0
ripples
10 -2
10 -2 10 0 10 2 10 4 10 6 10 8 10 10 10 12
hour day month year cent. mill.

temporal scale [days]
Figure 1.17: Coastal phenomena span a large range of time- and spatial scales, with time- and
spatial scales being closely related. For this figure, we followed the categorisation of coastal
phenomena by Dronkers (2005).

24 1. Overview
Figure 1.17 suggests that spatial and timescales are closely coupled; the larger the
spatial scale of a certain feature, the larger the timescale. As an example, smaller bed
forms (as mentioned in the last bullet above) not only have small spatial scales but also
small timescales; the time period in which significant changes occur is a few days or
less.
Examples of morphological timescales are given in Table 1.1. These examples are re-
lated to the observed morphological changes of the outer deltas of several Dutch tidal
inlet systems after abrupt changes in their flood basins. The figures support the gen-
eral idea that morphological timescales are directly coupled to the spatial scales of the
morphological units.
Table 1.1: Examples of morphological timescales.
Morphological unit Basin that is Surface area of flood basin Observed

abruptly closed after closure [km2 ] morphological
timescale [year]
Outer delta Haringvliet Haringvliet 120 11
(1970)
Zoutkamperlaag Lauwerszee 200 17
(1969)
Outer delta Texel Inlet Zuiderzee 680 32
(1932)
A timescale is generally interpreted as the period of time (e.g. in years) required for
typical morphological developments. Some researchers relate the timescale to the total
duration that a morphological system needs to reach a new equilibrium situation once
it has been disturbed by nature or by man (see Sect. 1.5.4). Equilibrium in this sense
means that the overall sediment balance of the morphological unit is maintained (no
erosion and no accretion).
A coastal engineer would generally not be interested in dynamics on very small scales
like the evolution of wave ripples on the seabed or even breaker bars in the surf zone
(see Sect. 1.4.2). The dynamics on these scales are generally oscillatory, which means
that they have no net effect on engineering timescales. By contrast, the natural beha-
viour on larger (human/engineering) scales generally shows a net trend. Examples are
depth reduction of an estuary due to infilling by sediments, the erosion or accretion
of a delta, and the migration of a tidal inlet. Human interventions have timescales
of years to decades and spatial scales of 1 km to 100 km. The coastal behaviour due
to human interventions (harbours, coastal defence works, land reclamation) interferes
with the natural behaviour of the coastal system on these time- and spatial scales.
Long-term sea level change acts on timescales larger than engineering scales (order of
ten thousand years). This long-term tendency needs to be taken into account in the
analysis on the engineering scale of the system; it provides boundary conditions for

processes and system evolution on the engineering scale. Coastal evolution on geolo-
gical timescales (such as plate tectonics) is responsible for the large-scale geographical
variation of coastal systems that we find around the globe nowadays. These large-scale
characteristics can be considered a given for any analysis on engineering timescales.
In Ch. 2 we discuss this distribution of coastal environments and – briefly – its origin.
1.5.4. Equilibrium concept

If a morphological system is not in equilibrium with the forcing (waves, tides, currents),
morphological adjustments start to take place immediately; the morphological system
reacts to disturbances. The rate of morphological adjustment has been observed to
depend on the magnitude of the still-existing disruption (the difference between the
actual situation and the equilibrium situation).
Sometimes, morphological changes are induced very abruptly, such as the closure of
parts of a tidal basin. Other changes take place more slowly, such as the response of the
shape of a cross-shore profile to sea level rise. Usually, the morphological response to
(sudden) changes shows a certain variation with time. The process of morphological
response will be fast at first and decelerates when the new equilibrium situation is
approached. Often such a morphological adjustment process can be approximated
with an exponential function. In terms of for instance sediment volume content of a
certain morphological unit we would then have:
𝑉 (𝑡) = 𝑉old + (𝑉new − 𝑉old ) (1 − 𝑒 −𝑡/𝜏 ) (1.2)
where:
𝑉 (𝑡) characteristic volume in morphological unit at time t m3

𝑡 time after the distortion yr
𝑉old the (equilibrium) volume before distortion m3
𝑉new the new equilibrium volume m3
𝜏 morphological timescale yr
E.g. Stive and De Vriend (1995) and Eysink (1991) use this type of approximation to
describe adaptation processes of units with large scales. Within a time 𝑡 equal to the
morphological timescale 𝜏 a good 63 % (namely: 1 − 1/𝑒) of the changes required to
reach new equilibrium (namely: 𝑉new − 𝑉old ) have taken place. The morphological
timescale 𝜏 is the time that would be required to reach equilibrium, if the rate of mor-
phological adjustment 𝑑𝑉 /𝑑𝑡 would remain equal to the rate at 𝑡 = 0:
(𝑉new − 𝑉old )
𝜏= (1.3)
(𝑑𝑉 /𝑑𝑡)𝑡=0
Equation 1.3 can be found by differentiation of Eq. 1.2 and evaluating the resulting
equation at 𝑡 = 0. Sediment transports are supposed to drive the morphological unit

26 1. Overview
to its new equilibrium and are therefore implicitly known. Such an implicit approach
to determine sediment transport rates based on the deviation from a predefined equi-
librium is very different from a more process-based approach to describe sediment
transport. The latter approach will be taken in Ch. 6.
1.5.5. Classification of coastal systems

Due to the intensive interaction between land and water, a large variety of coastal
ocean systems has developed. Many classifications of coastal systems have been pro-
posed that try to order the huge variety into classes with similar characteristics. In
fact, so many classifications have been proposed that it seems hard to make sense of
the large assortment of classifications. We will therefore not attempt to give an over-
view of all classification schemes, but focus instead on the points of similarity between
various classifications.
In order to characterise a given coast we need to describe a minimum of four terms:
1. Geological factors:
Any coast is the result of slow geological processes (like mountain formation)
that require millions of years and result in a certain initial state of the solid
boundaries (considered as given by a coastal engineer);
2. Nature and abundance of coastal ‘material’:
Is the material hard (rocks, coral) and/or soft? Soft material (mud, sand, gravel,
cobbles and carbonate sands) is present in depositional coastal features like
deltas, beaches and mud flats. They can host vegetation such as mangroves,
salt marshes, and dune vegetation;
3. Transgression or regression:
The gradual relative sea level changes (rise or fall) that have timescales of thou-
sands of years determine – in combination with the amount of sediment supply
to the coast – whether a coast during a certain period of time has advanced
(for instance a delta that has built out) or retreated (a drowned river valley). In
this terminology, transgression and regression are used to describe a horizontal
shift of the waterline. Hence, regression (of the sea) is equivalent to advance
of the coast. Similarly, transgression (of the sea) implies retreat of the coast.
Besides global effects there are regional and local effects (for instance bottom
subsidence);
4. Processes that construct and erode the coast:
Processes affecting sandy and rocky coasts are for instance wind and hydraulic
forcing by waves and tides; for coral coasts and in mangrove environments or
salt marshes chemical and biological processes are important. Waves and tides
give rise to significantly different shapes of depositional features consisting of
sand, mud and gravel. On a global scale these processes are affected by latitude
and climate, whereas on a local scale they are dependent on local bathymetry.

1.6. Important parties in the Netherlands 27
Coastal classifications are typically based on one or more of the above four identifi-
ers and are dependent on the scale: are we considering an entire continent or a small
section of a coastline? Are we looking at developments over thousands of years or
at smaller timescales? On regional (tens to hundreds of kilometres) and local scales
(only a few kilometres) coastal features are dependent on the forcing by processes such
as waves, tides and wind. The present book takes a mainly process-based approach
and hence focuses – from Ch. 3 onwards – on regional- and local-scale features. Un-
derlying these smaller-scale features are broader (or first-order) coastal features. The
latter cover large geographical distances (thousands of kilometres) and are linked to
the long-term geological process of plate tectonics and influenced by climate. On long
timescales and associated spatial scales, coastal development is also connected with
the expansion and retraction of ice-sheets and associated sea level changes.3 History’s
legacy to coasts is described in Ch. 2.
1.6. Important parties in the Netherlands

All over the world, many governmental and non-governmental organisations, insti-
tutes, universities, consultants and contractors are active in the field of coastal dy-
namics. Countries in the world in which coastal engineering and coastal morphology
receive significant attention are amongst others:
• Australia;
• Belgium;
• Brazil;
• China;
• Denmark;
• France;
• Germany;
• Italy;
• Japan;
• Singapore;
• South Africa;
• Spain;
• The Netherlands;
• UK;
• USA.
In the Netherlands coastal engineering has always been a very important topic for
the simple reason that the ground level of large parts of the Netherlands is below
MSL; see Fig. 1.18. Without dikes and dunes, people would not be able to live in these
parts of the country. The Dutch governmental organisations, consultants, contractors,
3
During the Quaternary (from 1.8 million years ago till present) the sea level fluctuated over more than
100 m vertically.

28 1. Overview
research institutes and universities that are active in the field of coastal engineering
and management are described below.
Amsterdam
Rotterdam
elevation [m + NAP]
< 0 (44%)
0 — +7 (17%)
> +7 (39%)
Figure 1.18: Elevation map of the Netherlands, illustrating the areas at risk of flooding. The
figure is generated using the SRTM30 dataset, which is comprised of a Shuttle Radar Topography
Mission (SRTM) flown in February 2000 (Farr & Kobrick, 2000) and the U.S. Geological Survey’s
GTOPO30 dataset. Data available on https://www.diva-gis.org/datadown.
Government
Various parts of the Directorate-General of Rijkswaterstaat (RWS) and the Directorate-
General Water, which are parts of the Ministry of Transport, Public Works and Water
Management, are active in the field of coastal engineering and management.
• Waterdienst Lelystad;
• Directorate North Sea;
• Regional Directorates of RWS.
The various Provinces bordering the North Sea, are – mainly at coordination level –
also involved in coastal zone management. Also, the regional Water Boards play an
important role in coastal zone management in the Netherlands.
Consulting companies/contractors
Many Dutch consulting companies and contractors are active in the field of coastal
engineering and management, either in the design phase or in the execution of specific
works. A non-exhaustive list of companies is (in alphabetical order):
• Arcadis;

1.6. Important parties in the Netherlands 29
• BAM;
• Boskalis;
• CDR International;
• Lievense | WSP;
• Royal HaskoningDHV;
• Svašek Hydraulics;
• Van Oord;
• Witteveen and Bos.
Research institutes
Dutch research institutes are positioned in between consulting activities and academic,
more fundamental research. Institutes active in the field of coastal engineering and
coastal (eco-)morphology are:
• NIOZ – The Netherlands Institute for Sea Research (Nederlands Instituut voor
Onderzoek der Zee);
• Deltares – an independent institute for Delta Technology (formed in 2008 out
of a merger between Delft Hydraulics, GeoDelft, parts of TNO and parts of Rijk-
swaterstaat).
Universities
The four Dutch universities strongly involved in education and research related to
coastal engineering and coastal morphology are:
1. TU Delft – Delft University of Technology;

2. UU – Utrecht University;
3. UT – Twente University;
4. WUR – Wageningen University.
In addition, the IHE Institute for Water Education offers hydraulic engineering educa-
tion to practising professionals from developing countries.
Research and management cooperation in the Netherlands

• NCK – Netherlands Centre for Coastal Research, a cooperation between TU
Delft, UU, UT, RWS-Water, TNO-NiTG, Deltares, NIOZ and the Netherlands
Oceanographic Institute;
• ENW – Expertise Network Water Defences (Expertise Netwerk Waterkeringen).

30 1. Overview
1.7. References
1.7.1. Lecture notes
These lecture notes are primarily developed as a student handbook and secondarily as
a reference book for getting acquainted with the field of coastal dynamics and coastal
engineering. More in-depth knowledge can be found in the overwhelming amount of
coastal engineering literature. At several places in the lecture notes reference is made
to selected literature. This is done by giving the author(s) and year of publication in
the main text, whereas the full reference is given in the bibliography at the end of this
book. It is of course recommended to consult other literature also. Below an overview
is given of relevant text books, journals and conference proceedings.
1.7.2. Textbooks
Background information on coastal engineering and coastal morphology is provided
in a number of handbooks that each have their own focus and approach. Several are
mentioned here, in alphabetical order:
• Dean, R. G. (2002). Beach Nourishment: Theory and Practice. (Vol. 18: Advanced
Series on Ocean Engineering). World Scientific Publishing
• Dean, R. G. & Dalrymple, R. A. (2004). Coastal processes with engineering applic-
ations. Cambridge University Press
• Dronkers, J. (2005). Dynamics of Coastal Systems (Vol. 25). World Scientific
• Fredsøe, J. & Deigaard, R. (1992). Mechanics of Coastal Sediment Transport. (Vol. 3).
World Scientific, Singapore
• Kamphuis, J. W. (2000). Introduction to Coastal Engineering and Management
(P. L.-F. Liu, Ed.; Vol. 16). World Scientific
• Komar, P. D. (1998). Beach Processes and Sedimentation (Second). Prentice-Hall,
Upper Saddle River, New Jersey
• Masselink, G. & Hughes, M. G. (2003). Introduction to Coastal Processes and Geo-
morphology. Hodder Arnold, London
• Nielsen, P. (1992). Coastal Bottom Boundary Layers and Sediment Transport. World
Scientific. https://doi.org/10.1142/1269
• Nielsen, P. (2009). Coastal and Estuarine Processes. World Scientific. https://doi.
org/10.1142/7114
• Soulsby, R. L. (1997). Dynamics of marine sands: a manual for practical applica-
tions. Thomas Telford, London
• Van Rijn, L. C. (1999). Principles of coastal morphology. Aqua Publications, Am-
sterdam, the Netherlands.
• Whitehouse, R. J. S. (1998). Scour at marine structures: a manual for practical
applications. Thomas Telford, London

1.7. References 31
1.7.3. Internet sources

• For translation of terminology in six languages consult waterdictionary.info;
• The Coastal Wiki (coastalwiki.org) is a fast-expanding source of information re-
garding coastal morphology, coastal engineering and coastal zone management;
• The Coastal and Hydraulics laboratory of the US Army Corps of Engineers
(USACE) gives on-line access to the Coastal Engineering Manual (CEM) – type
‘Coastal Engineering Manual’ in a search engine like Google:
“Coastal Engineering Manual (CEM)4 provides a single, comprehens-

ive technical document that incorporates tools and procedures to plan,
design, construct, and maintain coastal projects. This engineering
manual will include the basic principles of coastal processes, meth-
ods for computing coastal planning and design parameters, and guid-
ance on how to formulate and conduct studies in support of coastal
flooding, shore protection, and navigation projects. New sections are
being added on navigation and harbour design, dredging and disposal,
structure repair and rehabilitation, wetland and low-energy shore
protection, risk analysis, field instrumentation, numerical simulation,
the engineering process, and other topics.”
1.7.4. Interesting journals

Many coastal engineering topics are discussed in scientific journals. Examples of such
(mostly peer-reviewed) international journals are:
• Coastal Engineering. An international journal for coastal, harbour and offshore

engineers, Elsevier Science;
• Journal of Waterway, Port, Coastal and Ocean Engineering. American Society
of Civil Engineers (ASCE) Publications;
• Coastal Engineering Journal. World Scientific Publishing;
• Journal of Geophysical Research – Oceans. American Geophysical Union (AGU)
Publications;
• Ocean & Coastal Management. Elsevier Science.
• Journal of Coastal Research. Coastal Education and Research Foundation;
• Shore & Beach. American Shore and Beach Preservation Association (ASBPA);
• The Open Access Journal Water. Multidisciplinary Digital Publishing Institute
(MDPI).
4
The CEM used to be called Shore Protection Manual (Shore Protection Manual. 1984).

32 1. Overview
1.7.5. Conference proceedings

In the field of coastal engineering many conferences are organised where recent work
is first presented and subsequently collected in conference Proceedings. Four major
conferences are discussed below.
International Conference on Coastal Engineering

The International Conference on Coastal Engineering (ICCE) is the world’s premier
forum on coastal engineering and related sciences. The Institution of Civil Engin-
eers and the Coastal Engineering Research Council (CERC) of the American Society
of Civil Engineers (ASCE) organise this bi-annual conference on theory, measurement,
analysis, modelling and practice. As an illustration, the invitation brochure of ICCE
2004 mentions the following topics:
• Coastal Processes and Climate Change. Oceanography, meteorology, morpho-

dynamics and sediment processes, macro- and micro-tidal regimes, extreme events,
coastal waves, effects on coastal management;
• Flood & Coastal Defence Engineering and Management. Beach management
and nourishment, coastal and beach control structures, construction techniques
and performance;
• Flood Risk Management. Strategic planning, flood warning, forecasting and
coastal change monitoring, data management and exchange, risk and uncer-
tainty, decision making;
• Coastal Environment. Recreation, industrial activity, water quality, wetlands
and estuaries, sustainability, environmental economics;
• Ports and Harbours. Siltation, dredging and dredged material re-use, naviga-
tion channels, optimisation, wave-structure interactions, breakwater monitor-
ing, coastal interactions;
• Coastal Legislation, Planning and Cooperation. Government policy, funding,
collaborative projects, integrated coastal zone management, international co-
operation and conventions, law enforcement, effects of coastal hazards on land
use planning.
Coastal Sediments
Coastal Sediments is a multi-disciplinary international conference convened for re-
searchers and practitioners to discuss science and engineering issues of coastal sedi-
ment processes. The conference is organised every fourth year by the Committee on
Coastal Engineering of the Waterway, Port, Coastal and Ocean Division of the Amer-
ican Society of Civil Engineers. The conference provides a high-level technical forum
for the exchange of information on coastal engineering, geology, oceanography, met-
eorology, physical oceanography, and biology.

1.7. References 33
Coastal Dynamics
The Coastal Dynamics conferences are held under the auspices of the ASCE every four
years and are technical speciality conferences bringing together field and laboratory
experimentalists, theoreticians and modellers conducting research on coastal hydro-
dynamics and sediment transport. The proceedings of the multi-disciplinary confer-
ence are of interest to coastal engineers, coastal geologists, oceanographers, and re-
lated sciences.
Conference on Coastal and Port Engineering in Developing Countries

The mission of the Conference on Coastal and Port Engineering in Developing Coun-
tries (COPEDEC) conferences was originally to provide an international forum where
coastal and port engineers from developing countries can exchange know-how and ex-
perience amongst themselves and with their colleagues from industrialised countries.
In 1999, this was expanded to the following mission: To enable developing countries
to have a sustainable human resources pool of highly skilled coastal and port devel-
opment professionals. At the conference, which is held every four years, papers are
presented on many subjects with special reference to needs in developing countries;
viz.:
• Port and Harbour Infrastructure Engineering in Developing Countries. Port in-

frastructure design: choice of structures, design methods and techniques. Port
construction: choice of materials, dredging and construction techniques. Port
renovation: renovation and demolition techniques;
• Port Infrastructure Planning and Management in Developing Countries. Port
planning: economic forecasts, site selection, layout and nautical aspects. Eco-
nomic aspects: BOT, PPP, privatisation, containerisation. Operations and main-
tenance: performance, safety, management;
• Coastal Sediments and Hydrodynamics. Coastal stability, beach erosion, con-
trol and nourishment. Sedimentation, maintenance dredging of harbour basins
and approach channels. Waves, currents and tides: field survey and measuring
techniques;
• Coastal Zone Management in Developing Countries. Integrated coastal plan-
ning: development, implementation, evaluation. Impacts of coastal use: fish-
eries, infrastructure, tourism, recreation. Policy, regulations and guidelines for
coastal zone management;
• Coastal and Port Environmental Aspects. Environmental Impact Assessment:
pollution control and treatment. Sediment and dredging materials: isation, treat-
ment, disposal. Waste management: reception facilities, prevention, treatment.

35
2
Large-scale geographical variation of
coasts
2.1. Introduction
Present-day coasts show the imprint of both present-day and past processes. Coastal
morphology is thus partly inherited from the past. Section 2.2 introduces the long
timescales of geological processes and sea level changes and the concept of inheritance.
In Ch. 1 it was mentioned that the characterisation of a coastal system is dependent
on the scale that we consider. The broadest (or first-order) features of the coast cover
large geographical distances (thousands of kilometres) and are linked to the long-term
geological process of plate tectonics. Plate tectonic theory and the consequences for
coastal systems are discussed in Sect. 2.3.
Many of the geomorphologic features shaped or deposited during the Quaternary –
consisting of the Pleistocene and Holocene – are still clearly recognisable at present.
These coastal features are superimposed on the pre-existing geology that is controlled
by plate tectonics. The Pleistocene legacy of rocky coasts is briefly treated in Sect. 2.4,
whereas Sect. 2.5 discusses the effect of Holocene sea level changes in combination
with availability (supply or loss) of sediment.
Section 2.6 deals with large-scale variations in nature and abundance of coastal mater-
ial. This will be shown to be coupled to geological controls as tectonic plate setting
and glacial action, and to climate. The relevant climatic effects, such as global wave
height distribution and discharge of the world’s largest rivers are therefore also briefly
treated in Sect. 2.6. More specific information on global variation of wind, wave and
tidal characteristics and their effects on coastal developments are treated in Ch. 4.
On regional (tens to hundreds of kilometres) and local scales (only a few kilometres)
second- and third-order features become noticeable. In order to describe these regional

36 2. Large-scale geographical variation of coasts
and local features, a process-based approach will be introduced in Sect. 2.7, which will
be followed in the remainder of these lecture notes.
2.2. Cumulative evolution of coastal systems

2.2.1. Geological timescale
When we describe the physical history of the earth, we enter the realm of geology.
Geology has produced a timescale that covers the roughly 4 to 5 billion years since the
formation of planet earth. The geological timescale is divided in some main periods
(geological eras) and subdivided in a much larger number of sub-periods (called periods
and epochs). Since the more recent history is known in far greater detail than the initial
stages of development, the sub-periods that can be distinguished become progressively
shorter. Table 2.1 shows the more recent eras and periods.
The Quaternary is the present geological period and consists of the Pleistocene and
the Holocene. The old names for the Pleistocene and Holocene are Diluvium and
Alluvium respectively. These are the epochs of most concern to coastal engineers,
extending back a total of 1.8 million years before present.
Within the period encompassed by geological history, two aspects deserve special at-
tention in this coastal engineering textbook. First, the slow process of continental
separation that started 200 million years ago has greatly impacted the formation of
coastlines (Sect. 2.2.2). Second, the expansion and retraction of ice sheets and associ-
ated sea level changes1 has strongly determined coastal development during the Qua-
ternary (Sect. 2.3.2). Note that the modern continents essentially reached their present
positions during the Quaternary, having moved no more than 100 km relative to each
other since the beginning of the Quaternary.
2.2.2. Continental ‘drift’

Around 200 million years ago the world’s continents formed a primordial super-conti-
nent, called Pangaea (Greek for ‘all earth’). The continuity of geologic features across
the now widely separated continents supports this idea. The continental land masses
that formed Pangaea gradually drifted from their original positions (see Fig. 2.1). They
reached intermediate locations 135 million years ago, between the Jurassic and Creta-
ceous Periods. After almost 200 million years, the continents reached their present
positions, though we can observe that they are still drifting. Nowadays, it is known
that even before the formation of Pangaea, the continents were already drifting; there
have been a number of cycles of continental break-up, drift, and collision, each lasting
a few hundred million years.
1
During the Quaternary the sea level fluctuated over more than 100 m vertically.

2.2. Cumulative evolution of coastal systems 37
Table 2.1: Geological timescale with focus on the last 250 million years.
Era Period Epoch Time before present
Holocene
Quaternary 11,700 years
Africa, America separated

Pleistocene
(Ice age)
1.8 million years
extinction of dinosaurs
Pliocene
Cenozoic
Neogene
5.3 million years

Miocene
Tertiary
23.8 million years

Oligocene
Paleogene
33.7 million years

Eocene
54.8 million years
Paleocene
65 million years
Cretaceous
start of continent
Mesozoic
144 million years
separation
Jurassic
206 million years formation of planet

Triassic
±200 million years
248 million years
...
4.75 billion years
The hypothesis that continents ‘drift’ was fully developed by Wegener in the beginning
of the 20th century (Wegener, 1912, 1929). However, it was not until the development
of the theory of plate tectonics in the 1960s that a sufficient geological explanation
for that movement was found (see Sect. 2.3.1). This process of plate tectonics has had
an enormous impact on the formation of coastlines and determines the broadest fea-
tures of the coast. Important inherited aspects are the continental shelf configuration
(mainly width and slope of the shelf) and the lithology (see Sects. 2.3.2 and 2.3.3).

300
mil
lion
ye
ar
s
ag
o
50 m
illio
ny
ea
rs
a
go
1.5 m
illio
ny
ea
rs
a
go
deep ocean shallow sea land shelf edge
Figure 2.1: Continental drift is the movement of the earth’s continents relative to each other.
After Wegener (1929).

2.3. Tectonic control of coasts 39
2.2.3. Pleistocene inheritance

During the Pleistocene epoch (the most recent ice age), pronounced climatic fluctu-
ations occurred that resulted in at least eight cycles of glacials and interglacials. In
glacial periods the earth experienced severe cooling and the advance of glaciers over
land up to the 40th parallel in some places. Hence, glaciers periodically covered vast
areas of the continents and at maximum glacial extent roughly 30 % of the earth’s sur-
face was covered by ice sheets 1500 m to 3000 m thick. Since large volumes of water
were tied up in these continental ice sheets, global sea levels could temporarily drop
100 m or more. During subsequent interglacial periods, the global sea level again rose
due to ice retreat and melting. The Holocene , the current geological epoch, is also an
interglacial.
Pleistocene inheritance is found in coastal systems in the form of rocks formed by
glaciers (Sect. 2.4) and by sediments deposited in the Pleistocene. During interglacial
times in the Pleistocene drowned coastlines were common; the sea level rise allowed
temporary marine incursions and hence sediment deposits into areas that are now far
from the sea. In the next glacial period, these sediments were compacted by the ice
cover. Buildings in the Netherlands are often built on piles touching the hard Pleisto-
cene substrate. Many present-day morphological features date to the Holocene. They
are shaped by erosion and deposition on the one hand and sea level changes on the
other hand (Sect. 2.5). The sediments are either loose alluvial sediments supplied by
rivers during the Holocene (from around 12 000 years ago) or older (Pleistocene). Mar-
ine processes may rework these older deposits and feed the coastal system again from
the sea. This is called marine feeding.
2.3. Tectonic control of coasts

2.3.1. Plate tectonic theory
Our continents are part of the lithosphere, which is the uppermost layer of the earth
(containing the crust). By the 1960s, the scientific consensus was that the lithosphere
is divided into 12 large, tightly fitting plates and several small ones (see Fig. 2.2).
Since these so-called tectonic or crustal plates are riding on a semi-molten underlying
material (asthenosphere), the plates are in permanent motion, moving 1 cm to 10 cm
per year depending on the location. Six of the large plates bear the continents; the
other six are oceanic. Note that plate boundaries and continents do not exactly coin-
cide.
Where tectonic plates diverge (mainly mid-ocean) the semi-molten asthenosphere ma-
terial can be driven to the earth surface. As a result new (oceanic) earth crust is
formed, resulting in the so-called oceanic ridges (Fig. 2.3). At other places, instead of
divergence there is convergence. If an oceanic and continental plate meet, the denser
oceanic plate dives under the continental plate (Fig. 2.3). This process of convergence

18
80° 0° 0°
12
12
60°
0°
°
60
0°
60°
Eurasian Plate
1.0
40° North
2.3
American Anatolian Plate
Juan de Fuca
Plate Plate Iranian
Plate
20°
2.5 Arabian
Plate
Caribbean Philippine Pacific
African Plate
8.6 Plate Plate
Plate 3.0
0° Cocos
15.1 5.0 Plate
Pacific 3.5 4.4
Plate Nazca South
Plate American
20° 15.1 1.4 Indo-Australian Plate
5.9 Plate 3.5 7.2
7.5
9.4
40° 1.4 Antarctic Plate
Scotia Plate
60°
80°
divergent boundary conservative boundary rate of movement
8.6
convergent boundary suspected boundary (cm per year)
Figure 2.2: Movements of the tectonic plates: colliding or converging plates, diverging plates
and plates grinding past each other (conservative or transform boundaries).
creates mountains and oceanic trenches and is often accompanied by seismic and vol-
canic activity (see Fig. 2.4).
The oceanic ridge system consists of the Mid-Indian Ridge, Mid-Atlantic Ridge and
East Pacific Rise (see Fig. 2.4). The age of the crust on both sides of the mid-ocean
ridges increases with distance from the ridge. The rates of plate divergence range
Figure 2.3: Movement in the earth’s crust at convergent boundaries (subduction zones where
an oceanic plate dives under a continental plate) and divergent boundaries.

from 1 to 3.5 cm/yr at the Mid-Atlantic ridge to 9 to 15 cm/yr at the East Pacific Rise
in the southeastern Pacific. Nowadays, the movements are measured with the aid of
satellites using very accurate geodetic positioning systems (such as Differential Global
Positioning System (DGPS)).
Figure 2.5 shows the deep rift valley that runs along the axis of the Mid-Atlantic Ridge
and becomes visible at the surface on Iceland in a very spectacular way.
Figures 2.2 and 2.4 not only show convergent and divergent plate boundaries, but also
conservative or transform boundaries, where plates grind past each other. There are
no volcanoes at a conservative plate boundary, but earthquakes can be very destruct-
ive. An example is the San Andreas Fault in California where the Pacific and the
North American plates are sliding past each other, in the same direction but at differ-
ent speeds.
18
80° 0° 0°
12
12
60°
0°
°
60
0°
60°
40° Kuril
Mid- trench
Atlantic Japanese
Ridge
ridge trench
20°
Mariana
trench
0° East
East
Pacific
Pacific Mid-
rise
rise Peru- Indian Java
Chile ridge trench Tonga
20° trench trench
40°
60°
80°
earthquakes divergent boundary conservative boundary
high frequency of earthquakes convergent boundary suspected boundary
Figure 2.4: Global distribution of earthquake occurrence is associated with plate boundaries:
earthquakes are found near the oceanic ridge system at diverging boundaries and near the
trenches closer to continental edges at converging boundaries, with the greatest occurrence
at converging boundaries.
2.3.2. Tectonic plate setting of coasts

Inman and Nordstrom (1971) recognised that broad coastal characteristics such as shelf
width and coastal topography are related to the position on the moving tectonic plates.
Not only the proximity of a coast to a plate boundary, but also whether this boundary
is converging or diverging has a huge influence on the character of that coast. Inman

Figure 2.5: Mid-Atlantic Rift passing across Thingvellir National Park, Iceland. The North Amer-
ican Plate is shown to the left and the Eurasian plate to the right. Photo by Matthijs Buijs (‘Credits’
on page 575).
and Nordstrom (1971) classified coasts into three categories according to their tectonic
plate setting:
• Leading-edge or collision coasts: associated with the leading-edge of a crustal
plate viz. with converging plates. They are characterised by rugged, cliffed
coastlines, tectonic activity and a narrow continental shelf;
• Trailing-edge coasts or passive margins: coasts that are located away from plate
boundaries and are generally tectonically stable because the continent and ad-
joining ocean floor are part of the same plate. They have wide continental
shelves;
• Marginal sea coasts: tectonically stable coasts protected from the open ocean by
island arcs at converging plate boundaries.
Figure 2.6 shows the worldwide distribution of coastal types as classified by Inman.
Compare this figure with the tectonic plate setting of Fig. 2.2.
Figure 2.7 illustrates the formation of leading-edge coasts and trailing-edge coasts. It
can be thought of as a cross-section of North or South America. Along the west coasts
of the American continent, denser oceanic plates descend beneath the continental
edge of the North and South American plates (see the plate boundaries in for instance
Fig. 2.6). This collision has resulted in narrow shelves, earthquakes, coastal uplift, and
the formation of mountains immediately inland from the coast: the Andes Mountains
and the North American western mountain ranges. This is illustrated by Fig. 2.8, the
Big Sur coast in California, USA. The rising magma may also create volcanic activity.

18
80° 0° 0°
12
12
60°
0°
°
60
0°
60°
40° Kuril
Mid- trench
Atlantic
ridge Japanese
20° trench
Mariana
trench
0° East
Pacific Mid- Java
rise Peru- Indian
Chile trench Tonga
ridge
20° trench trench
40°
60°
80°
leading edge trailing edge marginal sea divergent boundary
neo convergent boundary
afro conservative boundary
amero suspected boundary
Figure 2.6: A tectonic-based classification of coasts. After Inman and Nordstrom (1971). The
three classes of trailing-edge coasts are discussed below.
In contrast, the east coasts of the American continent lie within the interior of tectonic
plates, and therefore experience little tectonic activity and are either stable or subside
rather than being uplifted.
Trailing-edge coasts are the result of plate divergence and are facing a spreading centre.
In plate tectonics, the formation of an ocean is a spreading process between two newly
formed plates. At first this process results in a rift valley (the African Rift Valley being a
recent example). In the next stage, the valley opens up until seawater enters; a present
day example is the Red Sea. The spreading process continues, the ocean becomes
wider and forms a mature ocean like the Atlantic. The mature trailing-edge coasts
bordering that ocean are far away from the mid-plate spreading centre. This means
that all coasts on both sides of the Atlantic Ocean in Europe, Africa, North and South
America are trailing-edge coasts (e.g. the Dutch coast, see Fig. 10.9). Other examples
of trailing-edge coasts are the west Australian coasts. Inman and Nordstrom have
categorised the new trailing-edge coasts formed near young spreading centres as neo-
trailing-edge coasts. Regarding the more mature trailing-edge coasts further from plate
boundaries, they discern Afro-trailing-edge coasts and Amero-trailing-edge coasts based
on differences in sediment supplies (see Sect. 2.3.3).
The leading-edge coasts as illustrated in Fig. 2.7 develop along the border of a land mass
where the oceanic edge of one plate converges with the continental edge of another
plate. They are called continental collision coasts. Other examples are found along the

Figure 2.7: Formation of leading-edge and trailing-edge coasts. This figure can be seen as a
cross-section of the American continent with the west coasts dominated by rocky coasts and
a narrow and steep continental shelf and the east coasts dominated by extensive sediment
deposits on a wide and flat continental shelf.
Figure 2.8: Leading-edge coast in the Big Sur region of California, USA. Photo by Marcel J.F. Stive
(‘Credits’ on page 575).

coast of Turkey and Greece. Besides these continental collision coasts, also island arc
collision coasts can be discerned. The latter form where two oceanic plates collide, see
Fig. 2.9.
Figure 2.9: Convergence of two oceanic plates and the associated volcanic activity leads to the
formation of island arc coasts.
The island arc is formed from volcanoes which erupt through the overriding plate as
the descending plate melts below it. Japan, the Philippines, the Caribbean island arc
and the Aleutian island arc in Alaska are examples of collision coasts located along the
margin of an island arc. Other examples of island arc coasts collision coasts are found
along the coasts of Borneo, Sumatra and New Guinea.
Marginal sea coasts occur in semi-protected environments including eastern Australia,
the Gulf of Mexico and eastern Asian shorelines. The left part of Fig. 2.9 shows a
marginal, inland sea enclosed between a land mass and a volcanic island arc at the
converging plate boundary. The marginal sea coast fronting the inland sea is therefore
protected from the open ocean by the volcanic island arcs (e.g. Korea) where a plate
collision is occurring. Although fairly close to the convergence zone, the marginal sea
coast is far enough away to be unaffected by convergence tectonics – it behaves like
a trailing-edge coast.
2.3.3. First-order coastal sedimentary features

Of the aforementioned tectonically-driven coastal characteristics, shelf width is the
major factor controlling coastal sedimentary features. Furthermore, differences in sed-
iment supply are of importance. Both aspects are discussed in this section.

Continental shelf width

The world distribution of continental shelf width is given in Fig. 2.10. As can be seen
from Figs. 2.6 and 2.10, there is a direct correlation between the tectonic plate setting
and the continental shelf width. For instance, some of the narrowest shelves are found
off the tectonically active west coasts of North and South America. Broad shelves on
the other hand are common to trailing margins. Since the shelf break (see Fig. 1.15)
is located at relatively constant water depths of 100 m to 200 m, narrow shelves imply
steep slopes, whereas wide shelves have low gradients.
18
80° 0° 0°
12
12
60°
0°
°
60
0°
60°
40°
20°
0°
20°
40°
60°
80°
shallow water (0 to 200 m depth)
Figure 2.10: Approximate world distribution of continental shelf width indicated by near-coast
regions up to 200 m water depth.
Because of their larger horizontal extent, wide and flat shelves facilitate the develop-
ment of extensive sedimentary features as opposed to narrow and steep shelves. Also
due to the smaller slope, the former permit more rapid coastal progradation. For this
reason, large deltas and barrier islands systems have the potential to develop on coasts
with wide shelves (as long as sufficient sediment is supplied). Hence, leading-edge
coasts and trailing-edge coasts are dominated, respectively, by erosional features like
sea cliffs, rocky headlands, etc., and depositional features like barrier islands.
The shelf width also has an effect on the hydrodynamic conditions, for instance storm
surge is augmented by shelf width. Storm surge is the piling up of water against the
coast due to a combination of onshore winds and low atmospheric pressure, see Ch. 5.
Narrow shelves do not have enough horizontal extent for water to pile up and have
lower potential surge elevations (see Fig. 5.50 and Eq. 5.90). Wind-wave heights are
higher at narrow shelves since the waves are permitted to keep their height; there is
no shallow sea bottom to interfere with and hence dampen the wave motion. Wider

shelves on the other hand have more frictional dampening of storm waves and thus
lower wave energy. Tidal amplitudes on the contrary are generally higher at wider
shelves, at least for semi-diurnal tides in non-polar areas. In those situations resonance
is approached, which amplifies the tidal amplitudes at the coast (see Sect. 3.8).
As pointed out above, large mid-ocean waves lose energy as they progress across the
gently sloping inner continental shelf of a trailing-edge coast. They consequently do
not inhibit deposition of sediment along the coast. As long as enough sediment is sup-
plied to the coastal system, the combination of reduced wave energy and the facilita-
tion of sediment accumulation ensures the development of large sedimentary features
on the wide and flat trailing-edge continental margins. An example are the extens-
ive mangrove swamps and tidal flats, which cover the low-relief Amero-trailing edge
coast near the mouth of the Amazon River in Brazil.
Similarly, marginal sea coasts have the potential to develop large sedimentary features
due to their wide continental shelves and low wave energy (see Fig. 2.11). They are low
wave energy coasts due to 1) the gentle slope and shallow waters of the continental
shelves in these areas; 2) the attenuation of wave energy due to the sheltering by
nearby or surrounding land masses or island arcs or ice; and 3) the restricted size of
the marginal seas and hence limited fetch, restricting the size of waves that develop
locally. The coastal plains of marginal sea coasts vary in width and may be bordered
by hills and low mountains (e.g. Vietnam).
Figure 2.11: Marginal sea coast at Jiangsu, China. Photo by Marcel J.F. Stive (‘Credits’ on
page 575).
In contrast, sediment delivered to leading-edge coasts is soon dispersed by the usually

large waves coming from the deep ocean. Moreover, because of the narrow continental
shelf, submarine canyon heads lie close to the shore. As a result, most of the sediment
deposited along the coast moves offshore through canyons and thus into deeper water

beyond the shelf. Hence, the coastal zone of leading-edge coasts has relatively little
sediment even if large amounts of sediments are transported to the coast by rivers.
Sediment availability
Trailing-edge coasts have typically been tectonically stable for many millions of years.
All these years, erosion processes have taken place and converted hills and cliffs into
coastal and submarine plains. Along these coasts, one can therefore find large deposits
of sediment. They are shaped and reshaped by currents, wind and waves into barrier
islands, deltas and other sedimentary shapes. For instance, the São Francisco Delta in
Brazil and the Senegal Delta in Africa have developed on trailing-edge coasts.
There are differences between the trailing-edge coasts of the African continent and
the east and Gulf coasts of the American continent, mainly as a result of differences in
sediment supply. The American trailing-edge coasts receive large quantities of sedi-
ments, since many of the rivers originate in mountainous regions and have very large
drainage basins. The latter is the result of the overall asymmetry of North America,
with the largest mountains and divides located closer to the west coast. The African
continent on the other hand occupies a position in the middle of a crustal plate. Due
to the absence of plate collisions, it lacks significant mountains. Therefore and be-
cause of climatic conditions, accordingly smaller quantities of sediment are delivered
to the coastal zone. The typical coastal features of the African continent have led to
the name Afro-trailing edge coast for a trailing-edge coast for which the coast on the
opposite side of the continent is also trailing. Although Afro-trailing edge coasts have
pronounced continental shelves and coastal plains, these features are usually smaller
than found at more mature coasts, and sedimentary features such as large deltas are
rare.
Amero-trailing edge coasts are represented by the east coasts of North and South Amer-
ica and are geologically the most mature coastal areas. Amero-trailing edge coasts
have a collision coast on the opposite side of the continent. In combination with the
moderate climate, this has led to the development of numerous large, meandering
river systems that for more than 150 million years have been carrying sediment across
a gentle incline, developing broad, low-relief coastal plains and subaqueous features.
Some of the largest deltas in the world have been formed on marginal sea coasts by
the great rivers of southeastern Asia and the Gulf region of the USA. Both areas have
a mild climate and abundant rainfall. Hence, the combination of relatively low-energy
coastal conditions and sizeable sediment loads leads to the formation of large deltas
and other coastal sedimentary deposits such as tidal flats, marshes, beaches and dunes.
Examples are the Mississippi River Delta in the Gulf of Mexico and the huge deltas of
the Pearl, Yangtze and Yellow Rivers that empty into the South China Sea.
The steep mountain slopes of leading-edge coasts hold rapidly flowing streams and
small rivers that quickly erode their beds. Because the watershed is at a high eleva-
tion near the coast, the rivers are short, steep, and straight. They may transport large

2.4. Pleistocene inheritance of cliffed coasts 49
quantities of sediments directly to the coastal areas, giving no opportunity for sedi-
ments to become entrapped in a meander, on a natural levee, or on a flood plain. The
rivers deposit their sediment loads into coastal bays or directly onto open beaches. The
coastal deposits are relatively coarse, since the material has to be transported over just
a short distance (see Sect. 2.6). However, even though mountain streams deposit large
amounts of sediment on the coast, they do not produce deltas (Davis Jr., 1994). In fact,
none of the world’s 25 largest deltas are found on leading-edge coasts. As pointed out
above, this is because this tectonic setting with narrow and steep continental shelves
does not have a shallow, nearshore area on which the sediment can accumulate. In-
stead, sediment will be lost to deeper areas beyond the shelf break through offshore
canyons. Furthermore, the dispersion of sediment by the usually large waves prevents
the development of extensive sedimentary features such as deltas.
Some trailing-edge coasts look like leading-edge coasts. This holds for instance for the
high-relief coast with coarse gravel beaches on the Sea of Cortez (or Gulf of California),
Mexico. This coast is a neo-trailing edge coast. Being only a few million years old, it
represents the first stage of coastal development and resembles a leading-edge coast.
2.3.4. Summary of tectonic classification

As noted by Inman and Nordstrom (1971), the tectonic setting of a coast controls the
physical nature of the coast, at least at first order (that is to say, to a first approxim-
ation). The leading-edge or collision coasts are all relatively straight and mountain-
ous and generally characterised by sea cliffs, raised terraces, and narrow continental
shelves.
The trailing-edge types of coasts are more variable. The Amero and Afro types have
low-lying depositional coastal forms such as barrier islands and the widest continental
shelves. The neo-trailing edge coasts are typically steep with beaches backed by sea
cliffs, so in many respects these neo-trailing edge coasts are similar to leading-edge
coasts. The marginal sea coasts have the greatest diversity of form. The land may be
low-lying or hilly, and the form of the coast can be dominated by local processes such
as the formation of river deltas.
2.4. Pleistocene inheritance of cliffed coasts

The larger part (about 75 %) of the world’s continental and island margins is lined
with cliffs, consisting of rocks or cohesive clays. As we have seen in Sect. 2.3.2, cliffed
coasts are commonly tectonically active convergent coasts, which produce a rocky,
high-relief border. Because they are formed on continental plate margins, under which
an oceanic plate is descending, virtually no continental shelf is present. The western
edges of North and South America are good examples of this type of coast, as discussed
above.

There are also cliffed coasts of which the origin is unrelated to plate tectonics. In these
cases the adjacent continental shelf is wide, with a gentle slope. First, Pleistocene
glaciers have had a hand in producing cliffed coasts. The moving ice masses gouged
out steep valleys, which were subsequently drowned as the sea level rose during an
interglacial period. Examples of these so-called fjords are found along the coasts of
Alaska, Scandinavia and Scotland. Although their profiles are similar, some fjords are
rocky and others are not. In Fig. 2.12, the rocky coast along a Norwegian fjord is shown.
It was carved by a glacier.
Figure 2.12: Geirangerfjord in Norway. Photo by Marcel J.F. Stive (‘Credits’ on page 575).
Still other cliffed coasts are formed of glacial till, sediment deposited by glaciers be-
neath and at the margins of the ice. The till is over 100 m thick in some places and
includes nearly any type of material, from stiff clays to sand, gravel and boulders.
Some of it is well-layered and some is massive, with essentially no internal coherence.
The accumulations known as end moraines tend to be linear and thick. When these
end moraines meet the sea, the waves sculpt steep bluffs. Coasts formed by glacial till
are also found in Scotland, Denmark and the United Kingdom.
Another variety of rocky and commonly cliffed coast is associated with areas where the
continental shelf and adjacent coast are dominated by carbonate sediments as a result
of skeletal shell and coral debris. In the Pleistocene, onshore winds blew the carbonate
sediment landward, where it accumulated in wide beaches and dunes. In a process
called lithification, the calcium carbonate grains were welded together by a cement
that is created as ocean spray or percolating ground water reacts with the calcium
carbonate. The evaporation of regularly wetted surfaces in arid climates enhances the
lithification of the sediments. The rapid cementation converts the dunes to a rock
called eolianite. The coast of North Africa is well known for its cemented sand, which

2.4. Pleistocene inheritance of cliffed coasts 51
hampers dredging operations because in surveys it appears to be sand, whereas it is

quite hard in reality. Similarly, on the beach we find the so-called beach rock, which
can be a nuisance for dredging operations.
An example of yet another type of cliffed coasts are the Cliffs of Moher (Fig. 2.13),
made up of layer upon layer of sand, silt and mud, compacted into solid rock. These
sediments were carried into an ancient sea by large rivers.
Figure 2.13: Cliffs of Moher, Ireland, showing horizontal layers of rock. Photo by Alejandra Gijón
Mancheño (‘Credits’ on page 575).
Erosion rates of cliffed coasts strongly depend on lithological factors. Typical erosion
rates for rock are 10−1 m/yr to 1 m/yr for chalk, 10−3 m/yr to 10−2 m/yr for limestone
and 10−3 m/yr for granite. Along rocky coasts, nearshore wave energy is often high
because the size of the waves is related to the nearshore bathymetry (see Sect. 2.3.3)
and to refraction patterns. The wave energy is focused on the headlands (see Fig. 5.6)
and dispersed in the bays, so the headlands erode as the intervening bays fill up. Wave
erosion of an indented coastline produces a straightened, cliff-bound coast, as shown
in Fig. 2.14. Wave-cut platforms and isolated stacks and arches may remain offshore
(see Figs. 2.15 and 2.16).
Our focus is on sandy (and to a lesser extent muddy) coastal systems and, therefore,
the processes of cliff and rock erosion are not considered any further in the remainder
of this book.

Figure 2.14: Wave-erosion effects. Along rocky coasts with headlands (a), the headlands erode
due to wave focusing (b), leaving behind isolated stacks and arches (c). After De Blij and Muller
(1993).
2.5. (Holocene) transgression versus regression

2.5.1. Geological sea level changes
Local mean sea level is defined as the height of the sea with respect to a land bench-
mark, with fluctuations caused by waves and tides smoothed out. Locally perceived
changes in mean sea level (also called relative sea level changes) can be the result of

2.5. (Holocene) transgression versus regression 53
Figure 2.15: Cliffs with stack and arc at Falaise d’Etretat, France. Photo by Blenda Gomes Rocha
Figure 2.16: Cliffs with stacks at the coast of Peniche, Portugal. Photo by Ascha Simons (‘Cred-
its’ on page 575).
either vertical movements of the land or an absolute (relative to the earth-centre) move-
ment of the sea level (Fig. 2.17). Absolute sea level changes are called eustatic changes
and can be of the same order of magnitude as the vertical land movements.
Eustatic sea level changes are commonly defined as absolute vertical movements of
the mean sea level, represented by the oceanic geoid2 . This oceanic geoid is the shape
that the ocean surface would take under the influence of the gravity of the earth, if
2
The geoid is a surface of equal gravitational attraction and is, within the ocean areas, roughly equal to
the mean sea level surface (if dynamic effects are ignored).

sea level rise

relative
sea level rise
land subsidence
Figure 2.17: Relative sea level rise is the combined effect of absolute sea level rise and land
subsidence or uplifting. It therefore is the locally perceived sea level change.
other influences such as winds and tides were absent. Therefore, it corresponds to the
mean sea surface at rest. Eustatic changes are caused by:
• Changes in the volume of the ocean water through:
– Changes in the amount of water as a result of glaciations and deglaciations
i.e. the advance and retreat of ice sheets and glaciers (glacio-eustasy);
– Changes in expansion of water (so-called steric changes due to temperature
or salinity changes);
• Changes in the volume of the ocean basins due to the very slow process of tec-
tonic plate divergence (tectono-eustasy), marine sedimentation or hydro-isostasy
(see below);
• Changes in the distribution of water due to changes in the earth’s gravitational
field and hence in the shape of the oceanic geoid (geoidal eustasy).
Of the listed causes, the most important cause of long-term variations in eustatic sea
level is changes in the amount of ocean water due to glaciations and deglaciations.
Such changes have taken place often during earth’s history under the influence of
global temperature change, and still occur. In the Quaternary, interglacial and glacial
climates alternated at the higher latitudes of the Northern Hemisphere. During the
cold periods the development of enormous ice sheets and glaciers resulted in a signi-
ficant lowering of the ocean level (more than 100 m) and large parts of the continental
shelf became dry land or coastal swamp. The melting down of the ice sheets during
the warmer periods resulted in the restoration of the ocean level.
The last listed cause should also be explained. Eustatic changes have long been be-
lieved to be worldwide and simultaneous. For that reason eustatic sea level rise was
also called global sea level rise. It was only discovered in the early 1970s that the
oceanic geoid is not uniform and constant. It has relief and is constantly deforming
under the influence of gravitational and rotational changes as a result of, amongst

other reasons, the formation and melting of ice sheets. The changing shape of the
geoid influences the worldwide distribution of water and therefore eustatic sea levels.
This explains why eustatic sea level changes during the last 6000 years, as reported by
several authors before this effect was taken into account, were rather divergent. The
role of the formation and melting of ice sheets can be explained as follows. Large ice
masses exert a gravitiational force on the ocean that results in a gradual tilt of the sea
surface toward the ice sheets. If the ice melts, the gravitational attraction and hence
the sea surface slope reduces. This yields a larger than average absolute sea level rise
far away from the ice sheets and a smaller than average absolute sea level rise or even
a sea level fall closer to the ice sheets.
Eustatic sea level changes result in deformations of the earth’s crust, which is known
as isostatic deformation (uplift or subsidence). The isostatic uplift or subsidence of the
earth’s surface can take two forms:
Glacio-isostasy is the loading and unloading by ice. It is related to the fact that the
weight of the ice sheet depresses the underlying land. When the ice melts away,
the land areas previously covered by ice (above about 40°N) slowly rise. Adjacent
areas that were never covered with ice tend to sink. This is known as isostatic
rebound and subsidence.
Hydro-isostasy is the loading and unloading by ocean water. It refers to the fact that
water from melted land ice creates an additional load on the ocean floor. To
attain isostatic equilibrium, the ocean floor is slowly depressed in response to
sea level changes. Simultaneously, the continents are flexed upward at their
margins. This ‘continental levering’ may be restored when the water is tied up
again in continental ice sheets.
Isostatic subsidence and rebound of the lithosphere not only impact relative sea levels,
but also absolute sea levels via changes in the volume of the ocean basins.
As a result of both geoidal eustasy and isostasy, the world-wide distribution of sea level
change is spatially non-uniform. Clark et al. (1978) were amongst the first to model
the spatially non-uniform deformation of the geoid and the earth’s surface in the last
6000 years as a result of viscoelastic isostatic responses to water and ice loads. They
distinguished various geographic regions based on differences in Relative Sea Level
(RSL) curves in response to the retreat of Northern Hemisphere ice sheets.
Figure 2.18 shows the results of more recent model predictions of the present-day rate
of worldwide sea level change in response to early Holocene deglaciation (Mitrovica
& Milne, 2002). The sea level change as predicted by Mitrovica and Milne (2002) is
due to isostatic adjustments to changes in ice and ocean loading (including interac-
tions with the earth’s gravitational field), assuming that Holocene deglaciation was
complete 5000 years ago. Hence, the patterns reflect ongoing sea level adjustments
during a time in which no meltwater is being added into the oceans. The saturated

red zones are previously glaciated areas that experience ongoing rebound due to gla-
cial unloading and thus RSL fall. Near these areas, the reduced gravitational attraction
associated with the ice mass and thus ocean unloading of the solid surface also play
a role. The saturated blue zones experience RSL rise due to isostatic subsidence; the
seafloor in these zones was previously levered up in response to the nearby depression
of the earth’s surface by mostly ice. After the decay of the ice sheets, the previously
up-levered areas tend toward their pre-ice age state, resulting in sea level rise. Note
also the sea level fall in equatorial basins. This is due to the redistribution of water to
fill the subsiding blue zones. Continental levering is evident in the thin bands of off-
shore sea level rise and onshore sea level fall along coastlines in the far field from the
late Pleistocene ice sheets; due to the additional water load on oceanic regions only,
the continents are flexed upward at their margins and downward offshore. The above
also implies that sea level curves will differ between oceanic islands and continental
margins.
The processes underlying the spatially non-uniform sea level change are further illus-
trated in Fig. 2.19 (see also Tamisiea et al., 2003). This figure shows both isostatic and
eustatic changes. The right-hand panel clearly shows that there may be a sea level fall
in the vicinity of a melted (or melting) ice sheet or glacier (cf. the saturated red areas in
Fig. 2.18). This spatial non-uniformity has implications also for future sea level change
in response to projected climatic change. If for instance the Greenland ice sheet were
to melt significantly in the next centuries, the sea level change around the Netherlands
for instance could just as well be zero.
In addition to eustatic and isostatic effects, there are other factors that locally and
regionally influence relative sea level changes, for instance:
• Regional subsidence can occur due to compaction of sediments and withdrawal
of subsurface fluids such as groundwater, oil, and natural gas;
• Tectonic activity can cause upward or downward movements of the shore. Areas
with significant tectonic movements are for instance the earthquake zones in
the western and southwestern Pacific and in the Mediterranean. In these areas
shorelines can be uplifted, submerged or tilted by these (sudden) earth move-
ments. In addition, most deltas are situated in slowly subsiding sedimentary
basins (tectonic subsidence between 2 and 10 cm per century).
All factors contributing to locally perceived, relative sea level changes are summarised
in Fig. 2.20.
Summarizing, curves of relative sea level rise versus time can show large local and
regional variations. This is especially true during the second half of the Holocene, the
last ca. 6000 years, when the land icecaps had already mostly vanished.
In Fig. 2.21 indications of Holocene sea level changes are given for a number of relat-
ively stable areas, i.e. without major subsidence or uplift (although in the Netherlands
the bottom is slowly subsiding).

0° 18
80° 12 0°
12
°
0°
60°
60
0°
60°
40°
20°
0°
20°
40°
60°
80°
80°
60°
40°
20°
0°
20°
40°
60°
80°
80°
60°
40°
20°
0°
20°
40°
60°
80°
_.7 _.6 _.5 _.4 _.3 _.2 _.1 .0 .1 .2 .3 .4 .5

mm/yr
Figure 2.18: Numerical prediction of the present-day rate of change of global sea level due to
isostatic adjustments (top) and the contributions to this change from ice loading (centre) and
ocean loading (bottom). After Mitrovica and Milne (2002).

melting ice
land rise due

change of geoid due to meltwater to changing
and lower attraction by ice mass ice load
gentle subsidence land rise
Figure 2.19: Spatially non-uniform sea level response. Dashed lines: the presence of a large
ice mass leads to raised water levels in the vicinity of the ice mass (as a result of gravitational
attraction) and a lowered solid surface. Solid lines: Melting of the ice sheet causes a fall of the
sea surface in a large area centred on the region of mass loss. The solid surface responds to
the changes in ice loading and ocean loading.
climate earth movements gravity, rotation
ocean level
ocean water ocean basin
distribution
volume changes volume changes
changes
glacial eustasy tectono-eustasy geoidal eustasy
ocean level
crustal
changes sea land
movements
“eustasy”
local
local changes
compaction
relative sea
meteorological
hydrological level changes
oceanographic
Figure 2.20: Relative mean sea level changes are the result of the three different types of eustasy,
vertical crustal movements due to tectonic activity and isostasy, and local compaction.
During the 20th century, the average rate of global sea level rise was about 1.5 mm/yr
to 2.0 mm/yr. Satellite measurements taken over the first decade of the 21st century
indicated that the rate of sea level rise had gone up to 3.1 mm/yr. Apparently, the
present-day rates of increase are in the order of magnitude of millimetres per year
which is significantly larger than the recorded rates over the past few thousands of
years (see Fig. 2.21). These accelerated rates of sea level rise have been interpreted as

0
Holocene
present sea level [m]

transgression
elevation w.r.t. -50
-100
-150
40 35 30 25 20 15 10 5 0
present sea level

present sea level [m]
0
elevation w.r.t.
-10
The Netherlands
Florida
Texas
-20
S.W.Louisiana
8 7 6 5 4 3 2 1 0
3
10 years before present
Figure 2.21: Indication of sea level changes in stable areas (no local subsidence or uplift).
evidence that atmospheric warming, since the industrial revolution, has led to thermal
expansion of the oceans, melting of mountain glaciers and icecaps (Gornitz & Lebedeff,
1987) and loss of ice from the Greenland and West Antarctic ice sheets.
2.5.2. Role of sea level rise in Holocene coastal evolution

Sea level changes can affect the coastal zone very strongly. The effect of sea level rise
can be understood from the well-known concept of Bruun (1954, 1962), which is briefly
explained here and treated in more detail in Ch. 7. The Bruun rule assumes that the
shoreface has a profile that is in equilibrium with the hydrodynamic forcing. Hence,
it states that the shore profile is vertically invariant in space and time relative to mean
sea level. Consequently, a sea level rise results in a water depth that is too large to be
in equilibrium with the forcing. In other words: extra space has become available for
sediment accumulation, the so-called accommodation space. In the absence of sediment
sources or sinks, equilibrium is again achieved by a landward and upward shift of the
profile: the shoreline retreats and a new equilibrium profile forms at the new shoreline
position by moving sediments to deeper water (Fig. 2.22).
The world’s coasts can be divided into two classes: transgressive coastal environments
and regressive coastal environments. As the shoreline moves seaward or landward in
response to sea level changes, it either exposes or inundates coastal areas and, in so
doing, causes the character of the coast to change. Additionally, the position of the
shoreline influences coastal processes that shape coastal environments. Inundation is

sea level rise
Figure 2.22: Bruun effect: the profile shape remains the same (the length of the vertical and
horizontal lines respectively is constant), but the profile moves up and landward as a result of
sea level rise. The volume of sediment eroded from the upper profile is equal to the deposited
volume in deeper water.
also called transgression (advance of the sea), whereas drying of the land is referred to
as regression (retreat of the sea). Transgressive coastal environments are characterised
by lagoons and estuaries. We have seen before that estuaries are semi-enclosed coastal
water bodies, which are on one side connected to the sea and on the other side have one
or more rivers or streams flowing into it. Rias and fjords are estuaries formed through
flooding of low-lying areas; rias are drowned river valleys (for instance Sydney har-
bour) and fjords are drowned valleys carved out by land ice (see Fig. 2.12). In regressive
coastal environments the shoreline moves seaward and estuaries and lagoons that are
present get filled in or abandoned.
The global distribution of transgressive and regressive systems is mainly determined
by the (late) Holocene relative sea level changes. Sea level changes are relative move-
ments and thus vary from place to place as can be seen from for instance Fig. 2.18.
Transgressive coastal environments are well developed in areas that have experienced
isostatic subsidence, viz. the middle and south of North America, the middle and south
of Europe and the Mediterranean region. In these regions relative sea level has risen at
an average rate of 1 mm/yr over the last 7000 years. On the contrary, regressive coastal
systems can be found (although not exclusively) in the regions that were located far
from glaciers during the last ice age. This includes most of Asia, Oceania, the middle
and south of Africa, and South America. In the beginning of the Holocene these areas
experienced a relative sea level rise, but this had changed to relative sea level fall later
in the Holocene, resulting in a relative sea level fall in the order of a few metres over
the last 7000 years. Lagoons and estuaries that have formed in the beginning of the
Holocene were filled or abandoned later, so that now evidence of regression is found
along those coasts.
Although the global distribution of transgressive and regressive coastal systems is
mainly controlled by sea level rise, the amount of sediment supply is very important
as well in controlling transgression versus regression. Clearly, sediment supply may
counteract the effect of shoreline retreat. For instance, if the rate of sediment supply

keeps up with the rate of sea level rise, the accommodation space created by the sea
level rise is filled by incoming sediments and the profile moves upward only; the po-
sition of the shoreline remains unchanged. Sediment can be supplied to the coast by
for instance rivers and erosion from coastal cliffs or headlands. Through alongshore
and cross-shore processes, waves and tides will rework the sediment supplied to the
coast, resulting at a certain location in either erosion or deposition of recycled (mainly
riverine) sediment. At a very high level of aggregation these sources and losses can
be summarised to be represented by the term ‘sediment availability’.
Figure 2.23 qualitatively summarises the effects of both relative sea level change and
sediment availability on the displacement of the coastline. This type of diagram is
called Curray’s diagram after Curray (1964).
erosion deposition
retrogradation
rate of accommodation space creation
due to relative sea level changes
n
sio
rise
es
gr n
ns sio
ra es
ort gr
at re
0 tre or
e re n ce
n a
eli dv
or ea
sh i n
el
or progradation
fall
sh
sink 0 source
rate of sediment supply
Figure 2.23: Factors controlling shoreline migration after Curray (1964). The effect of relative
sea level rise is to create more space to accommodate sediments. If the rate of sediment supply
equals the rate of space creation (the diagonal line), the shoreline remains stationary. If the rate
of sediment supply is larger than the rate of accommodation space creation (the region below
the diagonal), the shoreline advances; if it is smaller the shoreline retreats (the region above the
diagonal). The dashed lines indicate progradation or building out of the coast (Intermezzo 2.1).
Curray’s diagram shows that in the absence of a net sediment source or sink, regres-
sion occurs in the case of a falling sea level – called emergence in that case – and
transgression occurs in the case of a rising sea level or submergence. With a net source
of sediment, the transition between regression and transgression shifts towards low
levels of rising relative sea level, and alternatively, with a net sink of sediment, to-
wards low levels of falling relative sea level. In other words, in areas with erosion the
shoreline generally moves landward, inundating the coast (transgression), unless the
relative sea level fall is so large that the shoreline migrates seaward. Sediment supply
(deposition) results in regressive shorelines moving seaward, except when the regres-
sion is counteracted by a large relative sea level rise. This building out of the coast
is also called progradation. The various names that are used to describe the coastal
response are summarized in Intermezzo 2.1.

The main conclusion from Curray’s diagram is:
Whether a shoreline migrates seaward or landward is determined by the combina-

tion of relative sea level changes and amount of sediment supply or loss. Seaward-
migrating coasts are also referred to as regressive coastlines; landward-migrating
coastlines are termed transgressive coastlines.
Intermezzo 2.1 Classification on the basis of sea level change
On the basis of coastal response to substantial sea level changes, various classific-
ations have been proposed in literature. They are invariably based on the balance
between sea level change and sediment supply. Some terminology used to describe
the coastal response is given here (Fig. 2.24):
Regression seaward shift of the shoreline → former sea bottom exposed
Transgression landward shift of shoreline → inundation
Progradation sediment is deposited such that shoreline moves seaward
Retrogradation sediment is deposited but shoreline moves landward
Emergence land emerges out of the water due to relative sea level fall (e.g.
due to tectonic uplift: Chile)
Submergence inland regions are flooded due to relative sea level rise
The classification of Valentin (1952) is a clear example. He simply divides the
world’s coasts into coasts that have advanced (due to emergence and/or deposition)
and coasts that have retreated (due to submergence and/or erosion). The various
definitions are also indicated in Curray’s diagram.

progradation retrogradation
sea level sediment sea level

rise input rise sediment
input
effect sediment input > effect SLR effect sediment input < effect SLR
progradation retrogradation
shoreline advance shoreline retreat
sea level sea level sub-

emergence
fall rise mergence
emergence submergence
Figure 2.24: Classifications on the basis of sea level change and sediment supply. In the
case of sea level rise, the occurrence of progradation or retrogradation depends on the
amount of sediment input or loss. Note that the shoreline response in the sketches for
emergence and submergence does not take the Bruun effect into account (see Fig. 2.22).
If the amount of erosion due to sea level rise is small compared to sediment depos-
ition, the shoreline may migrate seaward even under conditions of relative sea level
rise. The Mississippi delta (see also Intermezzo 2.2) and the central Dutch coast are
good examples of regressive systems that have developed despite their location in re-
gions with Holocene relative sea level rise. In regressive coastal environments typic-
ally deltas and extensive strand plains (sandy beach and shoreface systems) and chen-
ier plains (muddy tidal flats) are found.
It is important to note that the characteristics of a coast show imprints of different
episodes in the history of its development. Australia for instance has a large number
of estuaries and lagoons, although the relative sea level fell during the late Holocene.
This is related to the dry climate and therefore very low sediment supply. The estuar-
ies that have formed during the rising of sea levels during the early Holocene can still
be found at present, since not enough sediments were supplied to fill them in during
the late-Holocene sea level fall. Therefore the present-day coastline still shows the
signs of early Holocene transgression and does not bear the typical characteristics of
regression. Classification apparently implies an integration of the developments over
a certain time period. At different moments in time a certain coast might occupy a

different position in the quadrant of Fig. 2.23. The classification class is therefore de-
pendent on the integration period. An example of a situation where progradation and
transgression have alternated over the geological history is the central Dutch coast.
Intermezzo 2.2 Geological delta formation
Most of the present active deltas are geologically very young features; some are
only a few hundred years old. Because a delta develops at the coast, its existence
is, in part, controlled by the sea level. It therefore was and still is vulnerable to sea
level rise, too. During periods of extensive glaciation, sea levels were much lower
and rivers traversed the present continental shelves, dumping their sediment loads
at or near the outer shelf edges. This suspended sediment cascaded down the
continental slopes in turbulent, high-density flows called turbidity currents. New
deltas did not form during this period, and deltas that had previously existed near
the positions of present-day coasts were abandoned and entrenched by rivers as
they flowed across the continental shelves. Melting glaciers brought a rapid rise
of sea level, and river mouths retreated so rapidly that deltas could not develop.
Finally, about 7000 years ago, the Holocene sea level rise slowed, and in some parts
of the world it stabilised at approximately its present position. Where conditions
were appropriate, deltas began to develop as large quantities of river sediment
accumulated.
Not all present-day deltas are only a maximum of a few thousand years old. Many
of them have formed on ancestral deltas built up during previous interglacial peri-
ods. A few, such as the Mississippi and Niger Deltas, are underlain by ancestral
deltas that formed tens of millions of years ago. The upper regions of these mature
deltas are also ancient, but their active delta lobes are only between 3000 and 6000
years old. The lower Mississippi Delta includes 16 detectable lobes. A new lobe
forms whenever the location of the river mouth changes. The channels of aban-
doned lobes fill up with sediment, contributed both by the river, by the waves and
by the tides of the coast. The present delta lobe of the Mississippi dates back only
600 years; its most active portion has developed since New Orleans was founded
in 1717.
2.5.3. More recent coastal development

A worldwide inventory by Bird (1985) indicates that approximately 70 % of the world’s
sandy coastlines have shown retreat over a period of decades; less than 10 % have
shown net progradation, while the remaining 20 % to 25 % have remained approxim-
ately stable. Recently, an analysis of satellite-derived shoreline data for 1984–2016
showed that 24 % of the world’s sandy beaches–sandy beaches making up 31 % of the
world’s ice-free shorelines–are eroding at rates exceeding 0.5 m/yr, while 28 % are ac-
creting and 48 % are stable (Luijendijk et al., 2018).

2.6. Nature and abundance of coastal material 65
While shoreline retreat on a geological timescale is undoubtedly connected with eu-

static sea level rise, it is more or less generally assumed that even the relatively small
sea level changes of the last century are driving this worldwide tendency of shore
retreat (cf. Vellinga & Leatherman, 1989).
The rate of sea level rise is likely to increase during the 21st century, although con-
siderable controversy exists about the likely size of the increase. The latest3 Interna-
tional Panel on Climate Change (IPCC) projections indicate a sea level rise range from
0.28 m to 1.02 m by 2100 relative to the 1995-2014 average (IPCC, 2021). Any future
rise in the mean sea level will result in the retreat of unprotected coastlines. Low-lying
countries without resources for extensive coastal defence are the most vulnerable to
this danger. Figure 2.25 shows Bangladesh, which lies in the Ganges-Brahmaputra
Delta on the Bay of Bengal. About 25 000 km2 (18 % of the total land area) will be in-
undated with a sea level rise of 1 m. A sea level rise of 0.5 m by the year 2100, which is
within the IPCC estimated range of ‘global’ sea level rise, would inundate more than
10 000 km2 .
It must be borne in mind though that there generally exists a complicated interaction
of agents affecting shore retreat (as illustrated in Fig. 2.26). For instance, both the
sea level rise relative to the land and its effects are rather site-specific. Besides the
eustatic sea level rise, local contributions can for instance be due to glacial rebound,
subsidence, compaction, and changes in ocean circulation. Additional (and often dom-
inating) causes of erosion can be due to alongshore or cross-shore losses, which in
their turn can be due to a variety of causes such as the physical geometry (e.g. head-
lands, submarine canyons), hydraulic boundary conditions (e.g. related to waves, tides,
wind) or human interference (e.g. harbours, erosion-mitigating structures). As seen in
Ch. 1, human interference at the land side can also influence the sediment budgets of
the coastal zone, for instance by the construction of dams, as has been done in many
rivers, especially in Asia.
2.6. Nature and abundance of coastal material

2.6.1. Sources of sediments deposits
This book focuses on depositional coastal environments. By that we mean coastal
environments consisting of soft or loose material (mainly sand or mud) that has been
deposited there at some point in the past. At present these materials might be subject
to either erosion or accretion due to the impact of waves and tides. The loose material
is present in the form of shorefaces, beaches and accompanying dunes and barriers.
Besides by dunes, beaches can also be backed by rocky material. The loose material
3
The Working Group I contribution to the 6th Assessment Report, 2021; the 5th Assessment Report was
finalized in 2014 (IPCC, 2014).

B r a hma p
ut ra
Bangladesh
India
G ange
India s
Myanmar
(Burma)
50 km
2006-2011
water depth > 5 m land elevation < 20 m 1 2 3 4 [m] sea level rise and
corresponding
water depth < 5 m land elevation > 20 m coastal inundation
Figure 2.25: The low-lying delta of the Ganges-Brahmaputra. The map shows the extent of the
projected flooding that is likely to occur for a sea level rise from 1 m up to 4 m as estimated from
elevation contours.
that we find on coasts can have various origins. We distinguish continental sediments
and carbonate sediments:
Continental sediments are the major type of sand of the coastal area and are formed
from weathered continental rock, usually granite. They are composed of silicate
minerals, with quartz and feldspar being the most abundant. Quartz is very res-
istant, feldspar easily weathers into fines. Granite and basalt are also the source
of heavy minerals (with a density larger than quartz). Streaks and patches of
dark-coloured grains – rich in heavy minerals – can be found on some beaches.
The majority of continental sediments found in coastal deposits is not from cur-
rent rivers, but consists of older Holocene sediments or Pleistocene sediments
reworked during the Holocene by marine processes.

riverine discharge temperature
sources
shoreline erosion evapotranspiration
onshore transport precipitation
aeolian processes climate
sediment coastal
budget processes wave climate
shoreline accretion longshore currents
storm washover riverine discharge
tidal inlets valley aggradation
coastal structures or incision
aeolian processes tides
sinks
offshore transport wind

resource extraction relative
human storms
sea level
activity
rise
subsurface fluid withdrawal coastal structures tectonic subsidence

river basin development artificial passes compactional subsidence
maintenance dredging dune alterations eustatic sea level changes
beach maintenance highway construction secular sea level changes
Figure 2.26: Interaction of agents affecting shore retreat and erosion according to Morton
(1977).
Carbonate sediments are formed from calcium carbonate. Most of these grains are
fragments of shells or remains of marine life. Marine carbonates are the second
major source of sands on the coast.
2.6.2. Sediment sizes

Not only the source but also the size of sediments varies. The size of sediments found
in a coastal system can range from large boulders with diameters exceeding 25 cm to
very fine material like clay. The size of the sediment grains can be used to classify the
sediment. An example of this classification is the Wentworth classification, often used
by geologists (see Table 2.2).
According to this classification, sand is defined as grains with a size between 63 µm
(lower limit very fine sand) and 2 mm (upper limit very coarse sand). Sand-sized ma-
terial consists mostly of quartz or carbonate and is the most abundant coastal material.
Sand is a non-cohesive material; the individual near-spherical grains do not stick to-
gether and are very resilient. Material with a grain size smaller than 63 µm can be
defined as either clay or silt. While silts tend to have larger particle sizes than clays,
silts and clays are mineralogically and chemically quite distinct, see Intermezzo 2.3.

Often both silt and clay particles are found together. A fluid-sediment mixture of (salt)
water, silt, clay and organic materials is named mud. Also, some very fine sand may
be present in mud.
Based on the size of the beach material, the world’s beaches can be classified as muddy
coasts, sandy coasts or gravel/shingle coasts. The main focus of this book is on sandy
coasts, although the finer sediments of muddy coasts are also considered.
Table 2.2: Unified Soils and Wentworth classifications of sediments on the basis of size. ASTM
mesh refers to the ASTM standard for sieve aperture sizes and 𝐷𝑛 is the nominal diameter. For
definition of the 𝜙 scale, see Sect. 6.2.2.
Unified Soils ASTM Dn ϕ Wentworth

Classification mesh [mm] value Classification
boulder
cobble
cobble
coarse gravel
fine gravel pebble
coarse sand
gravel
very coarse sand
medium sand coarse sand
medium sand
fine sand fine sand
very fine sand

silt
silt
clay
clay
colloid
Intermezzo 2.3 Properties of silt and clay
Silts are soils with fine, nearly spherical grains that do not include clay minerals.
Mineralogically, silt is composed of quartz and feldspar. Clay on the other hand
mainly consists of clay minerals that originate from the weathering of feldspar.
Chemically, clay minerals are layered metal silicates: thin plate-shaped particles
held together by electrostatic forces (Figs. 2.27 and 2.28). Hence, clays have strong
cohesive properties.

(a) (b) (c) (d)
Figure 2.27: Elementary particle arrangements of clay with (a) interaction between indi-
vidual particles in a ’card-house’ structure (flocculation); (b) dispersed individual particles;
(c) and (d) examples of group interaction showing also plate to plate contact.
Figure 2.28: 3D impression of a group of clay particles.
When the distance between particles is small enough, they can become stuck in
a ‘card-house’ structure. This process, called flocculation, is enhanced by ionic
constituents present in saline environments and the presence of organic material.
2.6.3. Geographical variation

The material present on beaches and shorefaces shows a latitudinal zonality. Fig-
ure 2.29 shows the relative frequency of inner continental shelf sediment type (up
to about 60 m water depth). It was first published by Hayes (1967) who described the
following correlations between coastal climatic zones and sediment types:
• Mud is most abundant off areas with high temperatures and high rainfall (humid
tropics);
• Sand is abundant everywhere and increases to a maximum in the subtropics
and lower mid-latitudes (between 20° to 40°). These are the intermediate zones
of moderate temperatures and rainfall. Sand includes both quartz and carbonate
sands;
• Gravel is most common off areas of low temperature (subpolar and polar);
• Coral is most common in areas with high (water) temperatures;
• Rock is generally more abundant in cold areas but its distribution is also strongly
controlled by plate tectonic setting (see Sect. 2.3.2);
• Shell distribution shows no strong latitudinal dependency.
These patterns are strongly controlled by climatic factors influencing the availability
of the material to the coast and the coastal processes.

100
shell
coral
80 rock & gravel ?
mud
percentage
60
40
?
sand
20
?
0
0° 20° 40° 60° 80°
latitude
Figure 2.29: Relative frequency of inner continental shelf (up to about 60 m depth) sediment
types by latitude according to Hayes (1967).
Availability of the material to the coast

The availability of the material to the coast is determined by:
1. Topographic relief and global precipitation pattern, and hence the presence or
absence of large rivers, determine the availability of sediments to the coastal
system (see Sect. 2.3.3). Figure 2.30 shows the global distribution of mechanical
erosion and the solid discharge of the world’s major rivers. This indicates that
the supply of sandy sediments to the coast is largest between 40°N and 40°S.
Figure 2.31 illustrates the influence of topographic relief and catchment size on
the beach material;
2. The type of weathering (mechanical or chemical) determines whether a coast
receives mainly sand or mud. Mud is produced through chemical weathering
which is enhanced for high temperatures in the source area and for high rainfall.
Hence chemical weathering is enhanced in the tropics. The ratio of mud to
sand supply to the coast increases as chemical weathering increases. If chemical
weathering is intense, the majority of the sediment reaching the coast is mud;
3. Pleistocene glaciation (see Sect. 2.4) has produced rocks and gravel, which are
thus restricted to subpolar and polar zones. Gravel is generally too coarse to be
transported to the coast by rivers (except on leading-edge coasts);
4. High water temperatures in the tropics are responsible for the growth of or-
ganisms responsible for the formation of coral and carbonate sands (see Inter-
mezzo 2.4). Figure 2.33 shows the global distribution of coral reefs and carbonate
sands. There is a latitudinal and longitudinal variation; at low latitudes and in
the west of the oceans the temperature is more favourable (higher) than in the
east. This is caused by the tropical trade winds blowing towards the west (see
Ch. 4). By contrast, most of the sands outside the tropics, e.g. in European
waters, consist of quartz minerals.

18
80° 0° 0°
12
12
60°
0°
°
60
0°
60°
40°
20°
0°
20°
40°
60°
80°
water-related mechanical erosion arid area solid discharge
[tons/km²/year] mainly wind erosion [106 tons/year]
0 50 100 240 2000 1000 500 250
Figure 2.30: Global distribution of mechanical erosion.
Figure 2.31: Spiaggia del Cannello, a 350 m long beach on the southernmost part of the island
Elba, Italy. It consists of white and black pebbles that get smaller nearer the water line. The
coarse beach material is supplied by short and steep rivers. Photo by Erik Mosselman (‘Credits’
on page 575).
Intermezzo 2.4 Coral reefs
The term coral reef refers to rigid subaqueous limestone formations consisting of
calcium carbonates. The limestone formations are accumulations of the cases that
polyps build around themselves (using calcium from the water) and that remain

after the polyps have died. The case of one dead polyp forms the foundation of the
case of a next polyp. The coral reef ecosystem is based on a closed energy cycle.
Warm water and the penetration of sunlight are essential to the development of
coral reefs. The Indo-Pacific region has the most extensive coral structures. In the
Atlantic region the corals are most often attached to other structures and therefore
are said not to be true coral reefs. Coral reefs are, like tropical rain forests, among
the most complex communities on the earth, and rock-producing reef communities
are among the most ancient life forms found in the fossil record. Because of their
complexity, the dynamics of coral reefs are not yet well understood.
Coral reefs are important for tourism, fisheries and shoreline protection. Many
low-income communities depend on reefs to protect their property against flood-
ing by high tides and wind set-up. Unfortunately, during the last centuries coral
reefs have been adversely affected by humans. Some of the most widespread
impacts are water pollution from various human activities, deforestation (and
thus soil erosion leading to increased turbidity), dredge and fill operations, over-
harvesting of fish and shellfish, and the harvesting of some corals for souvenirs.
All forms of stress on the coral retard its growth. Destruction has long-lasting
effects, since natural recovery would take thousands of years. Implantation exper-
iments have not been successful.
Where reefs border the coast they are termed fringing reefs; where they lie off-
shore, enclosing a lagoon, they are known as barrier reefs; and where they encircle
a lagoon, they are called atolls (Fig. 2.32). Barrier reefs and atolls form under the
influence of relative sea level rise, the reefs growing vertically upward with the
rising water level. Atolls form from fringing reefs surrounding an island. If the
land sinks or the sea rises, the polyps build upwards and seawards in a ring around
the island. When the island is completely submerged, a lagoon is formed with a
ring of atolls around it.
fringing reef barrier reef atoll
land
reef
Figure 2.32: Reef types.
Atolls are primarily found in isolated groups in the western Pacific Ocean and the
Indian Ocean (Maldives). Small low islands composed of carbonate sands may
form on these reefs. These islands are quite vulnerable to inundation, and to trop-
ical storms. Reef islands are naturally dynamic; carbonate sediment production,
erosion, deposition and cementation can occur concurrently.

18
80° 0° 0°
12
12
60°
0°
°
60
0°
60°
40°
20°
0°
20°
40°
60°
80°
coral reefs shelf carbonates 18° isotherm of the coldest month
Figure 2.33: Extensive coral reefs and known areas of major supply of shelf carbonate sands to
the coast are found at lower latitudes. Coasts at these latitudes are named low-latitude coasts.
The occurrence of coral reefs is to a large extent governed by temperature as indicated by the
18 °C isotherm (of air) of the coldest month.
Coastal processes
Besides availability of the material, coastal processes also play an important role in
determining the type of material found at a certain location:
• Erosion of rock into gravel by coastal processes;
• The winnowing out of muddy material by wave and tidal processes. In energetic
wave environments, fine muddy sediment cannot settle. Sand on the other hand
is deposited on the coasts in wave-agitated energetic conditions. This means
that even though fines dominate many of the world’s river sediment discharges,
they will only deposit in low-energy conditions, such as in estuaries and far off-
shore. On a global scale, the moderate wave climate in the tropics and subtropics
(see Ch. 4) enhances tropical mud deposition at low-latitude coasts;
• In the case of large concentrations of silt and clay particles, the concentration of
the particles is large enough that enhanced deposition through flocculation can
take place.
2.6.4. Muddy coasts

Muddy coasts can be found at all latitudes and on all continents, but are mostly found
in tropical areas, particularly in Asia, where they settle in quiet areas and near river
mouths. An example of a muddy coast is the Guyana coast in South America. This

1600 km long coast of Guyana, Surinam, French Guyana, and parts of Brazil and Venezuela
consists mainly of mud transported to the coast by the Orinoco and Amazon Rivers.
Along large parts of the coast, mangroves are found (see Sect. 2.6.6). Other examples
of mud coasts can be found near the mouths of the Mississippi River (USA) and the
Yellow River (China). Mud also plays an important role in parts of the coastal system
of the Netherlands. The low-lying parts of the Netherlands are mainly formed by sed-
iments carried by the rivers Rhine, Meuse and Scheldt. Especially the estuarine area
in the southwest (Zeeland) and the Wadden Sea are influenced by mud (see Fig. 2.34).
Besides, there is a large supply from the southern North Sea and the Channel.
Figure 2.34: Satellite image of the Wadden Sea from the Sentinel-2 mission of April 5, 2020.
Sentinel satellite data are retrieved from satellietdataportaal.nl (https://www.spaceoffice.nl/nl/
satellietdataportaal/). See also https://sentinel.esa.int/web/sentinel/missions/sentinel-2 for
more information about the Sentinel-2 mission.
2.6.5. Sandy coasts

Short (2005) lists the most extensive sand coasts on each continent:
North America the southeast and Gulf coasts, both low-gradient passive-margin coasts
supplied by numerous rivers including the Mississippi, with the sand reworked
onshore by a high-energy wave environment. However, coastal dunes are poorly
developed. Furthermore, parts of the west coast exposed to high shelf sediment
supply and winds to build dunes.
South America massive long-term sand supply to the entire east coast, leading to es-
sentially a sand barrier/dune coast from the Amazon south to Argentina. Long
beach-barrier-dune systems exist for much of the coast.

Africa beach, barrier and dune systems ring most of the continent, with sand supplies
by local rivers, including the Nile, Niger, and Orange, and on the most exposed
coasts there is shelf supply of quartz, and in the south also of carbonate sands.
Eurasia sand dominates most of the exposed western (Europe-Mediterranean), south-
ern (India), and eastern shores (Southeast Asia-China), with some substantial
river systems and deltas and extensive sand barriers in South India and Sri Lanka.
Australia 50 % of passive-margin coast consists of sand deposits, with low-energy
beaches in the north and high-energy beaches and dune systems across the
south. Supply comes from rivers and shelf in the north, from shelf quartz in
the southeast and from carbonate in the south and west.
2.6.6. Vegetation
Mangroves and salt marshes
On the fringes of estuaries and other basins – in the intertidal areas between high and
low water level – salt-resistant vegetation can establish a foothold. The vegetation
requires not only calm conditions but also a silty substratum to germinate. The silty
character keeps the soil moist, also during low tidal levels. The calm conditions and the
silty soil are closely related, since only in such conditions the finer particles can settle.
When conditions remain favourable, the number of species increases and the plant
cover gets denser. The type of vegetation depends strongly on climatic conditions.
While salt marshes are mostly found in moderate climate zones (hence on mid-latitude
coasts, e.g. along the southern North Sea), mangroves favour tropical and subtropical
climates (see Fig. 2.35).
A salt marsh may be a few metres wide or it may occupy the entire estuary except
for the tidal channels. In Fig. 2.36, a cross-section of a salt marsh is drawn. This
figure shows the different zones within the marshland, that can be distinguished by
the different species that are present. The upper limits of the salt marsh coincide
with the landward or upper limit of the spring high tide, the highest level of regular
inundation and sediment supply.
Individual marsh flats can develop into extremely valuable nature resorts. Apart from
many specific vegetation species, animals use them for breeding, feeding and during
seasonal migrations. Beautiful examples of such marshes in the Netherlands are the
Wadden Sea and the tidal flats of Eastern and Western Scheldt (see Fig. 2.37).
For hundreds of years, the Dutch, Germans, and Danes have been converting marshes
to farmland by draining them through a system of dams, dikes, and canals. This pro-
cess has now been stopped, mainly because the ecological value of the tidal wetlands
has been recognised and the Wadden Sea has been declared a nature reserve.
The marsh environment is quite similar to that of river and delta floodplains with
channels, meandering or not, cutting through the marshy plain. This system delivers

18
0° 0°
80° 12
12
°
60°
0°
60
0°
60°
40°
20°
0°
20°
40°
60°
80°
salt marshes 18° isotherm of the coldest month
mangrove swamps 10° isotherm of the hottest month
Figure 2.35: Global distribution of salt marshes and mangrove swamps. Mangroves are mostly
found at low-latitude coasts and salt marshes at mid-latitude coasts. The climate zones are
indicated by the isotherms of (air) temperature of the hottest and coldest months.
HAT
MHWS upland
high marsh
MHWN h
mars
1m
MSL
low
0.5
mudflats
0
0 50 100 m
Figure 2.36: Typical cross-section of a salt marsh. The tidal levels are explained in App. C.
sediment to the marsh in two ways: by regular but slow flooding of the marsh by turbid
water carried by sluggish currents that permit settling; and by storm tides that push
large amounts of sediment-laden water onto the marsh and deposit considerable sed-
iment in a short time. By reducing current velocities near the bottom, sedimentation
is enhanced by the vegetation. In addition to their positive role in catching sediment
to the substrate, marsh grasses are very important sediment stabilisers. They prevent
or inhibit currents and waves from removing sediment from the vegetated substrate,
partly by their root system (strengthening the soil), partly by reducing current velo-
cities. An extensive marsh is a sign of a natural estuary that has largely filled with

Figure 2.37: Lower salt marsh at Western Scheldt. Photo by Marcel J.F. Stive (‘Credits’ on
page 575).
sediment. Cord grass is a type of vegetation often introduced artificially to enhance

siltation and formation of new farmland.
Although a paradigm for marsh development has been given here, the present global
situation is one of eroding marshes due to sea level rise.
Mangroves are woody trees of various taxonomic groups (Fig. 2.38). Thick tangles of
shrub and tree roots, commonly called swamps but properly known as mangles, form
an almost impenetrable wall at about water level (Fig. 2.39). Most trees grow from 2 m
to about 8 m high, although some are much higher, depending on the species and the
environmental conditions – rare stands may be twice that height. The root systems of
mangroves are not only dense, but also diverse in appearance and function. In places
like tropical Australia or India, there are more than 20 species of mangroves.
Like salt marshes, mangroves accumulate and stabilise sediment and protect the coast-
line from erosion. They provide a resilient coastal defence against storms and hur-
ricanes as long as the conditions are not too rough. Furthermore, mangroves play an
essential role as the incubator of the coastal ecosystem. The thickets of mangrove roots
at the water line provide a sheltered habitat for a special community of organisms that
are adapted to an environment intermediate between land and water. Barnacles and
oysters encrust the roots and branches. Fish, snails, and snakes all find protection,
nesting sites, and food among the roots.
Because the mangrove swamps are rich in expensive seafood like shrimp, the shores
are often turned into artificial fish and shrimp farms. Destruction of mangrove forests
and their replacement by shrimp farms is a major factor responsible for the increase

Figure 2.38: Typical cross-section of a mangrove swamp with different species present at dif-
ferent elevations. The tidal levels are explained in App. C.
in the severity of flooding in many coastal areas, especially in India and Indonesia.
Water pollution and cutting of trees for firewood also hampers the continued growth of
mangrove forest. This means that mangrove vegetation is disappearing rapidly. Once
the forest has disappeared, it becomes clear how effective the forest was in preventing
erosion of the generally silty coastline. Reforestation is extremely difficult since the
young trees are quite vulnerable.
Dune vegetation
The beach above the high water line is another place where non-marine plant life can
survive (see e.g. Fig. 2.40). On the dry beach conditions are still quite harsh: the sand
has very little capacity to hold moisture, so that plants growing here must be drought-
resistant. When the first vegetation develops on the dry beach, it forms a nucleus for
the formation of dunes. The first vegetation provides some shelter against the wind,
which enables sand to accumulate and its roots stabilise the sand. In this way, the first
plants create small, slightly elevated undulations in the flat beach. These higher places
provide storage areas for fresh water, creating more favourable conditions for species

2.7. Process-based classification 79
Figure 2.39: Mangroves in the Mekong delta, Vietnam. Photo by Marcel J.F. Stive (‘Credits’ on
page 575).
such as the well-known Marram (or Helm in Dutch). Marram has a more extensive root
system and forms a dense cover with its stems and leaves. In this way, small dunes
are formed, and the higher the dunes, the better the conditions for a wider variety of
vegetation. In the dunes, freshwater is caught and this drains slowly to the lower parts
of the slopes where species requiring more water can establish themselves. Note that
part of the Dutch drinking water originates from dune water.
The vegetation cover of the dunes prevents the ‘wandering’ of dunes by aeolian (wind-
blown) transport, which is very important since dunes form an important part of the
sea defence, not only in the Netherlands, but also in other parts of the world. In the
Netherlands Marram is used extensively to provide artificial protection to young dunes
and to prevent wind erosion. The species are site-specific, in the sense that the com-
position of the soil and the climate play an important role in what species are able to
survive. This means that the use of vegetation to stabilise sandy shores must always
be based on observations of locally available and successful species.
2.7. Process-based classification

2.7.1. Dominance of fluvial, wave or tidal processes
Sections 2.3 and 2.5 described the classification of coastal systems on the basis of plate
tectonic setting and substantial sea level changes respectively. Superimposed on these
large-scale characteristics are regional- and local-scale variations in coastal landforms.
One way to distinguish between features at these scales is to look at the hydraulic

Figure 2.40: Primary vegetation development at the Sand Engine (in Dutch: De Zandmotor), The
Netherlands. Photo by Marcel J.F. Stive (‘Credits’ on page 575).
boundary forcing to the coastal system; the dominance of fluvial, wave or tidal pro-
cesses can be seen to influence the shape of coastal features.
A process-based classification on the basis of the relative importance of fluvial sedi-
ment supply and wave and tidal action is relevant to trailing-edge coasts (such as the
USA East Coast) and to the typical Dutch situation (being an example of a marginal
sea system; other ones being the Gulf of Mexico and the Chinese coast).
Such a process-based classification of coastal systems implies that we can distinguish
between various typical coastal morphologies on the basis of two criteria:
1. How important has fluvial sediment input been for shaping the coastal system?
2. Are sediment deposits being reworked by waves or tides (wave dominance or
tide dominance)?
Based on the answers to those questions, the following systems can be discerned (with
the numbers in the list referring to the numbers in Fig. 2.41 and the respective sketches
in Fig. 2.42):
1. Wave-dominated uninterrupted coastline (or pocket beaches) with tidal influ-
ence at the lower shoreface only
2. Wave-dominated deltaic (i.e. built out by fluvial sediment supply) coastline
3. Deltaic coastline with no significant wave or tidal influence
4. Tide-dominated deltaic coastline
5. Estuary with fluvial deposits shaped by tide
6. Estuary (tide-dominated)

7. Tidal inlet system dominated by tide with combined wave and tidal influence at
the inlet
8. Wave-dominated barrier system
3
deltas
4
inance
dom
er
riv
5
2
in a n c e
w av
om
stra
ed
rie s
ed
1 m 6
tid
nd
in
tu a
an
pla
ce
es
in
s
l
at
lf
s
go a
ti d
la
on
s s
in
b as
d al
d ti
8 w ave b a r ri e r s a n
7
Figure 2.41: The relative influence of fluvial, wave and tidal processes influences the shape of
coastal features. The numbers in the figure correspond to the typical morphologies of Fig. 2.42.
2.7.2. Ternary diagrams for progradation and transgression

Boyd et al. (1992) used the classification in prograding and transgressive coasts as a
starting point and added the effect of domination by waves or tides. In prograding
situations, the landside is on the winning hand, either because of a falling sea level rel-
ative to the land, or because of an excessive sediment supply. Transgression takes place
either because of a rise in sea level, or because of insufficient sediment supply. Note
again that the change in sea level is relative; subsidence of the land with a constant sea
level has the same effect. Since Boyd et al. (1992) only took depositional environments
into account, the distinction between prograding and transgressive coasts boils down
to the distinction between shoreline advance and shoreline retreat (see Fig. 2.23).
In the progressive case, deposition of river sediment leads to delta formation. When
wave power and tidal power are low, the sediment of the river will build up long nar-
row banks on both sides of its course. Due to the gradient of the river flow, water levels
at a fixed point along the river will gradually rise, since the distance of this point from
the actual river mouth is increasing. At a certain moment, most likely when river dis-
charge is high, the river starts overflowing the bank and it will erode a new, shorter
channel towards the sea. The same process is continuously repeated, which leads to

1. 2.
wave-dominated wave-dominated
uninterrupted deltaic
coastline coastline
3. 4.
deltaic coastline tide-dominated
(no significant deltaic
wave/tidal coastline
influence)
5. 6.
estuary with estuary
fluvial deposits (tide-dominated)
shaped by tide
7. 8.
tidal inlet system wave-dominated
with local wave barrier system
influence
Figure 2.42: Examples of different coastal systems; the figure shows typical coastal morpholo-
gies developing in response to the relative influence of fluvial sediment supply, tide action and
wave action.
an ‘elongate’ or ‘bird-foot’ delta (see also Sect. 2.7.3). Strong waves with longshore cur-
rents tend to stretch the delta coast parallel to the general orientation of the shoreline,
while strong tidal action usually creates patterns perpendicular to the shoreline. Out-
side the influence of the river, a strand plain develops when wave action is dominant
and tidal flats develop when tidal action is the strongest.
In the transgressive case, an estuary is the equivalent of a delta in the prograding case,
but now, the sediment supply is not enough to keep pace with the relative sea level

rise. The sediment is no longer merely fluvial, but also has a marine source, since the
flood tide or waves bring in sediment from the sea. A lagoon has a marine sediment
source only, as no river is flowing into it.
Based on the various processes, Fig. 2.43 gives a classification for prograding and trans-
gressive coasts. Figure 2.43a represents prograding situations, whereas Fig. 2.43b rep-
resents the transgressive case. Either ternary diagram presents the fluvial power on a
vertical axis from the top corner (100 %) to the bottom (0 %). The coastal powers are
represented on axes running from the lower two corners (100 %) to the middles of the
opposite corners (0 %); the left corner indicates 100 % wave power and the right corner
100 % tidal power. The top of the triangle represents deltas; the bottom strand plains
and tidal flats; estuaries are situated in between. A distinction can be made between
wave- and tide-dominated estuaries. This will be further elaborated on in Ch. 9. In the
ternary diagram, lagoons form the end member of the estuary spectrum.
The two diagrams together give an idea of the evolution of coastal systems over time,
relative to the change in sea level and sediment supply. For instance, with a rising sea
level, deltas change into estuaries (transgression). Strand plains and tidal flats vanish
and become shelf when the sea level rises.
river/fluvial source river/fluvial source

a f
rs
rs
yea
0s lly
ea
s o tly
g deltas h
00 ca
fy
of
t 1 ecen
t 1 stori
00
r
deltas
hi
i estuaries j
las
las
b c
k lagoons l
d strand plains tidal flats e
strand plains tidal flats
m n
waves/marine source tides waves/marine source tides/marine source
(a) high sediment supply, low rate of sea (b) no or reduced sediment supply, high rate
level rise, coastal progradation of sea level rise, marine transgression
Figure 2.43: Ternary coastal form classification diagrams for transgressive and prograding
coasts (reinterpreted from Boyd et al., 1992; Dalrymple et al., 1992). The letters correspond
to real-life examples listed in Table 2.3.
It is interesting to note that the triangles in Fig. 2.43 depicting deltas have been sep-
arately described by Galloway. Section 2.7.3 describes in detail how the various delta
shapes are influenced by the dominance of fluvial, wave or tidal processes.

2.7.3. Classification of deltas

The formation of a delta depends on the interaction between the river flow and sed-
iment supply on the one hand and the distribution of the river sediment by waves
and tidal currents on the other hand. William Galloway recognised the relative influ-
ence of these three major factors affecting delta development (river, waves and tide)
on the morphological structure. He proposed the triangular classification diagram of
Fig. 2.44, in which river-dominated, tide-dominated and wave-dominated deltas are
distinguished.
Wax
Mississippi
50 km
25 km
sediment input
ate
elo
Wax
ng
ng
ate
elo
Mississippi
Po
Ebro Yukon
river
25 km dominated
Mahakam
ate
Danube
lob
Mahakam
Ebro 25 km
Nile Orinoco
Niger Mekong
wave tide
dominated dominated
Rhone Fly
es
Kelanton
ate
tua
Brazos Senegal Colorado

sp
rin
cu
Sao Francisco Copper Ganges-Brahmaputra

e
waves tides
25 km Fly
Copper
Sao Francisco 25 km 25 km
Figure 2.44: William Galloway’s triangular delta classification diagram. The colors represent the
relative influence of waves (green), tides (blue) and fluvial sediment input (red). After Galloway
(1975).
As the water flows from the river mouth, its velocity decreases and it loses its capa-
city to carry sediment. Consequently, sediments accumulate in the river mouth area.
As the velocity of the outflowing river water decreases, the coarse material settles

first, followed by the finer sediments. At the seaward edge of the delta front, the sus-
pended sediment in the river water finally settles in deeper coastal water. This mud
accumulation is generally very thick and extends across part of the continental shelf.
The amount and configuration of sand accumulating in the delta-front depends on the
relative roles of the interacting river, wave, and tidal currents.
A common type of sand accumulation is a sandbar that forms just seaward of the chan-
nel mouth and typically causes the channel to bifurcate. Another is the formation of
banks along the sides of the channel. As the river deposits sand in the mouth, a situ-
ation can be reached in which the water level is affected by the sand deposits. The river
can then overflow its banks and then divide into distributary channels. Each distribu-
tary channel then continues to transfer massive amounts of fine-grained sediment to
the coastal area. When this new-born delta is situated in an environment with little
tide and wave action, it is categorised as being river-dominated. It can grow out into a
bird-foot type of delta. Examples are shown in Figs. 2.45 and 2.46 (Danube Delta and
Mississippi River, respectively).
Tide-dominated deltas develop where a large difference between the high and low tides
leads to strong tidal currents. On these coasts, the wave height is moderate to low,
and wave-induced currents along the coast are weak. These deltas resemble estuaries
because of their embayed setting of salt marshes, swamps, and tidal flats. An example
of such a tide-dominated delta is the delta of the river Fly on the south coast of Papua
New Guinea (Fig. 2.47).
If the wave climate is more severe, the bars at the river mouth are affected by the waves.
As a result of the wave action, sand is repositioned by alongshore and cross-shore
effects. Depending on the wave direction, this can lead to a delta that is symmetrical (in
the case of waves perpendicular to the coast) or asymmetrical (in the case of obliquely
incident waves) in shape.
Wave-dominated deltas typically have a rather smooth shoreline with well-developed
beaches and dunes. The delta plain tends to have few distributaries; some deltas of
this type have only a single channel. A wave-dominated delta is generally smaller
than other types, because the distributing power of the waves striking the delta front is
stronger than the carrying power of the river. When the wave climate is strong enough
to carry all the river sediment away, the delta will shrink and eventually disappear.
Two different shapes characterise these deltas. The general shape is symmetrically
cuspate. One of the best examples is the delta of the São Francisco in Brazil (Fig. 2.48).
The other shape is characterised by a strong longshore current. A sand spit (an elong-
ated and narrow accumulation of sediment that is attached to land at one end, see
Ch. 8) develops and protects the extensive wetlands that cover the delta plain. An
example of this is the Ebro delta (Fig. 2.49).

Kilia lobe
polder
Dan
u be
10 km
2002-2009
water depth > 5 m low-lying land beach/dune
water depth < 5 m terrace beach/dune deposits
Figure 2.45: Historically, the Danube delta has developed as a river-dominated delta, as can still
be recognised from the bird-foot shape of the northern Kilia Lobe. The southern part of the
present-day delta shows evidence of wave dominance.

ss issippi
Mi
New Orleans
Wax
Atchafalaya
25 km
2005-2010
water depth > 5 m inhabited land marsh beach/dune
water depth < 5 m low-lying land terrace
Figure 2.46: Mississippi delta system.The main branch (to the right) has developed as a bird foot
delta. Its present course, however, is strongly constrained by the man-made levee system. The
smaller Wax and Atchafalaya deltas (to the left) are clearer present-day examples of a bird-foot
delta.
Fly
25 km
2003-2005
water depth > 5 m water depth < 5 m land
Figure 2.47: Delta of the river Fly, Papua New Guinea.

o
ci s
c
an
S ã o Fr
5 km
2000-2010
water depth > 5 m land beach/dune deposits
water depth < 5 m beach/dune
Figure 2.48: São Francisco Delta, Brazil.
o
Ebr
5 km
2009
water depth > 5 m low-lying land beach/dune
Figure 2.49: Ebro delta, Spain, with a sand spit at the southern end.

2.8. Summary of coastal classification 89
2.7.4. Overview and examples of coastal forms

Table 2.3 not only gives an overview of the variety of coastal forms but also presents
real-life examples. The table distinguishes between 1) primary coastal forms (deltas,
estuaries, plains and flats, and barrier and lagoon coasts); 2) secondary coastal forms
based on the influence of wave action and tidal action and 3) sediment-rich and sediment-
poor systems. For each resulting class, an example is given and linked to the tern-
ary diagrams of Fig. 2.43. The latter includes the triangular classification diagram of
Fig. 2.44. The typical morphologies listed in Sect. 2.7.1 can also be positioned in the
ternary diagrams of Fig. 2.43.
2.8. Summary of coastal classification

Classification can be used as a means to inventory the large variety of coastal systems.
It may be based on:
• Material – hard/soft, origin, size;
• Tectonic controls – e.g. Inman and Nordstrom (1971);
• Sea level criterion – e.g. Valentin (1952);
• Dominant processes.
The diversity in coastal systems is the result of the simultaneous occurrence of all
coastal system determining factors in a nearly infinite number of combinations. That
is why there exists no truly unique classification system of coastal ocean systems and
why one encounters a large variety of classification systems in literature. Classification
is scale-dependent; the tectonic classification for instance describes only the broadest
features of a coast. Also in process-based classifications, the levels of aggregation can
vary.
The process-based classification on the basis of the relative importance of fluvial sedi-
ment supply and wave and tidal action comes closest to the approach in this book, that
discusses the characteristics of coastal systems based on the (hydrodynamic) boundary
forcing to the coastal system and resulting sediment transport processes.

Table 2.3: Classification of coastal forms with real-life examples. The examples are indicated by letters that correspond to the letters in Fig. 2.43.
90
Primary coastal Secondary coastal form Sediment rich Sediment poor

form
Yellow river, China (a) Mississippi, USA (ca. 1975–present) (f)
River-dominated deltas Mississippi, USA (until ca. 1975) (a)
Wax Lake, USA (a)
Deltas
Sao-Francisco, Brazil (b) Ebro, Spain (g)
Wave-dominated deltas
Senegal delta, Senegal (g)
Tide-dominated deltas Ganges (Megna estuary), Bangladesh (c) Amazon, Brazil (h)
Yangtze, China (poor in river sediment, tidal Pearl (Zhujiang River estuary), China (j)
Tide-dominated estuaries
Estuaries supply) (j)
Severn estuary (j)
Tide-dominated estuaries with Columbia River estuary (poor in river Mekong estuaries and coast, Vietnam (i)
wave-dominated ebb-delta sediment, wave supply) (i)
Western tips of Wadden Islands, The Eastern tips of Wadden Islands, The
Strand plains (beach ridges)
Plains and flats Netherlands (d) Netherlands (m)
Dune du Pilat, France (d)
Tidal plains Jiangsu coast (radial ridges in southern part), Jiangsu coast (northern part), China (n)
China (e)
Spatially varying dominance by Wadden Sea, Netherlands/Germany Redfish Pass, USA (between k and l)
Barrier & lagoon
waves & tide (between k and l)
coasts
Wave-dominated open and Albufeira lagoon, Portugal (wave supply) (k) Willapa Bay, USA (k)
seasonally closed lagoons
Tidal bays Baie de St. Michel, France (l) Hangzhou Bay, China (l)

2. Large-scale geographical variation of coasts
91
3
Ocean waves
3.1. Introduction
This chapter deals with ocean waves. By ocean waves we mean all oscillations of the
water surface generated in the ocean (see Sect. 3.2). The most important in shaping
the coastal zone are the short waves generated by wind and the longer tidal motion
generated by the attractive forces of the sun and the moon on the water masses of
the earth. In Sect. 3.3 we look at how waves can be measured. Statistical and spectral
representations of wind waves are treated in Sect. 3.4. Both wind wave generation
and the propagation away from the area of wave generation are treated in Sect. 3.5.
We see how ocean waves become longer and smaller when propagating away further
from their source due to the phenomenon of wave frequency dispersion and due to
frequency-dependent dissipation. Long term statistics are briefly discussed in Sect. 3.6.
The generation of the tide is explained and the origin of the different harmonic con-
stituents is treated in Sect. 3.7. We look into the propagation of the tide in the world’s
oceans and in doing so discuss propagation velocity, Coriolis forces and amphidromic
systems (Sect. 3.8). Tidal analysis and prediction are the topics of Sect. 3.9.
After having taken the TU Delft courses Ocean Waves (CIE4325), Hydraulic En-
gineering (CTB2410, in Dutch) and Open Channel Flow (CTB3350/CIE3310-09)
the larger part of this chapter on Ocean Waves should be familiar to you. This
particularly holds for the Sects. 3.2 to 3.6 on wind or short waves. You will find
that Sects. 3.7 to 3.9 treat the tidal generation and propagation more extensively
than CTB2410 does. In CTB3350/CIE3310-09 you will have encountered some as-
pects of tidal propagation as well. For those without any prior knowledge of wind
waves or tides in oceanic waters, this chapter deals with the main aspects that are
required in order to successfully follow Coastal Dynamics I (CIE4305).

92 3. Ocean waves
3.2. Oscillations of the ocean water surface

We can broadly define ocean waves as all sea surface variations on the timescale of
seconds to months as generated in the oceans. MSL is the sea level when these fluc-
tuations are averaged out. This definition of ocean waves includes wind waves, tides
and tsunamis. Wind and atmospheric pressure changes can also cause water level
variations.
The simplest way of representing a wave is by a sine or cosine function: a regular
variation of the water surface at a certain location. We will see later that such a sine
wave is a solution to the linearised equations describing the water motion. In such a
representation the temporal variation at one location is shown in the right-hand side
of Fig. 3.1.
c
L T
still
water
η a H level η a H
1 2 3
Figure 3.1: Simple representation of a wave. Left: the spatial variation is measured along the
direction of wave propagation at a single moment in time. Right: the temporal variation in water
level 𝜂 is measured over a certain time period at a single location (at location 2 and directly
following the recording of the spatial variation). A similar signal measured simultaneously at
location 1 (or 3) will have opposite sign.
The wave height 𝐻 is the vertical distance between the crest and the trough of the
wave and equals twice the amplitude 𝑎 for a sinusoidal variation. The wave period 𝑇
is the time the wave needs to pass the location, the inverse of which is the frequency
𝑓 , the number of waves passing a fixed location per unit time. When travelling in
the ocean at a certain moment in time, the wave can be seen as a similar sinusoidal
variation of the water surface, see the left-hand side of Fig. 3.1. The figure shows
the wavelength 𝐿 of the surface elevation deformation measured along the direction
of wave propagation 𝑥. The wavelength is the length a wave will travel in the wave
period 𝑇 . The ratio between wave height and wavelength is called the wave steepness
𝐻 /𝐿. The surface elevation 𝜂 can be described by:
𝜂 = 𝑎 sin(𝜔𝑡 − 𝑘𝑥) = 𝑎 sin 𝑆(𝑥, 𝑡) (3.1)
where 𝜔 is the angular frequency and 𝑘 is the wavenumber according to:
2𝜋 2𝜋
𝑘= ; 𝜔 = 2𝜋𝑓 = (3.2)
𝐿 𝑇

3.2. Oscillations of the ocean water surface 93
The units of 𝜔 and 𝑘 are rad/s and rad/m, respectively. They may also be written as
1/s and 1/m, respectively, since the radian is dimensionless, just like the degree.
The deformations propagate with the wave speed 𝑐:
𝐿 𝜔
𝑐= = (3.3)
𝑇 𝑘
In the linear (first-order) approximation, wave propagation is merely a matter of move-

ment of the wave form, not of mass. The water particles describe orbital or oscillatory
motions at particle speed and remain at the same position on average. In Sect. 5.5.1 we
will see that at second order there exists a non-zero net mass flux that becomes more
important for larger wave amplitudes.
There are various ways to discern between the different types of ocean waves, viz. by:
• the disturbing force, i.e. their mechanism of generation;
• the restoring force, i.e. the mechanism that dampens the wave motion;
• the length of the wave (represented by the wavelength, wave period or fre-
quency).
These different classifications are shown in Fig. 3.2 which shows the different ocean
waves as a function of their wave period. In this book we will discuss the waves
that are most important for our coastal system: wind waves (normal wind waves and
longer storm waves), tides, tsunamis and storm surges. The energy level as indicated
in Fig. 3.2 is a qualitative reflection of the amount of energy contained in these waves
and their relative frequency of occurrence. Storm surges and tsunamis contain a lot
of energy but are not as frequent as ordinary wind waves. Therefore the energy con-
tained in the wind wave range is higher.
Wind waves
Wind-generated gravity waves have periods ranging from 1⁄4 s to 30 s. They are called
gravity waves because gravity is the restoring force that dampens the wind wave mo-
tion by returning the particles to their average position in the water column. Local
wind fields generate relatively short (order 10 s, larger for larger storms) random and
irregular oscillations of the water surface that we call ‘sea’. These wind-generated
oscillations can travel large distances away from their area of generation. They will
then transform into longer, faster, lower and more regular ‘swell’ due to a process
called frequency dispersion and frequency dependent damping (see Sect. 3.5.2). Ca-
pillary waves ( < 1⁄4 s ) can be seen as small ripples on the water surface, but they
will die out very fast after the wind stops blowing. Their restoring force is surface
tension. Infra-gravity waves are longer gravity waves with periods up to 5 minutes.
Although deep-water infra-gravity waves do exist, infra-gravity waves are mainly a
shallow-water phenomenon. In shallow water (groups of) wind waves can generate

94 3. Ocean waves
trans- long- infra- ultra-

tidal period gravity gravity gravity capillary
wave waves waves waves waves waves waves
band
ordinary tidal waves
storms, storm surges, seiches

primary
disturbing sun,
force moon tsunamis wind
primary surface
restoring gravity tension
force
energy
period (s)
Figure 3.2: Sketch of the relative amounts of energy as a function of wave period in ocean waves.
The top section gives the classification based on wavelength, the section below the classifica-
tion based on the wave-generating force, and the bottom section the classification based on the
restoring force. After Munk (1950) and Kinsman (1965).
longer (30 s to 5 min) waves like surf beat, which has a period of a few minutes (cor-
responding to the length of a wave group). This will be treated in Ch. 5. Note that the
ratio of water depth to wavelength ℎ/𝐿 or alternatively 𝑘ℎ determines whether we are
dealing with deep water (or short waves) and shallow water (or long waves).
Wind-generated gravity waves (sea and swell together) are the major supplier of en-
ergy to the coastal system. In this chapter therefore due attention is given to the
generation and description of wind waves and their propagation in oceanic waters.
The transformation they undergo when entering more shallow coastal waters will be
treated in Ch. 5.
Tides
The tide is generated by the mutual gravitational attraction of the earth and the moon
and of the earth and the sun. The frequencies of the tide are governed by the well-
known movements of the earth, the moon and the sun and are mainly diurnal and
semi-diurnal (see the two spikes of ordinary tidal waves in Fig. 3.2) and not continu-
ous, in contrast to wind waves. The restoring force for long waves like tides is gravity,
although the waves are influenced by Coriolis (see Intermezzo 3.1). The tide also attrib-
utes to shaping our coastal systems. Due to the once- or twice-daily rise and fall of the
water level, the part of a coastal profile that is affected by waves changes during the

tidal cycle; at high tide the waves will attack more shoreward portions of the profile
than at low tide. In tidal basins intertidal areas between high and low water level act as
storage areas for the water brought in by the tide. The emptying and filling of the basin
can give rise to large tidal currents in the inlets and keep the inlets open. Large tidal
currents can also occur on open coasts around structures where convergence of the
tidal current can be expected. The generation and propagation of the tide in oceanic
waters will be treated in this Sects. 3.7 and 3.8. In Ch. 5 the specific processes in the
coastal zone will be treated.
Intermezzo 3.1 Coriolis acceleration
Since we are interested in water motion relative to the earth, our (numerical) mod-
els are mostly formulated in a frame of reference fixed to the earth’s surface. Gen-
erally, in such a reference frame, the 𝑧-axis is normal to the earth surface and
outward-directed: the 𝑥-axis points east and the 𝑦-axis points north (see Fig. 3.3).
Because the earth rotates around its own axis (see Sect. 3.7.2), the chosen frame of
reference also rotates (i.e. accelerates towards the centre of rotation).
ωₑ
y y
z x
z
x
Figure 3.3: As a result of the earth’s rotation any 𝑥, 𝑦, 𝑧-reference frame that is fixed to the
earth’s surface also rotates.
Newton’s equations of motion are valid in an inertial frame of reference: a refer-

ence frame that does not accelerate, e.g. a frame fixed to the distant stars. In order
to make Newton’s equations of motion valid in our rotating, non-inertial frame of
reference, we need to introduce ‘fictitious’ centrifugala and Coriolis forces. These
forces are called ‘pseudo-’ or ‘fictitious’ forces since they do not arise from any
physical interaction but from the choice of a non-inertial reference frame.
The Coriolis force is named after Gustave-Gaspard Coriolis who first described it
in the first half of the 19th century. As will be explained below, this force acts
on every moving particle in the rotating reference frame, at right angles to the
direction of motion, to the right in the Northern Hemisphere (NH) and to the left
in the Southern Hemisphere (SH). Hence, the Coriolis effect causes air and water

96 3. Ocean waves
currents in the Northern Hemisphere to deflect to the right, and in the Southern
Hemisphere to the left.
In the Northern Hemisphere the rotation of the reference frame implies an anti-
clockwise rotation of the 𝑥, 𝑦-plane, the horizontal plane tangent to the earth’s
surface (see Fig. 3.4). When observing motion in a horizontal plane that is turning
anti-clockwise, a particle travelling in a straight line (as seen from the stars) ap-
pears to be turning clockwise. To describe this deflection to the right, a Coriolis
force acting to the right must be introduced in Newton’s equations of motion.
Figure 3.4: Gnomonic map projection of latitudinal lines. In such a projection great circles
(the shortest distance between two locations on a spherical object) are displayed as straight
lines and directions are preserved. The 𝑥, 𝑦-plane turns anti-clockwise in the NH and clock-
wise in the SH.
a
The ‘fictitious’ centrifugal force (directed away from the axis of rotation) does not seem to be
invoked by any true force, since for an observer in the rotating reference frame the centripetal
acceleration of the earth (towards the rotation axis) is hidden.
In the Southern Hemisphere the 𝑥, 𝑦-plane has rotated clockwise after some time,
such that the Coriolis force acts to the left (or anti-clockwise). At the equator the
Coriolis force is zerob . This is because at the equator the earth’s surface lies in a
plane parallel to the earth’s axis of rotation. Hence, the earth’s rotation vector 𝜔𝑒
(see Fig. 3.4) has no component perpendicular to the earth surface and the angular
velocity of the horizontal plane (the rate at which the 𝑥, 𝑦 axes change direction)
is zero. Going from the equator to the poles, the rotation vector is less and less
parallel to the earth’s surface until at the poles it is entirely perpendicular to the

earth’s surface. Hence, the horizontal (i.e. parallel to the earth surface) Coriolis
forces are zero at the equator and largest at the poles.
At a latitude 𝜑, the earth-normal component of the rotation vector is 𝜔𝑒 sin 𝜑. It
represents the angular velocity of the horizontal plane (the rate at which the 𝑥, 𝑦-
coordinate axes change direction). The Coriolis force is proportional to 1) the
angular velocity of the horizontal plane 𝜔𝑒 sin 𝜑; and to 2) the velocity of the mov-
ing particle or object in the rotating reference frame. It can be shown that the
Coriolis acceleration (or Coriolis force per unit mass) to the right of the velocity
𝑉 reads:
𝑎𝑐 = 𝑓 𝑉 = 2𝜔𝑒 sin 𝜑𝑉 (3.4)
where:
𝑎𝑐 Coriolis acceleration m/s2

𝑓 Coriolis parameter 1/s
𝜔𝑒 angular velocity of the earth rad/s
𝑉 current velocity m/s
𝜑 latitude (positive in NH and negative in SH) °
The earth’s angular velocity 𝜔𝑒 = 72.9 × 10−6 rad/s is based on sidereal day, i.e.
the time needed by the earth to complete a rotation around its axis, 23 hours and
56 minutes. With the latitude and hence the Coriolis parameter positive in the
Northern Hemisphere and negative in the Southern Hemisphere, the direction of
the acceleration is to starboard in the Northern Hemisphere and to port in the
Southern Hemisphere. Its magnitude is largest at the poles where | sin 𝜑| = 1 and
zero at the equator where sin 𝜑 = 0. At mid-latitudes (say 𝜑 ≈± 45°) we have 𝑓 ≈
± 10−4 s−1 . However, in practice, in numerical models covering not too large areas,
the parameter 𝑓 can be assumed to be a constant.
Whether the Coriolis deflection is significant relative to inertia can be determined
by the Rossby number, which is given by:
𝑉
𝑅= (3.5)
|𝑓 |𝐿
with 𝐿 is the length scale of the motion. The Rossby number can be seen as the ra-
tio between the inertial forces and Coriolis forces, see also Sect. 3.8.3. For Rossby
numbers of order 1 and smaller, Coriolis deflection is important, which is for in-
stance the case for large-scale motions such as tides. To describe tidal motion, the

98 3. Ocean waves
Coriolis acceleration must be introduced in Newton’s equations of motion. How

this is done is discussed in Sect. 3.8.3.
b
Or more precisely: the horizontal components of the Coriolis force (in the 𝑥, 𝑦-plane) are zero.
Coriolis force acts in a direction perpendicular to the rotation axis of the earth, which is normal
to the earth’s surface at the equator.
Tsunamis
Also indicated in Fig. 3.2, tsunamis are a specific type of wave not caused by wind but
by large, impulsive displacement of the sea level. The disturbance of the water surface
is usually triggered by underwater earthquakes or underwater volcanic eruptions. The
surface disturbance travels away from its origin in a pattern comparable with the pat-
terns generated by the landing of a pebble in a pond. In deep water, tsunamis are not
visible because they are small in height and very long in wavelength (periods ranging
from 5 min to 60 min). They may grow to devastating proportions at the coast where
the water is shallow. This effect of tsunamis will be discussed in Ch. 5, whereas the
propagation of tsunamis through the oceans is treated in Sect. 3.5.2. Approximately
every 15 years a destructive, ocean-wide tsunami occurs.
Tsunami warning systems are in place, which recognise either seismic activity that
can potentially lead to a tsunami or unusual water level variations.
Storm surges
Storm surges are elevations of the water surface with time- and spatial scales equal
to those of the large storm fields that generate them. They are generated by the low
atmospheric pressure and high wind speeds in a storm field. The wavelength and
period are generally slightly shorter than those of tides. Storm surges can cause severe
flooding because the water will pile up against the coast when they approach the coast
(see Ch. 5). Examples are the flooding of New Orleans due to the storm surge associated
with hurricane Katrina in 2005, the regular cyclone-induced floodings of Bangladesh
and the 1953 storm surge in the Netherlands (see Sect. 1.4.6).
3.3. Measuring ocean surface elevations

Measurements of waves can be done in various ways, either at the location of the
waves itself (in situ) using for instance wave buoys or poles present in the water or
from some distance with remote sensing techniques like radar. The accuracy of most of
the instruments is ± 10 % or better. The easiest way to measure the vertical elevation
is by visual observation of individual waves from ships, a technique that has been
applied for a long time.

3.3. Measuring ocean surface elevations 99
The in-situ measurements are often aimed at obtaining the vertical elevation of the
sea surface at one location, either in deep water (buoys or poles) or in the surf zone.
Arrays of bottom-mounted pressure gauges can be placed just offshore of the surf
zone to determine the offshore wave heights and directions which act as boundary
conditions for the coastal system. Placed in deep water, it is suited to measure long
wave fluctuations like tides and tsunamis (the shorter fluctuations will not reach the
bottom, see Sect. 5.4).
Coastal tide gauges measure tide and mean sea level changes relative to points on land
rather than to an absolute reference. These coastal tide gauges are often based on floats
operating in a stilling well or on electrical measurements of water heights in a pipe.
In order to arrive at absolute mean sea level changes, the local vertical land motion
has to be measured and corrected for1 , see also Sect. 2.5.1. Differences in estimates
for historic mean sea level changes mainly result from the uncertainties related to the
local vertical motions. Therefore, although data is available for over 1750 stations
worldwide and sometimes going back as far as 1700 (for Amsterdam), there are a lot
of different interpretations of absolute mean sea level changes.
Nowadays the vertical movement of the land can be quite accurately recorded using
satellite data, Global Positioning System (GPS) receivers and altimeters based on abso-
lute gravity measurements.
The wave information from buoys, poles, radar or other instruments is important to
coastal engineers to gain an understanding of the longer-term wave climate (years)
or the short-term (storms and individual waves) characteristics of the waves in a cer-
tain region. As an example of information available for a certain engineering project,
Fig. 3.5 shows the permanent (viz. 85 % of the time) wave-measuring locations in the
Dutch North Sea that have provided directional wave data for over three decades.
In order to obtain wave information closer to the coast or at specific location, wave
models can be used which use the measured wave information at the buoy location as
boundary conditions. The buoys are located in relatively deep water, so that the waves
are not influenced by the (changing) bed. The elevation data – sampled a few times
per second – is translated to a radio signal and then transmitted to a post-processing
centre where data analysis and derivation of wave parameters takes place.
For the Dutch coast usually a combination of sea and swell is present in the record.
The swell can be recognised as the longer modulation in the record which becomes
clear when the two signals are separated in the shorter and irregular sea and longer
swell. A typical deep-water wave record from such a platform is given in Fig. 3.6. For
most of the Dutch measuring locations, this type of elevation data is available since
1
The local vertical land motions due to post-glacial rebound, tectonic effects, ground water or oil ex-
traction can be of the same order as the average global or absolute rise, which over the last century is
estimated at between 1 mm to 2 mm per year (IPCC, 2001, 2007)

100 3. Ocean waves
3 2
1. Schiermonnikoog Noord
2. Eierlandse Gat
3. K13A platform
4. IJmuiden Munitiestortplaats
5. Euro platform
Figure 3.5: The five buoy locations in the Dutch North Sea area permanently providing directional
wave data since at least 1989, two close to offshore platforms and the remaining ones at varying
distances from the coast or outer delta. For more information see https://www.rijkswaterstaat.
nl/water/waterdata-en-waterberichtgeving/waterdata & https://waterinfo.rws.nl, in Dutch.
1979. It can be used to arrive at short-term statistics and long-term wave climates. In
the following section the techniques to do that will be explained.
3.4. Short-term wave statistics

3.4.1. Description of wave characteristics
Figure 3.6 already showed that real ocean wind waves have an irregular character
(not-periodic, not repeating itself in time and space) as opposed to the single sinus-
oidal signal in Fig. 3.1. The waves are therefore called irregular or random waves. In
spite of the seemingly unpredictable (random) way in which the signal fluctuates, if
we describe the short-term variations in a statistical way by taking average paramet-
ers, it appears that the statistics can be considered constant in time (stationary). In
order for the averages to be representative of the sea state, the record should be short
enough to be statistically stationary (not changing in time). On the other hand the
record should be long enough to get reliable averages. At sea 15 to 30 minutes is used,
most commonly 20 minutes. On longer timescales the short-term mean values are vari-
able due to variations in mean wind velocity, tidal elevation or tidal currents which
change the wave characteristics. Thus wind waves are a random stationary process
for timescales up to half an hour. In practice one recording of for instance 20 minutes
is done every three hours. This record is thought to be representative for the entire

3.4. Short-term wave statistics 101
elevation [m]
1
0 sea + swell
-1
-2
-3
0 60 120 180
t [s]
3
2
elevation [m]
0 sea
-1
-2
-3
0 60 120 180
t [s]
3
2
elevation [m]
0 swell
-1
-2
-3
0 60 120 180
t [s]
Figure 3.6: Example of a combined swell and sea time series (upper panel), separated in a sea
component (middle panel) and a swell component (lower panel). The time series are generated
from the spectra as shown in Fig. 3.7.
time period of three hours. The duration of a storm is generally 6 to 8 hours during
which the conditions (mean wind speed) are more or less constant.
There are basically two ways to characterise a wave record in terms of its short-term
statistics:
1. Based on direct analysis of the time series and regarding it as a sequence of
individual waves, each with their own wave height and wave period (wave-by-
wave analysis);
2. Through a spectral analysis using the fact that the surface can be seen as a sum-
mation of an infinite number of sine waves with different heights, periods and
directions.

102 3. Ocean waves
The short-term statistical analysis based on the time series is treated in Sect. 3.4.2.
Spectral analysis is dealt with in Sect. 3.4.3. It appears that the short-term distribution
of wave heights can be described by a Rayleigh distribution, as long as we are dealing
with not too steep waves (wave slope 𝑎𝑘 small) in deep water. In these conditions the
parameters as determined by the wave-by-wave analysis and by the spectral analysis
can be related to each other by constant ratios. This is the topic of Sect. 3.4.4.
3.4.2. Analysis of the time series

The mean and the standard deviation are important statistical properties that can be
derived from an arbitrary time series. For a stationary surface elevation signal the
mean should be constant (zero for purely oscillatory signal), whereas the standard
deviation 𝜎 is a vertical measure that can be related to the wave height, as we will see
later. The variance of the surface elevation 𝜎 2 is an important quantity in the statistical
description of waves, since it is related to the mean wave energy per unit area 𝐸 which
is the sum of the potential and kinetic energy (𝐸 = 𝐸𝑝 + 𝐸𝑘 ). The potential energy is
𝐸𝑝 = 12 𝜌𝑔𝜎 2 . In the linear approximation (valid for small-amplitude waves compared
to the wavelength and water depth), the potential and kinetic energy are the same (or
𝐸𝑝 = 𝐸𝑘 = 12 𝐸) and hence:
𝐸 = 𝜌𝑔𝜎 2 (3.6)
The difference between variance and energy therefore is just a factor 𝜌𝑔.
Alternatively, when the short-term time record (order 20 minutes) is considered as a
series of individual waves with their own wave height and period, average parameters
can be taken of the series of wave heights and periods in order to characterise the
record. Remember that useful averages require a stationary record.
Before starting the wave-by-wave analysis, the mean water level should be subtracted
from the record. We then have a purely oscillatory surface elevation signal about the
mean. In the case of for instance a tidal variation it is possible that the record is not
entirely stationary but that there is a slight variation in the mean water level. For a
short record length, this trend will be approximately linear and can be determined by
regression analysis and then subtracted from the signal so that the signal is stationary
again.
When the mean or trend is removed from the signal, individual waves can be defined.
The wavelength and wave period are the distance and period respectively between two
subsequent downward or upward zero-crossings. The wave height of each individual
wave is the difference in elevation between the crest and the trough. An advantage
of using downward crossings is that it relates more directly to visual observations of
wave heights. The reason is that observers tend to define a wave as starting with the

trough, such that the wave height is taken to be the height of the crest relative to the
preceding trough.
Various average parameters can now be derived of which the most obvious probably
is the mean wave height. Nevertheless, the mean wave height is not used that often.
Of more practical use is the significant wave height 𝐻𝑠 or 𝐻1/3 . The significant wave
height is defined as the average height of the highest one third of the waves:
𝑁 /3
1
𝐻1/3 = ∑𝐻 (3.7)
𝑁 /3 𝑗=1 𝑗
where 𝐻𝑗 is the 𝑗-th wave (with 𝑗 = 1 the largest wave, 𝑗 = 2 the second largest etc.)
and 𝑁 is the total number of waves.
It is called significant wave height because it approximately corresponds to visual es-
timates of experienced observers at sea of a representative wave height. Apparently
observers tend to bias their estimates to the higher waves in the record. Its corres-
pondence with visual estimates and therefore easy quantification from ships and large
databases makes it a useful measure for coastal engineers.
Another often used parameter is 𝐻rms , the root-mean-square wave height, which is
obtained by taking the square root of the mean of the wave heights squared:
𝑁
1
𝐻rms = ∑ 𝐻𝑖2 (3.8)
𝑁 𝑖=1
√
It can be seen as a wave energy measure, since the wave energy is related to the wave
height squared (see also Sect. 3.4.4).
Other measures like 𝐻1/10 and 𝐻1/100 are also used and they are defined analogous to
𝐻1/3 as the average of the highest 1/10 and 1/100 of the waves respectively.
The mean of all wave periods is called the mean wave period or zero-crossing wave
period:
𝑁
1
𝑇0 = ∑ 𝑇𝑖 (3.9)
𝑁 𝑖=1
Similar to 𝐻1/3 the significant wave period is defined as the average wave period of
the highest one-third of the waves. The significant wave period is not correlated to
visual estimates and therefore has less physical meaning:
𝑁 /3
1
𝑇1/3 = ∑𝑇 (3.10)
𝑁 /3 𝑗=1 𝑗

104 3. Ocean waves
What values would typically be found at for instance the North Sea? For sea conditions
significant wave heights range from order 1 m during quiet periods (with mean wave
periods of order 5 s) to 10 m and more during storms (with mean periods of order 10 s).
North Sea swell has wave heights of 0.5 m to 1 m and wave periods of around 10 s.
3.4.3. Spectral analysis

An alternative way of arriving at a statistical representation of the sea state uses the
fact that the surface elevation at one location can be unravelled into various sine waves
with different frequencies of which the amplitudes and phases can be determined by
so-called Fourier analysis. Jean-Baptiste Fourier demonstrated that any signal can be
described by a sum of harmonic components, a so-called Fourier series. Under the
assumption of a stationary record, these sine waves have a constant amplitude and
phase per component in time. For a time record with finite duration, the Fourier series
can be written in terms of sine (or cosine) functions that fit an integer number of times
in the record duration 𝑇𝑟 .
The oscillatory surface elevation can be written as a Fourier series as follows:
𝑁
𝜂 = ∑ 𝑎𝑛 cos (2𝜋𝑓𝑛 𝑡 + 𝛼𝑛 ) (3.11)
𝑛=1
where
𝑛
𝑓𝑛 = for 𝑛 = 1, 2, …
𝑇𝑟
Although in nature the frequencies will be continuous, the frequencies in Eq. 3.11 are
discrete by necessity, because in practice the length of the wave series is restricted
to for instance 20 min (Δ𝑓𝑛 = 1/𝑇𝑟 ). The record length thus determines the smallest
frequency (the longest wave) that can be determined from the record: 𝑓min = 1/𝑇𝑟 .
Moreover, the time series is not continuous since the water level measurements are
performed with a certain sampling interval. The sampling interval determines the
1
highest frequency that can be determined from the record: 𝑓max = 2Δ𝑡 .
From the amplitudes of the various components the spectrum of wave energy over the
range of wave periods or frequencies can be calculated. With a bit of trigonometry it
can be found that for one harmonic component the variance is equal to 21 𝑎𝑛2 . Then for
a sum of harmonic components the corresponding variance is given by:
𝑁
1
∑ 𝑎𝑛2 (3.12)
𝑛=1 2

The variance density spectrum gives the variance density per unit frequency interval
for each frequency and is constant for Δ𝑓 → 0:
1/2𝑎𝑛2
lim = 𝐸 (𝑓𝑛 ) (3.13)
Δ𝑓 →0 Δ𝑓
By taking the integral of the spectrum the total variance is recovered again:
∞
2 2
∫ 𝐸(𝑓 ) d𝑓 = 𝜂 = 𝜎 (3.14)
0
Two conclusions can be drawn. First, in the spectrum the variance density is the con-
tribution of one component to the total variance. Second, the standard deviation 𝜎 of
the surface elevation signal can be estimated from the area under the spectrum. Note
further that from the variance density spectrum the energy density spectrum is readily
obtained, since variance and energy are coupled through Eq. 3.6.
The so-computed spectrum describes the time series under consideration but is only an
estimation of the spectrum representing the random process, since a next realisation
under the same condition will give a slightly different surface elevation. To get a
better estimate, averaging has to take place of spectra, either based on subdivisions of
the time series or over frequency bins.
In Fig. 3.7 energy spectra are shown with the corresponding time series. In the middle
and lower panels all energy is concentrated around the mid-frequencies. The middle
panel shows a spectrum for sea only. The narrower the spectrum, the more regular
the waves are. For larger, longer waves the spectrum will be shifted towards the lower
frequencies and contain more energy. For smaller, shorter waves the spectrum will be
shifted towards the higher frequencies and be lower. The lower panel shows a spec-
trum for swell only; it is narrower than the spectrum for sea, contains less energy and
the energy is concentrated around lower frequencies. Sometimes one can distinguish
between two adjacent or often partly overlapping parts of the spectrum. This means
that two distinct wave fields are present: swell and sea (upper panel). If the mean
frequencies of the two wave fields are close, then there will be so much overlap that
the spectrum is broad, but otherwise looks like a spectrum with only one wave field.
What about the phases of the different components? The distribution of the phases
over the frequencies is called a phase spectrum. Often the phase spectrum is not shown
since in not too steep waves and in deep water the phases seem to be independent of
each other and uniformly distributed between −𝜋 and 𝜋. This means that the different
components are not related through their phases and can be seen as individual waves
moving independently through the signal, as if they were alone. This is the case for lin-
ear small-amplitude waves. So then only the amplitude or variance/energy spectrum
remains to characterise the wave record.

106 3. Ocean waves
1.2 3 3
sea + swell
1 2 2
elevation [m]
0.8 1 1
E [m 2 /Hz]
0.6 0 0
0.4 -1 -1
0.2 -2 -2
0 -3 -3
0 0.1 0.2 0.3 0.4 0.5 0 200 400 600 0 60 120 180
f [Hz] t [s] t [s]
1.2 3 3
sea
1 2 2
elevation [m]
0.8 1 1
E [m 2 /Hz]
0.6 0 0
0.4 -1 -1
0.2 -2 -2
0 -3 -3
0 0.1 0.2 0.3 0.4 0.5 0 200 400 600 0 60 120 180
f [Hz] t [s] t [s]
1.2 3 3
swell
1 2 2
elevation [m]
0.8 1 1
E [m2 /Hz]
0.6 0 0
0.4 -1 -1
0.2 -2 -2
0 -3 -3
0 0.1 0.2 0.3 0.4 0.5 0 200 400 600 0 60 120 180
f [Hz] t [s] t [s]
Figure 3.7: Surface elevation time series and corresponding spectra for sea and swell combined
(upper panel), sea (middle panel) and swell (lower panel). To generate the sea component, a so-
called JONSWAP spectrum (a typical spectrum for sea, see Sect. 3.5.1) is used with 𝐻𝑆 = 1.8 m,
𝑇𝑝 = 5.5 s and a peak enhancement factor 𝛾 = 1 (see eq. (6.3.15) in Holthuijsen, 2007). For the
swell component, 𝐻𝑆 = 0.5 m, 𝑇𝑝 = 12 s and 𝛾 = 5 are used. For both sea and swell, the peak-width
parameter is 𝜎 = 0.07 for 𝑓 ≤ 𝑓𝑝 and 𝜎 = 0.09 for 𝑓 > 𝑓𝑝 . The sea and swell spectra are combined
in a bi-modal spectrum of swell and sea. The time series are generated from the spectra at a
sampling frequency of 25 Hz assuming random phases. Note that the timeseries in the right
panels correspond to Fig. 3.6.
Which parameters can be derived from the variance spectrum? First, the spectrum re-
veals the dominant frequencies in the wave record; most energy occurs at the spectral
peak and the corresponding wave period is called the peak spectral period 𝑇𝑝 . Other
typical average parameters can be expressed in terms of spectral moments:
∞
𝑚𝑛 = ∫ 𝑓 𝑛 𝐸(𝑓 ) d𝑓 for 𝑛 = … , −3, −2, −1, 0, 1, 2, 3, … (3.15)
0

𝑚0 is the area under the spectrum. Since 𝑚0 is the total variance integrated over all
frequencies, the standard deviation is given by 𝜎 = √𝑚0 (see Eqs. 3.13 and 3.14). In
Sect. 3.4.4 we will see how the zero-th moment 𝑚0 and the second moment can be used
to determine the zero-crossing period from the spectrum.
3.4.4. Short-term wave height distribution

In the previous section we have seen that in deep water and as long as the waves are
not too steep, the surface elevation can be considered as the sum of a large number of
components with random phases. In that case observations and theoretical consider-
ations have shown that the wave heights can be described by a Rayleigh distribution.
We will see below that by using the Rayleigh distribution, the parameters as determ-
ined by the wave-by-wave analysis and by the spectral analysis can be related to each
other.
The short-term distribution of wave heights known as the Rayleigh distribution is:
2
𝐻 −𝐻
𝑝(𝐻 ) = 2 𝑒 8𝜎 2 (3.16)
4𝜎
Although theoretically only valid for a narrow spectrum, observations have shown
that also for broader spectra the wave heights more or less obey a Rayleigh distribution.
In practice wave period variability is often ignored.
From the probability density function (Eq. 3.16) the probability can be derived that an
individual wave height 𝐻 ′ exceeds a specified wave height 𝐻 :
𝐻 2
−𝐻2
𝑃 (𝐻 ′ > 𝐻 ) = 1 − 𝑃 (𝐻 ′ < 𝐻 ) = 1 − ∫ 𝑝(𝐻 ) d𝐻 = 𝑒 8𝜎 (3.17)
0
where the only parameter 𝜎 is the standard deviation and thus a wave height measure.
As we have seen it can be estimated from either the time series or, since 𝜎 = √𝑚0 , from
the spectrum.
The wave height with a probability of exceedance 𝑃 follows from Eq. 3.17:
𝐻𝑃 = 2𝜎 √2 ln(1/𝑃) (3.18)
For instance, the height 𝐻2 % that is exceeded by 2 % of the waves is equal to 2𝜎√2 ln 50 =
5.59𝜎 and 𝐻1 % = 6.07𝜎 .
In Sect. 3.4.3 𝐻1/3 was calculated from the time series. We can also use the Rayleigh
distribution to calculate 𝐻1/3 . This can be shown to yield:
𝐻1/3 = 4𝜎 (3.19)

108 3. Ocean waves
The probability of exceedance of wave heights according to the Rayleigh distribution

can now also be expressed in terms of significant wave height:
2
−2( 𝐻𝐻 )
𝑃 (𝐻 ′ > 𝐻 ) = 𝑒 𝑠 (3.20)
which is often graphically represented as in Fig. 3.8.
100
0.5
exceedance probability
0.2
0.135
10-1
10-2
10-3
10-4
10-5
0 0.5 1 1.5 2 2.5
H/Hs
Figure 3.8: The exceedance probability 𝑃 (𝐻 ′ > 𝐻 ) (Eq. 3.20) as a function of 𝐻 /𝐻𝑠 shows as a
straight line when plotted using a vertical scale with a spacing proportional to √− ln 𝑃.
The probability that an individual wave height is larger than 𝐻𝑠 is 13.5 % (𝑒 −2 ), which
can be read from the graph or determined from Eq. 3.20. It also follows that 1 out of 100
waves will be higher than about 1.5 times 𝐻𝑠 . The strength of the storm considered is
apparently determined by just one value: 𝐻𝑠 . A stronger storm would lead to a steeper
distribution curve, which is again defined by a specific value of the significant wave
height.
When irregular waves enter shallow water, the highest waves will start breaking due
to the limited depth. This means that the Rayleigh distribution is no longer applicable
for 𝐻𝑠 > 0.3ℎ or so. Then, for the tail of the distribution, a Weibull distribution agrees
better.
Based on Eq. 3.19 we define 𝐻𝑚0 = 4√𝑚0 as an alternative formula to calculate the
significant wave height and use the subscript 𝑚0 to indicate that the wave height is
computed from the zeroth moment of the wave spectrum. By the way, from observa-
tions at sea it appears that a correction to this theoretical value gives 𝐻1/3 ≈ 3.8√𝑚0 .
Summarizing, one can state that the short-term distribution of wave heights, i.e. the
wave heights in a stationary sea state, exhibits some very typical relations:
The maximum individual wave height in a wave record depends on the length of the
record. Suppose we have a storm record with a duration of six hours. Suppose the

Table 3.1: Typical wave heights (non-breaking waves)
Description Notation 𝐻 / √ 𝑚0 𝐻 /𝐻𝑠

RMS height 𝐻rms 2√2 0.707
Mean height 𝐻 √2𝜋 0.63
Significant height 𝐻𝑠 = 𝐻1/3 4.004 1
Average of 1/10 highest waves 𝐻1/10 5.09 1.27
Average of 1/100 highest waves 𝐻1/100 6.67 1.67
mean wave period is 𝑇 = 10 s, so on average we have 6 × 60 × 60/10 ≈ 2000 waves

in the record. The maximum individual wave height can now be estimated by setting
𝑃 = 1/2000 in Eq. 3.18. We then find that in a storm the maximum individual wave
height 𝐻max ≈ 2𝐻𝑠 . This is handy method to quickly estimate 𝐻𝑠 from a wave record
with a certain number of waves.
For a regular wave, the energy content 𝐸 = 1/2𝜌𝑔𝑎2 = 1/8𝜌𝑔𝐻 2 . Using Table 3.1 we
2 . Apparently, the root-mean-
have for an irregular wave field 𝐸 = 𝜌𝑔𝑚0 = 1/8𝜌𝑔𝐻rms
square wave height 𝐻rms is the wave height representing the total energy content.
In a similar way, wave periods can be determined from the spectral moments. For
instance the zero-crossing period:
𝑚0
𝑇2 = (3.21)
√ 𝑚2
Theoretically this is equal to the zero-crossing period 𝑇0 as determined from the time
series. For narrow spectra such as for swell, 𝑇2 and 𝑇1/3 are approximately equal to the
spectral peak period 𝑇𝑝 . For a broader spectrum with a high frequency tail (a typical
sea spectrum) 𝑇1/3 is approximately equal to 0.9–0.95 𝑇𝑝 and 𝑇2 is roughly equal to 0.7
𝑇𝑝 . However, the value of 𝑇2 should be considered with care; because of the sensitivity
of the higher moments for higher frequencies, 𝑇2 will be sensitive to details in the
measurements and data processing.
The wave climate for the North Sea is not very extreme. The significant wave height
now and then reaches values of 8 metres in the northern part and 6 metres in the
southern part (off the Dutch coast). The mean wave period under those conditions is
around 10 seconds. The highest wave in the north is then approximately 15 metres
high with a period of 15 to 20 seconds. Averaged over the year 𝐻𝑠 = 1 m to 1.5 m and
𝑇 = 4 to 5 seconds.
So far we have not taken the directionality of waves into account. In reality, however,
the different harmonic components have different wave directions 𝜃. That is why you
will sometimes also find information on a mean wave direction 𝜃𝑚 and the amount of
spread around the mean.

110 3. Ocean waves
3.5. Wind wave generation and dispersion

3.5.1. Locally-generated sea
Waves are generated by local wind fields. At the area of wave generation, these waves
are relatively steep and short-crested. The latter means that there are no distinct
wave fronts because the waves are irregular and directional. An example is given
in Fig. 3.9. In this photo white-capping can also be seen, which is steepness-induced
wave-breaking when the wave height becomes too large compared to the wavelength
(𝐻 /𝐿 > 0.14).
The wave characteristics (height, period, propagation direction) and their duration
depend on the characteristics of the wind field (speed, duration and direction), the
fetch and the local water depth. Fetch is the maximum length of open water over which
the wind blows, which is determined by meteorological and geographical conditions.
Generally, the higher the wind speed and duration, the larger the wave height and
period. But only for ideal cases, wave heights can be estimated on the basis of wind
velocity, duration and fetch.
Although wind conditions cannot be predicted accurately long in advance, wind con-
ditions can be described statistically. The wind climate consists of both velocity data
and directional data. Velocities can be expressed by wind speed (when measured) or as
a certain number on the Beaufort scale (when visually observed, see Intermezzo 3.2).
These data can be found in meteorological yearbooks and in various atlases, for in-
stance specific hydrographical atlases that contain data collected at sea.
Figure 3.9: Wind-generated gravity waves at sea, probably Beaufort 5. Photo from Rijkswater-
staat (‘Credits’ on page 575).

3.5. Wind wave generation and dispersion 111
Parameterised wave spectra have been formulated which relate the wind field to spec-
tral density and can therefore be used to hindcast (estimate past events) wave para-
meters from known wind fields. An example is the Joint North Sea Wave Observa-
tion Project (JONSWAP) spectrum which is characteristic for (developing) wind sea
in oceanic waters. A typical JONSWAP spectrum was shown in the middle panel of
Fig. 3.7. It can be determined from the wind speed and the fetch. For a fully developed
sea (not limited by either fetch or duration, which in reality will hardly ever happen)
the so-called Pierson-Moskowitz spectrum is valid. This is a broader spectrum and
only dependent on the wind speed.
Intermezzo 3.2 Beaufort scale
The Beaufort wind speed scale (ranging from 0 to 12) relates wind speed to the
local sea state using descriptors such as wave height, wavelength, white capping,
amount of foam and spray. Beaufort, a British naval officer, introduced the Beaufort
wind scale in 1805. For tactical reasons the scale was intended to exchange ob-
jective information between sailing vessels of the British Navy. The lower scales
(2 to 4) refer to sailing speeds of the common naval vessel of that time (man-of-
war) under full sail. The intermediate scales (5 to 9) refer to conditions that re-
quired reefing of sail. The higher scales (10 to 12) deal with survival of ship and
crew. The Beaufort scale is summarised in Table 3.2 in the form that is used at
present. Bold-printed expressions refer to the official terms of the World Meteor-
ological Organisation (WMO). Pictures of typical sea states at various Beaufort
wind speeds are available to assist observers on board of sea-going vessels (see
also http://en.wikipedia.org/wiki/Beaufort_scale).
3.5.2. Wave dispersion

In principle the wave motion can be described by the continuity equation and the
Navier-Stokes equations of motion. Difficulties arise however when attempts are un-
dertaken to solve these equations. One of the complications is that the surface bound-
ary condition is the surface elevation that we try to solve. If we linearise this surface
boundary condition and assume a horizontal bottom, a simple solution to the equa-
tions is the single Fourier component that we described in Sect. 3.2. We then get the
Airy wave theory (App. A). Neglecting non-linearities gives a good approximation for
not too steep waves in deep water (𝑎𝑘 << 1 for 𝑘ℎ is large) or small-amplitude waves
in shallow water (𝑎 << ℎ for 𝑘ℎ is small). According to Airy wave theory, for a linear
sine wave the relation between frequency 𝜔 and wavenumber 𝑘 is given by:
𝜔 = √𝑔𝑘 tanh 𝑘ℎ (3.22)

112
Table 3.2: Beaufort scale
Beaufort Wind speed Phenomena observed on land State of the sea surface In Dutch as Wave
No. used by KNMI height
m/s kn m
0 0 – 0.2 0–1 Calm: Still; smoke will rise vertically Sea like mirror. Windstil 0
1 0.3 – 1.5 1–3 Light Air: Rising smoke drifts; weather vane is inactive Ripples with appearance of scales; no foam crests. Zwakke wind 0.1 – 0.2
2 1.6 – 3.3 3 – 6.5 Light Breeze: Leaves rustle, people feel wind on skin; Small wavelets; crests of glassy appearance, not breaking. Zwakke wind 0.3 – 0.5
weather vane is inactive.
3 3.4 – 5.4 6.5 – 11 Gentle Breeze: Leaves and twigs move around. Light-weight Large wavelets; crests begin to break; scattered whitecaps. Zwak tot 0.6 – 1.0
flags extend. matige wind
4 5.5 – 7.9 11 – 16 Moderate Breeze: Moves thin branches, raises dust and Small waves, becoming longer; numerous whitecaps. Matige wind 1.5
paper.
5 8 – 10.7 16 – 21 Fresh Breeze: Small trees begin to sway. Moderate waves, taking longer form; many whitecaps; some Vrij krachtige 2
spray. wind
6 10.8 – 13.8 21 – 28 Strong Breeze: Large tree branches move, open wires (such Larger waves forming; whitecaps everywhere; more spray. Krachtige 3.5
as telegraph wires) begin to whistle, umbrellas are difficult to wind
keep under control.
7 13.9 – 17.1 28 – 34 Near Gale: Large trees begin to sway, noticeably difficult to Sea heaps up; white foam from breaking waves begins to be Harde wind 5
walk. blown in streaks.
8 17.2 – 20.7 34 – 42 Gale: Twigs and small branches are broken from trees, Moderately high waves of greater length; edges of crests begin Stormachtig 7.5
walking into the wind is very difficult. to break into spindrift; foam is blown in well-marked streaks.
9 20.8 – 24.4 42 – 49 Strong Gale: Slight damage occurs to buildings, shingles are High waves; sea begins to roll; dense streaks of foam; spray Storm 9.5
blown off of roofs. may reduce visibility.
10 24.5 – 28.4 49 – 57 Storm: Large trees are uprooted, building damage is Very high waves with overhanging crest; sea takes white Zware storm 12
considerable. appearance as foam is blown in very dense streaks; rolling is
heavy and visibility is reduced.
11 28.5 – 32.6 57 – 65 Violent Storm: Extensive widespread damage. These Exceptionally high waves; sea covered with white foam Zeer zware 15
typically occur at sea, rarely inland. patches; visibility still more reduced. storm
12 >32.7 >65 Hurricane: Extreme destruction. Air filled with foam; sea completely white with driving spray; Orkaan >15
visibility greatly reduced.

3. Ocean waves
and is called dispersion relation. It is a function of the local water depth and the
restoring force 𝑔. The phase velocity 𝑐 = 𝜔/𝑘 is then given by:
𝑔𝑇
𝑐= tanh 𝑘ℎ = 𝑐0 tanh 𝑘ℎ (3.23)
2𝜋
The phase velocity (speed) is the rate at which any phase of the wave (for instance
the wave crest) propagates in space. It is also called propagation velocity (or speed) or
wave celerity. The nature of the hyperbolic tangent is shown in Fig. 3.10.
tanh(kh)
0.8
kh
tanh(kh)
0.6
0.4
≈0.31
0.2
shallow intermediate deep

0
0 1 2 3 4 5 kh
0 0.05 0.16 0.32 0.48 0.64 0.80 h/L
0.5
Figure 3.10: Nature of the hyperbolic tangent.
Given the fact that tanh 𝑘ℎ equals 1 for 𝑘ℎ ≫ 1, 𝑐0 represents the deep-water phase
velocity 𝑐0 ≈ 1.56𝑇 as function of 𝑇 . The deep-water (or short-wave) approximation
can be used without too many errors (order of 1 %) for 𝑘ℎ > 𝜋 or ℎ/𝐿 > 0.5. Wind
waves in oceanic waters can be considered short waves, such that their phase velocity
is linearly dependent on the wave period. Expressed in terms of the wavenumber in
deep water the phase velocity 𝑐0 = √𝑔𝐿0 /2𝜋 and is therefore proportional to the square
root of the deep-water wavelength. Hence, longer waves propagate faster than shorter
waves. Independent harmonic components of a wind wave field can be expected to
travel at different speeds. The separation of the different harmonic components due
to their different propagation speeds is called frequency dispersion. Oceanic wind
waves are highly dispersive.
Since tanh 𝑘ℎ equals 𝑘ℎ for 𝑘ℎ → 0, the dispersion relation reduces to 𝑐 = √𝑔ℎ for
𝑘ℎ → 0. Hence, if the wave is long enough (𝑘ℎ < 0.31 or ℎ/𝐿 < 1/20), the wave
celerity is only dependent on the local water depth. The shallower the water depth,
the smaller the propagation speed. Since the wave celerity is independent of the wave
period, the wave is called non-dispersive. This is the case for the tide and generally

114 3. Ocean waves
for tsunami waves as well. Wavelengths of tsunamis are easily 100 km or more2 which
is more than 20 times larger than the average water depth in the deep ocean, which
is about 4000 m deep. A tsunami then propagates at a velocity of 𝑐 = √9.81 × 4000 =
200 m/s or 700 km/h. As long as they do not dissipate their energy against a shore,
tsunamis can travel at these high speeds for a long period of time and lose very little
energy in the process.
If waves meet a current (in or against the wave propagation direction), the wavelength,
propagation velocity and wave height will be affected. The propagation velocity and
wavelength relative to a fixed reference frame will increase in the case of a current in
the propagation direction. The wave height will decrease. In the case of an opposing
current, it is the other way around.
3.5.3. Wave groups

Another interesting phenomenon in deep-water waves is that the wave crests move
faster than the wave energy. Let us first consider a series of regular waves moving
away from a disturbance into otherwise still water, such as a group of waves generated
by a wave board in a wave flume and now moving away from the wave board. It can
be seen that the wave front and the group of waves as a whole moves at a certain
speed. This speed is called the group speed and is the speed with which the energy of
the waves moves. The phase velocity, however, was defined as the velocity of a wave
crest. When following a wave crest, we can see how it originates at the rear of the
group, then moves forward through the group and finally disappears at the leading
edge of the group. The phase velocity of the wave crest is therefore larger than the
velocity of the wave front or the group (except in shallow water, where the group
velocity is equal to the phase velocity).
Now look at irregular waves. They appear more or less in groups, so that we now have
a series of groups as opposed to the one group as described above. The groupiness
is caused by interference between waves of different wavelength. To demonstrate
this, consider the limiting case of just two slightly different frequencies travelling in
the same direction. We have a water depth of 20 m and two regular waves with an
amplitude of 1.5 m each. The wave periods are 6.2 s and 7.0 s respectively. The two
wave trains will interfere with each other as shown in Fig. 3.11. The length and period
2
In Sect. 3.2, tsunami periods were said to range from 5 min to 60 min. As an exercise: use Table A.3 to
compute the corresponding wavelengths at a depth of 4000 m. Also compute the wave period range
for which this depth is classified as intermediate water.

of the groups can be computed from the differences in wavenumbers and frequencies
respectively:
2𝜋
𝑘group = Δ𝑘 = 𝑘2 − 𝑘1 → 𝐿group = (3.24a)
Δ𝑘
2𝜋
𝜔group = Δ𝜔 = 𝜔2 − 𝜔1 → 𝑇group = (3.24b)
Δ𝜔
2
elevation [m]
-1
-2
-3 A
0 0.5 1 1.5 2 0 0.5 1 1.5 2
x/L group t/T group
Figure 3.11: Wave grouping of two monochromatic (single-period) waves with a slightly different
period. Either regular wave has an amplitude of 1.5 m, whereas the wave periods are 6.2 s and
7.0 s. The water depth is 20 m. The left panel shows the grouping in space at 𝑡 = 0. The right
panel shows the grouping in time at location A. As an exercise: compute the ratio between the
number of waves in the group in time and the number of waves in the group in space and verify
this ratio from the figure.
The broader the spectrum, the more irregular the waves are and the less clearly the
groups can be distinguished. Because of the narrow spectrum, grouping is prominent
in swell (see Sect. 3.5.4) and more pronounced if the swell is dispersed more. As in
the case of disturbance, the energy of waves is carried in the group and not in each
individual wave.
The ratio between the group velocity and the higher phase velocity is called 𝑛 and
reads according to Airy wave theory:
𝑐𝑔 2𝑘ℎ
𝑛= = 0.5 (1 + ) (3.25)
𝑐 sinh 2𝑘ℎ
Figure 5.3 shows the dependency of 𝑛 on the water depth and (deep-water) wavelength.
For short waves in deep water 𝑛 reduces to 0.5. The group velocity therefore is half
of the propagation velocity of an individual wave crest. When calculating how long
it will take for a group of swell waves to cover a certain distance, the group velocity
needs to be taken into account. For long waves (or shallow water) 𝑛 equals 1 such
that the phase velocity is equal to the group velocity. The frequency dependency of
the group velocity is comparable to that of the phase velocity. Hence, the fronts of
longer-period waves travel faster.

116 3. Ocean waves
In Sect. 3.4.4 we have already seen that the energy per square metre of water surface
is 𝐸 = 81 𝜌𝑔𝐻rms
2 . This energy is propagated at the wave group speed 𝑐 , thus causing
𝑔
an energy flux 𝑈 , with
𝑈 = 𝐸𝑐𝑔 = 𝐸𝑛𝑐 (3.26)
Based on energy conservation, offshore wave conditions can be translated to the nearshore.
The energy balance therefore is an important concept in coastal engineering. It is dis-
cussed in detail in Ch. 5.
3.5.4. Sea versus swell waves

From Sect. 3.4 we know that an irregular wave train can be seen as a sum of sine waves
with various periods. For not too long periods of time (maybe an hour and distances
of tens of kilometres) and for small amplitudes, these sine waves have constant amp-
litudes and random phases. They travel in many different directions, all at their own
velocity, given by the so-called dispersion relation according to Airy or linear wave
theory.
Wave fields disperse (spread out) since the different harmonic components travel at
different speeds that depend on their frequency. In Sect. 3.5.2 we referred to this phe-
nomenon as frequency dispersion. From the dispersion relation it becomes clear that
longer waves travel faster than shorter waves. Also, we have seen that the group
velocity (the velocity of the front) is larger for longer-period waves.
At some distance from the storm centre, one would therefore first experience a long,
fast-travelling swell and later an increasingly shorter wave period. At long distances
from the storm centre the shorter waves are filtered out since dissipation processes
(due to currents and white-capping) more strongly affect the shorter waves3 . As a
result only a long, and fairly regular (as the various components travel at different
speeds) swell remains. Besides, the swell is uni-directional or long-crested, because
only waves travelling in a particular direction end up at a certain location away from
the storm centre. The spreading due to different directions of propagation is called
direction dispersion. Due to frequency and direction dispersion, the spectrum of swell
is narrow in frequency and direction respectively. As a result of spreading (and energy
dissipation) swell is relatively low. Figure 3.12 shows swell waves arriving at the coast
of Angola that have been generated in two different storms.
The characteristics of swell waves at a particular (coastal) location are determined by
the characteristics of the storm and the distance to the storm. Swell can travel the
oceans for thousands of kilometres. A 10 s swell wave travels at speed 𝑐0 ≈ 1.56𝑇 =
3
In addition to dissipation and dispersion, non-linear wave transfer plays a role; energy is moved from
the centre of the spectrum to both the higher and lower frequencies. The higher frequencies get sub-
sequently dissipated, whereas the lower frequencies gain energy.

Figure 3.12: Swell waves arriving at the Atlantic coast of Angola from two distinct directions
(southwest and northwest). This swell is generated by trade winds. Courtesy Stefanie Ross
1.56 × 10 = 15.6 m/s = 56 km/h. The group velocity will be about half that (in deep
water). In Fig. 3.13 a graph is shown with wave height and period as a function of
propagation distance from the storm centre.
In a first-order approximation,no mass transport is associated with short-wave propaga-
tion, such that the path of swell through the oceans is unaffected by the Coriolis effect.
Swell therefore travels the globe along great circles, the shortest distance between two
locations on a spherical object.
Some coasts around the world – for instance the west, east and south coasts of Aus-
tralia – mainly experience swell waves, which have been generated in storms far
away. A typical wave spectrum will then be narrow-banded. For other coasts, locally-
generated storm waves dominate the wave climate, as is the case for the Dutch coast.
The sea state then is irregular and short-crested. Most of the time, wave records off the
Dutch coast show both swell waves generated in distant storms and locally-generated
storm waves. Two distinct peaks can then be observed in the spectrum, see Fig. 3.14.
The swell can only come from the north and is usually not older than a day or some-
times two days.

118 3. Ocean waves
wave height [m] 4
3
period 14
13
period [sec]
2
12
height
1
11
0 10
0 2 4 6 8 10 12 14 16 18
decay distance [10 3 km]
Figure 3.13: The effect of swell decay on wave period and wave height in the case of a 6.1 m
high, 10-second wave according to J. L. Davies and Clayton (1980).
E(f,θ) 2D spectrum
storm swell
wind-sea
north
North Sea
θ f
local breeze
f
1D spectrum
wind-sea
E [m2 /Hz]
swell
f [Hz]
Figure 3.14: At a location off the Dutch coast, a northerly swell, generated by a storm off the
Norwegian coast, meets a southwesterly sea generated by a local breeze (left). The 2D spectrum
represents the spectral energy as a function of frequency and direction and is constructed by
combining JONSWAP spectral shapes with Gaussian directional distributions. The 1D spectrum
is obtained through integration over all directions. Even though in the 2D spectrum the swell peak
is higher than the sea peak, the sea peak is the larger peak in the 1D spectrum because of the
larger directional spreading for sea as compared to swell.

3.6. Long-term statistics and extreme values 119
3.6. Long-term statistics and extreme values

As indicated above, waves are measured at regular intervals of 3 to 6 hours during
a relatively short period during which the record can be considered stationary. This
results in a series of wave observations with a sampling interval of 3 or 6 hours. If
long enough (years or even decades, as in the case of the Dutch wave data), this series
can in its turn be considered a set of random data representing the long-term wave
climate of the location. Not only measured wave climates are used for this, but also
hindcasts based on archived wind fields. Note that for periods of decades, conditions
may not be stationary (due to for instance climate change impacts on wave climate).
For many engineering problems long-term statistics of average parameters are suffi-
cient. The long-term data can be represented in various ways:
• Histograms of 𝐻𝑠 present the percentage of occurrence of a certain significant
wave height.
• Scatter plots of 𝐻𝑠 versus wave period show the dependency of wave periods
and wave heights.
• Tables can also include information on wave period and wave angle, for instance
a table valid for a certain direction sector which gives the percentage of occur-
rence of 𝐻𝑠 versus 𝑇0 .
• Wave roses give the directional distribution of the wave heights.
Mostly, long-term distributions of significant wave heights are determined, which
present 𝐻𝑠 versus the percentage of exceedance. Long-term distributions of wave peri-
ods and wave angles can also be determined, but are often considered to be a function
of the long-term wave height distribution. Sometimes the long-term distribution func-
tions for 𝐻𝑠 are calculated for different wave angle classes.
Since extreme conditions are not always part of observed data, extrapolation is some-
times needed. For extrapolation of the probabilities, several distributions can be used.
The choice does not have a theoretical basis. Often used are the log-normal distribu-
tion and the Weibull distribution. These analyses will not give information on when
an event will happen but it is possible to determine how often it is likely to happen.
The percentage of exceedance can also be expressed as a return period. The return
period is defined as the average time between events with a (significant) wave height
larger than a certain value.
For an engineering project the wave climate is generally known at some distance from
the coastal site. We can then use a wave model to translate this offshore wave climate
to a climate representative for the project site. However, a full wave climate consists
of a multitude of wave conditions, which is often not practical. An example of a year-
averaged wave climate based on a wave study is shown in Table 3.3 and Fig. 3.15.
Instead of using all those conditions in morphodynamic computations, engineers of-
ten reduce such a climate to a representative set with fewer conditions. But what is

120 3. Ocean waves
representative in this context? That depends on the problem under consideration. For
an engineering problem which is governed by longshore sediment transport, a repres-
entative set of wave conditions means that at least the total longshore transport rate
reproduced by the smaller set is identical to the rate based on the full climate. After
such a wave climate schematisation, morphodynamic computations can be carried out
for fewer conditions, which is therefore faster.
Table 3.3: Wave conditions for validation case. From Mol (2007)
condition 𝐻𝑠 [m] dir [°N] 𝑇𝑝 [s] duration [%] days/yr

1 0.4 188 2.85 1.00 3.65
2 0.4 203 2.85 1.50 5.48
3 0.6 203 3.49 1.00 3.65
4 0.4 218 2.85 6.00 21.9
5 0.6 218 3.49 3.00 10.95
6 1.2 218 4.93 3.00 10.95
7 1.8 233 6.04 1.00 3.65
8 1.4 233 5.32 2.00 7.3
9 1.0 233 4.50 3.00 10.95
10 0.8 233 4.02 4.00 14.6
11 0.6 233 3.49 6.00 21.9
12 0.4 233 2.85 8.00 29.2
13 0.4 248 2.85 7.00 25.55
14 0.8 248 4.02 5.00 18.25
15 1.2 248 4.93 4.00 14.6
16 1.6 248 5.69 1.00 3.65
17 1.6 263 5.69 1.00 3.65
18 0.8 263 4.02 3.00 10.95
19 0.4 263 2.85 4.00 14.6
20 0.8 278 4.02 4.00 14.6
21 0.4 278 2.85 4.00 14.6
22 0.4 293 2.85 3.00 10.95
23 0.6 293 3.49 2.50 9.13
24 0.8 308 4.02 3.00 10.95
25 0.6 308 3.49 3.00 10.95
26 0.4 308 2.85 3.00 10.95
27 0.6 323 3.49 3.50 12.78
28 0.4 323 2.85 4.00 14.6
29 0.4 338 2.85 3.50 10.95
30 0.8 338 4.02 2.00 9.13
total 100 365

3.7. Generation of the tide 121
30%
20% HS [m]
10% 1.75 – 2
1.5 – 1.75
W E 1.25 – 1.5
1 – 1.25
0.75 – 1
0.5 – 0.75
0.25 – 0.5
Figure 3.15: Wave climate for the validation case in Table 3.3 presented as a wave rose.
For applications like determining the height of a deck of a platform, the probabilities of
individual wave heights are required. In such cases, the short-term and the long-term
expectations must be combined to obtain the probability of exceedance of individual
wave heights during a fixed period, for example the lifetime of the structure. This is
also often close to a Weibull distribution.
Because in shallow water there is a direct relation between maximum breaking wave
height and water depth, the wave height distribution is not independent of the occur-
rence of extreme water levels. Close to the shore, this could mean that the long-term
distribution of wave heights coincides with the distribution of extreme water levels.
3.7. Generation of the tide

3.7.1. Equilibrium theory of the tide
At the coast, the tide is most easily observed as daily water level variations. At most
places in the world, the tide is essentially semi-diurnal, i.e. having a period of ap-
proximately half a day. This means that two high waters and two low waters can be
observed daily. The tidal range between high water and low water can be 10 metres
or more depending on the location.
Isaac Newton was the first to explain the generation of the tide. After having identified
the tide-generating forces, he assumed the ocean water to respond instantly to these
forces. That is why his theory is called the equilibrium theory of tides. He also neg-
lected the presence of continents and assumed the earth to be entirely covered by wa-
ter. In the present section, we follow his theory to explain some well-known tidal phe-
nomena. Subsequently, in Sect. 3.8 the effects of land masses and non-instantaneous
response are discussed.

122 3. Ocean waves
3.7.2. Gravitational pull

The tide-generating forces find their origin in the gravitational pull of the moon and
the sun on the water in the oceans. The sun and the earth revolve around a common
centre of mass. They attract each other by a force that is proportional to their respect-
ive masses and inversely proportional to the square of their mutual distance. Similarly,
the moon and the earth revolve around their common centre of mass. Revolving about
a common centre of mass implies that the attraction forces act as centripetal forces, so
that the earth is constantly being accelerated towards the moon and the sun, as in free
fall, the direction of the free fall being everywhere parallel to the line connecting the
centres of mass. The difference in mass between the sun and the moon causes the
common centre of mass of the earth-sun system to be located inside the sun, and of
the earth-moon system within the earth. As a simplification, we can therefore say that
the earth orbits the sun and that the moon orbits the earth (see Fig. 3.16).
moon
earth
sun
Figure 3.16: The orbit of the earth around the sun and the smaller orbit of the moon. The earth-
moon system is orbiting the sun in the same direction as the moon is orbiting the earth.
At the same time, the earth rotates around its own axis in the same direction as the
earth orbits the sun and the moon orbits the earth. The periods of these motions are
explained in Intermezzo 3.3.
The gravitational pull of the sun on 1 kg of mass of the earth using the average distance
between the sun and the earth of 𝑑𝑠 = 1.5 × 1011 m is:
𝑀𝑠
𝑎𝑠 = 𝐺 = 6.0 × 10−4 𝑔 (3.27)
𝑑𝑠2
where 𝐺 = 6.6 × 10−11 N m2 /kg is the universal gravitational constant and 𝑔 = 9.81 N/kg
(or equivalently m/s2 ) is the gravitational pull of the earth itself at the surface of the

Table 3.4: Main facts about the sun-earth and earth-moon systems (approximate figures).
mass of the sun 1.99 × 1030 kg

mass of the earth 5.98 × 1024 kg
mass of the moon 7.35 × 1022 kg
distance between (centres of) the sun and the earth 1.50 × 108 km
distance between (centres of) the earth and the moon 3.84 × 105 km
radius of the earth 6.37 × 103 km
earth. Similarly, the gravitational pull of the moon on 1 kg of mass of the earth at an
average distance of 𝑑𝑚 = 3.84 × 108 m is:
𝑀𝑚
𝑎𝑚 = 𝐺 2
= 3.4 × 10−6 g (3.28)
𝑑𝑚
Note that Eqs. 3.27 and 3.28 represent the gravitational acceleration (force per unit
mass) for the centre of the earth. As such, they represent the attraction of the earth as
a whole towards the sun and the moon (assuming the mass of the earth is concentrated
at the earth’s centre of mass).
For different locations at the earth’s surface, the magnitude and direction of the grav-
itational acceleration slightly differ from each other and from the acceleration of the
earth’s centre, due to varying distances and angles to the centre of the attracting mass
(see Fig. 3.17).
Summarizing, the gravitational attraction provides the centripetal acceleration

that maintains the motion of the earth around the centre of mass of the earth-
sun (earth-moon) system. For every point on earth, this centripetal acceleration
of the earth in free fall towards the sun (moon) is directed parallel to the line
connecting the centres of mass of the earth and the sun (moon). Its magnitude
is determined by the gravitational acceleration 𝑎𝑠 (𝑎𝑚 ) of the sun (moon) on the
earth’s centre as given by Eq. 3.27 (Eq. 3.28). But nowhere at the surface of the
earth is the gravitational acceleration exactly equal in magnitude and direction to
the centripetal acceleration.
3.7.3. Differential pull or the tide-generating force

Equations 3.27 and 3.28 show that the sun’s gravitational pull is two orders of mag-
nitude larger than the moon’s gravitational pull. This makes sense, since the earth
orbits the sun and not the moon. Nevertheless, the sun contributes only about 30 %
of the tidal amplitudes in the oceans; the moon is responsible for the remaining 70 %.
The reason is that it is not the gravitational pull per se that is responsible for the tide-
generating forces.

124 3. Ocean waves
as or m, local
to the attracting mass

as or m Ms or Mm
Figure 3.17: The gravitational acceleration towards the sun 𝑎𝑠,local or moon 𝑎𝑚,local depends on the
distance to the attracting mass and on the angle to its centre. On the side of the earth closer to
the attracting mass, the acceleration is greater than on the far side of the earth. The attraction
of the earth as a whole towards the sun (moon) is found by considering the mass of the earth
concentrated at the earth’s centre of mass and is equal to 𝑎𝑠 (𝑎𝑚 ).
Above, we have seen that the gravitational attraction provides the centripetal acceler-
ation that maintains the motion of the earth around the centre of mass of the earth-
sun (earth-moon) system. With the gravitational attraction for the earth as a whole
(Eqs. 3.27 and 3.28) accounted for4 , the tide is generated by a much subtler effect, i.e.
the difference between the gravitational pull on ocean water masses that are located
at different distances from the sun and the moon.
Consider 1 kg of mass on the near side of the earth which therefore is the earth’s radius
𝑅 = 6.37 × 106 m closer to the sun than the centre of the earth is. The gravitational pull
of the sun on 1 kg of mass on the near side of the earth is Δ𝑎𝑠 greater than 𝑎𝑠 , i.e.:
𝑀𝑠 𝑀𝑠
Δ𝑎𝑠 = 𝑎𝑠,near side − 𝑎𝑠 = 𝐺 2
−𝐺 ≈
(𝑑𝑠 − 𝑅) 𝑑𝑠2
(3.29)
𝑀𝑅 2𝑅
2𝐺 𝑠3 = 𝑎𝑠 = 0.515 × 10−7 𝑔
𝑑𝑠 𝑑𝑠
This is known as differential pull. Note that the first-order approximation of Δ𝑎𝑠 is
inversely proportional to the cube of the distance. Similarly for the moon:
𝑀𝑚 𝑅 2𝑅
Δ𝑎𝑚 ≈ 2𝐺 = 𝑎𝑚 = 1.13 × 10−7 g (3.30)
𝑑𝑚3 𝑑𝑚
4
Alternative but equivalent explanations (as covered in TU Delft courses CTB2410 and CIE5317) take the
moving earth as a reference frame and introduce ‘fictitious’ centrifugal forces (see also Intermezzo 3.1).
Due to the revolution of the earth around the centre of mass of the earth-sun (earth-moon) system, there
is an outward-directed centrifugal force that is the same at every point on the earth and directed parallel
to the line of centres. The centrifugal acceleration (force per unit mass) balances the gravitational
acceleration of the earth’s centre towards the sun (moon).

The point furthest from the sun (moon) is in turn smaller than 𝑎𝑠 (𝑎𝑚 ) by the same
amount Δ𝑎𝑠 (Δ𝑎𝑚 ) which can be shown by a similar calculation. Carrying out the cal-
culation for all places on the earth leads to a differential pull on the earth, as indicated
schematically in Fig. 3.18 (compare with Fig. 3.17).
to Ms or Mm
Figure 3.18: The direction and relative magnitudes of the differential gravitational pull at different
locations on the earth’s surface. At each location, it is given by the vector subtraction of 𝑎𝑠 or 𝑚,local
and the part 𝑎𝑠 or 𝑚 that provides the centripetal acceleration. The depicted tidal bulges result
from the tangential components of the differential pull (see Fig. 3.19).
The differential pull is responsible for the tidal generation and is therefore also referred
to as tidal force. Since Δ𝑎 is proportional to 𝑀/𝑑 3 , the solar differential pull Δ𝑎𝑠 is only
0.46 times the lunar differential pull Δ𝑎𝑚 . The moon is responsible for 69 % of the tidal
mechanism, as Δ𝑎𝑚 /(Δ𝑎𝑚 + Δ𝑎𝑠 ) = 69 %.
The differential pull as sketched in Fig. 3.18 has components normal and tangential
(or parallel) to the earth’s surface. The normal components are many orders of mag-
nitude smaller than the earth’s own gravitational attraction. For example, for the point
closest to the sun, the differential acceleration is normal to the earth surface and with
a magnitude of 0.515 × 10−7 𝑔 it is negligible compared to 𝑔. The tangential or ‘hori-
zontal’ components are of the same order of magnitude as the normal components,
but since they are perpendicular to the earth’s gravity field, they cannot be neglected.
The tangential forces are demonstrated in Fig. 3.19.
The effect of the tangential forces is to shift water to the side of the earth facing the
sun (moon) and to the opposite side in tidal bulges (see Fig. 3.18). The piling up of
water due to the tangential forces is balanced by pressure gradients in the opposite
direction due to the sloping water surface. If the earth were completely covered by
water, its equilibrium configuration would be an ellipsoid (rugby ball or egg).
The rapid diurnal rotation of the earth around its own axis makes the earth rotate
underneath the tidal bulges, thus producing a semi-diurnal tide with two high and
two low waters passing the same point on the earth every day. The period of the solar
tide is exactly 12 h (our day is measured in terms of the sun, Intermezzo 3.3). The
period of the lunar tide is governed by the period between the moon phases or the
lunar day. The period of the semi-diurnal lunar tide is thus equal to half a lunar day =
12 hours and 25 minutes.

126 3. Ocean waves
60º
30º
body
n e rating
e-ge
e tid 0º
to th
30º
Figure 3.19: The horizontal component of the tidal force on the earth when the tide-generating
body (the sun or the moon) is above the equator at 𝑧.
Intermezzo 3.3
Movements of the earth, earth-moon and earth-sun system
The earth circles the sun in 365.25 days. The earth rotates around its own axis in
23 hours and 56 minutes (sidereal day, relative to the stars) in the same direction in
which the moon is revolving around the earth and the earth around the sun. While
the earth completes one daily rotation, the earth also revolves with respect to the
sun. As a result it is only after 24 hours or one (solar) day that the earth returns in
the same position with respect to the sun as seen by an observer on the earth. A
lunar day lasts even longer, viz. 24 hours and 50 minutes. This occurs because the
moon progresses in its orbit around the earth during earth’s daily rotation. The
extra 50 minutes are required for the earth to ‘catch up’ with the moon, as will be
explained below.
The period of the moon’s revolution around the earth is 27.3 days (sidereal month).
The moon takes 29.5 days to return to the same position relative to the sun as seen
by an observer on the earth. Such a lunar month is the time between successive
recurrences of the same phase; e.g., between full moon and full moon. During the
27.3 days of the sidereal month, the earth has moved along in its orbit around the
sun and now the moon must ‘catch up’ to this new position. It takes 2.2 days to
do so.

Furthermore, while the moon progresses in its orbit around the earth, the earth
rotates around its own axis. The earth must rotate a bit more than a full rotation
before any given location ‘catches up’ with the lunar bulge. During one rotation
of the earth around its axis, the moon has covered 1/29.5th of its total orbit around
the earth (with respect to the sun). Therefore, a lunar day is equal to (1 + 1/29.5)×
24 h = 24 h 50 min. Figure 3.20 shows the required additional rotation of the earth
to catch up with the moon.
moon
h
art
e
24/29 hours
Figure 3.20: A lunar day (the time between successive transits of the moon for an observer
on the earth) is 24 hours and 50 minutes.
3.7.4. Spring and neap tide

When the sun, the earth and the moon are in one line (at full and new moon), the solar
and lunar tides reinforce each other. The ellipsoid becomes more pronounced and the
tide gets a bigger amplitude; this is called spring tide. When the solar and lunar tides
are 90° out of phase, their effects cancel each other (in the first and last quarter). The
ellipsoid approaches a circle, and consequently the tide gets a smaller amplitude. This
situation is called neap tide (Fig. 3.21). Figure 3.22 shows an example of these tidal
variations.
The spring and neap tide cycle varies with moon phases and therefore with the lunar
month of 29.5 days. The ratio of spring and neap tide amplitudes according to equilib-
rium theory can be estimated from Eqs. 3.29 and 3.30 as:
(0.515 + 1.13)/(1.13 − 0.515) = 2.7 ∶ 1 (3.31)
This ratio is an approximation for the spring to neap tide ratio at open oceans. Any-
where else the ratio is affected by the presence of land masses (see Sect. 3.8).
3.7.5. Daily inequality

So far, we have either explicitly or implicitly assumed that the sun and the moon are
directly above the equator of the earth (or in other words that the orbits of the moon
and the earth lie in the equatorial plane). Under that assumption a certain place on the

128 3. Ocean waves
n
ne su
spr
w
ing
moon
tid
e
ne
ap
tid
r
arte
qu
e
t
firs
sp
rin
gt
ide
nea
p ti
de
ll
fu
m
oon
ar ter
lunar tide
qu
solar tide
ird
th
Figure 3.21: When the sun, the earth and the moon are in one line, the tidal forces of the sun and
the moon reinforce each other and spring tides occur. Neap tides occur when the effect of the
tidal forces of the sun and the moon on the tidal bulges cancel each other.
first quarter full moon third quarter new moon first quarter full moon
2
elevation [m +NAP]
-1
-2
9
/19
/19
/19
/19
/19
/19
1/1
/11
/10
/10
/10
/10
/10
/1
07
14
03
10
17
24
31
Figure 3.22: Tidal variation in Vlissingen during more than a month in 2019. Note the spring and
neap tidal variation and the daily inequality (Sect. 3.7.5). Data from https://waterinfo.rws.nl.

earth experiences two high waters and low waters per day of equal height as can be
seen from Fig. 3.23. The heights of those high waters and low waters depend on the
latitude.
earth
to the tide-generating body

A
Figure 3.23: The tidal bulges for a tide-generating body above the equator. Observers A and
B both experience two high and low waters a day of equal height. The heights depend on the
latitude.
In reality, the orbits of the moon about the earth and of the earth about the sun are not
in the equatorial plane, which complicates our simple picture of Fig. 3.23. While the
moon’s orbit and the earth’s orbit are approximately in the same plane (5° difference),
there is a time-varying declination angle between the equatorial plane and the earth-
sun and earth-moon connection lines. As the tidal bulges tend to align themselves
with the tide-generating body, the two high and low waters per day are not equal
(see Fig. 3.24). This phenomenon is referred to as daily inequality. Apparently, the
declination of the sun has a diurnal (daily) effect on the tides.
According to Fig. 3.24, the daily inequality is zero at the equator and increases with
latitude5 . At some higher latitudes, the daily inequality becomes so big that there is
only one high and one low water (diurnal tide, see Sects. 3.7.6 and 4.4.1).
The earth’s axis is tilted by 23.5° with respect to a line perpendicular to the plane of
the earth’s orbit around the sun. Due to the combination of this tilt and the orbiting
of the earth around the sun, the declination of the sun varies seasonally. It is zero at
the spring and autumn equinoxes (March 22 and September 22), positive during the
Northern summer and negative during the Northern winter. The minimum declination
of −23.5° (or 23.5° south) is reached on December 22 (Northern winter solstice) and
the maximum of 23.5° on June 22 (Northern summer solstice). The daily inequality
cycle of the solar tide therefore has a period of one year, with diurnal tides increasing
5
Some areas around the equator experience a large daily inequality. This is related to the presence of
land masses and will be further explained in Sect. 4.4.

130 3. Ocean waves
earth
equat to the tide-generating body

or pla
ne δ
A
Figure 3.24: Daily inequality due to a non-zero declination angle 𝛿 between the equatorial plane
and the line connecting the centres of the earth and the tide-generating body. For observer A
the two high and low waters have the same height, but observer B experiences unequal heights
in two successive high and low waters.
with increasing declination south or north and semi-diurnal tides maximum at zero
declination. An interesting consequence of the latter is that spring tidal ranges around
the equinoxes are usually higher than average spring tidal ranges (equinoctial tides).
The lunar daily inequality cycle has a period of 27.3 days. The north-south difference
of the declination of the moon during the month is twice 23.5° ± 5°. The deviation of
5° has a cycle of 18.6 years and is a result of the 5° difference in moon’s and earth’s
orbits. As with the sun, the largest daily inequalities correspond to the times that the
moon is furthest south or north (minimum and maximum declination).
3.7.6. Tidal constituents

The principal tidal constituents are shown in Table 3.5. We have seen in Sect. 3.7.3
that the main lunar tide has a period of 12.42 h and the main solar tide a period of 12 h
respectively. These tidal constituents (or tidal components) are called M2 and S2. The
influence of the sun is characterised by the letter S, the influence of the moon by the
letter M. The index 2 refers to phenomena that occur twice daily. The amplitudes and
phases of these two constituents vary with the location on the earth.
The tidal variations are in the order of decimetres only in the open oceans. According
to equilibrium theory, the amplitudes of M2 and S2 are 0.24 m and 0.11 m respectively.
This gives a ratio of S2/M2 = 0.46, which could also be calculated using Eqs. 3.29
and 3.30 as S2/M2 = 0.515/1.13 = 0.46. The M2 and S2 signals have a slightly different

Table 3.5: Principal tidal constituents with equilibrium amplitudes from Apel (1987)
Tidal constituents Name Equilibrium Period [h]

Amplitude [m]
Semi-diurnal
Principal lunar M2 0.24 12.42
Principal solar S2 0.11 12.00
Lunar elliptical N2 0.046 12.66
Lunar-solar declinational K2 0.031 11.97
Diurnal
Lunar-solar declinational K1 0.14 23.93
Principal lunar O1 0.10 25.82
Principal solar P1 0.047 24.07
Lunar elliptical Q1 0.019 26.87
Long period
Fortnightly Mf 0.042 327.9
Monthly Mm 0.022 661.3
Semi-annual Ssa 0.019 4383
frequency, which gives a so-called beating of the two signals resulting in spring-neap
tide variability: when the M2 and S2 components are in phase (the moon and the sun
are aligned) it is spring tide and when they are out of phase it is neap tide, as discussed
in Sect. 3.7.4. The beating of the two signals resulting in a variation of the amplitudes
over the lunar month is comparable to the beating of two wave trains with slightly
different frequencies, resulting in amplitude variation or modulation on a wave group
scale as treated in Sect. 3.5.3. Intermezzo 3.4 treats this amplitude modulation in more
detail.
The declination of the earth axis introduces semi-diurnal and diurnal tidal constituents
K1, K2 and O1 and P1. Diurnal components carry a subscript 1. K1 with O1 expresses
the effect of the moon’s declination, K1 with P1 the sun’s declination. K1, P1 and
O1 account for diurnal inequality (see Intermezzo 3.4) and, at extremes, diurnal tides
(where the semi-diurnal component has disappeared completely). The K2 constituent
modulates the amplitude and frequency of M2 and S2 for the declinational effect of
the moon and the sun, respectively. Other effects will generate other tidal constitu-
ents defined by an exact period, with their own amplitudes according to equilibrium
theory and with their own phases with respect to each other. For instance, the moon’s
distance from the earth varies because the moon’s orbit is elliptical and because the el-
liptical orbit is not fixed. This effect introduces semi-diurnal and diurnal constituents
N2 and Q1. The longest period is 18.6 years, which is the period of the 5° variation of
the lunar declination.

132 3. Ocean waves
For practical purposes, the tide can be seen as a sinusoidal semi-diurnal water level
variation modified with a fortnightly spring and neap tide amplitude variation and
with a daily inequality that varies with latitude and with the monthly and annual
cycle. In the extreme case the daily inequality is so large that the water level variation
is only diurnal.
M2 and S2 are the main but not the only constituents with frequencies near twice per
day (see Fig. 3.27). Note that the spectrum of equilibrium tides in Fig. 3.27 consists of
discrete lines; tides have precise frequencies determined by the orbits of the earth and
moon, and their spectrum is not continuous. We have seen before that ocean waves
have all possible frequencies, and their spectrum is continuous.
Intermezzo 3.4 Astronomical constituents
Section 3.7.6 described the ocean tide as a sinusoidal semi-diurnal water level vari-
ation modified with a fortnightly spring and neap tide variation and with a daily
inequality. The spring and neap tide variation is the result of the linear summa-
tion of principal components with a small differential frequency. Consider the M2
and S2 components. The linear combination of these two tidal components with
a small difference in frequency results in an amplitude variation with a period de-
termined by 𝜔S2 − 𝜔M2 , comparable to the beating of two short-wave trains with
slightly different frequencies (see Sect. 3.5.3). This is shown in Fig. 3.25.
1 M2
elevation [m]
S2
-1
beating period or group period
spring neap spring
1 M2+S2
elevation [m]
-1
-2 0 2 4 6 8 10 12 14 16
days
Figure 3.25: The interaction between M2 and S2 results in a spring and neap tide variation.
Verify that the beating period of M2 and S2 is 14.77 days (cf. the group period for wind
waves Eq. 3.24b).
The daily inequality for a predominantly semi-diurnal tide is the result of the sum-
mation of a semi-diurnal and a diurnal component and is demonstrated in Fig. 3.26

3.8. Propagation of the tide 133
for M2 and K1. The result is the succession of two symmetrical tides with differ-
ent tidal range (the difference between the elevations of the two successive high
waters is called daily inequality).
1 M2
elevation [m]
K1
-1
M2+K1
1
elevation [m]
-1
-2 0 2 4 6 8 10 12 14 16
days
Figure 3.26: The interaction between M2 and K1 leads to daily inequality between two suc-
cessive high waters and between two successive low waters. Also, a slower amplitude
variation can be observed.
3.8. Propagation of the tide

3.8.1. Dynamic theory of tides
So far, the earth was schematised as if it were completely covered with water, with the
earth rapidly turning through the slowly varying tide. For the semi-diurnal tide this is
equivalent to a propagating tidal wave covering the entire circumference of the earth
in one day. In reality, continents prevent the development of the tidal ellipsoid and the
land masses do not move through the tide but move the water masses along with them.
Furthermore, the limited water depth of the oceans prevents the development of an
equilibrium tide. This becomes clear when considering the propagation properties of
the tidal wave. The tidal wave is a long wave because the wavelength 𝐿 ≫ ℎ and has a
small amplitude (order 1 m on an average ocean water depth of 4000 m). Therefore, and
if friction can be neglected, the wave propagation speed follows from (see Sect. 3.5.2):
𝑐 = √𝑔ℎ (3.32)
where:

134 3. Ocean waves
10 2
M₂ S₂
N₂
K₂
10 1
amplitude [cm] 2N₂ μ₂ ν₂ L₂ T₂

0
10 ε₂ η₂
λ₂ R₂
3N₂ α₂ β₂
γ₂ 2T₂ ζ₂
10 -1
δ₂
10 -2
26 27 28 29 30 31 32
frequency [degrees/hour]
2
10
S₂
K₂
10 1
amplitude [cm]
T₂
10 0
R₂
2T₂
10 -1
10 -2
29.80 29.85 29.90 29.95 30.00 30.05 30.10 30.15 30.20
frequency [degrees/hour]
Figure 3.27: Spectrum of equilibrium tides with frequencies near twice per day. Upper: the spec-
trum is split into groups separated by a cycle per month (0.55 °/h). Lower: expanded spectrum
of the S2 group, showing splitting at a cycle per year (0.04 °/h). The finest splitting in this fig-
ure is at a cycle per 8.847 years (0.0046 °/h). (From Richard Eanes, Center for Space Research,
University of Texas in Stewart, 2008, see also https://github.com/introocean/introocean-en).
𝑐 wave propagation speed m/s

𝑔 gravitation acceleration m/s2
ℎ water depth m
Note that this phase velocity can also be derived directly on the basis of continuity
and by balancing acceleration and pressure gradient, see Intermezzo 3.5.
For the development of the equilibrium tide, the semi-diurnal tide should cover half of
the circumference of the earth in one period of 12 hours and 25 minutes. The circum-
ference at the equator is 2𝜋𝑅 = 2𝜋 × 6.37 × 103 km = 4.00 × 104 km and therefore the
required propagation speed is 𝑐 = 4.00 × 104 km/(2 × 12.42 h) = 447 m/s. This means
that the tidal wave requires a water depth of 20 km – much more than the depth of
the oceans – to travel fast enough (use Eq. 3.32). The travel distance reduces towards
the poles and therefore the required propagation speed and water depth as well. The
result is that the water depth becomes less of a limitation at higher latitudes.
Only in the Southern Hemisphere at a latitude of about 65°S, an equilibrium tide can
more or less exist. To the south of Africa, South America and Australia the earth is
circled by an uninterrupted band of water such that the tidal wave can travel around

the earth. Besides, at this latitude the water depth is not so much of a limitation for
the propagation velocity, so that the original ellipsoid can develop.
From around 65°S the tidal wave propagates to the north into the Atlantic, Indian and
Pacific Oceans. Because it takes time for the tidal wave to progress through the oceans,
the tidal constituents at a particular location away from the area of tidal generation
(around 65°S) lag behind the theoretical constituents from equilibrium theory. The
further the location is away from the South Pole, the longer the time shift between the
celestial event and its appearance in the form of the tide.
Assuming an average depth of the ocean of 4000 m, the propagation speed of the tidal
wave is about 200 m/s and for 𝑇 = 12.42 h the wavelength 𝐿 = 9000 km. At a speed
of 200 m/s (or 720 km/h) and if travelling in a more less straight line, it takes the tidal
wave less than a day to travel from 65°S to say, Scotland. To reach the Netherlands,
the tidal wave needs to cross the shallow North Sea basin. The latter takes another
day or so, since the propagation speed strongly decreases in the shallow North Sea.
This means that in the Netherlands spring and neap tide occur about two days after
the corresponding moon configurations (see Fig. 3.22).
On its way, the tidal wave is distorted by local differences in water depth and – due
to land masses – by restriction of the width or reflection. The period does not change,
only the length of the wave changes via changes in propagation speed. The latter can
be illustrated as follows. For the long tidal wave the wave celerity is proportional to
the square root of the water depth. When the tidal waves reaches shallower water, the
celerity will decrease which results in a concentration of energy and thus an increase
in tidal amplitude (comparable to the shoaling effect of wind waves in shallow water;
see Sect. 5.2.2). In shallower water along the ocean coasts and shallow seas like the
North Sea, we have for instance a wavelength of 𝐿 = 450 km (with ℎ = 10 m and
𝑐 = 10 m/s). Depending on the geometry of a bay, sea or ocean basin, resonance may
occur, leading to amplification of the tidal amplitude at the coast (standing waves or –
in combination with Coriolis – rotary standing waves, see Sect. 3.8.2).
In deep water, tidal current velocities are very small. For an ocean depth of 4000 m and
a tidal amplitude of for instance 0.25 m, we get with Eq. 3.39 a tidal velocity amplitude
of 1.2 cm/s. In shallower water this is quite different: tidal amplitude and tidal current
velocities are larger. If 𝑎 = 1 m and ℎ = 10 m, the tidal current velocity is 1 m/s, hence
two orders of magnitude larger.
Intermezzo 3.5 Tidal propagation in the open oceans
The tide can be described by the shallow-water equations (shallow-water approx-

imation to the Navier-Stokes equations). For the tidal propagation in open oceans
we can simplify these equations by neglecting advection, friction, horizontal dif-
fusion and short-wave effects. We also assume a small tidal amplitude and bottom

136 3. Ocean waves
slope. The resulting momentum equation is a balance between acceleration (iner-

tia) and the pressure gradient:
𝜕𝑢 𝜕𝜂
= −𝑔 (3.33)
𝜕𝑡 𝜕𝑥
where 𝑢 is the horizontal velocity and 𝜂 the oscillatory surface elevation. We fur-
ther have the vertically integrated continuity equation:
𝜕𝜂 𝜕𝑢
= −ℎ (3.34)
𝜕𝑡 𝜕𝑥
where ℎ is the water depth. Eqs. 3.33 and 3.34 together describe the tidal propaga-
tion.
𝜕 𝜕
Taking 𝜕𝑡
of Eq. 3.34 and subtracting ℎ 𝜕𝑥 of Eq. 3.33 gives the classical wave equa-
tion:
𝜕 2𝜂 𝜕 2𝜂
= 𝑔ℎ (3.35)
𝜕𝑡 2 𝜕𝑥 2
Substitution of a progressive sine (or equivalently: cosine) wave 𝜂 = 𝑎 sin (𝜔𝑡 − 𝑘𝑥)
into Eq. 3.35 yields:
𝜔 2 = 𝑔ℎ𝑘 2 (3.36)
and thus (only taking the positive solution for 𝑐):
𝜔
𝑐= = 𝑔ℎ (3.37)
𝑘 √
The velocity is in phase with the surface elevation:
𝑔𝑎𝑘
𝑢= sin(𝜔𝑡 − 𝑘𝑥) (3.38)
𝜔
Evidently, for progressive waves the maximum velocities are found under the crest
of the wave. The velocity amplitude is:
𝑔𝑎𝑘 𝑎𝜔 𝑔
𝑢̂ = = =𝑎 (3.39)
𝜔 𝑘ℎ √ℎ

Compare this expression for the velocity amplitude with the expression according
to linear wave theory in the shallow-water approximation (see Eq. 5.23). As you
can see, it is identical.
3.8.2. Amphidromic systems

The propagation of the tide is influenced by friction and resonances determined by the
shapes and depths of the ocean basins and marginal seas. Because of the large scale of
the tidal motion, the tidal propagation is also influenced by Coriolis acceleration (see
Intermezzo 3.1). Generally we can neglect the effect of Coriolis for waves shorter than
a few kilometres. In the previous section we found tidal wavelengths of the order of
thousands of kilometres in the oceans and hundreds of kilometres in shallow seas.
Since the movement of the tides is deflected by Coriolis and blocked by land masses,
rotary movements are formed in oceans basins, bays and seas that are counter-clockwise
in the Northern Hemisphere and clockwise in the Southern Hemisphere. Such rotary
systems are called amphidromic systems. In an amphidromic system the wave pro-
gresses about a node (no vertical displacement) with the antinodes (maximum vertical
displacement) rotating about the basin’s edges (see Figs. 3.28 and 3.29). The water can
be seen to be sloshing around the basin. The node, where the amplitude of the vertical
tide is zero, is called an amphidromic point.
Figure 3.28: Normal standing wave.
It is possible to visualise the propagation of the tidal wave by mapping the lines of
simultaneous high water (occurrence of High Water (HW) in sun hours after moon
culmination) and the lines of equal tidal range (vertical distance between HW and
Low Water (LW) in m). The lines of simultaneous HW are called co-tidal lines or,
since these lines connect points of equal phase, co-phase lines. They often radiate
away from a node and are not equally spaced, since the propagation speed depends on
the water depth ℎ. Co-range lines connect points experiencing the same tidal range.
They often form irregular concentric circles about a node. This is illustrated in Fig. 3.30
and Fig. 3.31.
An amphidromic point is said to be degenerate when its centre appears to be located
over land rather than water. Examples are found at the southern tip of Norway and

138 3. Ocean waves
Figure 3.29: Tidal oscillation of a basin (Northern Hemisphere). The frequency of the oscillation
is determined by the size of the basin and its depth.
northwest of Bournemouth (along the southwest coast of England, southeast of Bristol,

see Fig. 3.31).
Evidently, the local tide at a coastal location is dependent on the size, shape and depth
of the basin. Every place along the coasts of the world has its own specific tidal curve.
If the tidal forcing is in resonance with an oscillation period for the sea or bay, the tidal
range is amplified and can be enormous. At some locations, the difference between
high and low water is up to 12 m (compare that to a few decimetres at the open oceans!).
3.8.3. Kelvin waves

As coastal engineers we are interested in the tidal propagation and the tidal range along
the boundaries of oceans and seas. To better understand the propagation along closed
coastal boundaries, we need a further understanding of the rotary wave forming the
amphidromic system. These waves depend on the existence of a closed boundary. A
definition sketch is given in Fig. 3.32. Wave propagation is in the positive 𝑦-direction
in the Northern Hemisphere (NH), hence 𝑐 is positive. Such a wave would deflect from
this eastern boundary in the Southern Hemisphere (SH) because of Coriolis. Therefore,
for the SH we expect a wave propagating in the negative 𝑦-direction (𝑐 negative).
As before, we assume a more or less horizontal bottom slope and a small tidal amplitude
compared to the water depth. Balancing inertia, Coriolis, the pressure gradient and

60°N
8h
h
10
6
5h
h
4h
3h
40°N 12 h
0.6
m 2h
m
8
0.
0.8
0h
1.6
m
1.2
6
0.
m
m
1h
12
0.4 m h
11
h
10
h
20°S
9h
0.6 m
m 1.0 m
0.8 1.2 m
8h
1.4 m
1.6 m
1.0
0° h
6h
7
5h
m
4h
3h
m
0
1.
2h
20°S
0 .8 m
m
0.6
m
4 m
0.
2
6h
0.
7h
8h
1h
40°S 0.4 m
0.8 m
0h h
12
h
10 0.6 0.6 m
m
11
h
10 0.4 m
h 0.2
m
100°W 80°W 60°W 40°W 20°W 0° 20°E
Figure 3.30: Propagation of the M2 tide in the Atlantic Ocean. The solid lines are co-tidal lines
of simultaneous HW, the dashed lines are the co-range lines of equal tidal range. Model results
obtained using the FES2014 tide model. FES2014 was produced by Noveltis, Legos and CLS and
distributed by Aviso +, with support from CNES (https://www.aviso.altimetry.fr/).

140 3. Ocean waves
60˚N
9h
10
h
10
h
11 h
12 h
0h
3h
1h
57˚N
2h 2h
0.5
m
1m
3h
1h
2
m
4h 0h
3m
12 h
5h
11
h
10
m
h
2
54˚N
9h
8h
6h
7h
6h
1m
7h 5h
4m
4h
9h 3h
h 2
h
11
1h
0h
h
12
3m
2m
51˚N 4m
3m
4m
5m
1m
6m
11
h
0˚ 3˚E 6˚E
Figure 3.31: Propagation of the M2 tide in the North Sea with co-tidal lines radiating away from
the amphidromic points and co-range lines encircling them. The co-tidal lines show that the
phase increases counter-clockwise around the amphidromic point (typical of NH amphidromes).
The co-range lines show the tidal range increasing away from the node. Model results obtained
using the FES2014 tide model. FES2014 was produced by Noveltis, Legos and CLS and distrib-
uted by Aviso+, with support from CNES (https://www.aviso.altimetry.fr/).

Figure 3.32: Closed eastern boundary with direction of tidal wave propagation (NH).
bed friction leads to the following reduced shallow-water equations in two horizontal
dimensions:
𝜕𝑢 𝜕𝜂 𝜏𝑏,𝑥
− 𝑓 𝑣 = −𝑔 − (3.40a)
𝜕𝑡 𝜕𝑥 𝜌ℎ
𝜕𝑣 𝜕𝜂 𝜏𝑏,𝑦
+ 𝑓 𝑢 = −𝑔 − (3.40b)
𝜕𝑡 𝜕𝑦 𝜌ℎ
where:
𝑢, 𝑣 velocity in 𝑥-, 𝑦-direction m/s

𝑓 2𝜔𝑒 sin 𝜑 is the Coriolis parameter with the earth’s angular 1/s
velocity 𝜔𝑒 = 72.9 × 10−6 rad/s and the latitude 𝜑 positive in the
NH and negative in the SH
𝜂 oscillatory water level variation m
𝜏𝑏,𝑥 bottom shear stress in 𝑥-direction N/m2
𝜏𝑏,𝑦 bottom shear stress in 𝑦-direction N/m2
𝜌 water density kg/m3
ℎ water depth m
The Coriolis acceleration has been introduced in the above equations of motion in such
a way that it makes a right angle with the particle velocity and acts towards the right
(to starboard side) in the NH and towards the left (to port) in the SH (see Intermezzo 3.1).
Check for yourself the signs of the Coriolis terms in Eqs. 3.40a and 3.40b. In doing
so, note that the latitude 𝜑 and hence the Coriolis parameter 𝑓 are assumed to have
positive values in the NH and negative values in the SH.

142 3. Ocean waves
Continuity requires:
𝜕𝜂 𝜕𝑢 𝜕𝑣
+ ℎ( + ) = 0 (3.41)
𝜕𝑡 𝜕𝑥 𝜕𝑦
If we neglect friction and take into account that the velocity 𝑢 at the closed boundary
(the coastline) is zero, the balance equations reduce to:
𝜕𝜂
−𝑓 𝑣 = −𝑔 (3.42a)
𝜕𝑥
𝜕𝑣 𝜕𝜂
= −𝑔 (3.42b)
𝜕𝑡 𝜕𝑦
𝜕𝑣 𝜕𝜂
ℎ =− (3.42c)
𝜕𝑦 𝜕𝑡
Note that the cross-shore (𝑥) momentum balance (Eq. 3.42a) is geostrophic: there is a
balance between Coriolis force and the pressure gradient due to water level differences.
This is comparable to the influence of Coriolis on the flow in a confined channel (see
Intermezzo 3.6). The alongshore momentum balance, Eq. 3.42b, is the same as that
for shallow-water gravity waves (the progressive tidal waves from Intermezzo 3.5).
Equation 3.42b combined with the reduced continuity equation Eq. 3.42c yields a set
of equations that is equivalent to Eqs. 3.33 and 3.34 and thus has an equivalent solution.
Now substitute this solution 𝜂 = 𝜂̂ cos (𝜔𝑡 − 𝑘𝑦) and 𝑣 = 𝑔/𝑐 𝜂 with 𝑐 = ±√𝑔ℎ in the
geostrophic flow equation, Eq. 3.42a. The resulting equation now gives a solution
known as a Kelvin wave:
𝑓𝑥
( )
𝜂(𝑥, 𝑦, 𝑡) = 𝜂0 𝑒 𝑐 cos(𝜔𝑡 − 𝑘𝑦) (3.43a)
𝑐 (𝑓𝑥 )
𝑣(𝑥, 𝑦, 𝑡) = 𝜂0 𝑒 𝑐 cos(𝜔𝑡 − 𝑘𝑦) (3.43b)
ℎ
The Kelvin wave propagates along the coast (in the 𝑦-direction) at the shallow-water
speed. The alongshore velocity is in phase with the water level (as for the propagating
wave of Intermezzo 3.5). The amplitude is maximum at the coast (𝜂0 ) and then decays
with distance from the coast (in the negative 𝑥-direction). The scale of the decay is
𝑐/𝑓 in which the variables are the latitude and the water depth:
𝑐 1 √ℎ
= √𝑔ℎ −4
= 21453 (3.44)
𝑓 1.46 × 10 × sin 𝜑 sin 𝜑
At 45° latitude this amounts to about 1900 km for the deep oceans, with an average
water depth of 4000 m, and about 200 km for a shallow sea with a typical water depth
of 50 m.

The value of 𝑐/𝑓 cannot be negative, since this would make the sea level grow expo-
nentially offshore. This implies, as expected, that the Kelvin wave propagates in the
positive 𝑦-direction in the NH (where 𝑓 is positive) and in the negative 𝑦-direction in
the SH (where 𝑓 is negative).
Intermezzo 3.6 Geostrophic momentum balance
If the flow takes place in a confined conduit or channel that prevents a deviation
of the course (i.e. a steady current), the Coriolis acceleration causes a pressure
gradient across the conduit:
1 𝜕𝑝
= 2𝜔𝑒 𝑉 sin 𝜑 (3.45)
𝜌 𝜕𝑛
where:

𝑝 water pressure N/m2
𝑛 normal to and directed to starboard of the current 𝑉 −
In open channel flow, the pressure gradient becomes visible as a gradient of the
water surface:
1 𝜕𝑝 𝜕𝜂
=𝑔 (3.46)
𝜌 𝜕𝑛 𝜕𝑛
Note: upon comparing the above two equations (Eqs. 3.45 and 3.46) with the full
depth-integrated shallow-water equations, it can be seen that only the pressure
gradient due to the water level surface and the Coriolis acceleration are retained.
As an example, we compute the sea level difference across the Strait of Florida.
The Florida Current is located at latitude 26°N; the current velocity is about 1 m/s;
the width of the Strait of Florida is about 80 km.
1 𝜕𝑝
= 2 ⋅ 0.729 × 104 ⋅ sin 26° ⋅ 1 = 6.4 × 10−5 m/s2 (3.47)
𝜌 𝜕𝑛
The elevation difference over 80 km is computed as follows:
1 𝜕𝑝 Δ𝑥 6.4 × 10−5
Δ𝜂 = = ⋅ 80 × 103 = 0.52 m (3.48)
𝜌 𝜕𝑛 𝑔 9.81
The observed value is 0.45 m, which is close to our estimate. (Similar computations
can be made for e.g. the Western Scheldt, the British Channel or the Dutch Texel
Inlet).

144 3. Ocean waves
The Kelvin wave is a coastally trapped wave; it needs a coastline. In the NH, the wave
propagates poleward along an eastern boundary and equatorward along a western
boundary with its maximum amplitude at the boundary. It thus forms a wave trapped
to the boundary and rotating counter-clockwise around an amphidromic point (like a
standing wave, but now rotating). The rotation is clockwise in the SH.
In the North Sea multiple amphidromic points can be observed (see Fig. 3.31). The
Kelvin wave enters the North Sea basin from the north. Some of the energy is dissip-
ated in the basin and some is reflected from the shallow areas in the southern part of
the North Sea. This reflected wave forms its own amphidromic system.
For the ideal Kelvin wave, friction is not taken into account. Neglecting friction can be
a good approximation for deeper water. Near the coast inertia is relatively unimport-
ant, but bottom friction needs to be taken into account. The velocity is then governed
by the balance between the bottom friction and the alongshore water level gradient.
This will be treated in more detail in Ch. 5.
3.9. Tidal analysis and prediction

Because the tide is caused by regular astronomical phenomena, it can be predicted
accurately a long time ahead (although not including meteorological effects such as
storm surges). The method used for tide prediction is harmonic analysis. Analogous
to the treatment of wind waves, the water level at a certain location as a function of
time is expressed by the following formula:
𝑁
𝜂(𝑡) = 𝑎0 + ∑ 𝑎𝑛 cos (𝜔𝑛 𝑡 − 𝛼𝑛 ) (3.49)
𝑛=1
where:
𝜂𝑡 measured (or predicted) tidal level with reference to a fixed level m

𝑎0 mean level m
𝑎𝑛 amplitude of component number 𝑛 m
𝜔𝑛 angular velocity of component number 𝑛 1/h
𝑎𝑛 phase angle of component number 𝑛 −
𝑡 time h
𝑁 number of harmonic components −
Contrary to the traditional harmonic analysis, the frequencies 𝜔𝑛 are known here, hav-
ing been derived from astronomical considerations. The phase angles 𝛼𝑛 have to be
derived from observations as they are extremely site-specific. This applies to the amp-
litudes 𝑎𝑛 as well. Tidal analysis for a certain location therefore is the determination

3.9. Tidal analysis and prediction 145
of amplitudes and phases. Note that the phase angles are a function of the adopted
time origin.
The length of the analysed water level record determines the number of constituents
that can be determined. A year’s length can unravel the main constituents, except
for the constituent with the period of 18.6 yr. The effect of this can be introduced
by adjusting amplitudes and phases according to the position in that long-term cycle.
Instead of Eq. 3.49 we then get:
𝑁
𝜂(𝑡) = 𝑎0 + ∑ 𝑓𝑛 𝑎𝑛 cos (𝜔𝑛 𝑡 − 𝛼𝑛 + 𝛽𝑛 ) (3.50)
𝑛=1
Here 𝑓𝑛 is the so-called nodal factor that captures the effect of the 18.6 yr-cycle on the
tidal amplitudes. The correction to the phase is given by 𝛽𝑛 . The nodal modulation
can also be used to reduce the number of constituents in a tidal analysis; since the tidal
constituents are gathered in groups with similar frequencies (see Fig. 3.27), the effect
of smaller amplitude constituents in a group can be taken into account via corrections
to the amplitude and phase of the main, larger amplitude constituent in the group.
When the tidal constituents are known for a certain location, they can be used to
predict future tides. For many ports in the world, the tidal constituents are known and
publicly available. If for a project local tidal constituents are not known, two solutions
can be chosen. The first is to collect local data for a short period (for instance a month)
from which the most important constituents can be determined. Another possibility
is to use data from a nearby station and use a model for tidal propagation to determine
the amplitudes and phases for the project site.
An example of the result of a harmonic analysis for some ports along the Dutch coast
is presented in Table 3.6. This table shows the main harmonic components used for
prediction of the astronomic tide. Each component has an internationally agreed ab-
breviation. The most important constituents have already been discussed in Sect. 3.7.6.
Besides those principal constituents, each constituent may also have higher harmon-
ics, generated by non-linearities. Higher-order components carry a subscript 3, 4 or
higher. From the table one can see that the ratio of the effects of the sun and the moon
is approximately 1 to 4 along the Dutch coast (ratio S2/M2 ≈ 1/4).
Nota bene: in Table 3.6, the mean level is denoted by 𝐴0 and gives the mean difference
between Normaal Amsterdams Peil (NAP) – the fixed reference level for height in the
Netherlands – and MSL. MSL is the mean sea level as determined from measurements
and is the level without tides and averaged meteorological effects. Close to the coast
this difference can be neglected; but if one looks at a river farther upstream, the river
gradient influences the mean sea level. The difference between NAP and MSL changes
a little during the year, as can be seen from the small amplitude of component SA. The
angular velocity of this component (0.041) leads to a period of 365 d.

146 3. Ocean waves
Contrary to the Dutch tide tables, in other such tables 𝐴0 represents the difference
between a Chart Datum and MSL. Chart Datum is then defined as a low level that is
exceeded rarely, for instance Lowest Astronomical Tide (LAT) or Mean Lower Low
Water (MLLW). LAT is defined as the lowest tide level which can be predicted to occur
under average meteorological conditions and under any combination of astronomical
conditions. MLLW is the average height of the lower of the two daily low waters over
a long period of time. When only one low water occurs on a day, this is taken as the
lower low water.
Because of this site-specific definition of the Datum level, the Datum plane is not ne-
cessarily horizontal. Utmost care is required when performing hydraulic calculations
in this case (see also App. C). The Datum level used by different countries for the same
waterway can also be different, which leads to different depth figures for the same loc-
ation. This is the case for the Western Scheldt, where Dutch and Belgian charts show
such differences.
Tidal levels like MLLW are long-term averaged tidal levels based on measurements and
therefore include averaged meteorological influences on the water level. The period
of averaging is preferably 18.6 years, so that the tidal component with this duration is
properly taken into account. Commonly used long-term averaged tidal levels are for
instance Mean Low Water (MLW) and Mean High Water (MHW). Other tidal levels
are defined in App. C.
The tidal range can be defined using different tidal levels. The normal tidal range is
defined as MHW – MLW. But also the spring tidal range (Mean High Water Spring
(MHWS) – Mean Low Water Spring (MLWS)) or neap tidal range (Mean High Water
Neap (MHWN) – Mean Low Water Neap (MLWN)) can be used.
In morphodynamic modelling preferably a complete spring-neap tide cycle is taken
into account. This is, however, computationally demanding. To reduce computational
costs, a so-called morphological representative tide can be used. This is a single tidal
cycle that is expected to have a similar effect on the morphology as the total spring-
neap tidal cycle.

Table 3.6: Main constituents of the tide at several places in the Netherlands. Amplitude (cm) and phase lag in ° ref. to CET (UTC+1)
Component Angular Vlissingen 51°27′ N Euro Platform 52°0′ N H. of Holland 51°59′ N Rotterdam 51°55′ N IJmuiden 52°28′ N 4°35′ E Delfzijl 53°20′ N 6°56′ E
Velocity in 3°36′ E 3°17′ E 4°7′ E 4°30′ E
°/h
𝐴0 ref. to NAP −1 0 7 24 2 7
in cm
a [cm] 𝜑 [°] a [cm] 𝜑 [°] a [cm] 𝜑 [°] a [cm] 𝜑 [°] a [cm] 𝜑 [°] a [cm] 𝜑 [°]
SA 0.041 7 216 9 213 8 222 7 241 10 220 9 219

SM 1.016 4 33 3 31 3 32 3 43 3 22 4 32
Q1 13.399 3 133 4 126 3 131 3 148 4 133 3 179
O1 13.943 11 195 11 188 11 191 9 209 11 193 9 247
P1 14.959 3 353 3 340 3 346 2 11 3 346 3 48
K1 15.041 7 10 8 358 8 359 6 17 8 358 8 43
3MS2 26.952 3 281 2 288 2 312 2 344 2 338 4 167
MNS2 27.424 3 143 1 154 2 182 2 211 2 210 3 33
3.9. Tidal analysis and prediction
NLK2 27.886 4 354 2 1 2 26 2 58 2 54 4 245

𝜇2 27.968 13 161 6 174 8 200 8 232 9 227 15 55
N2 28.440 29 35 12 26 12 59 10 95 10 108 21 310
NU2 28.513 9 26 4 25 5 52 5 86 4 88 8 288
MPS2 28.943 3 110 1 107 1 170 1 206 2 205 5 27
M2 28.984 175 59 74 54 79 86 72 121 68 129 136 333
𝜆2 29.456 6 76 3 80 3 110 3 144 3 142 5 348
2MN2 29.528 13 257 6 261 7 290 7 325 7 323 12 168
S2 30.000 48 117 18 111 19 147 17 184 17 198 34 46
K2 30.082 14 117 5 111 6 147 5 184 5 198 10 43
2S M2 31.016 4 348 2 358 2 25 2 61 3 54 4 270
2M K3 42.927 3 162 1 141 1 191 1 225 1 263 1 120
MK3 44.025 2 316 1 281 1 288 1 349 0 279 1 278
3MS4 56.952 2 196 1 193 2 235 2 303 3 268 4 216
MN4 57.424 4 94 4 105 6 137 5 204 7 157 5 118
M4 57.968 13 120 10 130 17 165 15 230 20 186 17 145
MS4 58.984 9 181 7 185 11 222 9 291 12 246 10 224
MK4 59.066 2 178 2 184 3 221 3 290 4 244 3 222
2MN6 86.408 5 82 2 64 2 95 2 211 2 269 4 321
M6 86.952 9 109 4 92 5 128 4 243 4 290 7 352
2MS6 87.968 9 161 4 146 4 188 4 302 5 343 7 61
M8 115.936 3 115 1 142 2 230 1 358 3 330 1 217
3MS8 116.952 5 166 2 194 4 281 2 51 4 23 2 276

147
149
4
Global wave and tidal environments
4.1. Introduction
In this chapter we look into the global variation in the main processes that shape the
coast: wind, waves and tides. Based on large-scale observations such as the latitude
and the continent, a general idea of the wave, wind and tidal conditions at a project
site can be obtained. Questions that can be answered are for instance:
• What is the wind system we are dealing with at this latitude and what is the
dominant direction?
• Are locally generated waves important or are we mainly dealing with swell
waves?
• What wave heights can we expect?
• Does the wave climate exhibit seasonality?
• Can we expect a large tidal range in this part of the world?
• Is there a diurnal or a semi-diurnal tide?
This chapter starts with a treatment of the zonal wind systems (Sect. 4.2). Knowledge
of these global wind patterns is helpful in identifying the prevailing wind conditions
for a project site, as well as the wave climate. In Sect. 4.3, the global wave climate
is discussed and some generalisations are made about the coastal impact of different
wave conditions. Subsequently, global tidal environments and coastal characteristics
are discussed in Sect. 4.4. Here it is emphasised, moreover, that it is the relative effect of
waves and tides rather than the absolute tidal ranges and wave heights that determines
the coastal character.
Please bear in mind that wind, waves and tides vary not only globally, but also re-
gionally and locally. An example of variation due to regional geographic variation
is the sheltering of the southern part of the Florida coastline from waves due to the
presence of the Bahamas. On a local scale, the location of a land reclamation project

150 4. Global wave and tidal environments
may be chosen such that persistent swell cannot arrive at the site. These smaller-scale
variations are considered from Ch. 5 onwards.
4.2. Zonal wind systems and ocean circulation

4.2.1. Solar radiation and temperature distribution
Winds and ocean currents develop as a consequence of uneven distribution of heat over
the earth’s surface. This heat imbalance is largely explained by the fact that different
parts of the earth’s surface receive different amounts of solar radiation.
The main source of thermal energy for the earth is the electro-magnetic radiation emit-
ted by the sun. As the radiation passes through the earth’s atmosphere, it is depleted
by reflection from the top of clouds and absorption by clouds and atmospheric gases.
The incoming radiation that actually reaches the earth’s surface may be absorbed there,
be transmitted downwards if it encounters a material which is transparent to it, or be
reflected. The absorption of radiation leads to heating. The heat may be transmitted
downwards by conduction or, in the case of fluids, by convection.
The earth in turn emits electro-magnetic radiation into space. The low-frequent ter-
restrial radiation is readily absorbed by gases in the atmosphere such as water vapour,
carbon dioxide and ozone. These gases in turn emit long-wave radiation in all direc-
tions. As a result, they act as a layer of insulation around the earth analogous to the
glass of a greenhouse; their effect on earth temperatures has been called the green-
house effect. Human-induced emissions of greenhouse gases may increase the ability
of the atmosphere to absorb radiation and lead to a gradual warming of the earth and
atmosphere. This is referred to as global warming.
Ignoring any change in the earth’s mean annual temperature from one year to the
next and taking mean annual values, a balance must exist between incoming solar ra-
diation and outgoing terrestrial radiation. However, the amount of incoming radiation
is strongly dependent on the latitude. The total annual incoming radiation is greatest
at the equator and decreases towards the poles. Figure 4.1 shows the long-term aver-
aged incoming and outgoing radiation intensity as a function of the latitude. At high
latitudes the incoming radiation is less than the outgoing radiation: a net loss of heat
by radiation is found. Near the equator there is a net gain. The changeover from a sur-
plus to a deficit in the net annual radiation balance occurs at about 37° latitude N and S.
The uneven distribution of heat over the earth’s surface requires transfer (advection)
of heat. For that reason, both winds (Sect. 4.2.2) and ocean currents (Sect. 4.2.3) are
generated that are responsible for advective heat transport. About 60 % of the advect-
ive heat transport can be attributed to the movement of air and the remaining 40 % to
ocean currents.
The amount of incoming solar radiation is determined by factors such as the average
distance between the sun and the earth, the daily sunlight duration, the transparency

4.2. Zonal wind systems and ocean circulation 151
300
radiation intensity [Wm -2]
250
200
150 heat transfer heat transfer
100
50 outgoing deficit
incoming surplus
0
10
20
30
40
50
60
70
90
-90
-70
-60
-50
-40
-30
-20
-10
latitude (scaled proportional to area) [°]
Figure 4.1: Zonal mean incoming and outgoing top-of-the-atmosphere radiation from CERES
satellite radiation measurements in the period 2001--2014.
of the atmosphere and the angles at which the sun’s rays strike the earth. These factors
vary not only with latitude but have a seasonal component as well. Another factor gov-
erning the heat distribution is the fact that different surfaces absorb and store energy
at different rates. Land surfaces heat rapidly during the day and cool down fast during
the night, whereas an ocean responds more slowly to changes in incoming radiation.
This is because:
• in water, the solar radiation penetrates further than in land;
• water has a roughly four times greater specific heat capacity than land (four
times more energy is required to raise the temperature of water);
• water has a big storage possibility for heat by the process of mixing and evapor-
ation.
Figure 4.2 shows the effect of the seasons and the differential warming of the oceans
and the land on the air temperature distribution over the earth surface. The figure
shows the temperature distribution in January (Northern winter, Southern summer)
and July (Northern summer, Southern winter), which for most places on earth repres-
ent the extreme conditions. In winter the oceans remain warmer than the land and in
summer the land heats up more than the oceans. As a result, in January the isotherms
over the NH oceans bend towards the North Pole and the isotherms over the SH oceans
bend towards the equator. In July this situation is reversed. In winter the isotherms
are more closely spaced (larger thermal gradients) than in summer. Furthermore, iso-
therms are more closely spaced over land masses than over open oceans.

Figure 4.2: Air temperatures at ℎ = 2 m above the surface in January (upper plot) and July (lower
plot). Isotherms join places with similar temperature conditions. Generally, the isotherms follow
the latitudes, but isotherms over the oceans bend towards the equator in summer and towards
the poles in winter. Data from the ECMWF ERA5 model reanalysis dataset (ECMWF, n.d., and
https://apps.ecmwf.int/codes/grib/param-db/?id=167).

4.2.2. Atmospheric circulation and wind patterns

Near the equator, where the average solar radiation is greatest, the air is warmed at the
surface and rises. It attains a maximum vertical altitude of about 14 kilometres (top of
the troposphere1 ) and then begins flowing horizontally to the North and South Poles.
This creates the band of low air pressure known as the Intertropical Convergence Zone
(ITCZ). Because air moves from high surface pressure toward low surface pressure, the
ITCZ draws in surface air from higher latitudes, resulting in surface winds towards the
equator.
If the earth did not rotate, and if its surface would be entirely uniform with respect
to transparency to solar radiation, heat capacity and thermal conductivity, then we
might expect a simple convection cell circulation to exist within the troposphere in
each hemisphere (Fig. 4.3). Each cell would have a horizontal dimension of the order
of 104 km, with a vertical dimension of only some 10 km.
pole pole
equator
equator
Figure 4.3: Convection cell circulation on a non-rotating uniform earth. The left and middle pan-
els show the one cell per hemisphere circulation pattern when only solar radiation is taken into
account. The right panel shows the analogy with a fire heating the air and creating a circulation
cell in the room.
In reality, due to the rotation of the earth, the cells on the Northern and Southern
Hemispheres break up in three smaller cells each (Fig. 4.4).
The earth rotation results in the so-called Coriolis effect (see Intermezzo 3.1). It causes
currents and atmospheric flows to deviate to the right (starboard side) on the NH and to
the left (port side) on the SH. Due to the Coriolis effect, the air in the upper atmosphere
is deflected when moving away from the equator.
At about 30° of latitude (north and south) the air begins to flow from west to east (this
is the subtropical jet stream), causing an accumulation of air in the upper atmosphere.
To compensate for this accumulation, some of the air in the upper atmosphere sinks
back to the surface, creating the subtropical high pressure zone.
From this zone, the surface air travels in two directions:
1
Troposphere is the upper part of the atmosphere, where the temperature decreases with increasing
altitude.

60º
30º
0º
Figure 4.4: Convection cell circulation on a rotating uniform earth. Due to the rotation of the
earth, the cells on the Northern and Southern Hemispheres break up in three smaller cells each.
1. Back towards the equator, creating the trade winds or tropical easterlies (Fig. 4.5).
These trade winds are deflected by the Coriolis effect, resulting in the northeast
trades (NH, right deflection) and the southeast trades (SH, left deflection);
2. Towards the poles, producing the westerlies that are also deflected by Coriolis.
POLAR HIGH
Polar Easterlies
60°N
LOW
Westerlies 30°N
SUBTROPICAL HIGH
N.E. Trades
LOW 0°
S.E. Trades
SUBTROPICAL HIGH
30°S
Westerlies
LOW
60°S
Polar Easterlies
POLAR HIGH
Figure 4.5: Schematic presentation of pressure belts and prevailing wind systems at the earth’s
surface. The earth is encircled by several broad prevailing wind belts, which are separated by
narrower regions of either subsidence (highs: poles and about 30°N and S) or ascents (lows:
ITCZ and about 60° N and S). The direction and location of these wind belts are determined by
solar radiation and the rotation of the earth.
Roughly at 60° north and south latitude, the subtropical westerlies collide with cold
air travelling from the poles (polar easterlies). This collision results in the uplift of air
and the creation of subpolar low pressure zones (and associated mid-latitude cyclones).

After it reaches the top of the troposphere, part of this lifted air is directed towards the
polar highs and part is directed towards lower latitudes.
In Fig. 4.5 we can thus clearly distinguish regions with mainly westerly winds at latit-
udes between 30° and 60°, which we know extremely well in the Netherlands. These
are strong and variable winds. Also the regions with mainly NE and SE trade winds
between the equator and 30° are clearly visible. Trade winds are moderate but per-
sistent throughout the year. They mainly occur over the oceans, since near the con-
tinents they are generally overruled by tropical and subtropical seasonal winds called
monsoons (see below). The polar easterlies are moderate as well and blow over land
(Antarctica) or ice (Arctic area) for the larger part of the year. The area near the ITCZ
where the wind climate is predominantly calm is called doldrums. However, the trop-
ics can experience tropical storms which develop over sea and ocean areas with high
surface temperatures. Depending on the location, these tropical storms are called hur-
ricanes (near the Americas), cyclones (near India and Africa) or typhoons (near SE-
Asia and Australia). The storms follow a path which is only partly predictable, and
stop only after crossing into a continent. Another cyclonic source are the east coast
cyclones that form off the east coasts of USA, Australia, Brazil and Africa between 25°
and 35° N and S.
When the non-uniformity of the earth’s surface is introduced, the situation becomes
considerably more complex. Due to the presence of land masses, the large-scale pres-
sure belts are broken up into several areas of low and high pressure (see Fig. 4.6). Both
the topography of a certain area (mountains affect the pressure distribution) and the
differential warming of the oceans and the land play a role. The resulting typical wind
patterns for the months of January and July are also shown in Fig. 4.6. We can recog-
nise westerlies and trade winds as well as seasonally reversing winds called monsoons.
The westerlies are the strongest winds, especially around 50° to 60° N and S. Their
seasonality is largest in the Northern Hemisphere where the differential warming of
oceans and continents strongly influences the location of Highs and Lows (due to the
larger presence of land masses in the NH). Especially the Asian land mass causes sig-
nificant deviations from the large-scale pressure and wind belts (compare Fig. 4.5 and
in Fig. 4.6).
It can , moreover, be clearly observed that some tropical areas are dominated by trade
winds (blowing in the same direction throughout the year), whereas other areas in
the tropics and subtropics are dominated by seasonally reversing monsoons. These
moderate but persistent winds are the result of the larger amplitude of the seasonal
cycle of land temperature compared to that of nearby oceans. This is evident in for
instance SE Asia, where the Asian continent warms up in July, thus creating a Low
above China, causing a SW wind blowing from the sea to the land. In January, when
the water of the Indian Ocean maintains a higher temperature than the continent, the
situation is reversed, causing a NE wind. The SW summer monsoon (blowing from
the sea to the land) is warm and humid and the NE winter monsoon (blowing from the

Figure 4.6: The air pressure at mean sea level and the global wind patterns, in January (top panel)
and July (bottom panel), indicating the main wind systems. Pressure data from the ECMWF
ERA5 model reanalysis dataset (ECMWF, n.d., and https://apps.ecmwf.int/codes/grib/param-
db/?id=151).

land towards the sea) is relatively cold and dry2 . The above holds for large continental
land masses (like Asia), but also N and S America and Africa have a monsoon. For
minor land masses onshore winds may be experienced in winter.
Summarising, the zonal wind systems are determined by large-scale pressure belts as
a result of latitude-dependent heating of the earth. We can distinguish between:
• Polar easterlies at high latitudes (>70°);
• Strong westerlies at mid-latitudes (30°–70°);
• Extensive, but moderate trade winds in the subtropics (10°–30°);
• Quieter doldrums around the equator (10°N–10°S).
These winds blow in the same direction throughout the year, but vary spatially and
temporally with the seasons. The trade winds and in particular the westerlies are the
most important in supplying energy to the coastal system. Although the trade winds
are not as strong as the westerlies, they blow over large areas throughout the year.
Regional and local effects are:
• Seasonally reversing monsoons due to changes in heating of continents and
oceans;
• Cyclones (tropical and east coast cyclones);
• Land and sea breezes that arise from differences in temperature of land and sea
4.2.3. Oceanic circulation

Besides atmospheric circulation, ocean water circulation also contributes to the con-
tinuous re-distribution of excess heat from the equatorial zone. This thermohaline
ocean circulation is density-driven and redistributes not only heat, but salt and dis-
solved gases as well. It is sometimes called the great or ocean conveyor belt. In a
simplistic view of the great conveyor belt (see Fig. 4.7), warm, salty surface water is
chilled in the North Atlantic and eventually sinks to flow south towards Antarctica.
There, it is cooled further to flow outward at the bottom of the oceans into the At-
lantic, Indian, and Pacific basins. After upwelling primarily in the Pacific and Indian
Oceans, the water returns as surface flow to the North Atlantic, again supplying heat
to the polar zones. The surface flow is primarily wind-driven and is confined to a layer
of typically 50 m to 100 m of well-mixed water.
2
Similarly, differences between day and night land-sea temperatures can locally generate an onshore
breeze during the day (sea breeze) and an offshore breeze (land breeze) at night.

Figure 4.7: The global current patterns (‘great conveyor belt’) consisting of a wind-driven surface
flow (red) and the density-driven deep ocean currents (blue). Indicated are locations of upwelling
(U), downwelling (D) and mixing (M).
4.3. Large-scale variation in wave environments

4.3.1. Wave height variation
Ocean waves are generated by wind. It can therefore be expected that the global wind
systems (Sect. 4.2.2) determine the global wave environments. Global wave environ-
ments are those zones of the seas and the oceans that have similar general wave char-
acteristics, such as similar year-averaged significant wave heights and similar season-
ality. Besides, the shape and orientation of the oceans determine the fetch and hence
influence the propagation of waves.
Figure 4.8 shows annual mean values for the significant wave height, as well as monthly
means for January and July. For this analysis, mean significant wave heights were used
obtained by visual observation of individual waves from ships for the period of 1958
to 1997. The significant wave heights vary mostly between 1 m and 5 m.
From Fig. 4.5, Fig. 4.6 and Fig. 4.8, the following general conclusions can be drawn:
• Wave heights are highest at mid-latitudes (north and south). This is the result
of the westerlies, which are the strongest winds. These winds (and embedded
mid-latitude cyclones) are the source of relatively large waves;
• The mid-latitude wave climate in the North Pacific and North Atlantic is espe-
cially seasonal, with much larger wave heights in the Northern winter than in
the summer. This is the result of the strong seasonality in the NH westerlies,

4.3. Large-scale variation in wave environments 159
Figure 4.8: World-wide significant wave height of combined wind waves and swell (in m) based
on model reanalysis data for the period 1979–2019 (ECMWF, n.d.).Top: annual mean values.
Middle: monthly mean values for January. Bottom: monthly mean values for July.

due to the presence of land masses, especially the large Asian land mass (see
Sect. 4.2.2);
• The Southern Ocean is characterised by an almost unlimited fetch and vast re-
gions with high waves. Although the waves are highest in the Southern winter,
the seasonality is much smaller than for the Northern Hemisphere (as a result
of the smaller seasonality of the westerlies in the absence of vast land masses);
• Wave heights in the subtropics associated with the gentle trade winds are mod-
erate;
• In the tropics and subtropics, sources of larger waves are either swell propagat-
ing from higher latitudes (originating from the westerlies), or seasonal winds
(e.g. monsoons and tropical storms);
• Monsoons have a regional impact. It appears that the highest wave heights in
the Arabian Sea coincide with the SW monsoon (summer). This is because the
SW monsoon blows from sea to land. The NE monsoon blows from the land,
except for the Malaysian Peninsula, which is exposed during the NE monsoon;
• Tropical and east coast cyclones can generate large waves but are limited in
extent and too seasonal to greatly impact longer-term wave climates.
The wave climate (in terms of wave height) is generally characterised by the mean
significant wave height 𝐻𝑠 on a yearly average basis:
• Low wave energy 𝐻𝑠 < 0.6 m;
• Medium wave energy 0.6 m < 𝐻𝑠 < 1.5 m;
• High wave energy 𝐻𝑠 > 1.5 m.
4.3.2. Wave environments

J. L. Davies and Clayton (1980) identified four major deep-water wave environments
(Fig. 4.9):
• Storm wave environments;
• West coast swell environments;
• East coast swell environments;
• Protected sea environments.
Besides, they identified trade and monsoon influences and tropical cyclone influences.
The global wave environments are strongly linked to the zonal wind systems (Sect. 4.2)
and have the following characteristics (see also Short, 2005):
Storm wave climate

• The most energetic wave environment;
• Locally generated by westerlies and associated mid-latitude cyclones;
• Located between 40° and 60° N and S;

4.3. Large-scale variation in wave environments 161
18
80° 0° 0°
12
12
60°
0°
°
60
0°
60°
40°
20°
0°
20°
40°
60°
80°
west coast swell environment protected sea environment
east coast swell environment tropical cyclone influences
storm wave environment trade and monsoon influences
Figure 4.9: World-wide distribution of wave environments. Classification according to J. L. Dav-

ies and Clayton (1980).
• Operates year-round in the Southern Hemisphere and in winter in the Northern

Hemisphere;
• Generally a combination of sea and swell is present at a certain location;
• Waves are steep, short-crested, irregular and multi-directional (sea);
• Direction is predominantly westerly (to southwesterly in the Northern Hemi-
sphere) impacting west-facing and south-facing (NH) coasts;
• Deep water significant wave heights are 5 m to 6 m 10 % of the time. The South-
ern Ocean has the most persistent higher waves with heights of 2 m to 3 m 90 %
of the time;
• Wave periods are about 5 s, longer during storms
West coast swell climate

• Stems from storm waves generated by westerlies in the Northern and Southern
storm wave belt (highest and most persistent is the westerly swell generated
at 55°S). Note that swell in the Northern Hemisphere can be generated in the
Southern storm wave belt and vice versa;
• Located between 0° to 40° N and S;
• Year-round in the Southern Hemisphere, seasonal (in winter) in the Northern
Hemisphere (like the storm waves from which they originate);
• West coast swell reaches west coast of Americas, Africa, Australia and New Zea-
land;

• Swell tends to arrive mostly from the northwest when generated in the North-
ern storm wave belt and from southwest when generated in the Southern Hemi-
sphere;
• Consists of persistent and long-period waves (typical period of 10 s);
• Waves are uniform in direction, shape and size and wave heights are moderate
to high (typical wave heights are for instance 1 m to 2 m off swell-dominated
coasts);
• There is not much variation in wave heights around the mean (only as a result of
tropical storms occurring for instance once a month in Queensland, Australia);
• In the tropics (for instance Angola, see Fig. 3.12) swell can also stem from trade
winds.
East coast swell climate

The east coast swell climate has many of the characteristics of west coast swell. East
coast swell originates from storms in the same storm wave belts as the west coast swell,
but is directed such that it reaches east-facing coasts. They are generally lower than
west coast swell and arrive less frequently than west coast swell.
Protected wave environments

Protected wave environments are areas protected from the arrival of swell and with
irregular, low-amplitude waves from local winds. These are areas shielded by ice (in
the polar zones), reefs (in the tropics), island archipelagos or land masses. An example
of the latter are enclosed or semi-enclosed seas such as the Mediterranean.
4.3.3. Coastal impact of different wave conditions

The world-wide distribution of wave characteristics has strong implications for coastal
engineering. The wave climate at a particular site is dependent on regional and local
factors such as basin characteristics and local geometry. It is therefore almost im-
possible to classify coasts on the basis of the global wave climates. Nevertheless, the
following broad generalisations can be made for a wave-dominated coastal system (see
also Mangor, 2004):
• On open coasts, a storm wave climate is characterised by waves which are highly
variable in height, period and direction. The waves are continuously reshaping
the coastal profile, which results in a dynamic, sandy coastal profile with bars
and wide sandy beaches backed by dunes. The profile often has an offshore
storm bar, which is formed by offshore transport by the larger waves associated
with storms, which in the NH mostly occur in winter. These larger waves break
at relatively deep water, so the littoral zone (the active coastal zone, Fig. 1.14) ex-
tends to relatively large water depths. Since the slope of the profiles tends to be

4.4. Large-scale variation in tidal characteristics 163
flatter for steeper waves, the result is a wide littoral zone. Since the wave influ-
ence decreases offshore, there is an offshore fining of sediments. The breaking
waves tend to be of the spilling type (see Sect. 5.2.5);
• Since swell waves are relatively low and long (low steepness) and have a more
or less constant wave height year-round, a swell climate gives a relatively nar-
row sandy littoral zone. The breaking waves tend to be of the plunging type
rather than the spilling type (see Sect. 5.2.5). The transition from coarser sandy
sediment in shallower waters to finer sediments in deeper waters is quite abrupt.
The gently sloping outer part of the littoral zone is dominated by finer sediments.
The low and long waves tend to move sand onshore;
• The monsoon climate of Southeast Asia gives a seasonal wave climate with the
highest waves in summer under the influence of the SW monsoon. The summer
waves are moderate in height and relatively constant in direction and height.
The corresponding profiles therefore are similar to the swell climate profiles: a
fairly narrow sandy inner littoral zone, shifting to a gently sloping outer part of
the littoral zone dominated by finer sediments;
• Cyclones give rise to very high waves and storm surge. When they hit, they
greatly impact the coastal profile, causing erosion and storm bars. But because
of their relatively low frequency of occurrence (for instance one or two per year
that make landfall at a particular location), the coastal morphology will first and
foremost be determined by the normal wave climate (either a monsoon or swell
climate).
4.4. Large-scale variation in tidal characteristics

4.4.1. Global tidal environments
Not only wave characteristics, but tidal characteristics also vary globally. The two
main variables on the basis of which tidal environments can be classified are:
• Magnitude of the tide, which can be characterized by the tidal range, i.e. the
vertical distance covered by the tide;
• Tidal character, which can be determined by the importance of diurnal versus
semi-diurnal components.
As we have seen in Sect. 3.8 the tidal wave is distorted by local differences in water
depth (and thus influenced by the slope and width of the continental shelf) and by the
location and shape of land masses and large embayments. This results in a global vari-
ation in tidal range controlled by the large-scale coastal configuration and indicated
in Fig. 4.10. The categories that form the basis of this figure are:
• Micro-tidal regime: mean spring tidal range < 2 m;
• Meso-tidal regime: mean spring tidal range 2 m to 4 m;
• Macro-tidal regime: mean spring tidal range > 4 m.

18
80° 0° 0°
12
12
60°
0°
°
60
0°
60°
40°
20°
0°
20°
40°
60°
80°
mean spring tidal range [m]
0 2 4
Figure 4.10: World distribution of mean spring tidal range (MHWS − MLWS). Tidal classification
according to J. L. Davies and Clayton (1980).
Figure 4.10 shows that semi-enclosed seas possibly enhance tidal amplification (see
Sect. 3.8.2) and therefore often experience a macro-tidal range. For open coasts and
fully enclosed seas a micro-tidal regime can generally be observed.
The tidal character is defined by the form factor 𝐹 . The form factor is determined as
the ratio of the amplitudes of the sum of the two main diurnal components K1 and O1
and the sum of the two main semi-diurnal components M2 and S2:
𝐹 = (K1 + O1 ) / (M2 + S2 ) (4.1)
with the symbols of the constituents in this case indicating their respective amplitudes.
Based on the form factor, four categories are distinguished (see Table 4.1). Examples
of tidal curves per category are given in Fig. 4.11.
Table 4.1: The tidal character expressed by the form factor 𝐹
Category Value of 𝐹
Semi-diurnal 0 – 0.25
Mixed, mainly semi-diurnal 0.25 – 1.5
Mixed, mainly diurnal 1.5 – 3
Diurnal > 3

Playas Negras, El Salvador

1.5
1
elevation [m] 0.5
0
-0.5
-1
F = 0.14
-1.5
Cabo San Lucas, Mexico
1
0.5
elevation [m]
-0.5
F = 0.62
-1
Phan Rang, Vietnam
1
0.5
elevation [m]
-0.5
F = 2.60
-1
Pattaya, Thailand
1
0.5
elevation [m]
-0.5
F = 8.27
-1
01/01/20 11/01/20 21/01/20 31/01/20
date
Figure 4.11: Examples of tidal curves illustrating all four different tidal characters. The time
series, at offshore locations indicated by the nearest large city, were generated by accessing
the TPXO7.2 model created by Oregon State University (OSU) (Egbert & Erofeeva, 2002) using
the Tide Model Driver (TMD) toolbox provided by Earth and Space Research (ESR) (Erofeeva et
al., 2020).
Global variations in the form factor arise due to a combination of geography and lat-
itude. As explained in Sect. 3.7.5, diurnal components are introduced due to the de-
clination of the earth axis. The combination of diurnal and semi-diurnal components
manifests itself as daily inequality (one high water is higher than the other). The daily
inequality increases with latitude and is further influenced by the presence of land
masses which can locally magnify the larger tide. The latter may occur when the di-
urnal tidal component excites one of the resonance modes of the basin or bay. Due to

this resonance phenomenon, many areas around the equator (e.g. Vietnam) experience
a mainly diurnal tidal regime, in spite of their low latitude.
The world’s distribution of the tidal character is shown in Fig. 4.12. The figure shows
that most of the world’s coastlines experience either semi-diurnal or mainly semi-
diurnal mixed tides. Nevertheless the extent of the areas with diurnal and mainly
diurnal mixed tides is still significant. When comparing Fig. 4.10 and Fig. 4.12, it can
be seen that many of the areas where diurnal tides dominate have a micro-tidal regime
and none a macro-tidal regime. Apparently, areas with diurnal and mainly diurnal
mixed tides tend to have smaller tidal ranges than semi-diurnal systems.
18
80° 0° 0°
12
12
60°
0°
°
60
0°
60°
40°
20°
0°
20°
40°
60°
80°
diurnal semi-diurnal
mixed, mainly diurnal mixed, mainly semi-diurnal
Figure 4.12: Tidal environments of the world. Note that the transitions between the tidal types
are progressive and not abrupt. The attribution of the tidal type follows J. L. Davies and Clayton
(1980).
4.4.2. Coastal impact of tide and classification

The semi-diurnal or diurnal rise and fall of the water level creates a so-called intertidal
zone that is exposed during low water and submerged during high water. In the ab-
sence of waves (or with only very low wave energy), tide-dominated coasts develop
wide, low-gradient tidal flats (see for example Fig. 2.11) in the intertidal zone and sub-
tidal zone (the area only infrequently exposed, during extreme low tides). Tidal cur-
rents in combination with horizontal translation of the water line determine the mor-
phology of these tidal flats. Due to the low-energy conditions, tide-dominated coasts
generally consist of relatively fine sediments. The sediment distribution patterns are
exactly opposite to those on wave-dominated coasts: since tidal currents increase in

strength for larger water depths, the finest sediments occur on the often muddy flats
and in the wetlands of the upper intertidal zone. The coarser sandy sediments occur in
the lower intertidal zone and further seawards. From the upper part of the intertidal
zone to the supratidal zone (only submerged during spring tides or storm surges) salt
marshes are well developed. In tropical to subtropical regions, mangroves occupy the
intertidal zone.
Tide-dominated flats occur for large tidal ranges and small wave heights. This is the
case for macro-tidal coasts, but tide-dominated coasts can also be found for micro-tidal
regimes, as long as the wave energy is very low (for instance in an estuary). It is rather
the relative importance of tide and waves than the absolute tidal range that determines
the coastal character. Hayes (1979) and Davis Jr. and Hayes (1984) distinguish five
classes, based on a combination of tidal range and wave energy classification (Fig. 4.13).
low medium high wave energy

6
macrotidal
gh
5 hi
(
d
te
ina
w)
(lo
om
d )
4
te ted
e-d
mean tidal range [m]
ina
tid
om
a
in
e-d
om
(tid mesotidal
e-d
3 gy
er
tid
en
d ted)
ixe -domina
m ve
(wa
2
rgy
d ene
xe
mi
microtidal
1
-dominated
wave
0
0 0.5 1 1.5 2 2.5
mean wave height [m]
Figure 4.13: Relationship between mean tidal range and wave height according to Davis Jr. and
Hayes (1984) and Hayes (1979), delineating different fields of wave and tide dominance. Note
that the convergence of the fields for low wave and tidal energy means that very small differ-
ences in tide or waves may result on a different dominance and corresponding morphology.
A useful parameter to distinguish between wave and tide influence is the relative tidal
range 𝑅𝑇 𝑅 as introduced by Masselink and Short (1993):
𝑅𝑇 𝑅 = 𝑀𝑆𝑇 𝑅/𝐻𝑏 (4.2)
where 𝑀𝑆𝑇 𝑅 is the mean spring tidal range and 𝐻𝑏 is the wave height just before
breaking. For 𝑅𝑇 𝑅 < 3 we find the wave-dominated beaches as described in Sect. 4.4.2.
For 𝑅𝑇 𝑅 > 15 the beaches gradually approach the pure tidal flat situation. For the

intermediate range, we find beaches shaped by waves with some distinct tidal charac-
teristics, such as a wide intertidal zone. The effect of the tide is that the zone of wave
attack shifts with the tidal phase. This means that there is not enough time for wave-
dominated bar morphology to develop. In other words: tides flatten out the beach mor-
phology. Tides develop wide, low-amplitude parallel or sub-parallel bars (tidal ridges)
on the intertidal beach, especially for macro-tidal regimes. During storms, waves flat-
ten these ridges out. Tide-dominated coastal features are treated more extensively in
Ch. 9.

169
5
Coastal hydrodynamics
5.1. Introduction
This chapter deals with the nearshore hydrodynamics that are important for sediment
transport. It treats mean and oscillatory water levels and currents induced by waves,
wind and tides. Quite a lot of attention is paid to waves and wave-induced currents
because of their effectiveness in transporting sediment in the surf zone.
The following aspects of waves are described:
• Linear wave propagation effects in shoaling waves (until wave-breaking): in-
creasing wave heights, decreasing wavelengths and refraction towards normal
incidence (Sect. 5.2);
• Non-linear transformation of wave shapes from initially symmetric, sinusoidal
profiles, to the asymmetric, pitched-forward profiles typical for near-breaking
waves (Sect. 5.3);
• Wave dissipation in the wave boundary layer and its effect on wave-orbital ve-
locities, bed shear stress and net wave-induced flow (called Longuet-Higgins
streaming) close to the bed (Sect. 5.4);
• Wave-induced water level changes in breaking waves such as set-up (raising
of the water level) at the coast, and wave-induced flow in breaking waves: a
circulation current in the cross-shore direction (with, in the lower part of the
water column, an offshore-directed undertow) as well as a longshore current
along the coast (Sect. 5.5).
Subsequently Sect. 5.6 describes wind-generated set-up and currents. Section 5.7 is
dedicated to tidal propagation in coastal waters. Last, Sect. 5.8 discusses some other
long-wave phenomena in coastal waters, viz. seiches and surf beat.

170 5. Coastal hydrodynamics
5.2. Wave transformation

5.2.1. Energy balance
When waves propagate from deep into intermediate and shallow water, the waves
transform, i.e. wave height, length and direction change until the waves finally break
and lose their energy. Wave transformation takes place because the waves are affected
by the seabed through processes such as refraction, shoaling, bottom friction and wave-
breaking.
When the water depth becomes less than about half the wavelength (see Sect. 3.5.2 for
deep- and shallow-water criteria), the waves start to be affected by the bottom and slow
down. A certain harmonic component retains its frequency, but the propagation speed
𝑐 decreases through Eq. 3.23 and the wavelength 𝐿 decreases correspondingly. The
effect on the wave height can be imagined as follows: as the first wave in a wave train
is slowed down due to decreasing water depth, the following wave is still in slightly
deeper water and is thus moving at a higher speed. This wave tends to ‘catch up’ with
the wave in front of it, which is being slowed. This results in a concentration of wave
energy and an increase in wave height. This process is called shoaling (Sect. 5.2.2).
Changes in water depth and thus propagation speed can also occur along a wave crest.
This forces an obliquely incident wave (at an angle with the coastline) to bend toward
normal incidence (refraction, Sect. 5.2.3). Diffraction is wave transformation due to
sheltering by obstructions like islands or breakwaters.
Various software packages are available to translate offshore wave conditions to wave
conditions in the nearshore. The effects of shoaling, refraction, bottom friction and
wind can be incorporated in these models. Examples are HISWA and SWAN, both
developed at Delft University of Technology. The simpler packages are based on for
instance a spectrally integrated energy balance. Numerically (or analytically) solving
the energy balance yields information on the wave transformation (i.e. the changes in
𝐻 , 𝐿, 𝑐 and wave direction 𝜃) of a wave field, while the waves approach the shore. In
the presence of a current, energy is not conserved any longer, since transfer of energy
between waves and currents is possible. In that case another wave quantity, wave
action 𝐸/𝜔, will be conserved and the wave action balance rather than the energy
balance should be solved. In the absence of a current, the wave action balance reduces
to the energy balance. Since in this book the focus is on conceptual understanding, we
will only consider the energy balance and treat a few simplified situations in which
wave transformation is separated into the processes of shoaling, refraction, diffraction
and breaking.

5.2. Wave transformation 171
Integrating over all frequencies and directions in an irregular wave field, the following
energy conservation equation can be composed:
𝜕𝐸 𝜕 𝜕
𝜕𝑡
+ 𝜕𝑥
(𝐸𝑐𝑔 cos 𝜃) + 𝜕𝑦
(𝐸𝑐𝑔 sin 𝜃) = 𝑆−𝐷 (5.1)
⏟ ⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟ ⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟ ⏟
change of energy import of energy in 𝑥-direction and in 𝑦-direction gain of energy
In this equation, 𝜃 is the wave direction with respect to the 𝑥-axis (see Fig. 5.1), 𝑆 is the
generation term and 𝐷 is the dissipation term. Underlying the spectral integration are
the assumptions that: 1) the irregular wave field at one location can be represented by
a single value for 𝜃; and 2) that the total energy 𝐸 = 1/8𝜌𝑔𝐻𝑟𝑚𝑠 2 is propagated at the
wave group speed 𝑐𝑔 (see Eq. 3.26). This only holds for a narrow-banded spectrum,
and therefore not for a wide spectrum or a spectrum of combined swell and storm.
For wave action conservation to reduce to energy conservation, we must assume a
spatially constant peak period. rs
ntou
o
th c
dep
n s
ropagatio
wave p
φ θ
wave ray
t
fron
ave
w
y wave front
ray
ve
wa
coastline
x
Figure 5.1: Definition of wave angles for a wave propagating along a wave ray 𝑠. The wave
direction 𝜃 is the angle with the 𝑥-axis. It generally differs from the (local) angle of incidence 𝜑
with respect to the depth contours. For a uniform coastline with the 𝑥-axis perpendicular to the
coast we have 𝜃 = 𝜑.
The deep-water group velocity is independent of location and can therefore be pulled
out of the derivative. In intermediate and shallow water, the group velocity is de-
pendent on the location (the water depth) and can therefore not be pulled out of the
derivative. 𝑆 can be additional input of energy due to wind, which we generally neg-
lect for the relatively small nearshore zone. Dissipation of wave energy 𝐷 results in a
decrease of wave height while waves approach the shore. Various processes can con-
tribute to the dissipation term on the right-hand side. The most efficient wave energy
dissipation mechanism is wave breaking. This occurs mainly in the surf zone, but also
in deeper water (where it is called white-capping). Other mechanisms include bottom

friction – especially over large areas with shallow water – and interaction with vegeta-
tion (mangroves, salt marshes). 𝐷 (and 𝑆) are not known well and all sorts of empirical
formulas exist which allow for the equation to be solved.
If we assume that the wave conditions are stationary (do not change in time), the term
𝜕𝐸/𝜕𝑡 on the left-hand side equals zero and the energy balance in the coastal zone can
be written as:
𝜕 𝜕
(𝐸𝑐𝑔 cos 𝜃) + (𝐸𝑐𝑔 sin 𝜃) = −𝐷𝑓 − 𝐷𝑤 (5.2)
𝜕𝑥 𝜕𝑦
2 being propagated at the wave group speed

with the total wave energy 𝐸 = 1/8𝜌𝑔𝐻𝑟𝑚𝑠
𝑐𝑔 in the wave propagation direction 𝜃. Wave dissipation due to wave-breaking is
denoted 𝐷𝑤 , and wave dissipation due to bottom friction is denoted 𝐷𝑓 . The direction
𝜃 can change and we therefore need information on 𝜃 to find a solution. For simple
cases (alongshore uniform coast), Snell’s law gives this information (see Sect. 5.2.3).
Eq. 5.2 can also be written along a wave ray 𝑠:
𝑑
(𝐸𝑐𝑔 ) = −𝐷𝑓 − 𝐷𝑤 (5.3)
𝑑𝑠
Be aware that the wave ray is not a straight line due to 𝜃 variations (see Fig. 5.1).
In order to close the equation, expressions for 𝐷𝑓 and 𝐷𝑤 need to be formulated. 𝐷𝑓
is a function of the shear stress due to (mainly) wave orbital motion near the bed (see
Sect. 5.4.3) and is relatively small. Battjes and Janssen (1978) express 𝐷𝑤 based on an
analogy with a bore model from the observation that after breaking, waves behave like
a bore or a moving hydraulic jump.
5.2.2. Shoaling
Consider a linear (single-harmonic) long-crested wave propagating in water that be-
comes gradually shallower (an alongshore uniform, sandy coast with parallel depth
contours). The wave is normally incident, viz. the wave crest is parallel to the depth
contours, Fig. 5.2.
The wave propagation speed will be affected by the bottom when the water depth
becomes less than about half the wavelength. A decreasing water depth yields a de-
creasing wave speed and wavelength, according to the dispersion relation as intro-
duced in Sect. 3.5.2. Figure 5.3 shows 𝑐/𝑐0 and 𝐿/𝐿0 – where the subscript 0 refers
to deep-water conditions – as a function of local water depth ℎ divided by 𝐿0 . To
find a relation between the wave height 𝐻 and the water depth ℎ we have to examine

wave front
wave ray
depth contours
Figure 5.2: Normally incident waves with parallel depth contours (𝜑 = 𝜃 = constant = 0).
the energy balance. Outside the breaker zone the dissipation is approximately zero
(neglecting bottom friction and white-capping) and integration of Eq. 5.3 yields:
𝑈 = 𝐸𝑐𝑔 = 𝐸𝑛𝑐 = constant (5.4)
where:
𝑈 wave power or energy flux per unit wave crest width (see also J/(m s)
Eq. 3.26)
𝐸 wave energy per unit surface area J/m2
𝑐𝑔 wave group velocity m/s
𝑐 wave celerity m/s
𝑛 ratio 𝑐𝑔 to 𝑐 −
Eq. 3.25 and Fig. 5.3 show the dependency of 𝑛 on the water depth and (deep water)
wavelength. The energy flux 𝑈 is also called the wave power and is the rate at which
energy is transmitted in the direction of wave propagation across a vertical plane per-
pendicular to the direction of wave propagation and extending over the entire depth.
Since 𝐸 = 1/8𝜌𝑔𝐻 2 , Eq. 5.4 can be used to relate the wave heights at two arbitrary
locations (locations 1 and 2):
𝑈2 = 𝑈1 → 𝐸2 𝑛2 𝑐2 = 𝐸1 𝑛1 𝑐1 → 𝐻22 𝑛2 𝑐2 = 𝐻12 𝑛1 𝑐1 (5.5)
or:
𝐻2 𝑐 𝑛
= 1 1 (5.6)
𝐻 1 √ 𝑐 2 𝑛2
where the subscripts indicate the location at which the parameters are evaluated.

If we choose location 1 in deep water where the wave properties are more easily eval-
uated (𝑛1 = 𝑛0 = 1/2), we find the following formula for the wave height (subscripts 2
are dropped):
𝐻 1 1
= = 𝐾𝑠ℎ (5.7)
𝐻0 √ tanh 𝑘ℎ 2𝑛
The parameter 𝐾𝑠ℎ is called the shoaling factor and is purely a function of 𝑘ℎ (𝑛 is a
function of 𝑘ℎ only). In Fig. 5.3 it is shown as a function of ℎ/𝐿0 . It is 1.0 in deep water,
then decreases slightly with water depth to 0.91 and subsequently rises to infinity. In
reality the wave height increase in the shoaling zone is limited by dissipation due to
wave-breaking. Breaking and the limits of breaking are discussed further on in this
chapter (Sect. 5.2.5). Note that the theory of shoaling is equally valid for decreasing
as for increasing water depth. This means that a wave that passes over a local shoal
resumes its original height, as long as no breaking has occurred.
5.00
H/H 0
n
1.00
0.50
L/L 0
c/c 0
0.10
0.05
0.001 0.005 0.010 0.050 0.100 0.500
h/L0
Figure 5.3: The shoaling factor 𝐾𝑠ℎ = 𝐻 /𝐻0 (Eq. 5.7), 𝑛 (Eq. 3.25) and 𝑐/𝑐0 = 𝐿/𝐿0 = tanh 𝑘ℎ (see
Eq. 3.22) as a function of ℎ/𝐿0 .
In order to determine the shoaling coefficient, it may be helpful to use standard tables
containing the values of the hyperbolic functions. These tables can be found in, for
instance, the CEM (see Sect. 1.7.3). An extract of these tables is given in App. A.

Shoaling of tides and tsunamis

The phenomenon of shoaling has been explained in this section for wind waves. How-
ever, the intermediate- and shallow-water propagation of other waves such as tsuna-
mis and tides is affected by depth variations as well1 . For example, at sea the wave
heights of a tsunami are only of the order of a metre. When tsunamis travel into
progressively shallower water, however, their energy is concentrated by shoaling and
possibly tunnelling2 , causing them to steepen and rise to many metres in height.
5.2.3. Refraction
Instead of a normally incident wave, consider now an obliquely incident linear wave
approaching at a deep-water angle 𝜑0 to the shore. The wave is again long-crested
and the bottom contours are essentially straight and parallel as shown in Fig. 5.4. The
wave is in the shoaling region outside the breaker zone.
depth contours
ay φ
ve r
wa
φ0
wave front
surf zone
Figure 5.4: Obliquely incident waves propagating on alongshore uniform depth contours.
When a wave approaches underwater contours at an angle, it is evident that the sec-
tions of the crest in the deeper parts travel faster than those in the shallower sectors.
This causes the wave crest to turn towards the depth contour. This bending effect
is called refraction, and is analogous to similar phenomena in physics (light, sound).
The effect is shown in Fig. 5.5. It takes place in addition to the effects of shoaling and
continues up to the shoreline.
In analogy with refraction of light, the direction of the wave rays changes proportion-
ally to the wave propagation speed according to Snell’s law3 :
sin 𝜑2 sin 𝜑1
= (5.8)
𝑐2 𝑐1
1
Note that shallow water is a relative measure depending on the wavelength of the wave; in Sect. 3.8.1
for instance, it was demonstrated that the oceans already constitute shallow water for the tidal wave.
2
Tunnelling is the concentration of energy due to width restriction. See also Sect. 3.8.1 and Sect. 5.7 for
examples.
3
Named after the Dutch astronomer Willebrord Snellius (born Willebrord Snel van Royen, 1580–1626).

Thus, along the wave ray, sin 𝜑/𝑐 is constant and equal to its deep-water value sin 𝜑0 /𝑐0 .
Eq. 5.8 holds for breaking as well as non-breaking waves, but for straight, parallel depth
contours only.
By applying Snell’s law, it is possible to construct a field of wave rays over a given
bottom configuration for a given wave direction and wave period. The wave angle
can thus be considered known in the energy balance, Eq. 5.2.
wave orthogonals or rays wave crests
b0
a φ0
20
wa
ve
p ro
pa
depth contours [m]
ga
15 tio
nd
ire
ct
ion
10
b φ
a
5
0
coastline
Figure 5.5: Wave refraction over straight and parallel depth contours (i.e. at a uniform coast),
showing every second wave crest. The waves refract from 𝜑0 = 60° offshore towards shore-
normal according to Snell’s law (Eq. 5.8). The wave celerity as a function of the water depth is
computed using the dispersion relation (Eq. 3.22) and 𝑇 = 6 s.
Figure 5.5 shows that the distance 𝑏 between the wave rays varies. If we assume that for
long-crested waves no wave energy moves laterally along the wave crest and thus that
the energy remains constant between wave rays (normal to the wave crest), energy
conservation between two wave rays requires that (cf. Eq. 5.3):
𝐸𝑛𝑐𝑏 = const → 𝐻22 𝑛2 𝑐2 𝑏2 = 𝐻12 𝑛1 𝑐1 𝑏1 (5.9)
The wave height at two locations therefore relates as:
𝐻2 𝑐 𝑛 𝑏
= 1 1 1 (5.10)
𝐻1 √ 𝑐2 𝑛2 √ 𝑏2

Using 𝑛0 = 1/2 (valid in deep water) the wave height 𝐻 at any location can be related
to the wave height in deep water:
𝐻 1 𝑐0 𝑏0
= = 𝐾𝑠ℎ 𝐾𝑟 (5.11)
𝐻0 √ 2𝑛 𝑐 𝑏
where 𝐾𝑠ℎ is the shoaling coefficient according to Eq. 5.7 and
𝑏0
𝐾𝑟 = (5.12)
√𝑏
is the refraction coefficient used to calculate the change in wave height when a wave
approaches at an angle to the shore. Since, for parallel depth contours, every wave ray
refracts in the same way, the distance between given wave rays, measured parallel to
the depth contours, remains constant (distance 𝑎 in Fig. 5.5) and is equal to:
𝑏
𝑎= = const (5.13)
cos 𝜑
and thus:
cos 𝜑0
𝐾𝑟 = (5.14)
√ cos 𝜑
This result for the wave height variation in the case of parallel depth contours may
also be found directly from Eq. 5.2. Assuming an alongshore uniform situation (𝑦-
derivative equal to zero) outside the surf zone (dissipation negligible), integration
yields:
𝐸𝑛𝑐 cos 𝜑 = const (5.15)
which directly results in Eq. 5.10.

The effect of refraction on the wave height in the present example is to reduce the in-
crease in wave height due to shoaling. In a real-life situation with a more complicated
pattern of depth contours, two basic calculating techniques are available for refraction
patterns: graphically and numerically (see above). A description of the first method is
given in the CEM (see Sect. 1.7.3). Fundamentally, all methods of refraction analysis
are based on Snell’s law and conservation of wave energy. A refraction diagram is
given in Fig. 5.6 as an example of the results of a refraction study. If the wave rays
converge, there is an accumulation of energy and relatively high wave heights can be
expected. In contrast, if wave rays diverge, the energy is spread over a larger part of
the wave crest, so the wave height is reduced.

Depth refraction in obliquely incident shoaling waves was explained from differences
in water depth and hence wave celerity along a wave crest. Refraction may also occur
due to mean currents, in which case it is called current refraction. When short waves
interact with a current, amongst others the celerity of the waves is affected. Current
refraction takes place if the current velocity varies along a wave crest, e.g. in tidal
entrances (Sect. 9.4.1), in major ocean currents, or in harbour entrance channels.
divergence convergence divergence
Figure 5.6: Wave refraction diagram. Wave energy converges in the case of convex depth con-
tours (as seen from the sea) and diverges in the case of concave depth contours.
5.2.4. Diffraction
If obstructions to the wave propagation (an offshore island, a breakwater, a headland)
or abrupt changes in the bottom contours are present, there is a large (initial) variation
of wave energy along a wave crest, which leads to transfer of energy along the wave
crests. This phenomenon is called diffraction. Figure 5.7 shows the diffraction of an
incident wave train in case there are no depth changes. A part of the wave front is
blocked by the breakwater and is reflected seaward. The remainder of the wave front
will bend around the obstacle and thus penetrate into the zone in the lee of the obstacle
(shadow zone). The diffracted wave crests will form concentric circular arcs with the
wave height decreasing along the crest of each wave.
Figure 5.7 also shows the wave ray that separates the shadow or diffraction zone from
the wave zone. Due to the lateral transfer of wave energy into the shadow zone, the
wave height along this ray is lower than the incident wave height; in the case of con-
stant depth (and thus constant celerity), the wave height is 50 % of the original wave
height according to linear theory and about 70 % for irregular directional waves. The
wave heights decrease deeper into the shadow zone. Further from the breakwater into
the wave zone, the wave height gradually approaches the incident wave height.

incident wave train
zo ne
io n
ct
fra
f
di
wave zone
Figure 5.7: Diffraction of an incident wave train.
The extent of energy penetration in the area landward of an obstacle depends on the
ratio of a typical lateral dimension of the obstacle, e.g., the length of a single detached
breakwater 𝜆 to the wavelength 𝐿. When a thin pile is located in waves with a large
wavelength, 𝜆/𝐿 ≪ 1, wave energy spreads behind the entire pile. In the case of a
long, detached breakwater, 𝜆/𝐿 ≫ 1, diffraction occurs around each breakwater head,
but wave energy will not spread in the entire zone behind the breakwater.
Figure 5.8 shows waves passing through a gap between two detached breakwaters.
Diffraction occurs in the lee of the breakwaters on both sides of the gap. For large
enough gap lengths (relative to the wavelength), these diffraction patterns are inde-
pendent. Interaction of the two patterns may occur for smaller gap lengths.
Figure 5.8: Typical diffraction pattern through a gap between two detached breakwaters in
Pesaro, Italy. Photo by Roberto Lo Savio (‘Credits’ on page 575).

The theory of wave diffraction is solved mathematically by application of the ‘Cornu

spiral’ (see for instance Battjes (2006)). This graphical method yields spatially vary-
ing diffraction coefficients, defined as the ratio of the diffracted wave height to the
incident wave height, assuming that the latter is not disturbed by the obstacle. The
CEM Chapter II-7 and (in more detail) the Shore Protection Manual Volume I Chapter
2 present diffraction coefficients as a function of position (relative to a semi-infinite
rigid impermeable breakwater) and as a function of the breakwater gap width.
The disadvantage of the above methods is that they assume a constant water depth.
In reality there will generally be a sloping bottom or an uneven bed and the results
will therefore be influenced by this bottom. Numerical models can take into account
shoaling, diffraction, refraction, reflection (and breaking) simultaneously.
5.2.5. Wave-breaking
Section 5.2.2 demonstrated how shoaling would increase the wave height until infinity,
at least in the absence of a physical limit to the steepness of waves. A wave crest
becomes unstable and starts breaking when the particle velocity exceeds the velocity
of the wave crest (the wave celerity). This breaking condition corresponds to a crest
angle of about 120° (see Fig. 5.9).
120°
Figure 5.9: Maximum crest angle.
Miche breaking criterion and breaker index

Miche (1944) expressed the limiting wave steepness based on the Stokes wave theory
(a non-linear expansion of the linear Airy theory that better describes steeper waves,
see Intermezzo 5.1):
𝐻
[ ] = 0.142 tanh(𝑘ℎ) (5.16)
𝐿 max
In deep water Eq. 5.16 reduces to:

𝐻0 1
[ ] = 0.142 ≈ (5.17)
𝐿0 max 7
When the deep-water steepness exceeds this limit, steepness-induced wave-breaking

(called white-capping) occurs. Only a limited part of the wave energy is dissipated
through white-capping. The steepness 𝐻𝑠 /𝐿0,𝑝 of wind waves is often less than 0.05.

Note that 𝐿0,𝑝 is the deep-water wavelength corresponding to the peak period. Con-
sider for example a deep-water wave height 𝐻𝑠 = 1.5 m and a wave period of 𝑇𝑝 = 5 s.
These are average conditions for the Dutch coast. We find 𝐿0,𝑝 = 1.56𝑇𝑝2 = 39 m and
𝐻𝑠 /𝐿0,𝑝 = 1.5/39 = 0.04. Under these conditions, even for the higher waves in the
record, little white-capping is expected.
In shallow water Eq. 5.16 becomes:
𝐻 2𝜋ℎ ℎ
[ ] = 0.142 ≈ 0.88 (5.18)
𝐿 max 𝐿 𝐿
This is equivalent to:
𝐻 𝐻
𝛾 =[ ] = 𝑏 ≈ 0.88 (5.19)
ℎ max ℎ𝑏
with 𝛾 is the breaker index, 𝐻𝑏 is the breaking wave height and ℎ𝑏 is the water depth
at the breaking point. Using solitary wave theory – a non-linear wave theory valid for
shallow water – gives a slightly different value 𝛾 ≈ 0.78.
The breaker index shows that in the shallow nearshore zone wave-breaking of indi-
vidual waves starts when the wave height becomes greater than a certain fraction of
the water depth. This is called depth-induced breaking, since the limiting wave height
is governed by a water depth limitation.
The water depth limitation can be explained as follows. The decrease in water depth
and the increase in wave heights due to shoaling result in a significant increase in
horizontal particle velocities with respect to the wave celerity. When waves approach
the shore, the wave celerity is reduced. In addition, we have seen that the wave height
increases due to shoaling and the orbital motion within the waves is also changing (see
also Sect. 5.4): although the orbit in deep water is a circle, in shallow water the orbit
becomes an ellipse with a horizontal axis longer than the vertical axis. The vertical
axis of the orbit at the water surface is equal to the wave height. Since shoaling causes
an increase in the wave height, the vertical motion of the water particles at the surface
must also increase. In addition to this, the horizontal movements grow in relation
to the vertical movements, which means that there must be a significant increase of
the particle velocity near the surface, until the horizontal particle velocity exceeds the
wave celerity.
The Rayleigh distribution demonstrated that the maximum wave height 𝐻max in a
wave record is equal to 2𝐻𝑠 . The maximum value of 𝐻𝑠 /ℎ for which the largest waves
are breaking is therefore half of the value of the breaker indices. Hence 𝐻𝑠 /ℎ ≈ 0.4−0.5
based on the Miche criterion.
The breaking parameters are shown in Fig. 5.10. The point where the wave height
suddenly decreases because the largest wave in the wave field starts breaking is called
the breaker point.

wave height [m] 1 breaking point
0
0
1
depth [m]
0 50 100 150 200 m
Figure 5.10: Wave-breaking parameters 𝐻𝑏 and ℎ𝑏 .
The above wave-breaking criteria have been derived for a horizontal bottom. In reality
the bottom will be sloping, which will affect the breaker index. Before we look at the
dependency of the breaker index on the angle of the slope, we first consider the process
of breaking on a slope.
Effect of bed slope on breaking process

Depending on the wave properties and the angle of the bed slope, the process of break-
ing takes place in various different ways. Battjes (1974) showed that the Iribarren
parameter guides this process. It is defined as follows:
tan 𝛼
𝜉 = (5.20)
√𝐻0 /𝐿0
where:
tan 𝛼 steepness of the beach −

𝐿0 wavelength in deep water m
The Iribarren parameter 𝜉 represents the ratio of the slope steepness tan 𝛼 and the
wave steepness. The latter is represented by the deep-water steepness instead of the
steepness at the breaker point. In irregular waves the Iribarren parameter is often
computed with 𝐻𝑠 and 𝐿0 based on 𝑇𝑝 . A distinction is made between spilling, plunging
and surging breakers, depending on the value of 𝜉 (Fig. 5.11). The transition from
surging to plunging breakers is often referred to as collapsing breakers. The typical

value of the Iribarren parameter 𝜉 for the breaker type is also given in the figure. These
values are indicative and the transition between the various breaker types is gradual.
Spilling breakers are usually found along flat beaches. These waves begin breaking
at a relatively great distance from shore and break gradually (over a distance of 6
to 7 wavelengths) as they approach progressively shallower water. During breaking,
a foam line develops at the crest and leaves a thin layer of foam over a considerable
distance. There is very little reflection of wave energy back towards the sea. Practically
all wave energy is dissipated in the breaking process.
surging: ξ > 5 collapsing: 3.3 < ξ < 5
plunging: 0.5 < ξ < 3.3 spilling: ξ < 0.5
Figure 5.11: Breaker types.
A plunging breaker is a type that is found on moderately-sloped beaches. The curling

top is typical of such a wave. When the curling top breaks over the lower part of the
wave, a lot of energy is dissipated into turbulence. Some energy is reflected back to
the sea, and some is transmitted towards the coast, while forming a ‘new’ wave.
Surging breakers occur along rather steep shores for relatively long swell waves. The
waves surge up and down the slope with minor air entrainment. The breaker zone
is very narrow and more than half of the energy is reflected back into deeper water.
The breakers form like plunging breakers, but the toe of each wave surges upon the
beach before the crest can curl over and fall. It is debated whether surging breakers
are actually breakers or rather standing waves (caused by interference of the incoming
and reflected wave).
A collapsing breaker is between a plunging and a surging breaker and thus in between
breaking and non-breaking.
The parameter 𝜉 indicates that, for a slope, the notions ‘steep’ and ‘gentle’ are relative.
A beach slope of 1:100 is usually thought to be gentle, at least for a wind wave with a
period of a few seconds. For this beach slope, the wave conditions discussed above –
a deep-water wave height 𝐻𝑠 = 1.5 m and a wave period of 𝑇𝑝 = 5 s – lead to 𝜉 = 0.05.
The breakers are thus of the spilling type. For a tidal wave, however, such a beach slope
is rather steep; a tidal wave on a beach does not break and is completely reflected.
How does the bottom slope influence the breaker index 𝛾 ? Experimental results and
theoretical considerations suggest that waves need time to break, so that at steeper

slopes they will break at smaller water depths, which results in a larger breaker index.
This means that 𝛾 increases with increasing 𝜉 from ca. 0.6 to 0.8 for spilling breakers
to 0.8 to 1.2 for plunging breakers. The average value is ca. 0.8, corresponding to
the values found for a horizontal bottom, see e.g. Eq. 5.19. Battjes and Janssen (1978)
therefore adapted the Miche equation (Eq. 5.16) to account for the influence of bottom
slope.
Surface roller
Breaking waves generate a layer of air-water mixture, which moves in a landward
direction in the upper parts over the water column. This so-called surface roller is
thought to act as a temporary storage of energy and momentum. Instead of being
dissipated immediately after the breakpoint, organised wave energy is converted into
turbulent kinetic energy first (which can be seen from the development of a roller at
the face of a breaking wave), before being dissipated ultimately via the production of
turbulence. To take this into account, sometimes an additional energy balance equa-
tion is used: the roller balance equation. The roller energy 𝐸𝑟 represents the amount
of kinetic energy in a roller propagating at the shallow-water speed 𝑐 = √𝑔ℎ.
5.3. Wave asymmetry and skewness

In the previous sections, linear wave propagation from deep to shallow water was
discussed. We considered the propagation and conservation of a bulk quantity, the
energy, in the shoaling region and breaker zone, without considering any exchange of
energy (or momentum) between different wave components due to non-linear inter-
actions. The local wave characteristics were described by linear wave theory.
However, waves propagating towards the shore become more and more asymmetric,
until the point of wave-breaking. Apart from an increase in wave height, the shoaling
process is typically characterised by:
• gradual peaking of the wave crest and a flattening of the trough (this asymmetry
relative to the horizontal axis is called skewness);
• relative steepening of the face until breaking occurs, resulting in a pitched-forward
wave shape (this asymmetry relative to the vertical axis is often simply called
asymmetry).
These non-linear effects cannot be described by linear theory, and the closer we come
to the shore the more apparent the deviations from linear theory become. Numer-
ous non-linear theories (Stokes theory, cnoidal wave theory, Boussinesq equations;
see Intermezzo 5.1 for an overview) have been developed to take into account these
complicated non-linear processes. The non-linear effects are crucial in determining
the magnitude of the wave-induced transport (see Ch. 7). In the following we will
therefore subsequently consider skewness and asymmetry.

5.3. Wave asymmetry and skewness 185
Skewness
Figure 5.12 shows a wave with a long, flat trough and narrow, peaked crest, as can be
observed in shallow water.
η/ηmax [-]
phase [rad]
Figure 5.12: A skewed wave (𝐻 /ℎ = 0.5 and 𝐿/ℎ = 30) with on the horizontal axis the phase
𝑆(𝑥, 𝑡) = 𝜔𝑡 − 𝑘𝑥. The shallow-water wave is computed using fifth-order cnoidal wave theory
(Intermezzo 5.1). Alternatively, it can be computed using a Fourier approximation, for which
20 harmonic components are required to converge towards the cnoidal solution (see Fig. 5.18).
Produced using software by Fenton (https://johndfenton.com/Steady-waves/Fourier.html).
This asymmetric profile (relative to the horizontal axis) can only be described by a
sum of sinusoidal waves with higher harmonics (frequencies that are a multiple of
the basis frequency: cos 𝑆, cos 2𝑆 etc. with 𝑆 = 𝜔𝑡 − 𝑘𝑥). Let us illustrate this using
Stokes second-order theory. The second-order equation for the surface elevation can
be written as:
𝜂 = 𝜂̂1 cos(𝜔𝑡 − 𝑘𝑥) + 𝜂̂2 cos 2(𝜔𝑡 − 𝑘𝑥) (5.21)
The amplitude of the second-order correction is small compared to the first-order com-
ponent (as long as 𝜂̂1 = 𝑎 is sufficiently small). The Stokes wave is sketched in Fig. 5.13.
It can be seen that the resulting Stokes wave profile 𝜂1 + 𝜂2 has crests which are nar-
rower and more peaked than those of a cosine profile and troughs that are wider and
flatter; the profile is skewed. The second term 𝜂2 represents the Stokes second-order
wave travelling at the same speed as the first-order wave 𝜂1 (hence, it does not obey
the linear dispersion relation). Since the two components travel at the same speed,
the wave form does not change; it has a permanent form. Similarly in energy spectra
of shoaling waves, one can find harmonics of the spectral peak which are coherent
in phase with the spectral peak. Note that taking into account higher-order terms in
Eq. 5.21 will make the profile approach the profile of Fig. 5.12 more and more (see also
Fig. 5.18).
Since the higher harmonics remain phase-locked and in phase with the primary har-
monic, Stokes waves have a permanent form. This is different from the deep-water
situation we described in Ch. 3. There we considered the irregular sea state at a certain

η=η1
η=η2
η=η1+η2
[-]
Figure 5.13: A second-order Stokes wave (solid line) is composed of a first-order (dashed line)
and a second-order (dashed-dotted line) component, according to Eq. 5.21 with 𝜂̂2 /𝜂̂1 = 0.2. The
phase on the horizontal axis is the phase 𝑆 = 𝜔𝑡 − 𝑘𝑥 of the primary harmonic. For more back-
ground on Stokes theory see Section 5.6.2 in Holthuijsen (2007).
location to be the sum of linear, freely moving waves with random phases. Because of
the different phase speeds of the different components (satisfying the linear dispersion
relation), we cannot have a permanent form.
A statistical indicator for the skewness is the normalised cube of the surface elevation
⟨𝜂3 ⟩/𝜎 3 : the time-averaged value of the cube of surface elevation normalised with the
cube of the standard deviation 𝜎 3 . The brackets denote time-averaging. Figure 5.14
shows the quantity 𝜂3 for a sinusoidal wave and for a second-order Stokes wave. It can
be seen that for a linear wave4 the time-averaged value ⟨𝜂3 ⟩ is zero, while for a Stokes
wave it is not, since the crests weigh more heavily than the troughs. The sign defines
the ratio of crests to troughs. If it is positive – as is the case for the Stokes wave – then
the crests are bigger than the troughs. Note that Stokes type wave forms have positive
values of skewness, but zero asymmetry (around the vertical axis), since they are not
pitched forward.
Linear theory could be used to estimate the variance (energy density) of nearshore
waves. However, it cannot be used to estimate skewness, since a linear superposition
of random waves has by definition zero skewness over the ensemble.
Of course not only the surface elevation, but also the orbital velocities become skewed;
the orbital velocities in the direction of wave movement (‘onshore’) become higher,
and the orbital velocities against the wave direction (‘offshore’) become smaller. This
naturally means that the duration of the onshore-directed orbital motion becomes
smaller and the duration of the offshore-directed orbital signal becomes larger.
4
Similarly, in an irregular wave field approximated by a linear superposition of freely moving waves,
the skewness value is zero.

[-]
0 π 2π 0 π 2π
Figure 5.14: Quantity of 𝜂3 /𝜂̂31 compared to 𝜂/𝜂̂1 for a sinusoidal wave (left) and a second-order
Stokes wave (right) as a function of the phase 𝑆 = 𝜔𝑡 − 𝑘𝑥 of the first or primary harmonic. Note
that for the sinusoidal wave, the time-average ⟨𝜂3 ⟩ = 0, whereas for the second-order Stokes
wave ⟨𝜂3 ⟩ > 0.
Asymmetry
The pitched-forward wave shape is the result of the fact that in shallow water the
wave crest moves faster than the wave trough. This is clear because the propagation
speed of non-linear shallow-water waves 𝑐 = √𝑔(ℎ + 𝜂). Note that for small-amplitude
shallow-water waves this reduces to 𝑐 = √𝑔ℎ (see Intermezzo 3.5). The wave crest of a
harmonic wave with amplitude 𝑎 has a higher propagation velocity 𝑐crest = √𝑔(ℎ + 𝑎)
than the trough which propagates with 𝑐trough = √𝑔(ℎ − 𝑎). This leads to a pitched-
forward profile or relative steepening of the face of a wave that can be represented
by the inclusion of a second harmonic that is forward phase-shifted with respect to
the primary harmonic (see the left panel of Fig. 5.15). A Stokes wave train cannot
exhibit wave asymmetry; because of the small amplitude approximation of the Stokes
theory, the higher harmonic remains phase-locked to the primary harmonic. The right
panel of Fig. 5.15 shows a time series of a pitched-forward (sawtooth-like) wave at one
location, namely location A (𝑥/𝐿 = 0 ) in the left panel of Fig. 5.15.
The left panel of Fig. 5.15 shows a rapidly rising and slowly falling surface elevation.
Note that for the sawtooth-like wave of Fig. 5.15, the phase shift between the first and
the second harmonic is such that the asymmetry is maximum and the skewness is zero.
The sawtooth-like wave of Fig. 5.15 can also be depicted as a function of 𝑆 = 𝜔𝑡 − 𝑘𝑥.
This results in Fig. 5.16. Compare this figure to the skewed wave of Fig. 5.13.
Figure 5.17 compares the quantity 𝜂3 for a skewed wave and an asymmetric wave as a
function of the phase 𝑆 = 𝜔𝑡 − 𝑘𝑥 of the first or primary harmonic. It can be seen that
for the asymmetric wave the time-averaged value ⟨𝜂3 ⟩ is zero, while for the skewed
wave it is not, since the crests weigh more heavily than the troughs.
Shoaling waves first become gradually more skewed while remaining reasonably sym-
metric about the vertical axis – as an example see the flume measurements of Fig. B.2

η=η1 η=η2 η=η1+η2

c crest
A
[-]
c trough
x/L t/T
Figure 5.15: Asymmetric wave (solid black line), computed as the superposition of a first-order
(dashed line) and a 90° forward phase-shifted second-order signal (dashed-dotted line). The
phase on the horizontal axis is the phase 𝑆 = 𝜔𝑡 − 𝑘𝑥 of the first-order component (or primary
harmonic). The left figure shows the spatial variation at 𝑡 = 0 and the right figure shows the
temporal variation of the water surface elevation at A (𝑥/𝐿 = 0).
η=η1
η=η2
η=η1+η2
[-]
Figure 5.16: A second-order asymmetric wave, generated using two harmonic components, of
which the second one is 90° forward phase-shifted.
[-]
0 π 2π 0 π 2π
Figure 5.17: Quantity of 𝜂3 /𝜂̂31 compared to 𝜂/𝜂̂1 for a skewed wave (left) and an asymmetric wave
(right). Note that for the skewed wave, the time-average ⟨𝜂3 ⟩ > 0, whereas for the asymmetric
wave ⟨𝜂3 ⟩ = 0.

in App. B. Closer to the surf zone, phase-shifting of the harmonic(s) leads to an in-
crease in wave asymmetry and – eventually – to a decrease in wave skewness as well.
Ultimately the pitching forward results in wave-breaking.
The Ursell parameter 𝑈 = 𝐻 𝐿2 /𝐻 3 can be used as an indicator for skewness and asym-
metry. Doering and Bowen (1995) presented parameterisations of velocity skewness
and asymmetry based on the Ursell parameter. They mainly focused on the skew-
ness and asymmetry of the orbital velocities, because of their importance for sediment
transport.
Intermezzo 5.1 Non-linear wave theories
Linear wave theory is obtained by linearising the free-surface boundary condi-

tions; the boundary conditions are applied to the mean water surface 𝑧 = 0, in-
stead of the instantaneous water surface 𝜂. Non-linearities can be neglected for
not too steep waves in deep water (𝑎𝑘 ≪ 1 for 𝑘ℎ is large) or small-amplitude
waves in shallow water (𝑎 ≪ ℎ for 𝑘ℎ is small).
η/ηmax [-]
x/h [-]
Figure 5.18: Surface elevation records generated for 𝐻 /ℎ = 0.5 and 𝐿/ℎ = 30, which amounts
to 𝑇 √𝑔/ℎ = 26.29 or ℎ/(𝑔𝑇 2 ) = 1.45 × 10−3 and 𝐻 /𝑔𝑇 2 = 7.23 × 10−4 in Fig. 5.19. Both fifth-
order cnoidal wave theory and a numerical Fourier approximation method (with 5, 10 and
20 harmonic components) are used. About 20 harmonic components are required for the
Fourier approximation to converge towards the cnoidal solution. Produced using software
by Fenton (https://johndfenton.com/Steady-waves/Fourier.html).
For a non-linear solution, the free-surface boundary conditions have to be applied

at the free surface 𝜂. The complication is that 𝜂 is unknown. Amongst others, the
following non-linear theories have been developed:

• Stokes series expansion. Stokes used the result of the linear theory to find
a first approximation to the neglected non-linear terms. This results in a
second-order correction to the first (linear) approximation of the solution.
Note that the second-order correction can be used to obtain a third-order
correction and so on. Non-linear terms consist of quadratic terms, so instead
of the linear solution 𝑎 cos 𝑆 (with the phase 𝑆 = 𝜔𝑡 − 𝑘𝑥), the second-order
solution has an additional term proportional to 𝑎2 cos2 𝑆 that results in a
term proportional to 𝑎2 cos 2𝑆, a harmonic with half the period of the linear
solution. A Stokes series thus has the form 𝜂 = 𝜂̂1 cos 𝑆+𝜂̂2 cos 2𝑆+𝜂̂3 cos 3𝑆+
… in which the first term is the linear solution. For convergence, every next
coefficient should be small compared to the lower-order coefficient, which
will only be the case for sufficiently small amplitudes 𝜂̂1 = 𝑎. Stokes theory
does not converge in shallow water; if the Ursell parameter 𝑈 = 𝐻 𝐿2 /ℎ3 is
too large, the series diverges.
• The stream function theory is an alternative to the Stokes theory. It also
adds higher harmonics to the linear solution.
10 -1
deep-water (Miche)
breaking criterion
th or 5th ORDER
KES 4
STO
KES 3 ORDER
rd
shallow-water STO
10 -2 breaking criterion
(solitary wave)
KES 2 ORDER
nd
STO
SE
AV
(1.45×10-3,7.23×10-4)
W
2
RY
-3
H/gT
10
TA
LI
SO
LINEAR WAVE THEORY

(AIRY)
CNOIDAL
-4 WAVES
10
shallow-water intermediate-depth deep-water

waves waves waves
10 -5
10 -4 10 -3 10 -2 10 -1 10 0
2
h/gT
Figure 5.19: The relative depth on the horizontal axis and the wave steepness on the vertical
axis delineate the validity of various waves theories. The black dot indicates the wave of
Fig. 5.18. (Based on Kamphuis, 2000; Le Méhauté, 1976).

5.4. Wave orbital velocity, pressure and bed shear stress 191
• Applicable in shallow water is the cnoidal wave theory. Solutions are given
in terms of elliptic integrals of the first kind; the solution in deep water
is identical with linear wave theory and in shallow water to solitary wave
theory. The latter is a single wave without a trough and a mass of water
above mean water level moving entirely in the wave propagation direction.
• Boussinesq models can be seen as an extension of the shallow-water equa-
tions (see for instance TU Delft course CTB3350/CIE3310-09). The shallow-
water equations describe non-linear waves that all travel at the same speed
𝑐 = √𝑔(ℎ + 𝜂). Orbital velocities are depth-uniform and the pressure is hy-
drostatic. Since there is no vertical variation of the wave field, the vertical
does not have to be resolved. Boussinesq models account – to some ex-
tent – for the non-hydrostatic pressure distribution and depth-dependency
of orbital velocities. The advantage over the shallow-water equations is that
Boussinesq models include frequency dispersion. Since the equations can
still be integrated over the vertical, the equations are computationally effi-
cient.
A review of the validity of various wave theories (until wave-breaking) is given

by Le Méhauté (1976). See also Fig. 5.19.
5.4. Wave orbital velocity, pressure and bed shear stress

5.4.1. Wave orbital velocities
So far we have focused on the wave surface elevation. Underneath the wave sur-
face, there is a fluid motion associated with the motion of the water surface; the fluid
particles describe an orbital path. To first order (within the limitations of Airy small-
amplitude wave theory), the orbits are closed5 circles in deep water and closed ellips-
oids in water of finite depth, with the ellipsoids becoming flatter near the bottom. At
the bottom vertical velocities are zero per definition. The further from the surface,
the smaller the orbital diameter becomes. In deep water the orbital diameter has been
reduced to only 4 % of the value at the surface at a depth of half a wavelength. In
the case of shallow water, the water particles under a propagating wave of finite amp-
litude describe an elliptical orbit. From the surface down to the bottom, the vertical
displacement of the water particles reduces to zero, while the horizontal displacement
remains almost constant (Fig. 5.20).
5
At second order, particle paths are no longer closed orbits and there is a drift or mass transport in the
direction of wave propagation (Sect. 5.5.1).

Figure 5.20: Water particle movement in finite-amplitude wave in intermediate to shallow water
depth.
Now consider waves of infinitesimal amplitudes. According to linear theory, the hori-
zontal orbital velocity varies harmonically with an amplitude 𝑢̂ equal to:
cosh 𝑘(ℎ + 𝑧)
𝑢(𝑧)
̂ = 𝜔𝑎 (5.22)
sinh 𝑘ℎ
where:
𝜔 angular frequency (2𝜋/𝑇 ) rad/s

𝑎 wave amplitude m
𝑘 wavenumber (2𝜋/𝐿) rad/m
𝐿 wavelength m
The 𝑧-axis is defined positive upward with 𝑧 = 0 at the surface and 𝑧 = −ℎ at the
bottom. The velocity 𝑢 is in the wave propagation direction. Using the cosh 𝑘ℎ and
sinh 𝑘ℎ approximations for shallow and deep water (App. A), the horizontal velocity
profiles can be drawn schematically as in Fig. 5.21.
In shallow water (𝑘ℎ ≪ 1; in practice 𝑘ℎ < 𝜋/10 or ℎ/𝐿 > 1/20), the depth-uniform
velocity amplitude is given by (cf. Eq. 3.39):
𝜔𝑎 𝑎 𝐻
𝑢̂ = = 𝑐 = √𝑔ℎ (5.23)
𝑘ℎ ℎ 2ℎ

shallow intermediate deep
(a) (b) (c)
Figure 5.21: Schematic drawing of vertical profiles of the velocity amplitude 𝑢.̂
The particle excursions (i.e. the horizontal and vertical displacements of the particles)
are the time integrals of the oscillatory horizontal and vertical flow velocities respect-
ively. This means that the amplitude of the horizontal particle excursion is given by:
𝑢̂
𝜉̂ = (5.24)
𝜔
Note that Sect. 5.3 described how in shoaling waves the surface elevation becomes
gradually more skewed and asymmetric. Of course the wave orbital velocities near the
bed will also become skewed and asymmetric. Although the linear approximation will
not describe the changing wave form, the orbital velocity magnitude can be estimated
reasonably well using linear wave theory.
5.4.2. Dynamic pressure

Pressure gradients in coastal waters are mainly due to mean water level variations (hy-
drostatic pressure) and fluctuations of pressure due to waves. The hydrostatic pressure
due to mean water level variation is 𝑝0 = −𝜌𝑔𝑧 and hence linearly increases from zero
at the water surface 𝑧 = 0 to 𝑝0 = 𝜌𝑔ℎ at the bottom 𝑧 = −ℎ. In the case of waves, the
total pressure is the sum of this hydrostatic pressure 𝑝0 (from 𝑧 = −ℎ to 𝑧 = 0) plus the
wave-induced or dynamic pressure 𝑝wave (from 𝑧 = −ℎ to 𝑧 = 𝜂). Wave-induced pres-
sure oscillations are different under wave crest and wave trough and – in intermediate
and deep water – reduce with depth below the free surface (Fig. 5.22a).
According to linear theory, the wave-induced pressure varies harmonically (in phase
with the surface elevation 𝜂) with amplitude:
𝜌𝑔𝐻 cosh 𝑘(ℎ + 𝑧)

𝑝̂ = (5.25)
2 cosh 𝑘ℎ

wave crest
wave trough
actual
hydrostatic
pressure
pressure
actual
pressure
(a) deep water

wave crest
wave trough
actual
hydrostatic
pressure
pressure
actual
pressure
(b) shallow water
Figure 5.22: Wave-induced pressure oscillations combined with the hydrostatic pressure in still
water give the actual pressure.
which reduces in shallow water to:
𝜌𝑔𝐻
𝑝̂ = (5.26)
2
Hence, in shallow water the hydrostatic dynamic pressure varies linearly with the
free-surface elevation 𝑝wave = 𝑝 ̃ = 𝜌𝑔𝜂 (Fig. 5.22b). The tilde indicates the purely
oscillatory character.
To derive Eq. 5.25 the amplitude was assumed to be very small in order to linearise the
free-surface boundary condition. It is therefore not valid in the region between the
trough and the crest elevation. We can, however, assume that the dynamic pressure is
hydrostatic between wave trough and wave crest: 𝑝 ̃ = 𝜌𝑔𝜂.
5.4.3. Wave boundary layer

Most wave theories are valid from the water level to a small distance from the bed
(Fig. 5.23), where the flow still is unaffected by the boundary. Closer to the bed, in a
thin layer called the wave boundary layer, vorticity (rotation) can be generated, which
is not included in linear wave theory or in most other (irrotational) wave theories for
that matter. The distance denoted as 𝛿 in Fig. 5.23 is the thickness of the wave boundary
layer, the transition layer between the bed and the layer of ‘normal’ oscillating flow.
The thickness is generally between 1 cm and 10 cm for short-period waves (𝑇 < 10 s).

The reason for this small thickness is that there is not sufficient time for the layer to
grow out in the vertical direction, because the current regularly reverses.
Figure 5.23: The velocity at the boundary is zero. At some distance above the boundary, the velo-
city reaches a constant value, 𝑢̂0 called the free stream velocity. The region of velocity variation
in the vertical or shear is called the boundary layer. The height or thickness of the boundary layer
is 𝛿.
It is typical for oscillating boundary layers that the maximum flow velocity near the
bed is somewhat larger (a few percent) than the so-called free stream velocity. The
free-stream velocity amplitude 𝑢̂0 according to linear theory is given by Eq. 5.22 for
𝑧 = −ℎ:
𝜔𝑎
𝑢̂0 = (5.27)
sinh 𝑘ℎ
The flow in the wave boundary layer is generally turbulent due to the presence of
roughness elements on the bed. The water moving along the bed incurs a shear stress
on the bed. This becomes clear when imagining that – as a result of viscosity and
turbulence – the flow sticks to the wall (no-slip condition). Hence, the orbital velocity
increases from zero at the bed to the undisturbed free-stream velocity at the top of the
wave boundary layer 𝑧 = 𝛿. Because of the thin boundary layer, the velocity gradients
perpendicular to the bed are large and give rise to large stresses in the wave boundary
layer. The friction in the wave boundary layer results in dissipation of wave energy.
In coastal waters, turbulent stresses – arising from turbulent fluctuations of the ve-
locity – are much larger than viscous stresses arising from small-scale erratic move-
ments of molecules. Now think of the total horizontal velocity u and vertical velocity
w to be composed of a mean, a wave and a turbulent part, hence 𝑢 = 𝑈 + 𝑢̃ + 𝑢 ′ and
𝑤 = 𝑊 + 𝑤̃ + 𝑤 ′ . Turbulent shear stress is defined as the stress introduced when
averaging over the turbulent motion:
𝜏 (𝑧) = 𝜌𝑢 ′ 𝑤 ′ (5.28)

The detailed modelling of the wave boundary layer is a complex research field. It in-
volves the modelling of the turbulence in the wave boundary layer that is induced
by roughness of the bed. Such a wave boundary layer model gives the detailed time-
dependent velocity distribution over the wave boundary layer. Besides a purely oscil-
latory flow, a non-zero wave-averaged horizontal flow – called streaming – is found. It
was first explained by Longuet-Higgins (1953) who demonstrated that for linear waves
the streaming is directed in the wave propagation direction (see Fig. 5.26). It is there-
fore potentially important for transporting sediment onshore. The disturbance of the
wave motion due to the wave boundary layer leads to additional stresses in the wave
boundary layer when averaging over the organised wave motion. In analogy with
Eq. 5.28, these stresses are given by 𝜌 𝑢̃ 𝑤̃ (the overbar now represents averaging over
the short-wave motion). Above the wave boundary layer, this term can generally be
neglected. In Intermezzo 5.5 it is shown that the term in the wave-averaged horizontal
momentum equation responsible for the streaming is (proportional to) 𝜕 𝑢̃ 𝑤/𝜕𝑧.
̃ It acts
as a (depth-varying) force in the wave boundary layer, pushing the flow forward. The
formula found by Longuet-Higgins for the mean streaming velocity at the top of the
boundary layer is:
2
3 𝑢̂
𝑈0 = 0 (5.29)
4 𝑐
For practical purposes it is generally enough to consider the following aspects of the
wave boundary layer:
• The water moving along the bed incurs a shear stress on the bed. The orbital
motion under waves, even without the presence of a uniform current, gives a
time-varying shear stress at the bed, which can set sediment grains into motion
(see Ch. 6);
• Due to bed friction, the wave boundary layer dissipates energy from the flow
above (this is the term 𝐷𝑓 in Eq. 5.2);
• The wave-induced streaming (Eq. 5.29) should be taken into account for net-
sediment transport computations.
In Sect. 5.5.6 it is briefly indicated how the vertical structure of the mean flow (includ-
ing the wave boundary layer) can be resolved using a relatively simple eddy viscos-
ity turbulence model. Until then, we will only consider depth-integrated momentum
equations that hence contain depth-invariant quantities. For instance, we consider the
bed shear stress in a simple parameterised way, so that we do not have to bother with
distributions of the shear stress over the depth.
Bed shear stress

If we refrain from turbulence modelling, (bed) friction is a major unknown that has to
be determined using (empirical) friction laws. This introduces coefficients that need
to be calibrated, which makes the calibration of these models important.

To determine the bed shear stress, Jonsson (1967) introduced the concept of a wave
friction factor in analogy with the current friction factor (see Intermezzo 5.2). The
current friction factor relates the bed shear stress to the depth-averaged current velo-
city, whereas the wave friction factor relates the bed shear stress to the free stream
velocity. For a current only, the magnitude of the bed shear stress is 𝜏𝑐 = 𝑐𝑓 𝜌𝑈 2 (with
flow aligned with 𝑥-axis, see Intermezzo 5.2). Under waves, the bed shear stress varies
in time and reverses with the direction of the orbital velocities.
For linear waves with a free stream velocity 𝑢 = 𝑢̂0 cos 𝜔𝑡, Jonsson defined the friction
factor 𝑓𝑤 through the following formula for the magnitude of the maximum bed shear
stress:
𝜏𝑤̂ = 0.5𝜌𝑓𝑤 𝑢̂02 (5.30)
For a rough bed and turbulent flow, it is not easy to determine the friction coefficient
as it cannot be measured directly. The friction coefficient will generally depend on the
bed material and the bed forms (e.g. ripples). The following variables can be found in
expressions for 𝑓𝑤 for rough turbulent flow:
• the bed roughness 𝑘𝑠 (Nikuradse roughness) or 𝑟 of the wall; the bed roughness
represents the size of the roughness elements, for instance the grains;
• the particle excursion amplitude close to the bed 𝜉0̂ = 𝑢̂0 /𝜔 (see Eq. 5.24 and
App. A).
Intermezzo 5.2 Current-only bed shear stress
Assume a current-only situation with the depth-mean flow velocity 𝑈⃗ = (𝑈 , 𝑉 ) in

the (𝑥, 𝑦)-direction. The bottom shear stress acts in the direction of the current
and can be described by a quadratic friction law:
𝜏𝑏 = 𝜌𝑐𝑓 |𝑈⃗ |𝑈⃗ (5.31)
The friction factor 𝑐𝑓 is a dimensionless coefficient relating the bed shear stress to
the square of the velocity. Linear friction laws are rarely used since they have no
physical justification.
𝑐𝑓 can be expected to depend on the bed material and bed forms (bed roughness).
For uniform flow in a canal driven by a small slope of the mean water surface,
Chézy derived theoretically a Chézy coefficient 𝐶 = 18 log 12ℎ/𝑟, where 𝑟 is the
bottom roughness and ℎ is the water depth. The Chézy coefficient 𝐶 relates to 𝑐𝑓
as follows:
𝑔
𝑐𝑓 = (5.32)
𝐶2

Alternatives are to prescribe a Nikuradse roughness height 𝑘𝑁 or an empirical

Manning value 𝑛. The Nikuradse roughness and the Manning value can be related
to 𝑐𝑓 . The choice between these methods affects the depth dependency of 𝑐𝑓 and
can therefore have important consequences for computed flow fields.
Let us consider two cases in which 𝑢̂0 is the same, but 𝑇 differs. For the case with a
large value for 𝑇 , the value of 𝜉0̂ is larger than in the case with the lower value for 𝑇 .
Assuming an equal value for 𝑟, this means that the case with the larger value for the
wave period gives lower values for the wave friction factor. This can be understood by
considering that the boundary layer thickness varies with time. In case of a larger wave
period, more time is available to develop the boundary layer, which therefore reaches
a larger maximum thickness. Consequently, the velocity gradients in the boundary
layer are smaller, leading to a smaller maximum shear stress and friction factor.
A frequently applied formula for 𝑓𝑤 is that of Jonsson (1967), rewritten by Swart (1974)
into:
−0.194
𝜉0̂
𝑓𝑤 = exp [−5.977 + 5.213 ( ) ] (5.33a)
𝑟
𝜉0̂
𝑓𝑤 = 0.30 for ( ) < 1.59 (5.33b)
𝑟
Many simpler formulas exist, such as (Soulsby, 1994):
−0.52
𝜉0̂
𝑓𝑤 = 1.39 ( ) (5.34a)
𝑟/30
𝑓𝑤,max = 0.3 (5.34b)
For the roughness value 𝑟 the so-called Nikuradse roughness 𝑘𝑠 is often used, which is
normally set as a function of the grain size.
Equation 5.33 is shown graphically in Fig. 5.24 and compared to the simpler formula-
tion of Eq. 5.34. As expected, the friction factor 𝑓𝑤 increases if 𝜉0̂ /𝑟 decreases. The
upper limit of 𝑓𝑤 has been questioned by various researchers: some suggest that there
is no upper limit and that the friction factor remains proportional to 𝜉0̂ /𝑟.
In case of irregular waves, the near-bed orbital velocity amplitude in the above formu-
las should be based on the root-mean-square wave height and the wave orbital excur-
sion parameter near the bed on the root-mean-square wave height and peak period.
Example 5.1 presents an example of a bed shear stress computation.

0.5 Swart (1974)

0.3 Soulsby (1994)
0.2
fw [-]
2 5
ξˆ0 /r [-]
Figure 5.24: Friction factor 𝑓𝑤 as a function of the particle excursion at the bed 𝜉0̂ divided by
the bed roughness 𝑟 = 𝑘𝑠 . Equation 5.33 (after Swart, 1974, black solid line) is compared to the
simpler Eq. 5.34 (after Soulsby, 1994, gray dashed line).
Example 5.1 Bottom shear stress under waves only
Input parameters
Water depth ℎ = 3 m
Roughness height 𝑟 = 0.06 m
Wave height 𝐻 = 1.18 m
Wave period 𝑇 = 8 s
Required Bottom shear stress
Output The amplitude of the velocity near the bed can be found from linear
wave theory:
𝜔𝐻 1 𝑢̂0 𝑇
𝑢̂0 = = 1 m/s and 𝜉0̂ = = 1.27 m (5.35)
2 sinh 𝑘ℎ 2𝜋
For a value of 𝜉0̂ /𝑟 > 1.59 the friction factor equals:
−0.194
𝑓𝑤 = exp [−5.977 + 5.213 (𝜉0̂ /𝑟) ] = 0.045 (5.36)

The maximum bottom shear stress follows from:
1
𝜏𝑤̂ = 𝜌𝑓𝑤 𝑢̂02 = 22.5 N/m2 (5.37)
2
Discussion For a maximum near-bed orbital velocity of 1 m/s, the bottom shear
stress due to waves equals 22.5 N/m2 . Now consider the situation of a mean current
of 1 m/s. Assume that the roughness height and water depth are the same as above.
Using 𝐶 = 18 log 12ℎ/𝑟 = 50 m1/2 /s we find:
𝑔 2
𝜏𝑐 = 𝜌 𝑈 = 3.9 N/m2 (5.38)
𝐶2
The bed shear stress due to waves with a maximum near bed orbital velocity of
1 m/s is almost 6 times the value for the bed shear stress due to an average cur-
rent velocity of 1 m/s. This is a direct consequence of differences in boundary
layer thickness and hence near-bed velocity gradients. The wave boundary layer
is limited in thickness, whereas the current boundary layer generally occupies the
entire water column.
5.5. Wave-induced set-up and currents

5.5.1. Wave-induced mass flux or momentum
Propagating waves not only carry energy across the ocean surface, but momentum as
well. Momentum is defined as the product of mass and velocity. It can be thought
of as mass in motion or a mass transport or flux: a water particle has a mass, and if
the particle is moving, it has momentum. Momentum per unit volume can thus be
written as the product of the mass density 𝜌 and the velocity 𝑢⃗ = (𝑢𝑥 , 𝑢𝑦 , 𝑤) of the
water particles. Momentum (per unit volume) 𝜌 𝑢⃗ = (𝜌𝑢𝑥 , 𝜌𝑢𝑦 , 𝜌𝑤) is a vector quantity,
a quantity which is fully described by both magnitude and direction. The direction of
the momentum vector is the same as the direction of the velocity vector.
The total amount of wave momentum per unit surface area in the direction of wave
propagation is obtained by integration over the depth. Averaged over time this gives
(with 𝑢 the horizontal orbital velocity in the wave propagation direction):
𝜂
𝑞 = ∫ 𝜌𝑢 𝑑𝑧 (5.39)
−ℎ
There is only a contribution to the momentum from the wave trough level to the wave
crest level, since below the wave trough the velocity varies harmonically in time (see

5.5. Wave-induced set-up and currents 201
Sect. 5.4.1), giving a zero time-averaged result. If we measure the velocity at some
point above MSL, we will only record velocities during part of the wave period and all
of the recordings will be positive (and in the wave propagation direction). Between
the wave trough level and MSL, we will record velocities for a larger part of the wave
period and, although a part of the recording will be negative, the wave-averaged mean
velocity will still be positive. The momentum 𝑞 can thus be interpreted as a net flux of
mass between wave trough and wave crest associated with wave propagation.
We can compute the integral of Eq. 5.39 for a single harmonic component (non-breaking)
by substituting the velocity according to linear wave theory (see Eq. 5.22) at 𝑧 = 0 and
integrating from 𝑧 = 0 to the instantaneous surface elevation 𝜂 = 𝑎 cos 𝜔𝑡. We then
find, for the mean momentum in a plane perpendicular to the wave propagation dir-
ection per unit surface area:
𝑎 cos 𝜔𝑡
𝑎𝜔 𝑎𝜔
𝑞non-breaking =∫ 𝜌 cos 𝜔𝑡 𝑑𝑧 = 𝑎 cos 𝜔𝑡𝜌 cos 𝜔𝑡 =
0 tanh 𝑘ℎ tanh 𝑘ℎ
(5.40)
𝜌𝑎2 𝜔 𝜌𝑔𝑎2 𝐸
= =
2 tanh 𝑘ℎ 2𝑐 𝑐
This formula shows that 𝑞 is a non-linear quantity in the amplitude 𝑎. The result is
second-order accurate in the amplitude (linear wave theory is first-order accurate). Ap-
parently, wave-induced mass flux occurs even for a perfect sinusoidal orbital motion,
but is a second-order effect. In the linear, small-amplitude approximation, 𝑞 is zero
and wave propagation is merely a matter of movement of the wave form, not of mass.
In relation to net mass flux associated with wave propagation, the term Stokes drift is
often used (see Intermezzo 5.3).
Equation 5.40 is valid outside the surf zone. In the surf zone the mass flux is substan-
tially larger than outside the surf zone. It is assumed to consist of two parts, one due
to the progressive character of the waves (Eq. 5.40) and the other due to the surface
roller in breaking waves:
𝐸 𝛼𝐸𝑟
𝑞drift = 𝑞non-breaking + 𝑞roller = + (5.41)
𝑐 𝑐
In this equation, 𝐸𝑟 is the roller energy. The first part of the right-hand side is the mass
flux for non-breaking waves, whereas the second part accounts for the contribution of
the mass of the surface roller. Various authors have argued values for the factor 𝛼 in
Eq. 5.41 to be in the range of 0.22 to 2 (Nairn et al., 1990; Roelvink & Stive, 1989). Their
arguments are too complex to discuss here, so we assume that 𝛼 is in the order of 1.
In the case of a closed boundary, like a coastline, there is a zero net mass transport
through the vertical, as otherwise water would increasingly pile up against the coast.

This means that there must be a net velocity below the wave trough level to com-
pensate for the flux above the wave trough level: a return current. The cross-shore
depth-mean velocity below the wave trough level must compensate for the mass flux
perpendicular to the shore and is therefore given by:
𝑞drift,𝑥 𝑞drift cos 𝜃

𝑈below trough = − =− (5.42)
𝜌ℎ 𝜌ℎ
In breaking waves, the mass transport towards the coast between wave crest and wave
trough may be quite large, resulting in rather large seaward-directed velocities under
the wave trough level (see Fig. 5.25). The large return current in the surf zone is called
undertow.
circulation current
surf zone
Figure 5.25: The undertow is a return current below the wave trough level to compensate for the
onshore mass flux in the surf zone.
Also in the two-dimensional case of a laboratory wave flume, the same mass of water
has to return to the ‘sea’ again. In the lower part of the water column, this gives a
return flow (see Fig. 5.26). The figure also shows Longuet-Higgins streaming close to
the bed (see Sect. 5.4.3). In Fig. 5.25, we have assumed that in the surf zone the steady
Longuet-Higgins streaming may well be overridden by the undertow. The distribution
over the depth of the return current or undertow (and the streaming) can be solved
using a horizontal momentum equation (not depth-averaged!), see Sect. 5.5.6.
wave board 0.17 m/s
0.3 m
Stokes drift
return current
0.08 m/s
Longuet-Higgins streaming
wave-damping structure
Figure 5.26: Velocities averaged over a wave cycle under propagating waves in a wave flume.
Even in this case of non-breaking waves, there is a small return current. Note that the term
‘undertow’ is only used for the larger return current under breaking waves.

Appendix B provides an example of wave flume experiments of periodic and random

waves on a gently sloping beach (Stive, 1985). Figure B.2 shows the measurements of
return currents in shoaling and breaking periodic waves. In non-breaking waves there
is relatively small return current. In breaking waves, the mass transport towards the
coast between wave crest and wave trough may be quite large, resulting in rather large
seaward-directed time-averaged velocities under the wave trough level.
Intermezzo 5.3 Stokes drift
We explained mass transport from an Eulerian point of view in the text above
(by placing a measuring pole in a fixed cross-section and concluding that above
the wave trough level the recorded velocity has a non-zero time-averaged result
in the wave propagation direction). It can also be explained from a Lagrangian
point of view, viz. by following a water particle moving in its orbital motion (see
Fig. 5.20). Since the horizontal movement is in general smaller closer to the bed
(see Eq. 5.22 and Fig. 5.21), the water particle moves faster in the wave propagation
direction when it is located under the wave crest, and then it runs backward when
under the trough of the wave. As a result, particle paths are not entirely closed
orbits and there is a residual motion in the wave propagation direction over one
wave period. This residual motion is referred to as Stokes drift and gives rise to a
net mass transport in the direction of wave propagation. When we integrate the
Lagrangian mass transport over the vertical, we get the same result as the Eulerian
mass transport above the wave trough level.
The undertow is important for seaward sediment transport, because of the relatively
high offshore-directed velocity in the lower and middle part of the water column in
a zone with relatively high sediment concentrations (due to wave-breaking). The un-
dertow is thought to be responsible for the severe beach erosion during heavy storms.
Sediment transport due to return currents is also important for shallow areas that have
a deeper area between them and the coast. Examples are shallow areas on the ebb-tidal
deltas of coastal inlet systems, where the mass flux is not (entirely) compensated by
undertow, as the water can flow away at the back of these flats (where often tidal
channels are present). This is further discussed in Sect. 9.4.1.
5.5.2. Radiation stress

Newton’s second law states that the rate of change of momentum of a fluid element
equals the forces on the element. Waves can change the momentum through net ‘in-
flow’ or ‘outflow’ of momentum, either by net inflow or outflow of momentum with
the particle velocity or via a net wave-induced pressure force. Radiation stress is the
name given to the depth-integrated and wave-averaged flow (or flux) of momentum
due to waves. It was first defined by Longuet-Higgins and Stewart (1964) as the excess
momentum flux due to the presence of waves.

If there is change in wave-induced momentum flux (radiation stress) from one location
to another, wave forces act on the fluid, impacting mean water motion and levels.
These wave forces are responsible for:
• lowering the mean water level in the shoaling zone (set-down);
• raising the mean water level in the surf zone (set-up);
• driving a longshore current in the case of waves obliquely approaching the shore
(1 m/s under some conditions).
Definition of radiation stress components

The wave-induced horizontal flux of momentum through a vertical plane at a given
location consists of:
• the transfer of momentum 𝜌 𝑢⃗ through that plane, with the particle velocity nor-
mal to that plane;
• the wave-induced pressure force acting on the plane due to the wave-induced
pressure 𝑝wave in the water.
First assume that the plane under consideration is perpendicular to the wave propaga-
tion (Fig. 5.27). The wave-induced pressure 𝑝wave acts by definition normal to the
plane. Besides, at every height above the bed, the particle velocity 𝑢 transports mo-
mentum 𝜌𝑢 through the plane (per unit crest length).
Figure 5.27: Horizontal transport of wave-induced momentum through a vertical plane of unit
width perpendicular to the wave propagation direction.
Now consider a coordinate system according to Fig. 5.28. The wave propagates at an
angle with the 𝑥-axis. The particle velocity has a component 𝑢𝑥 in the 𝑥-direction
and a component 𝑢𝑦 in the 𝑦-direction. The particle velocity 𝑢𝑥 transports both 𝑥-
momentum 𝜌𝑢𝑥 and 𝑦-momentum 𝜌𝑢𝑦 . The transport of 𝑥-momentum by the particle
velocity 𝑢𝑥 through a vertical plane perpendicular to the 𝑥-axis (per unit time and per
unit area) is (𝜌𝑢𝑥 )𝑢𝑥 and the transport of 𝑦-momentum through the plane is (𝜌𝑢𝑦 )𝑢𝑥 .
The momentum fluxes can be seen in the left part of Fig. 5.29.
Figure 5.29 visualises the momentum transport and radiation stress components in
𝑥-direction and 𝑦-direction at a certain point in 𝑥, 𝑦-space. The momentum transport
through the entire plane per unit crest length is obtained by integration over the depth

orbital motion
Figure 5.28: Coordinate system and velocity vector.
Figure 5.29: Schematic of the momentum transport and radiation stress components at a cer-
tain point in 𝑥, 𝑦-space for obliquely incident waves. The radiation stresses are obtained by
integration of the momentum transport over the water column and averaging over time.
from bottom to instantaneous water surface. Time-averaging yields, for the total wave-
averaged transport of 𝑥-momentum in the 𝑥-direction or the radiation stress 𝑆𝑥𝑥 :
𝜂 𝜂
𝑆𝑥𝑥 = ∫ (𝜌𝑢𝑥 ) 𝑢𝑥 𝑑𝑧 + ∫ 𝑝wave 𝑑𝑧 (5.43)
⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟
−ℎ0 ⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟
−ℎ0
part due to advection by pressure part
horizontal particle velocity
𝑆𝑥𝑥 acts normal to the considered plane and is therefore a normal component of the
radiation stress. It is equivalent to a normal stress acting in the 𝑥-direction.
The radiation stress component 𝑆𝑥𝑦 (the shear component of the radiation stress) is
defined as the transport of 𝑥-momentum in the 𝑦-direction. It acts as a shear stress
on the plane (it is directed in the 𝑥-direction, and works on the plane normal to the
𝑦-direction) and is given by:
𝜂
𝑆𝑥𝑦 = ∫ (𝜌𝑢𝑥 ) 𝑢𝑦 + 𝜏𝑥𝑦 𝑑𝑧 (5.44)
−ℎ0

The shear stress due to waves is zero for an irrotational ideal fluid, so that we have:
𝜂
𝑆𝑥𝑦 = ∫ (𝜌𝑢𝑥 ) 𝑢𝑦 𝑑𝑧 (5.45)
−ℎ0
It consists of a contribution due to advection by the horizontal orbital velocity only.

Note that for the special case of normally incident waves, the x-direction is the wave
propagation direction, and 𝑢𝑦 and hence 𝑆𝑥𝑦 are zero.
Considering a plane normal to the 𝑦-direction, we find for the momentum fluxes in
the 𝑦-direction, the normal component:
𝜂 𝜂
𝑆𝑦𝑦 = ∫ (𝜌𝑢𝑦 ) 𝑢𝑦 𝑑𝑧 + ∫ 𝑝wave 𝑑𝑧 (5.46)
⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟
−ℎ0 ⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟
−ℎ0
part due to advection by pressure part
horizontal particle velocity
and the shear component:
𝜂
𝑆𝑦𝑥 = ∫ (𝜌𝑢𝑦 ) 𝑢𝑥 𝑑𝑧 (5.47)
−ℎ0
In the special case that the 𝑥-direction is the wave propagation direction 𝑢𝑦 = 0, 𝑆𝑦𝑦
reduces to the pressure part and 𝑆𝑦𝑥 = 𝑆𝑥𝑦 = 0.
Radiation stress formulas using linear wave theory

By using linear (first-order) wave theory, formulas for the radiation stress can be
obtained that are valid to second order. The complete derivation can be found in
Holthuijsen (2007). Intermezzo 5.4 gives an excerpt of this derivation.
Intermezzo 5.4
Derivation of (normal) radiation stress from linear wave theory
It is easiest to start from a wave propagating in positive 𝑥-direction so that 𝑢𝑥 = 𝑢

and 𝑢𝑦 = 0. If we substitute 𝑢 = 𝑢̂ cos(𝜔𝑡 − 𝑘𝑥), with 𝑢̂ according to Sect. 5.4.1,
in the particle velocity part of Eq. 5.43, we find (see Holthuijsen (2007) for the
derivation):
𝜂̄ 0
𝑆𝑥𝑥, horizontal particle velocity = ∫ (𝜌𝑢 2 ) 𝑑𝑧 ≈ ∫ 𝜌𝑢 2 𝑑𝑧 = 𝑛𝐸 (5.48)
−ℎ0 −ℎ0
where:

𝑆𝑥𝑥, hor. part. vel. radiation stress in the wave propagation direction N/m
due to advection of momentum by the horizontal
orbital motion
𝑛 ratio of group velocity and phase velocity −
𝐸 wave energy in the water column per m2 J/m2
In the case of irregular waves, this equation can be applied using 𝐸 = 1/8𝜌𝑔𝐻𝑟𝑚𝑠2 .
For the mass flux, we found a zero contribution for every level below the wave
trough because 𝑢 = 0. The momentum flux, however, is non-zero for the entire
water depth, since 𝑢 2 ≠ 0. It varies in principle over the depth. Only in shallow
water, where the horizontal orbital velocity is uniformly distributed, the particle
velocity part of the radiation stress is uniformly distributed.
The pressure for any level below the trough would give a zero time-averaged result
according to linear theory. There is, however, a contribution to the time-averaged
wave-induced pressure 𝑝wave due to the vertical flux of momentum by the vertical
fluid motion. This may be thought of as the vertical oscillatory fluid motion, help-
ing carry the weight of the water column, and is given by −𝜌𝑤 2 , in which 𝑤 is the
vertical orbital velocity. The contribution to the radiation stress can be found by
substituting the linear equation for 𝑤 and integrating to the mean water level. We
find:
𝜂̄ 0
𝑆𝑥𝑥, pressure, 1 = −∫ 𝜌𝑤 2 𝑑𝑧 ≈ − ∫ 𝜌𝑤 2 𝑑𝑧 = (𝑛 − 1)𝐸 (5.49)
−ℎ0 −ℎ0
Furthermore, there is a contribution to the radiation stress due to the pressure

fluctuations between wave trough and wave crest level. In Sect. 5.4.2 we assumed
that the pressure between wave trough and crest level fluctuates as 𝑝 ̃ = 𝜌𝑔𝜂. This
gives the following net contribution to the radiation stress:
𝜂̄
1 1
𝑆𝑥𝑥, pressure,2 = ∫ 𝜌𝑔𝜂 𝑑𝑧 = 𝜌𝑔𝜂2 = 𝐸 (5.50)
0 2 2
This gives for the pressure part of the radiation stresses:
𝑆𝑥𝑥, pressure = (𝑛 − 1)𝐸

⏟⏟⏟⏟⏟⏟⏟⏟⏟ + 1/2𝐸
⏟ = (𝑛 − 1/2) 𝐸 (5.51)
contribution due to (depth-varying) contribution at top
vertical momentum flux of water column
The term 𝑤 2 varies in principle with the water depth, whereas the other part (from
Eq. 5.50) is located at the top of the water column. In shallow water (𝑛 = 1) the
vertical orbital velocities are zero and the pressure part of the radiation stresses

reduces to a contribution 𝑆𝑥𝑥,pressure = 1/2𝐸 at the top of the water column. In deep
water, where 𝑛 = 1/2, the pressure component to the radiation stress is (𝑛 − 1/2)𝐸.
The total radiation stress in the wave propagation direction (𝑥-direction) is now:
𝑆𝑥𝑥 = 𝑆𝑥𝑥, pressure + 𝑆𝑥𝑥, horizontal particle velocity = (𝑛 − 1/2)𝐸 + 𝑛𝐸 (5.52)
The stresses normal to the wave propagation direction 𝑆𝑦𝑦 consist of the pressure
part (𝑛 − 1/2)𝐸 only. The shear stresses are zero.
From Intermezzo 5.4 – or more specifically from Eq. 5.52 – we conclude that:
1. The pressure part of the radiation stress is equal to (𝑛 − 1/2)𝐸. Since pressure is
a scalar, this term is part of radiation normal stresses in all directions.
2. The magnitude of the advective part of the radiation stress, the part due to trans-
port of momentum by the particle velocity, is 𝑛𝐸. It is by definition in the wave
propagation direction.
With this result, we can now deduce the more general formulas for the radiation stress
components for waves travelling in a direction 𝜃 relative to the positive 𝑥-direction
(see Fig. 5.28). If the 𝑥-axis is not the propagation direction we have 𝑢⃗ = (𝑢𝑥 , 𝑢𝑦 ) =
(𝑢 cos 𝜃, 𝑢 sin 𝜃) and with Eqs. 5.43, 5.45 and 5.47, we find:
1
𝑆𝑥𝑥 = (𝑛 − + 𝑛 cos2 𝜃) 𝐸 (5.53a)
2
1
𝑆𝑦𝑦 = (𝑛 − + 𝑛 sin2 𝜃) 𝐸 (5.53b)
2
𝑆𝑥𝑦 = 𝑆𝑦𝑥 = 𝑛 cos 𝜃 sin 𝜃𝐸 (5.53c)
In the situation of an alongshore uniform coast, the positive 𝑥-direction is generally

taken as the shore-normal direction and the positive 𝑦-direction is parallel to the coast-
line. The angle of incidence of the wave 𝜑 with respect to the depth contours is now
equal to the angle 𝜃 with the positive 𝑥-axis. For that situation, the radiation stress
components are sketched in Fig. 5.30.

coastline
Figure 5.30: Radiation stress components according to linear wave theory. Situation sketch for
alongshore uniform coast with depth contours parallel to 𝑦-axis. ‘A’ indicates the pressure part,
‘B’ indicates the part due to advection by the horizontal particle velocity.
Note that if the 𝑥-axis coincides with the wave propagation direction (𝜃 = 0), the above
Eqs. 5.53a to 5.53c reduce to:
𝑆𝑥𝑥 = (2𝑛 − 1/2) 𝐸 (5.54a)
𝑆𝑦𝑦 = (𝑛 − 1/2) 𝐸 for waves propagating in the 𝑥-direction (5.54b)
𝑆𝑥𝑦 = 𝑆𝑦𝑥 = 0 (5.54c)
This is in accordance with the earlier findings that:

• The normal stresses perpendicular to the direction of wave propagation consist
of the pressure part (𝑛 − 1/2)𝐸 only;
• The normal stresses in the direction of wave propagation consist of the pressure
part (𝑛 − 1/2)𝐸 plus the part due to momentum transfer by the horizontal orbital
motion 𝑛𝐸;
• The shear stresses are zero if the 𝑥-direction is the direction of wave propagation.
In deep water, where 𝑛 = 1/2, the pressure component to the radiation stress is (𝑛 −
1/2)𝐸 = 0 and the fluid particle part becomes 𝑛𝐸 = 1/2𝐸 in the direction of wave propaga-
tion. Thus, in the case of wave propagation in the 𝑥-direction, we find 𝑆𝑥𝑥 = 1/2𝐸, and
𝑆𝑦𝑦 = 0. In shallow water, on the other hand, where 𝑛 = 1, it follows that 𝑆𝑥𝑥 = 3/2𝐸,
and 𝑆𝑦𝑦 = 1/2𝐸. Here both contributions add to the total radiation stress 𝑆𝑥𝑥 and 𝑆𝑦𝑦
is equal to the pressure part. The radiation (normal) stress in shallow water is clearly

larger than in deep water. Moreover, it has two components in shallow water: parallel
and perpendicular to the direction of wave propagation.
Radiation stress in breaking waves

In breaking surf zone waves, the effect of the surface roller will be to delay the mo-
mentum release from the wave-breaking. In practice this means that the equations for
the radiation stresses will contain a term due to roller energy too. Different models
exist for this term, enhancing either the pressure part or advection part of the radi-
ation stress or both. Svendsen (1984) proposes an additional contribution to the radi-
ation stress in the propagation direction due to the velocity in the roller by an amount
𝑞roller 𝑐 = 𝛼𝐸𝑟 (see Eq. 5.41). This contribution is concentrated near the water surface,
where the wave-breaking takes place. Under the assumption of shallow water, the ra-
diation stress below the wave trough level is dominated by the part due to advection
of momentum by the horizontal orbital motion (verify this from Intermezzo 5.4). Since
in shallow water the horizontal orbital velocity is uniformly distributed, the advective
part of the radiation stress is uniformly distributed as well.
5.5.3. Wave-induced forces

The above equations describe the wave radiation stresses acting in a vertical plane of
water. Horizontal gradients in the radiation stresses give rise to a net wave-induced
force on the water in a particular direction.
The net force in the 𝑥-direction is described by (Fig. 5.31):
𝜕𝑆𝑥𝑥 𝜕𝑆𝑥𝑦
𝐹𝑥 = − ( + ) (5.55)
𝜕𝑥 𝜕𝑦
where the first term represents the effect of variations in the 𝑥-directed radiation nor-
mal stresses and the second term the effect of variations in the 𝑦-direction of the 𝑥-
directed radiation shear stress.
In coastal engineering practice, it is common to work with alongshore and cross-shore
orientated axes. The 𝑦-axis is defined parallel to the shoreline, while the 𝑥-axis is
perpendicular to the shoreline. 𝐹𝑥 is thus the force in the cross-shore direction.
The second (shear) term in Eq. 5.55 is zero for an alongshore uniform coastline (no
gradients in the 𝑦-direction). We then have:
𝑑𝑆𝑥𝑥
𝐹𝑥 = − for an alongshore uniform coast (5.56)
𝑑𝑥

Figure 5.31: Wave-induced forces. An increase of momentum transport (increase in radiation

stress) in 𝑥- or 𝑦-direction is equivalent to a loss of momentum, hence, to exerting an opposite
force on the water body.
In the alongshore direction the force is:
𝜕𝑆𝑦𝑦 𝜕𝑆𝑦𝑥
𝐹𝑦 = − ( + ) (5.57)
𝜕𝑦 𝜕𝑥
The first term in Eq. 5.57 can be non-zero if gradients in wave height occur along the
coast. For an alongshore uniform coast this term is zero such that Eq. 5.57 reduces to:
𝑑𝑆𝑦𝑥
𝐹𝑦 = − for an alongshore uniform coast (5.58)
𝑑𝑥
Remember that 𝑆𝑦𝑥 – the radiation shear stress on a plane perpendicular to the 𝑥-axis
– represents the transport of 𝑦-momentum in the 𝑥-direction. A cross-shore gradient
herein gives a net force in the 𝑦-direction.
Variations in the radiation stresses occur due to changes in 𝑛, 𝐸 or 𝜃 . Further offshore,
the wave-induced forces are relatively small, but in the nearshore zone large forces
occur due to wave transformation as a result of large gradients in water depth.
In the shoaling region, the wave height and hence wave energy increase up to the edge
of the surf zone, from where the wave height and wave energy decrease again. In the
shoaling region the value of 𝑛 gradually increases from its deep-water value 𝑛 = 1/2,
to its shallow-water value 𝑛 = 1. The wave angle 𝜃 gradually decreases from deep
water to shallow water. The combined result is a positive gradient of 𝜕𝑆𝑥𝑥 /𝜕𝑥 and
an offshore-directed force in the shoaling region, and a negative gradient of 𝜕𝑆𝑥𝑥 /𝜕𝑥
and an onshore-directed force in the surf zone. Section 5.5.4 shows how this results
in wave set-down in the shoaling region and wave set-up in the surf zone. We will
further see that the force in the 𝑦-direction is zero outside the surf zone but non-zero

in the surf zone, where it drives a longshore current. It is of course only present in
situations where the waves approach the coastline under a certain angle (such that
there is a component of wave orbital velocity in the alongshore direction).
We have expressed the depth-integrated wave forces in terms of the radiation stresses,
which is quite common in coastal engineering applications. Other equivalent expres-
sions can also be found in model descriptions (Intermezzo 5.5).
Intermezzo 5.5 Wave forces in momentum equations
In momentum equations that are averaged over the wave motion, wave forces
appear. These wave forces have two contributions: due to pressure fluctuations
and due to momentum transfer by the particle velocity. They may be formulated
in terms of radiation stresses (as in Eq. 5.55-Eq. 5.58) or they may equivalently be
expressed in terms of the wave orbital motion and wave-induced pressure. In the
latter case the wave forces in three dimensions may be expressed as:
𝑅𝑥 𝜕 ⟨𝑢̃𝑥 𝑢̃𝑥 ⟩ 𝜕 ⟨𝑢̃𝑦 𝑢̃𝑥 ⟩ 𝜕 ⟨𝑤̃ 𝑢̃𝑥 ⟩ 1 𝜕 ⟨𝑝wave ⟩

=− − − − (5.59a)
𝜌 𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜌 𝜕𝑥
𝑅𝑦 𝜕 ⟨𝑢̃𝑥 𝑢̃𝑦 ⟩ 𝜕 ⟨𝑢̃𝑦 𝑢̃𝑦 ⟩ 𝜕 ⟨𝑤̃ 𝑢̃𝑦 ⟩ 1 𝜕 ⟨𝑝wave ⟩

=− − − − (5.59b)
𝜌 𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜌 𝜕𝑦
𝑅𝑧 𝜕 ⟨𝑢̃𝑥 𝑤⟩
̃ 𝜕 ⟨𝑢̃𝑦 𝑤⟩
̃ 𝜕⟨𝑤̃ 𝑤⟩
̃ 1 𝜕 ⟨𝑝wave ⟩
=− − − − (5.59c)
𝜌 𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜌 𝜕𝑧
Note that the wave forces 𝑅𝑖 are not integrated over the depth. The brackets ⟨ ⟩ de-
note time-averaging over the short-wave period. The wave orbital motion is now
denoted with (𝑢̃𝑥 , 𝑢̃𝑦 , 𝑤),
̃ with the tilde to distinguish the wave motion from the
mean motion. Note the pressure contribution due to the non-zero time-averaged
wave pressure and the horizontal momentum transfer terms in 𝑅𝑥 and 𝑅𝑦 that are
directly related to the radiation stresses. In 𝑅𝑧 we recognise the term −𝜌 𝑤̃ 2 that
contributed to Eq. 5.49. We can also recognise the term 𝜕⟨𝑢̃ 𝑤⟩/𝜕𝑧
̃ that was said to
act, in the wave boundary layer, as a horizontal force pushing the flow forward
(causing streaming, Sect. 5.4.3).
5.5.4. Cross-shore balance: wave set-up and set-down

Wave forces have an effect on the mean flow; they induce mean water level variations
(set-down, set-up) and mean currents (a longshore current in the case of waves ob-
liquely approaching the shore).

First we consider the wave force in the cross-shore direction (given by Eq. 5.55). For
illustration purposes, we consider the simplified situation of a long-crested wave nor-
mally incident to straight and parallel depth contours (parallel to the 𝑦-axis, 𝜃 = 𝜑 = 0).
This means that there are no gradients in the alongshore direction: the situation is
alongshore uniform (all 𝑦-derivatives are zero) and the wave force reduces to Eq. 5.56.
Cross-shore mass balance (alongshore uniform coast)

In a stationary case, the cross-shore current averaged over the entire water column must
be zero (since water neither piles up higher and higher against the coast nor flows
towards deeper water). At every point in the cross-shore profile, the onshore- directed
mass flux near the water surface is therefore compensated by an offshore-directed
return current at lower elevations, such that the net depth-averaged flow through each
cross-section is zero (see Fig. 5.25). The offshore-directed depth-mean velocity under
the wave trough level can be found from Eq. 5.42.
Cross-shore momentum balance (alongshore uniform coast)

The wave force is determined by the cross-shore gradient of 𝑆𝑥𝑥 . The magnitude of
the radiation stress 𝑆𝑥𝑥 = (2𝑛 − 1/2)𝐸 in the wave propagation direction depends on
the wave height, the water depth and the wavelength. Based on energy conservation,
Sect. 5.2.2 showed that the wave energy (wave heights) tends to increase when the
waves approach the surf zone (after a small initial decrease). Since 𝑛 increases in inter-
mediate water depths, 𝑆𝑥𝑥 increases in the shoaling region. Further offshore in really
deep water, the radiation stress 𝑆𝑥𝑥 is constant. An increasing 𝑆𝑥𝑥 in the landward
direction means that, at a water column, a resulting force due to radiation stresses is
acting in the seaward direction (see Figs. 5.31 and 5.32). A (small) difference in water
level at both sides of the water column (lower towards the coast) ensures that equilib-
rium of forces is achieved again. This phenomenon is called wave set-down, which
means that outside the breaker zone in intermediate water depths, the water level
at the landward side of a water column is a bit lower than at the seaward side (see
Figs. 5.32 and 5.33, top figure). Inside the surf zone, the magnitude of 𝑆𝑥𝑥 decreases
rapidly due to wave-breaking while moving towards the waterline. The decrease in
𝑆𝑥𝑥 is equivalent to a force in the landward direction (see Fig. 5.32). To achieve equilib-
rium again, the water level at the landward side of the column should be higher than
at the seaward side (wave set-up, see Fig. 5.32) creating a seaward-directed pressure
force (see Fig. 5.33, upper plot).
The lower figure of Fig. 5.33 illustrates the cross-shore balance of momentum between
two (arbitrary) points 1 and 2. The net pressure force (per unit alongshore distance) is
𝑃𝑥 Δ𝑥 = −𝜌𝑔ℎ 𝑑𝜂/𝑑𝑥 Δ𝑥. It consists of:
• the hydrostatic force 1/2𝜌𝑔ℎ2 at point 1; minus
• the hydrostatic force 1/2𝜌𝑔(ℎ + 𝑑ℎ/𝑑𝑥 Δ𝑥)2 ≈ 1/2𝜌𝑔ℎ2 + 𝜌𝑔ℎ 𝑑ℎ/𝑑𝑥 Δ𝑥 at point
2; minus

set-up
set-down
point of breaking
shoaling zone surf zone
Figure 5.32: In the shoaling zone, 𝑆𝑥𝑥 increases in landward direction, resulting in a wave force
acting in seaward direction and a lowering of the mean water level towards the breaking point.
Wave-breaking in the surf zone causes 𝑆𝑥𝑥 to decrease in landward direction, such that an on-
shore wave forces raises the water level towards the shore (set-up).
set-up
set-down
1 2
surf zone
shoaling
zone
1 2
Figure 5.33: The control volume for a simple cross-shore momentum balance between two loc-
ations 1 and 2. The radiation stress is 𝑆𝑥𝑥,1 at location 1 and 𝑆𝑥𝑥,2 at location 2. The triangles and
quadrangle indicate the pressure forces. If locations 1 and 2 are in the surf zone, the onshore-
directed wave force (onshore since 𝑆𝑥𝑥,1 > 𝑆𝑥𝑥,2 ) is balanced by a net offshore pressure force
(through a raising of the water level towards the coast).

• the horizontal component ≈ 𝜌𝑔ℎ(𝑑ℎ0 /𝑑𝑥 Δ𝑥) of the hydrostatic force along the
bottom.
The equilibrium between the radiation stress gradient and pressure term due to the
water level slope (see Fig. 5.34) yields the following first-order differential equation:
𝑑𝑆𝑥𝑥 𝑑 𝜂̄ 𝑑 𝜂̄
𝐹𝑥 = − = 𝜌𝑔ℎ = 𝜌𝑔 (ℎ0 + 𝜂)̄ (5.60)
𝑑𝑥 𝑑𝑥 𝑑𝑥
where:
𝑥 co-ordinate axis pointing from land to sea m

ℎ0 still-water depth at point 𝑥 (in absence of waves) m
𝜂 wave-induced water level set-up at point 𝑥 m
Figure 5.34: The force balance between wave force and pressure force reads 𝐹𝑥 + 𝑃𝑥 = 0. In the
surf zone 𝐹𝑥 is positive (landward-directed) and 𝑃𝑥 is negative (seaward-directed).
Eq. 5.60 is the depth-integrated cross-shore momentum equation for a stationary along-
shore uniform situation. This balance in the cross-shore direction holds both for ob-
liquely and normally incident waves. We have not included bottom friction in this
equation, because it is generally assumed to be an order of magnitude smaller than
the other terms in the equation6 . Furthermore, the Coriolis term is assumed to be
small compared to the wave forcing. The equation is valid inside and outside the surf
zone under the conditions listed above.
Note that inside (outside) the surf zone the term 𝑑𝑆𝑥𝑥 /𝑑𝑥 is negative (positive) yielding
a positive (negative) water level gradient (𝑑 𝜂/𝑑𝑥),
̅ see Figs. 5.32 and 5.34. Therefore,
the still water surface becomes lower in the shoaling region (set-down). Moving inside
the surf zone towards the waterline, the still water surface becomes higher (set-up). In
order to calculate the wave set-up (and set-down), we need to assess the radiation
stress terms inside and outside the surf zone and integrate Eq. 5.60 with respect to 𝑥.
6
Apotsos et al. (2007) showed that this assumption may result in underprediction of the set-up. For
simplicity, however, bottom stress is often neglected, also in this book.

With Eq. 5.53a we can write Eq. 5.60 as:
𝑑 1 𝑑 𝜂̄
− [(𝑛 − + 𝑛 cos2 𝜃) 𝐸] = 𝜌𝑔ℎ (5.61)
𝑑𝑥 2 𝑑𝑥
The spatial wave energy variation must be solved from the conservation of energy,
Eq. 5.2, which reduces to 𝑑(𝐸𝑐𝑔 cos 𝜃)/𝑑𝑥 = −𝐷𝑤 under the present assumptions.
Simple model for wave set-down in the case of normally incident waves
For waves normally incident to an alongshore uniform coast (𝜃 = 𝜑 = 0) and under the
assumption of shallow water (𝑛 = 1), the momentum balance Eq. 5.61 can be written
as:
𝑑 3 𝑑 𝜂̄
− ( 𝐸) = 𝜌𝑔ℎ (5.62)
𝑑𝑥 2 𝑑𝑥
Section 5.2.2 gave us the following energy balance for a normally incident shoaling
wave:
𝐻2 𝑐𝑔,1
= (5.63)
𝐻1 𝑐𝑔,2
√
Amongst others, Longuet-Higgins and Stewart (1962) derived a formula for the set-
down for regular waves by integration of Eq. 5.62 using the energy balance of shoaling
waves Eq. 5.63. The integration constant is determined from 𝜂 = 0 in deep water. It
can be shown that, in the shallow-water approximation, the set-down then equals:
1 𝐻2
𝜂=− (5.64)
16 ℎ
For waves propagating without dissipation towards the shore, the wave height in-
creases and the water depth decreases. Hence, the set-down gradually increases in
magnitude to a maximum just outside the breaker zone, at the point where the waves
start breaking. At the breaker point the set-down equals:
2
1 𝐻𝑏 1
𝜂𝑏 = − = − 𝛾 𝐻𝑏 (5.65)
16 ℎ𝑏 16
where:
𝜂𝑏 water level change at the point of wave-breaking m

𝛾 wave-breaking index (𝐻𝑏 /ℎ𝑏 ) −
𝐻𝑏 wave height at point of breaking m
ℎ𝑏 still water depth at point of breaking m

Thus, with 𝛾 put equal to 0.8 the set-down at the point of breaking is 4 % of the local
water depth.
Analytical model for set-up in the case of normally incident waves

Inside the surf zone, the dissipation due to wave-breaking needs to be included in
the energy balance. In the following, an analytical equation for wave set-up is de-
rived using a very simple dissipation model and assuming normally incident waves.
The derivation considers equilibrium of forces for the entire breaker or surf zone (see
Fig. 5.35).
A simple model for the energy dissipation due to wave-breaking assumes that the wave
height everywhere in the surf zone is proportional to the local water depth: every-
where in the surf zone 𝐻 = 𝛾 ℎ (see Eq. 5.19). If we further use the shallow-water
approximation (𝑛 = 1), then it follows that 𝑆𝑥𝑥 = 3/2𝐸 is decreasing from the breaker
line towards the coastline. Following similar reasoning as in the previous section, this
causes landward-directed resulting forces acting on a water column. This will be bal-
anced by a water level increase (in the landward direction) over the water column. If
we substitute 𝑆𝑥𝑥 = 3/2𝐸 = 3/16𝜌𝑔𝐻 2 = 3/16𝜌𝑔𝛾 2 ℎ2 in Eq. 5.62, we find the following
balance:
𝑑 3 2 2 𝑑 𝜂̄
− [ /16𝛾 ℎ ] = ℎ (5.66)
𝑑𝑥 𝑑𝑥
and thus:
𝑑 𝜂̄ 𝑑ℎ
= −3/8𝛾 2 (5.67)
𝑑𝑥 𝑑𝑥
Substituting ℎ = ℎ0 + 𝜂 in Eq. 5.67 gives:
𝑑 𝜂̄ 3/8𝛾 2 𝑑ℎ0
=− 2
(5.68)
𝑑𝑥 (1 + 3/8𝛾 ) 𝑑𝑥
Hence, because the water depth decreases in the surf zone (𝑑ℎ0 /𝑑𝑥 < 0), the set-up
increases (𝑑𝜂/𝑑𝑥 > 0). With a constant bottom slope and a constant breaker index, the
water level slope inside the breaker zone is constant.
Equation 5.67 – or equivalently Eq. 5.68, which gives a bit more complex math – can be
integrated from the breaker point to the water line to get the wave set-up throughout
the surf zone (using 𝜂 = 𝜂𝑏 for ℎ = ℎ𝑏 to find the integration constant 𝐶):
3
𝜂 = − 𝛾 2 ℎ + 𝐶 = 𝜂𝑏 + 3/8𝛾 2 (ℎ𝑏 − ℎ) (5.69)
8

At the shoreline, the water depth ℎ is zero and we find for the set-up at the water line:
𝜂shore = 𝜂𝑏 + 3/8𝛾 2 ℎ𝑏 = 𝜂𝑏 + 3/8𝛾 𝐻𝑏 (5.70)

The water level difference Δ𝜂 between the breaker line and the point of maximum
water level rise (wave set-up) equals:
3
Δ𝜂 = 𝛾 𝐻𝑏 (5.71)
8
With 𝛾 = 0.8 this amounts to 0.3𝐻𝑏 . Wave set-up can thus be quite significant, as is
further illustrated in Example 5.2. Since the maximum set-down at the breaker line
is 1/16𝛾 𝐻𝑏 , the maximum wave set-up relative to the still water level is 5/16𝛾 𝐻𝑏 (see
Fig. 5.35).
max. max.
set-down set-up
Figure 5.35: Set-down and set-up and equilibrium of forces in the entire breaker zone.
Example 5.2 Maximum wave set-up relative to still water level
Input parameters
Offshore wave height 𝐻0 = 5 m
Wave period 𝑇 = 12 s
Offshore wave direction 𝜑0 = 0° (normally incident)
Breaker index 𝛾 = 0.7
Required Maximum wave set-up relative to the still water level
Output The first step is to compute the wave height at the breaker line. 𝐻𝑏 de-
pends on shoaling and (though not in this case) refraction. Computer programs
can be used, or linear wave theory using the following iterative steps (for an arbit-
rary wave angle):

1. Guess a breaker depth, ℎ𝑏 and compute ℎ𝑏 /𝐿0

2. Use Table A.3 in App. A to determine the shoaling coefficient 𝐾𝑠ℎ and the
ratio of wave speeds 𝑐/𝑐0
3. Determine the angle 𝜑 using sin 𝜑 = 𝑐/𝑐0 sin 𝜑0
4. Compute the wave height at water depth ℎ𝑏 from 𝐻 = 𝐻0 𝐾𝑠ℎ √(cos 𝜑0 / cos 𝜑)
5. Check whether 𝐻 /ℎ𝑏 = 𝛾 = 0.7. If yes: o.k.; if no: return to step 1 with a
better guess of ℎ𝑏 .
In this case of normally incident waves, step 3 can be omitted. It follows that
𝐻𝑏 = 5.5 m. The maximum wave set-up then becomes 5/16𝛾 𝐻𝑏 = 1.2 m above still
water level.
Conclusion If we assume an average beach slope of 1:50, then almost 60 m of

beach width is ‘lost’ by the effect of wave set-up.
Summary
• In the shoaling zone, 𝑆𝑥𝑥 increases in the positive 𝑥-direction (the cross-shore
direction); the positive gradient in 𝑆𝑥𝑥 (𝜕𝑆𝑥𝑥 /𝜕𝑥 > 0) is equivalent to a force
𝐹𝑥 directed in the offshore direction. This is compensated for by an onshore-
directed pressure force due to a set-down (lowering of the water level). The
set-down can be approximated as 1/16𝛾 𝐻𝑏 ;
• In the surf zone, 𝑆𝑥𝑥 decreases in the positive 𝑥-direction; the resulting negative
gradient in 𝑆𝑥𝑥 (𝜕𝑆𝑥𝑥 /𝜕𝑥 < 0) is equivalent to a force 𝐹𝑥 directed in the onshore
direction. This is balanced by an offshore-directed pressure force due to wave
set-up (raising of the water level towards the water line). The maximum wave
set-up relative to the still water level is 5/16𝛾 𝐻𝑏 .
Concluding remarks
• The wave height increase in the shoaling zone as well as the wave height decay
in the surf zone can be predicted reasonably well using a wave action or energy
balance. When assuming (as we did) that inside the surf zone local wave height
and local water depth are directly related, the build up of the wave set-up starts
at the (first) point of breaking. In reality, there is a delay in the transfer of mo-
mentum from the wave motion to the mean flow. This means that the set-up
(and also the start of the build-up of the longshore current) is shifted in the land-
ward direction. The delay may be caused by the temporary storage of energy
and momentum in surface rollers of breaking waves. In this way the dissipation
process is delayed, shifting the region of wave set-up in the shoreward direction.
This temporary storage can be modelled using a roller model and a roller energy
balance;

• Waves are irregular. For the computation of the wave set-up for irregular waves,
the root-mean-square wave height 𝐻𝑟𝑚𝑠 is generally applied;
• In an irregular wave field, wave heights vary at the wave group scale (Sect. 5.8.2).
That means that the magnitude of the radiation stress varies on the wave group
scale as well. This time-varying force generates water level fluctuations with a
timescale much longer than the wave period (long waves);
• Waves generally approach the coastline under a (small) angle that decreases in
smaller water depths. Therefore 𝑆𝑥𝑥 , the cross-shore gradients of 𝑆𝑥𝑥 and the
set-up are smaller than for normally incident waves.
5.5.5. Alongshore balance: longshore current

In the alongshore direction, the transfer of momentum from the wave motion to the
mean flow gives rise to a longshore current. Note that besides this wave-induced
(due to wave dissipation in the breaker zone) current, tidal and wind forces can also
generate a current along the coast. These are discussed in Sect. 5.6 and Sect. 5.7.2.
The longshore current (velocity magnitude and cross-shore distribution) is an import-
ant input parameter in longshore sediment transport computations.
Alongshore momentum balance (straight, parallel depth contours)

We consider again the 2D situation (see Fig. 5.4) of a long-crested wave obliquely in-
cident to an alongshore uniform coast (𝜃 = 𝜑 and all 𝑦-derivatives are zero). The
cross-shore rate of variation of the shear component of the radiation stress 𝑆𝑦𝑥 acts
as a driving force (Eq. 5.58). In the cross-shore direction, the balancing force was sup-
plied by a hydraulic pressure gradient. However, for an infinitely long uninterrupted
coastline, no such hydraulic pressure gradient can develop in the alongshore direction.
The counterforce, which restores the equilibrium, therefore must be supplied by bed
shear stresses that develop when a longshore current is generated. The bottom shear
stress restrains the current and is non-zero only in the presence of a current. Note that
in the cross-shore direction the bed shear stress associated with the mean current was
assumed to be small, as compared to the pressure force.
We only consider the stationary situation. The alongshore component of the mo-
mentum balance for a steady state and alongshore uniformity can be written as:
𝑑𝑆𝑦𝑥
𝐹𝑦 = − = 𝜏 𝑏,𝑦 (5.72)
𝑑𝑥
The balance between the driving force and the resisting or retarding force is shown in
Fig. 5.36.

surf zone
Figure 5.36: Forces acting on the water column (plan view). The longshore current direction is
from left to right in the figure.
Wave force in alongshore uniform situation

Let us first assess 𝐹𝑦 = −𝑑𝑆𝑦𝑥 /𝑑𝑥 in deeper water, i.e. outside the breaker zone. In
Sect. 5.5.2 we found 𝑆𝑦𝑥 = 𝐸𝑛 sin 𝜑 cos 𝜑. The wave angle changes as a function of 𝑥,
but does not vary with 𝑦. The latter implies that Snell’s law for regular waves is valid
(Eq. 5.8): sin 𝜑/𝑐 = constant. Conservation of energy (Eq. 5.2) requires 𝑑(𝐸𝑐𝑔 cos 𝜑)/𝑑𝑥 =
−𝐷𝑤 under the present assumptions. For linear waves we can now write the wave force
Eq. 5.58 as:
𝑑𝑆𝑦𝑥 sin 𝜑 𝑑 𝐷𝑤
𝐹𝑦 = − =− 𝐸𝑐𝑔 cos 𝜑 = sin 𝜑0 (5.73)
𝑑𝑥 𝑐 𝑑𝑥 𝑐0
Apparently the alongshore driving force is a function of the dissipation of the wave
energy. Outside the surf zone, the dissipation of wave energy can be neglected and
hence the energy flux 𝐸𝑐𝑔 = (𝐸𝑐𝑔 cos 𝜑, 𝐸𝑐𝑔 sin 𝜑) is constant. We have in the absence
of dissipation: 𝑆𝑦𝑥 is constant and 𝐹𝑦 = 0. So, although outside the surf zone the wave
conditions change with 𝑥 (wave height due to shoaling; wave direction due to refrac-
tion), the radiation shear stress is constant. Therefore, since the alongshore forcing is
only present when the waves are breaking, the longshore current is confined to the
surf zone (see also Intermezzo 5.6).
Intermezzo 5.6 Energy dissipation and mean currents
The finding that a mean current is only driven in the case of energy dissipation is
universally valid, also in less simplified situations. The wave force consists of two
parts:
• A part in the wave propagation direction that is related to the energy dis-
sipation. This term is concentrated near the water surface, where the dis-
sipation actually takes place. This 𝐷/𝑐 part is rotational and can therefore
induce mean currents or circulation currents in the case of a closed bound-
ary. It is a result of the vorticity generated by wave-breaking and stems

from the effect of both velocity and pressure variation on the momentum
flux;
• An irrotational part that is depth-invariant. This implies that there is no
vertical imbalance. Thus, this part can be balanced by set-up and set down
without driving currents. It only affects the wave-induced current indirectly,
through its effect on the mean depth. In the surf zone, the latter influence
on the mean current is small compared with that of the dissipation.
Analytical model for alongshore wave force in the surf zone

To find an easy analytical equation for the alongshore wave force, we assume again
the simple model for wave dissipation due to breaking that relates the wave height
to the local water depth: 𝐻 = 𝛾 ℎ for every water depth in the surf zone. An altern-
ative would be solving the energy balance numerically. Snell’s law and the simple
dissipation model result in:
sin 𝜑 𝑑 sin 𝜑0 𝑑 1
𝐹𝑦 = − 𝐸𝑐𝑔 cos 𝜑 = − /8𝜌𝑔𝛾 2 ℎ2 𝑐𝑔 cos 𝜑 (5.74)
𝑐 𝑑𝑥 𝑐0 𝑑𝑥
In shallow water 𝑐𝑔 = 𝑐 = √𝑔ℎ and since, due to refraction, 𝜑 is small (generally around
10° to 15°), we assume cos 𝜑 ≈ 1. We now have:
sin 𝜑0 1 𝑑 5 5 sin 𝜑0 𝑑ℎ
/8𝜌𝑔 /2 𝛾 2 ℎ /2 = − 𝜌(𝑔ℎ) /2 𝛾 2
3 3
𝐹𝑦 ≈ − (5.75)
𝑐0 𝑑𝑥 16 𝑐0 𝑑𝑥
As mentioned above, the radiation shear stress 𝑆𝑦𝑥 is constant seaward of the border
of the breaker zone (thus the wave force is zero). It decreases inside the breaker zone
to zero at the waterline (see Fig. 5.37).
For many day to day wave conditions, this gradient in 𝑆𝑦𝑥 causes alongshore stresses
(forces) which are in the same order of magnitude as the bottom shear stress in rivers:
in the order of 1 N/m2 to 10 N/m2 . The cross-shore gradient in the alongshore radiation
stress 𝑆𝑦𝑥 is therefore an important driving force in the littoral zone.
Quadratic friction law for bed shear stress

The next step in determining an analytical equation for the alongshore current is the
formulation of the bed shear stress. In Sect. 5.4.3 we have already discussed the quad-
ratic friction law for wave- and current-only situations. For currents the turbulence
is present in the entire water column. The turbulence due to waves is present in the
wave boundary layer only and significantly increases the bed shear stress. This is not
only due to the significant increase in friction factors, but also due to the velocity being
generally higher in the wave motion. For the vertical profile of the longshore current,

1.2
1.1
radiation shear stress Syx [103 N/m]

1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0 50 100 150 200 250 300 350
distance from shore [m]
Figure 5.37: Radiation shear stress 𝑆𝑦𝑥 for 𝐻0 = 2 m, 𝑇 = 7 s, 𝜑0 = 30°, constant bottom slope
1:100 and a breaker index 𝛾 = 0.8.
we need a quadratic resistance law of combined current and wave action. Such a law
is non-trivial since:
• Due to the quadratic friction law, the addition of the effects of waves and cur-
rent is non-linear. Due to the non-linear addition, the combined shear stress
averaged over the wave period is generally larger than the linear summation of
the current-only and wave-only bed shear stress;
• The waves and currents may not have the same direction. The waves are incid-
ent with a certain angle, whereas the longshore current is parallel to the coast.
As a result, the direction and magnitude of the bed-shear stress during the wave
cycle varies continuously. The time-averaged bed shear stress depends on the
angle between the waves and current;
• As discussed earlier, for currents the turbulence is present in the entire water
column and for waves in the wave boundary layer only. It therefore makes
sense to relate the bed shear stress due to currents to a depth-mean velocity and
due to waves to the free-stream velocity outside the wave boundary layer. This
raises a question, however: the velocity at which height should be chosen for
the combined wave current motion?
• Note that the above descriptions deal with the time-averaged bottom shear stress;
they disregard the momentary (or intra-wave) shear stresses that may occur
within the wave period. This seems reasonable when determining the wave-
induced currents. However, the intra-wave velocity fluctuations may be import-
ant for the net (wave-averaged) sediment transport (see also Chs. 6 and 7).
For the reasons mentioned above, many different models exist for the bed shear stress
under waves and currents. Depending on the model, the relative contributions of
waves and currents to the bed shear stress vary. Some models describe the time-
averaged bed shear stress, others the instantaneous bed shear stress. The description

of the sediment transport requires an accurate formulation of bed shear stress (intra-
wave or time-averaged, depending on the sediment transport model). Therefore in
Sect. 6.5 bed shear stress is treated in more detail.
For the determination of the time-averaged bed shear stress in the alongshore direction,
necessary to compute the longshore current, we take a fairly simple approach:
• The wave motion is described by shallow-water theory (constant orbital amp-
litude outside the wave boundary layer);
• The angle of incidence is very small, such that for the wave motion (𝑢𝑥 , 𝑢𝑦 ) =
(𝑢̂ cos 𝜔𝑡, 0);
• The bed friction vector is related to the depth-averaged velocity vector. The
latter is the sum of the depth-averaged longshore current velocity and the wave
orbital motion: 𝑢⃗ = (𝑢̂ cos 𝜔𝑡, 𝑉 ), see Fig. 5.38;
• The time-varying bed friction is written as:
𝜏⃗𝑏 = 𝜌𝑐𝑓 |⃗
𝑢 |⃗
𝑢 (5.76)
• The enhancement of the friction factor (compared to a current-only situation)

due to the small height of the wave boundary layer as compared to the current
boundary layer is not further specified.
Figure 5.38: The magnitude of the instantaneous velocity vector in the surf zone, with 𝑉 the
longshore current velocity and 𝑢̂ cos 𝜔𝑡 the orbital velocity (in the cross-shore direction).
In the cross-shore direction, the time-averaged (not the instantaneous) bed shear stress
is zero. With the above approximations, the time-averaged bed shear stress in the
alongshore direction reads:
𝜏 𝑏,𝑦 = 𝜌𝑐𝑓 |⃗𝑢|𝑉 = 𝜌𝑐𝑓 √𝑉 2 + 𝑢̂ 2 cos2 𝜔𝑡𝑉 (5.77)
If we further assume that 𝑉 ≪ 𝑢,̂ this can be simplified to:
2
𝜏 𝑏,𝑦 = 𝜌𝑐 𝑢𝑉
̂ (5.78)
𝜋 𝑓

With 𝑢̂ in shallow water given by Eq. 5.23 and with a constant ratio of wave height
over water depth across the entire surf zone we find:
1 𝐻
𝜏 𝑏,𝑦 = 𝜌𝑐𝑓 √𝑔ℎ 𝑉 (5.79)
𝜋 ℎ
Analytical model for longshore current (no lateral dispersion)

For steady conditions, the alongshore velocity follows from the balance between the
driving force and the resisting friction force (Eq. 5.72). This yields with Eq. 5.73 and
Eq. 5.79:
𝐷𝑤 1 𝐻 𝜋 sin 𝜑0 𝐷𝑤 (𝑥)
sin 𝜑0 = 𝜌𝑐𝑓 √𝑔ℎ 𝑉 ⇔ 𝑉 (𝑥) = ℎ(𝑥) (5.80)
𝑐0 𝜋 ℎ 𝑐𝑓 𝜌 √𝑔 𝑐0 𝐻 (𝑥) √
The magnitude of the depth-averaged longshore current velocity varies in the surf
zone as a function of the dissipation, wave height and water depth. The dissipation
and wave heights can be modelled using a wave model (with roller model). In our
simple dissipation model 𝛾 = 𝐻 /ℎ = constant and we can write the force balance
(Eq. 5.72) as:
5 sin 𝜑0 𝑑ℎ 1
𝜌(𝑔ℎ) /2 𝛾 2
3
− = 𝜌𝑐 𝑔ℎ𝛾 𝑉 (5.81)
16 𝑐0 𝑑𝑥 𝜋 𝑓 √
which leads to:
5 𝛾 sin 𝜑0 𝑑ℎ
𝑉 (𝑥) = − 𝜋 𝑔 ℎ (5.82)
16 𝑐𝑓 𝑐0 𝑑𝑥
For a constant beach slope tan 𝛼 = −𝑑ℎ0 /𝑑𝑥 and for 𝑑ℎ/𝑑𝑥 ≈ 𝑑ℎ0 /𝑑𝑥, the current
velocity is proportional to the depth with a maximum at the breaker line (where ℎ =
ℎ𝑏 ):
5 𝐻𝑏 sin 𝜑0 ℎ
𝑉 (𝑥) = 𝜋 𝑔 tan 𝛼 (5.83)
16 𝑐𝑓 𝑐0 ℎ𝑏
A longshore current profile according to Eq. 5.83 is shown in Fig. 5.39. The larger the
wave height 𝐻𝑏 at breaking, the larger the maximum longshore current velocity and
the wider the littoral zone; a factor 2 larger wave height would then result in a factor
8 larger discharge in the surf zone. Eq. 5.83 also allows us to get an idea of the effect
of the beach slope tan 𝛼. A steeper slope, on the one hand, results in (linearly) higher
velocities. On the other hand, the width of the surf zone becomes linearly smaller. The
discharge through the entire surf zone is then constant to a first approximation (we
have, amongst other factors, neglected the effect of the beach slope on the breaking

parameter, Sect. 5.2.5). If the wave angle is small, the longshore current velocity at
a specific water depth within the surf zone becomes a linear function of 𝜑0 (since:
sin 𝜑0 = 𝜑0 for small 𝜑0 ).
1.2
1.1
longshore current velocity V [m/s]
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0 50 100 150 200 250 300 350
distance from shore [m]
Figure 5.39: Alongshore velocity distribution (regular wave field, 𝐻0 = 2 m, 𝑇 = 7 s, 𝜑0 = 30°, bot-
tom slope 1:100, 𝛾 = 0.8, roughness height 𝑟 = 0.06 m). The deviation from the linear distribution
stems from cross-shore variation in friction factors. The current velocity is zero outside the surf
zone.
Turbulent forces redistributing momentum

So far, the effect of lateral dispersion of momentum by turbulence has been ignored.
Including turbulence in the momentum equation tends to smooth out velocity gradi-
ents (including the unrealistic velocity gradient at the breaker point).
Let us first look at turbulence and turbulence modelling in a bit more detail. In Sect. 5.4.3,
the total velocity vector was said to be composed of a mean, a wave and a turbulent
part. The turbulent shear stress was subsequently defined as the stress introduced
when averaging over the turbulent motion. In analogy with the modelling of viscous
stresses, in a turbulent flow the shear stress is generally related to velocity gradients
through a turbulent or eddy viscosity 𝜈𝑇 . The molecular viscosity 𝜈 seems to make
water sticky and resist flowing and may be thought of as a measure of viscous fluid
friction. Similarly 𝜈𝑇 is a measure of turbulent fluid friction (in coastal waters 𝜈𝑇 ≫ 𝜈).
The eddy viscosity 𝜈𝑇 [m2 /s] depends on a characteristic spatial scale and on a charac-
teristic velocity. In the littoral zone, both are related to wave motion. The wave orbital
motion, for instance, can be regarded as a measure for the characteristic velocity. For
vertical mixing, the characteristic (mixing) length is the depth. Horizontal mixing is
not restricted by water depth. For that reason, in nearshore modelling the horizontal
eddy viscosity 𝜈𝑇 ,𝐻 is often taken much larger than 𝜈𝑇 for vertical mixing. Typical
values for the eddy viscosity 𝜈𝑇 are 10−2 m2 /s.

We have seen that the driving force for the longshore current is 𝜕𝑆𝑦𝑥 /𝜕𝑥. The shear
component of the radiation stress 𝑆𝑦𝑥 was defined through Eq. 5.47, in which the velo-
city components are due to the orbital motion. In analogy with Eq. 5.47, we can write
for the turbulent force:
𝜂
′
𝑆𝑦𝑥 =∫ (𝜌𝑢𝑦′ 𝑢𝑥′ ) 𝑑𝑧 (5.84)
−𝑘0
where the overbar now represents averaging over the turbulent motion (indicated with
primes). This shear stress or friction force per unit surface area, acts on a surface
parallel to the coast. It can be modelled as:
′ ≅ ℎ𝜌𝜈 𝑑𝑉
𝑆𝑦𝑥 𝑇 ,𝐻 (5.85)
𝑑𝑥
The eddy viscosity 𝜈𝑇 ,𝐻 [m2 /s] is also referred to as horizontal diffusivity.

The momentum equation in the alongshore direction now reads:
𝐷𝑤 𝑑 𝑑𝑉
sin 𝜑0 + (ℎ𝜌𝜈𝑇 ,𝐻 ) = 𝜏 𝑏,𝑦 (5.86)
𝑐0 𝑑𝑥 𝑑𝑥
The effect of turbulent forces, smoothing the longshore current profile, is indicated
in Fig. 5.40. Since the largest velocity gradient occurs at the breaker line, the max-
imum transfer of horizontal momentum will occur here. This leads to a reduction in
the maximum velocity, a landward shift of the position of maximum velocity and to
a situation where, also outside the breaker zone, longshore current velocities occur.
The cross-shore distribution of the eddy viscosity now also determines the velocity
distribution.
breaker line
Figure 5.40: Effect of turbulence on the velocity profile.

Roller momentum
Section 5.5.4 discussed that the measured onset of set-up occurs closer to the shore
than predicted. This spatial lag was attributed to the roller momentum, which had
not been taken into account. Similarly longshore current velocity profiles show an
onshore shift in the maximum longshore current velocity. This can be modelled by
including the roller contribution in the alongshore momentum equation.
Irregular waves
Until now we have only considered regular waves. In reality, of course, waves are
irregular and there is no sharply defined breaker line. The effect of wave irregular-
ity is therefore to smooth out the velocity distribution, very similar to the effect of
turbulence, giving a wider and less sharply peaked velocity distribution. This is also
illustrated in Fig. 5.41, which shows the output of a computation with the computer
model Unibest-CL+.
2 wave height
longshore velocity [m/s]

0.75
wave height [m]
0.5
1
0.25
longshore velocity
0 0
0
depth [m]
10
Figure 5.41: Wave-driven velocity computed with the computer model Unibest-CL+ (https://www.
deltares.nl/en/software/unibest-cl) using as model input: 𝐻𝑠,0 = 2 m; 𝛾 = 0.7; 𝜑0 = 30°; 𝑇𝑝 = 8 s).
Profile with breaker bar

If we have a coastal profile with a breaker bar, then the determination of the velocity
distribution becomes more complicated. In a simplified approach, we can distinguish
between a breaker zone at the seaward side of the breaker bar (if waves indeed break
at the corresponding water depths) and a breaker zone near the shoreline where the
remaining wave energy is dissipated. This simplification would lead to a zero long-
shore current velocity in the deeper section between both breaker zones. Also in such
a situation, both lateral transfer of horizontal momentum and the irregularity of the
waves will smooth out the velocity distribution. An example of the longshore current
distribution is given in Fig. 5.42. Wave-breaking tends to concentrate on the bars and
at these locations a longshore current exists. For the highest wave condition, wave-
breaking already occurs at the outermost bar, whereas for the lowest wave condition
wave-breaking occurs only at the innermost bar and close to the shoreline.

Hs = 1.0 m Hs = 2.0 m Hs = 3.0 m

Tp = 5.0 s Tp = 7.0 s Tp = 9.0 s
H [m]
φ [°]
V [m/s]
2012
z [m]
x [m]
Figure 5.42: Longshore current profiles, calculated with the Unibest-CL+ (https://www.deltares.
nl/en/software/unibest-cl), for Jarkus transect 7003850 (JARKUS, n.d.), which is near Egmond,
the Netherlands. The profile was measured in 2012 and shows the effect of both the autonom-
ous bar cycle (Sect. 7.3.4) and – in deeper water – shoreface nourishments. Three different
deep-water wave conditions are used: 𝐻𝑠 = 3 m and 𝑇𝑝 = 9 s (breaking occurs on all bars),
𝐻𝑠 = 2 m and 𝑇𝑝 = 7 s (no substantial breaking takes place at the outermost bar), 𝐻𝑠 = 1 m
and 𝑇𝑝 = 5 s (breaking takes mainly place at the innermost bar and close to the shoreline). In all
cases 𝜑 = 30° in deep water.

5.5.6. Vertical structure of the wave-induced currents

The previous sections explained that in an alongshore uniform situation, the (depth-
integrated) cross-shore radiation stress balances the wave set-up. The alongshore ra-
diation stress gradient, on the other hand, was seen to be balanced by a time mean bed
shear stress associated with a longshore current. Since the wave forces were vertically
integrated, depth-dependent variation was not considered. However, due to the ver-
tical non-uniformity of the driving forces in the nearshore zone, secondary currents
are driven. The imbalance between the cross-shore wave radiation stress gradient and
the pressure gradient due to the set-up drives a 2D-vertical (2DV) circulation current
inside the breaker zone, with a surface current running towards the coast (mass flux)
and a seaward current below the wave trough level (the undertow).
The formula for the depth-mean value of the undertow under wave trough level was
already determined in Sect. 5.5.1 based on a computation of the mass flux above the
wave trough level and continuity considerations. The vertical imbalance between the
forces (per unit mass) is indicated in Fig. 5.43.
z
x MWL
Figure 5.43: Vertical distribution of stresses in undertow. The imbalance in the vertical between
the radiation stresses (left figure) and pressure (middle figure) drives a circulation current in the
surf zone (right figure).
The pressure gradient due to set-up, the 𝑔𝜕𝜂/𝜕𝑥 term, is the same at all levels, whereas
the radiation stress term 𝜕𝑠𝑥𝑥 (𝑧)/𝜕𝑥 is not evenly distributed over the water column.
Instead it is particularly large near the water surface due to the strong pressure and
velocity variations above wave trough level (in the surface roller of breaking waves).
A fluid particle below mean trough level will therefore experience a net force averaged
over the wave period which is directed seaward.
To determine the velocity distribution of the cross-shore circulation current and the
longshore current, the horizontal momentum balances without integration over the
vertical need to be solved. To avoid computationally demanding 3D computations, a
quasi-3D approach can be taken. This involves solving the depth-averaged horizontal
momentum equations to obtain the depth-averaged quantities, and subsequently at
every position solving a momentum balance that resolves the vertical, but neglects
spatial gradients. By neglecting the advective terms, locally a horizontal bottom is
assumed. The computation is simplified by only considering the area below the wave

trough level. In a steady situation, the locally applied momentum balance in the 𝑖-
direction, 𝑖 = 𝑥 or 𝑦 then is of the form:
𝜕𝜏𝑖
= 𝑅𝑖 (5.87)
𝜕𝑧
where 𝑅𝑖 (see Intermezzo 5.5) is the forcing that below the wave trough level is domin-
ated by the pressure gradient. The trough-to-crest layer, containing the moving water
surface, is accounted for via an effective shear stress at the trough level, compensating
for the momentum decay above it, and via the condition that the net mean flow below
trough level must compensate for the mass flux in the surface layer. The modelling of
the surface layer is thus reduced to the formulation of the effective shear stress and
the mass flux. The effects of wind can be added by applying a surface shear stress due
to wind (see also Sect. 5.6). If the appropriate forcing term - related to the dissipation
in the wave boundary layer - is taken into account, the near-bottom streaming in the
wave boundary layer will be resolved as well (Sect. 5.4.3).
The shear stress can be related to the velocity gradients by:
𝜕𝑢𝑖
𝜏𝑖 = 𝜌𝜈𝑇 (5.88)
𝜕𝑧
The eddy viscosity 𝜈𝑇 represents the turbulence in the water column. Often a parabolic
distribution of the eddy viscosity is used; for a slope-driven current the turbulence is
restricted by the bed and the surface and is largest in the middle of the water column.
Furthermore, turbulence from different sources (wind-driven current, wave-breaking
and increased turbulence in the wave boundary layer) can be taken into account. For
wind-induced- and wave-breaking-induced turbulence, a maximum can be expected
near the surface (which can be modelled by for instance a half-parabolic distribution).
Integration of Eq. 5.88 using an appropriate eddy viscosity distribution yields the velo-
city distribution over the vertical. For a parabolic eddy viscosity, we find a logarithmic
velocity distribution.
Up to this point, we have considered the undertow and alongshore current in isolation.
In reality, the wave-induced current has a 3D structure, as can be seen in Fig. 5.44.
5.5.7. 3D effects
So far, we considered situations without alongshore variations. For various reasons
however, wave conditions can vary along a certain stretch of coast; for instance due
to wave refraction on a non-uniform nearshore region or due to wave diffraction in
the lee side of structures. Wave-induced forces will vary according to changes in the
wave conditions and the terms 𝜕𝑆𝑦𝑦 /𝜕𝑦 and 𝜕𝑆𝑥𝑦 /𝜕𝑦 may be non-zero in Eqs. 5.55
and 5.57. Variations in wave height along a coastline create variations in cross-shore

sh
swa
king
brea swash
motio
ing n
shoal
ay s
ave r w
ting w flo
refrac rn
r e tu
w
f
lo
e
or
gsh
lon
Figure 5.44: 3D structure of the wave-induced current profile in the surf zone, composed of the
undertow and the alongshore current.
wave forces along the coast and hence in wave set-ups. Consequently, pressure gradi-
ents 𝜌𝑔ℎ 𝜕𝜂/𝜕𝑦 occur along the coast and 3D current patterns are the result. We will
briefly discuss the following situations:
• Eddy formation in the shadow zone of structures;
• Creation of rip currents;
• 3D current patterns around shoals.
Eddy formation in the shadow zone of structures

In the case of wave sheltering due to for instance groynes or detached breakwaters,
wave set-up can be expected to be less in the sheltered area than in the unsheltered re-
gion. This generates local nearshore currents towards the sheltered area. Figure 5.45
shows a situation at a groyne. In the shadow zone of the groyne, due to set-up dif-
ferences a current runs towards the groyne, until it is diverted outward along the
structure, creating an eddy.
In Fig. 5.46 set-up differences create nearshore currents towards the sheltered zone
from both sides of the detached (emerged) breakwater. Continuity requires that wa-
ter leaves the area as well. Therefore, a return flow will be present in deeper water,
resulting in the development of two eddies. A characteristic flow pattern is generated
(see Fig. 5.46).

eddy longshore current
Figure 5.45: Current pattern in lee side of groyne (eddy formation). Note also that upstream (or
updrift) of the breakwater, the longshore current diminishes slowly because of the decreasing
angle of incidence.
large small large surf zone

wave set-up wave set-up wave set-up
Figure 5.46: Current patterns behind a detached, emerged breakwater.
Rip currents due to convergence or divergence of wave energy

Rip currents are strong, narrow currents that flow seaward from the surf zone. They
are fed by longshore-directed surf zone currents that turn seaward to form a rip cur-
rent. The longshore currents are generated by set-up differences and run from the
position of the highest set-up towards the position of the lowest set-up. The along-
shore variation in wave set-up can be generated by convergence and divergence of
wave energy due to depthrefraction or sheltering effects due to for instance headlands.
An undulating coastline gives concentration of wave energy towards the undulations
due to depthrefraction (Fig. 5.47b). The longshore currents run from the position of
the highest waves and hence highest set-up to the position of the lowest waves. In
this way, a rhythmic pattern of rip currents is generated. Such a situation will only be
able to develop for nearly normal incidence; in the case of obliquely incident waves,
a longshore current driven by gradients in radiation shear stress (Fig. 5.47a) will over-
shadow the more subtle effects of set-up differences. A combination of the two effects
may occur for slightly oblique wave incidence (Fig. 5.47c).

surf zone
(a) typical longshore current distribution in case of obliquely incident waves

surf zone
(b) rip current pattern for undulating coastline in case of approximately normal incidence
surf zone
(c) combination of previous two cases for slightly oblique waves
Figure 5.47: Nearshore circulation patterns for different angles of wave incidence 𝜑𝑏 .
Wave-induced currents around shoals (and submerged breakwaters)

In case of a complex topography with interrupted breaking, the pattern of wave-induced
currents is more complicated. A still relatively simple example concerns a shoal on
which waves are breaking (Fig. 5.48). Due to refraction, the waves will tend to con-
verge toward the top of the shoal. As the waves break on the seaward slope of the
shoal, they generate a dissipation-related wave force. At the top of the shoal, there is
no closed boundary that requires a zero mean flow. Instead, the water will flow over
the shoal in the direction of the force. The water flows over the shoal until it reaches
the channel behind the shoal, where water level gradients will deflect it and drive it to

5.6. Wind-induced set-up and currents 235
the sides of the shoal. There it has room to flow seawards again, thus closing the cir-
culation. In tidal inlets systems the wave-driven currents around shoals on the outer
delta can be so strong that they dominate the tidal residual currents (see Sect. 9.4.1).
Figure 5.48: Wave-induced forces (white wider arrows) and currents (dark thinner arrows)
around a shoal.
In a similar fashion, the interruption of wave-breaking can generate rip currents along
a stretch of coast. Non-homogeneous wave-breaking can occur in the case of a non-
homogeneous alongshore bar system or a (series of) submerged breakwater(s) on which
waves break, see Fig. 5.49. (Partial) wave-breaking on the bar induces a set-up over
the bar as well as an onshore flow. Water level gradients and continuity force the
flow to deflect to the sides and return seaward in between the breakwaters or bars in
a concentrated rip current.
onshore flow over bar
0 + + 0 + + 0 + + 0
concentrated return flow

+ set-up
0 zero set-up
Figure 5.49: Rip currents in case of submerged breakwaters (or an interrupted bar system).
5.6. Wind-induced set-up and currents

Moving air exerts a shear stress 𝜏 on the water surface that can be modelled by, again,
a quadratic friction law:
𝜏wind = 𝐶𝑑 𝜌𝑎 𝑊 2 (5.89)

where:
𝜏wind wind shear stress N/m2

𝐶𝑑 drag coefficient depending on wind velocity: −
𝐶𝑑 = (0.63 + 0.066𝑊 ) × 10−3 for 2 < 𝑊 < 21 (Smith & Banke,
1975)
𝜌𝑎 density of air (1.25 kg/m3 ) kg/m3
𝑊 wind velocity at the water surface m/s
Due to this wind shear stress, the upper parts of the water layers will start to move
more or less in the same direction as the wind direction. When the wind is seaward
directed, a seaward-directed current is generated in the upper water layers, the velocity
of which is determined by the duration of the particular wind condition and its force.
Similarly, a landward-directed wind will induce a landward current in the upper layers.
However, a coastline forms a barrier for this landward current and in an equilibrium
situation the mean onshore directed flow should be zero (just as with wave-induced
flow). Therefore, to compensate for the landward- (or seaward-) directed water mass
movement in the upper layers, an opposite-directed water mass transport follows in
the lower water layers. In addition to these ‘compensation currents’, a water level
set-up or set-down develops near the coast to balance the wind-induced shear stresses
(see Fig. 5.50)
SWL
circulation current
Figure 5.50: Wind set-up balancing wind shear stress in the case of onshore wind on a shelf.
In the deep ocean, the water set into motion in the upper layers can easily spread out. Vertical
exchange of water makes this possible. On the shelf there is no escape due to limited water
depth and the surface water is driven ashore.
The equilibrium condition is given by:
𝑑𝜂
𝜌𝑔ℎ = 𝜏wind,𝑥 (5.90)
𝑑𝑥
Equation 5.90 shows that the wind set-up is inversely proportional to the water depth.
Hence, in the shallow coastal zone (up to the shelf break) water can pile up to great
heights (storm surge).

5.6. Wind-induced set-up and currents 237
When the wind is directed parallel to a (long) coast, a longshore current is generated.
It takes some time (order of magnitude one day) for the wind-driven current to develop
completely. In the equilibrium situation and in the absence of other driving forces, the
wind shear stress and the bed shear stress are equal, as was the case for the wave-
induced longshore current.
The vertical distribution of the wind-generated current differs substantially from the
current generated by a water level gradient (Fig. 5.51). The highest flow velocities
occur at the water surface, with usually a rapid decrease in the downward direction
(much more than with the logarithmic velocity profile).
wind-driven velocity distribution
logarithmic velocity distribution
Figure 5.51: Typical velocity distribution for wind-driven current.
During storms, the wind stress may have an important effect on the residual longshore
current. Often, however, the effect of wind stress on the longshore current in the lit-
toral zone can be neglected. Furthermore, the morphological impact is limited due to
the relatively small velocities near the bottom, where the highest sediment concentra-
tions occur.
Wind-driven currents are of greater importance in, for instance, coastal lagoons. The
Dutch Wadden Sea is an example of a series of tidal basins in which wind-driven cur-
rents may play an important role. With dominant SW winds, a significant volume of
water and sediment can be transported towards the east. This may lead to a gradual,
but over the years continuing, shift of the watersheds between the barrier islands and
the mainland. FitzGerald and Penland (1987) hypothesise that this primarily wind-
generated transport is the main force behind inlet migration in the German Wadden
Sea.
In Sect. 1.4.6, the phenomenon of storm surge was introduced. A storm surge is like
a raised dome of water due to the combination of strong onshore winds (causing the
wind set-up described above) and lowered atmospheric pressure in a storm centre.
The lower atmospheric pressure also raises the water level. The coastal topography
of the North Sea (shallow and funnel-shaped) and its predominantly westerly winds
make it very susceptible to storm surges. The severity of a storm surge depends on

its timing relative to the tidal cycle and the duration of the storm system. A storm
system with a long duration (a few days) coinciding with spring tide can result in
severe flooding of coastal areas. For example, the storm surge height in the southern
North Sea, resulting in the 1953 flood event, was 2.9 m and the storm duration was
more than two days. In contrast, storm surges in the Bay of Bengal build up from
tropical cyclones. Bangladesh experienced storm surge heights of around 9 m in 1970.
Due to the lack of appropriate coastal protection, this surge resulted in a very large
number of casualties (500 000).
5.7. Tidal propagation in coastal waters

5.7.1. Definitions
The vertical rise and fall of the water level is called the vertical tide or simply tide. High
tide means high water levels, whereas low tide means low water levels (see Fig. 5.52).
The rising period is the time it takes for the water level to get from the lowest elevation
to the highest elevation, the falling period is the time it subsequently takes to reach the
lowest level. The associated horizontal movement back and forward is the horizontal
tide or tidal current. We speak about flood currents if the current velocity is in the
tidal wave propagation direction. Ebb currents are directed against the propagation
direction.
Although this seems clear, there is the danger of confusion. Ebb or ebb tide or ebb
period may be used to indicate falling water levels as well as ebb currents. Due to the
complex phase relationship between vertical and horizontal tide (see Sect. 5.7.2 and
further), a falling tide does not necessarily coincide with ebb currents. To avoid any
confusion, we use the words ebb and flood solely to refer to the horizontal tide.
Slack water is the name used for tidal flow reversal. We will see in Sect. 5.7.2 that
the velocity generally leads the surface elevation in shallow water. As a consequence,
flow reversal from ebb to flood occurs around low water and is therefore called Low
Water Slack (LWS), see Fig. 5.52. High Water Slack (HWS) around high water occurs
for flow reversal from flood to ebb. The term slack water period refers to the duration
of slack water (i.e. the period of time during which current velocities are below some
threshold)
5.7.2. Tidal propagation along the shore

Figure 5.53 shows measured water level (vertical tide), current velocity (horizontal
tide) and current directions for a location a few kilometres off the Dutch coast.
The flood-tidal current runs in the northward direction along the Dutch coast and
the ebb current runs southward, in accordance with the direction of the rotary wave
in the ocean basins and seas in the Northern Hemisphere (Sect. 3.2). The flood and

5.7. Tidal propagation in coastal waters 239
HW LW
1
0.5
elevation [m]
-0.5
-1
0.5
velocity [m/s]
-0.5
HWS LWS
-1
0 3 6 9 12 15
t [hours]
Figure 5.52: Vertical and horizontal tide in Rotterdam, where High Water Slack (HWS, flow re-
versal from flood to ebb) occurs shortly after HW, and Low Water Slack (LWS, flow reversal from
ebb to flood) follows LW with a slightly larger time difference.
1.5
water level
[m + NAP]
1
0.5
0
-0.5
1
magnitude
0.75
velocity
[m/s]
0.5 flood ebb flood ebb

0.25
360 northward
[degrees rel. to N]
currnet direction
270 westward
180 southward
90 eastward
0 northward
0 6 12 18 24
t [hours]
Figure 5.53: Water level (top panel), current velocity (middle panel) and current direction (lower
panel) at the measuring location ‘Stroommeetpaal IJmuiden’ on the 21st of February 2020 (data
source: https://waterinfo.rws.nl/). The location is about 500 m offshore from the entrance of
the IJmuiden harbour.

ebb velocities are at a maximum around high and low water respectively. The latter
is typical of the propagation of the tide in relatively deep water, where bed friction
has relatively little effect on the propagation. As explained in Sect. 3.8, under those
circumstances, the tidal wave has a progressive character and water level and velocity
are in phase (just as for wind waves). Figure 5.53 further shows that the tidal record
deviates from an ideal symmetrical oscillation, which will be discussed later on in this
section.
The effect of bottom friction

As previously, an alongshore uniform coast is considered with the 𝑦-axis defined paral-
lel to the shoreline and the 𝑥-axis perpendicular to the shoreline. If the tidal elevation
at lowest order is equal to 𝜂(𝑡) = 𝑎 cos(𝜔𝑡 − 𝑘𝑦), the alongshore tidal velocity can be
written as 𝑣(𝑡) = 𝑉 cos(𝜔𝑡 − 𝑘𝑦 − 𝜑). For the M2 tide 𝜔 ≈ 1.4 × 10−4 s−1 .
For the Kelvin wave we found 𝜑 = 0 for the propagation alongshore (Sect. 3.8). With
the flood velocity defined as positive, 𝜑 = 0 means that velocity and elevation are
in phase (progressive wave). The Kelvin wave was found by solving the momentum
balance in the 𝑥- and 𝑦-directions, Eqs. 3.42a and 3.42b, and the continuity equation
Eq. 3.42c. We assumed that friction was very small compared to inertia.
In coastal engineering applications, we generally consider the tidal flow in a zone
relatively close to the coast (order 10 km). In that case we cannot neglect friction. If
we neglect the inertia term 𝜕𝑣/𝜕𝑡 in Eq. 3.42b, but add a friction term, the momentum
equation (in the alongshore 𝑦-direction) becomes:
Z 𝜕𝜂 𝜏𝑏𝑦
Z 𝜕𝑣 = − 𝑔 − (5.91)
Z 𝜕𝑡 𝜕𝑦 𝜌ℎ
Z⏟
Z
⏟⏟⏟⏟⏟⏟⏟ ⏟⏟⏟⏟⏟⏟⏟⏟⏟
inertiaZ alongshore water friction
acceleration)
(local ZZ
level gradient
In this equation the alongshore pressure gradient 𝜕𝜂/𝜕𝑦 is constant in the cross-shore
direction (for the narrow coastal strip under consideration). Although a quadratic
friction law is more appropriate, for simplicity we assume that the friction is linearly
dependent on the alongshore tidal velocity: 𝜏𝑏𝑦 = 𝜌𝑐𝑓 𝑣|𝑣| ≈ 𝜌𝑟𝑣. Eq. 5.91 now reads:
𝜕𝜂 𝑟
𝑔 =− 𝑣 (5.92)
𝜕𝑦 ℎ
This equation suggests that the local nearshore water level gradient in a certain tidal
phase is balanced by (linear) bed friction. Hence, the tidal velocity is not in phase with
the tidal elevation, but with the negative alongshore water level gradient. Or, to put
it simply: at any point in time the water flows from a location with high water to a
location with low water (not different from river flow, with the difference that the tide
reverses direction). In this example the phase difference 𝜑 = −𝜋/2.

Figure 5.54 shows that for 𝜑 = −𝜋/2 = −90° the velocity leads the elevation by a
quarter period or about 3 hours for the M2 tide. Figure 5.54 shows also that for the
special case of 𝜑 = −𝜋/2:
• during the entire time it takes for the water to reach the lowest elevation (the
falling period), the velocities are negative (ebb current);
• during the time it takes to reach the highest elevation (the rising period), the
velocities are positive (flood current).
Thus: in case that the velocity leads the surface elevation by 90°, the falling period
coincides with the ebb duration and the rising period with the flood duration.
η/a u/û
1
0.5
-0.5
-1
0 π/2 π 3π/2 2π 5π/2 3π 7π/2 4π

phase [rad]
Figure 5.54: Tidal elevation and tidal velocity divided by their respective amplitudes. With flood
velocities defined as positive, 𝜑 = −𝜋/2 = −90° means that the velocity leads the elevation by
about 3 hours in the case of the M2 tide.
In general, the phase relationship between vertical and horizontal tide is very complex.
Not only friction but also (partial) reflections of the tidal wave introduce phase differ-
ences between velocity and elevation. Generally, the phase difference 𝜑 in coastal
waters and basins varies between zero and 𝜑 = −𝜋/2. If there is a phase difference
between velocity and tidal elevation, it will be such that the velocity peaks before the
tidal elevation.
The effect of friction is not only to introduce a phase difference between elevation
and velocity, but to reduce their magnitudes as well, compared to the same frictionless
wave. In Sect. 5.7.3, the effects of friction and reflection are examined in more detail
for the propagation into tidal basins.
Alongshore differences
Along the Dutch coast the tidal wave propagates northwards. We saw that at lowest
order the tidal elevation and velocity along the coast (𝑦-direction) can be described by
𝜂 = 𝑎 cos(𝜔𝑡 − 𝑘𝑦) and 𝑣 = 𝑉 cos(𝜔𝑡 − 𝑘𝑦 − 𝜑). This means that the phase 𝑘𝑦 of the
tide increases (or in other words: the wave form is delayed) from the delta area in the
south towards the Wadden area in the north. This can be seen from Fig. 5.55.

14 October 2019 01:35

3
3 IJmuiden
2 4 Den Helder
1 Vlissingen
elevation [m + NAP]
5 Harlingen
2 Hoek van Holland
1 6 Delfzijl
-1
-2
-12 -6 0 6 12 18 24
time after moon culmination [h]
Figure 5.55: Tidal curves at six stations along the Dutch coast (data source: https://waterinfo.
rws.nl/). The station numbers rise from south to north. The time in hours on the horizontal axis
is relative to the moon culmination on October 14th 2019 01:35 (calculated using https://www.
mooncalc.org). Note that MSL is approximately equal to NAP; current MSL is less than 0.1 m
above NAP.
The figure shows that not only the phase 𝑘𝑦 differs along the coast, but the shapes and
amplitudes as well. Figure 5.56 clearly shows that the tidal range is largest at Vlissin-
gen (on average 3.8 m) and smallest around Den Helder (on average 1.3 m). Along the
Holland coast, the central part of the Dutch coast, the average tidal range is 2 m at
most7 . Nevertheless, this tidal range is significantly larger than the tidal range from
equilibrium theory. The alongshore differences are related to the position of the am-
phidromic points in the North Sea (see Fig. 3.31).
As opposed to the stochastic wind waves, the tidal motion is deterministic, viz. inde-
pendent of weather or climatic conditions. Once the tidal constituents (amplitudes and
phases) are known at certain measurement locations, the tide can be forecast. Using
a model based on momentum and continuity equations and data from measurements
made simultaneously at different points along the coast, the tide at locations different
from the measurement locations can be forecast as well.
Skewness and asymmetry

Figures 5.55 and 5.56 indicate that the tidal curves deviate from an ideal symmetrical
tide. We see that:
• At some of the locations, high water is further above the mean than low water
is below it; this also implies a shorter duration of positive water levels than
negative water levels; this type of asymmetry is equivalent to the skewed wind
7
The spring tidal range, however, is everywhere at least 2 m. Therefore, the Dutch coast qualifies as a
meso-tidal regime, see Sect. 4.4.1.

elevation [m + NAP]
SW delta coast Holland coast Wadden coast
alongshore position from south to north
Figure 5.56: Average tidal range along the Dutch coastline. MHW and MLW are computed for
each station as the yearly average of all recorded high waters and low waters in 2019 (data
source: waterinfo.rws.nl). The 𝑥-axis denotes the alongshore position north of Vlissingen.
waves in the shoaling zone, with long, flat troughs and narrow, peaked crests
(see Sect. 5.3). In Vlissingen the skewness of the tidal elevation is positive and
in Den Helder negative.
• the time it takes for the water to reach the lowest elevation (the falling period)
is not equal to the time it takes to reach the highest elevation (the rising period).
The resulting shape of the tidal curve is asymmetric about the vertical axis, com-
parable with the asymmetry in wind waves just before breaking.
Figure 5.57 shows in more detail an average tidal curve for Vlissingen. It shows a
falling period of 6.28 h and a rising period of 5.57 h. Although the exact ratios vary,
for all stations along the Dutch coast the falling period is longer than the rising period.
For instance, in IJmuiden the falling period and rising period are 8.03 h and 4.22 h
respectively. This phenomenon is often referred to as tidal asymmetry.
The longer falling period can be explained from the phase velocity for shallow water
𝑐 = √𝑔(ℎ + 𝜂). For high tide (𝜂 positive), the propagation velocity is larger than for low
tide (𝜂 negative). Since the high tide (the wave crest) propagates faster than the low
tide (the trough), the rising period is smaller than the falling period. This generally
holds for the open coast, but in basins this may be different.
When it is said that the (vertical) tide is flood-dominant, this refers to a shorter rising
than falling period. Vice versa, ebb dominance indicates a shorter falling period. For
the reasons mentioned in Sect. 5.7.1, the terms flood and ebb dominance in relation

to the vertical tide may be confusing. We will therefore use flood (or ebb) dominance
only to indicate the direction of net sediment transport (see Sect. 9.7) as a result of
asymmetries of the horizontal tide.
daily
0:57
0:57
0:57
inequality
+190 +190
+182 8
NAP NAP
MSL −10
culmination of the moon
12:25
daily
−178 inequality
−194 16 −194
7:25
7:25
7:25
falling period rising period

6:28 5:57 LLWS −250
Figure 5.57: Average tidal curve in Vlissingen (adapted from Rijkswaterstaat, 1949). Time (ho-
rizontal axis) in hours and height (vertical axis) in cm. Note that MSL is shown as being below
NAP, which is nowadays not the case any longer.
Like the vertical tide, the horizontal tide may display both skewness and asymmetry:
• the average peak flood current may be stronger than the average peak ebb cur-
rent, which leads to a shorter flood duration than ebb duration (or vice versa);
• the velocity signal may be asymmetric around the vertical, which means that
the rate at which the velocity changes around slack water (i.e. flow reversal) is
different when changing from ebb to flood than when changing from flood to
ebb (see for instance Fig. 5.52).
Figure 5.53 shows that if the horizontal and vertical tide are in phase, the skewed and
asymmetric tidal elevation directly translates to similar characteristics for the tidal
velocity. In basins the relationship between the vertical and horizontal tide can be
more complex. Tidal asymmetry will be discussed in more detail in Sect. 5.7.4.

Cross-shore distribution of the tidal velocity

Surface waves refract towards the coast and generate a longshore current along the
coast. Tidal propagation tends to be along a coast or channel. Hence, both wave-
induced longshore currents and tidal currents are mainly parallel to the coast (see
Fig. 5.58). The direction of the tidal current reverses during the tidal cycle. Alongshore
tidal velocities in shallow water can be anything from a few decimetres per second to
several metres per second.
C D
tide
A B
shoreline y
Figure 5.58: Tidal current along the shore. The negative 𝑥−axis and positive 𝑦−axis are indic-
ated.
In very shallow water, the inertia effect is small and the velocity at any moment in the
tide is governed by the balance between alongshore water level gradient and friction.
According to Eq. 5.92, the velocity magnitude is, at every moment of the tidal phase,
linearly dependent on the water depth and the alongshore water level gradient. If we
use a quadratic friction law with a constant friction factor, the velocity magnitude is
proportional to the square root of the product of water depth and alongshore water
level gradient:
𝜕𝜂
𝑣∝ ℎ (5.93)
√ 𝜕𝑦
If the tidal velocity at one particular water depth is known, for instance from meas-
urements, then a very practical method for finding the cross-shore distribution of the
velocities is by using Eq. 5.93. This leads to:
ℎ2
𝑣2 = 𝑣 1 (5.94)
√ ℎ1
where:
𝑣1,2 tidal current velocities at cross-shore positions 1 and 2 m/s

ℎ1,2 still water depth at points 1 and 2 m

Note that in this approach it is assumed that the alongshore water level gradient does
not vary in the cross-shore direction (the water level difference between D and C in
Fig. 5.58 is equal to the difference between B and A). Following this simple approach,
a tidal alongshore velocity of 0.7 m/s at a water depth of 10 m yields a tidal velocity
of 0.3 m/s at a water depth of only 2 m. In this shallow region, with 2 m water depth,
waves can be very effective in increasing bed shear stress (the wave boundary layer
acts as an extra resistance for the flow). Consequently, the tidal velocities in the wave-
influenced littoral zone will even be smaller.
Effect of tide on wave-generated longshore current

The tide influences the wave-generated longshore current in the breaker zone. One
complicating factor here is that the direction of the tidal velocity changes twice during
one tidal cycle. The appropriate way to include the effect of tides is first to combine
the driving forces (during ebb and flood periods) and then to calculate the velocity.
Simply adding the tidal velocity to the wave-induced longshore current is not correct
(that would only be possible if the velocity is a linear function of the driving force,
which is not the case).
Figure 5.59 shows the velocity distribution for the combination of wave-generated cur-
rents with ebb and flood tidal velocities. The maximum tidal velocity occurs outside
the breaker zone, but the effect inside the breaker zone can be quite substantial. If the
tidal force is in the same direction as the wave-driven current, the maximum along-
shore flow velocity increases (and shifts towards the breaker line). If the tidal force
is in the opposite direction, the maximum velocity decreases (and shifts towards the
shoreline). The longshore current in the breaker zone, although reduced in velocity,
might then be in the opposite direction as the tidal current present further seawards.
If the wave-induced alongshore driving forces are large enough relative to the tidal
forces, the tidal current in the breaker zone may be overshadowed by the wave-induced
longshore current. Flow reversal with the tide may not occur.
Tidal currents may significantly complicate the determination of the longshore current
in the breaker zone. In practical cases, like for instance for the Dutch coast, it will be
necessary to establish the longshore current throughout the entire tidal cycle. Only in
areas with weak tidal forces (for instance the Mediterranean) the effect of tides on the
water movement in the breaker zone can be neglected.
Tidal currents around structures

Structures tend to divert the tidal flow especially when they extend far seaward. The
harbour moles of IJmuiden, the sea port of Amsterdam, were extended to approxim-
ately 2500 m in the period 1962 to 1968. The convergence of the tidal flow (contraction
of the streamlines) around this breakwater led to relatively high velocities in front of
the harbour entrance, where subsequently a scour hole developed. Another example
is the long dam that was constructed at the Dutch Wadden island of Texel in 1995.

tidal wave-induced
longshore longshore
current current
surf
zone
tidal wave-induced
longshore longshore
current current
surf
zone
Figure 5.59: Combination of wave-induced and tidal longshore currents at two different phases
of the tide: (a) tidal current; (b) wave-induced longshore current and (c) combination of the two.
In order to compute the combination, the driving forces rather than the velocities must be added.
Figure 5.60 schematically shows the deflection of tidal currents by IJmuiden harbour.
Due to the tidal variation, the flow field is not stationary. The depicted flow field
represents the ebb flow. Near the harbour entrance a flow contraction can be noticed,
while downstream from the harbour moles, an eddy is visible. A simple rule of thumb
states that the alongshore length of the eddy should be – in a stationary situation –
around six times the length of the harbour mole. Due to the tide reversal, however, the
growth of the eddy is restricted. The flow patterns have implications for the sediment
transport and resulting morphology.
In the computation that this schematic was based on, the harbour basin itself was not
included. The tidal flow passing the harbour entrance can drive an eddy in the harbour
basin. This can lead to an exchange of water and sediment between the harbour and
the area outside.

eddy
1 m/s
Figure 5.60: Schematic representation of the ebb flow field in the vicinity of IJmuiden harbour.
5.7.3. Tidal propagation into basins

In tidal basins we generally find a main channel, which transports the majority of the
ebb and flood discharge. It is along this main channel that the tidal wave is primarily
propagated. If we choose to align the 𝑥-axis with the channel-axis, we may approx-
imate the tidal propagation with a set of one-dimensional equations. These are the
balance equations for mass and momentum in the 𝑥-direction.
To derive these equations, we schematise the cross-section 𝐴(𝑥, 𝑡) of the basin with a
width 𝑏 into two parts. One part represents the flow carrying part of the cross-section,
discharging all the flow; the other part is the storage part, which only stores water
without discharging, see Fig. 5.61. The storage part represents the tidal flats that are
covered during higher water and are exposed during lower water.
b
bs
As
Figure 5.61: Definition of cross-sectional parameters.
The flow-carrying cross-section has an area 𝐴𝑠 (𝑥, 𝑡), which is the product of a flow-
carrying width 𝑏𝑠 (𝑥, 𝑡) and a representative instantaneous depth ℎ(𝑥, 𝑡). This depth
ℎ(𝑥, 𝑡) is the sum of the representative flow-carrying water depth ℎ0 (𝑥, 𝑡) relative to
a horizontal mean water level and the tidal level 𝜂(𝑥, 𝑡). The mean flow velocity in
the flow-carrying cross-section (averaged over the total cross-section) is indicated as
𝑢𝑠 (𝑥, 𝑡). The flow-carrying cross-section can be written as 𝐴𝑠 = 𝑏𝑠 ℎ = 𝑏𝑠 (ℎ0 + 𝜂). The
tidal discharge through this cross-section is 𝑄 = 𝐴𝑠 𝑢𝑠 .

The one-dimensional balance equations are derived for an infinitesimal short section
in the basin. The mass balance (continuity equation) yields:
𝜕 𝜕𝜂 1 𝜕 𝜕𝜂
(𝐴 𝑢 ) + 𝑏 =0 ⇔ (𝑏𝑠 ℎ𝑢𝑠 ) + =0 (5.95)
𝜕𝑥 𝑠 𝑠
⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟ ⏟ 𝜕𝑡 𝑏 𝜕𝑥 𝜕𝑡
volume change volume change
due to in- and due to water
outgoing transport level change
The mass balance equation can be simplified with 𝑏 = 𝑏𝑠 (no intertidal storage areas),
and under the assumption of a prismatic channel (cross-sectional shape and size and
bottom slope are constant along the channel, hence 𝑏 is independent of 𝑥). This gives:
𝜕 𝜕𝜂
(ℎ𝑢) + =0 (5.96)
𝜕𝑥 𝜕𝑡
Note that for convenience we have dropped the subscript 𝑠 for 𝑢. The total water depth
can be taken outside of the 𝑥-derivative for a small ratio of tidal amplitude over water
depth and a prismatic channel. Compare this equation to Eq. 3.41.
The momentum balance reads:
𝜕 𝜕 𝜕𝜂 𝑏𝑠
(𝐴𝑠 𝑢𝑠 ) + (𝐴𝑠 𝑢𝑠2 ) + 𝑔𝐴𝑠 + 𝜏𝑏 = 0 (5.97)
𝜕𝑡
⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟ 𝜕𝑥
⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟ 𝜕𝑥 ⏟⏟⏟⏟⏟⏟⏟⏟⏟
⏟⏟⏟⏟⏟⏟⏟ 𝜌
momentum in- and outflow pressure
bottom
change of momentum gradient friction
With 𝑏 = 𝑏𝑠 and under the assumption of a prismatic channel (so that 𝐴𝑠 does not vary
with 𝑥), we find the 1D shallow water tidal propagation equation:
𝜕𝑢 𝜕𝑢 𝜕𝜂 𝜏
+𝑢 +𝑔 + 𝑏 =0 (5.98)
𝜕𝑡 𝜕𝑥 𝜕𝑥 𝜌ℎ
Here, we have again dropped the subscript 𝑠 for 𝑢.

By conducting a scale analysis, it can be shown under what conditions which terms are
of similar order of magnitude and which terms may be neglected. For typical orders of
magnitude of the tidal amplitude to water depth ratio 𝑎/ℎ = 𝒪(10−1 ) and ℎ ≈ 𝒪(10 m)
the advective (second) term on the left-hand side can be neglected.
For simplicity we assume a linear friction law 𝜏𝑏 = 𝜌𝑐𝑓 |𝑢|𝑢 ≈ 𝜌𝑟𝑢. We get:
𝜕𝑢 𝜕𝜂 𝑟
+𝑔 + 𝑢=0 (5.99)
𝜕𝑡 𝜕𝑥 ℎ
Compare the above equation to Eqs. 3.40a and 3.40b and Eq. 5.91.

As in Intermezzo 3.5, we take the 𝜕/𝜕𝑡 of the continuity equation (Eq. 5.96) and subtract
ℎ𝜕/𝜕𝑥 of the momentum equation (Eq. 5.99). For a small tide (𝑎/ℎ small) this leads to:
𝜕 2𝜂 𝜕 2 𝜂 𝑟 𝜕𝜂
− 𝑔ℎ 2 + =0 (5.100)
𝜕𝑡 2 𝜕𝑥 ℎ 𝜕𝑡
This equation has elements of the classical wave equation (Eq. 3.35) and of a diffusion
equation (last two terms of the left-hand side).
We can implement further simplifications depending on the relative magnitude of the
inertia (local acceleration) and friction term. In many shallow tidal basins, the flow is
friction-dominated. The inertia term may then be neglected to a first approximation.
Friction-dominated flow
Consider a shallow estuary for which friction dominates inertia. The basin is long, i.e.
all tidal energy is dissipated in the basin before the tide reaches the end of the basin.
In a straight basin Eq. 5.100 is valid. For friction-dominated flow, Eq. 5.100 reduces to
a diffusion equation (the last two terms remain):
𝜕𝜂 𝜕 2𝜂
=𝐷 2 (5.101)
𝜕𝑡 𝜕𝑥
in which 𝐷 = 𝑔ℎ2 /𝑟 is the tidal diffusion coefficient. The solution to this equation can
be shown to be:
𝜔
𝜂(𝑥, 𝑡) = 𝑎𝑒 −𝑘𝑥 cos(𝜔𝑡 − 𝑘𝑥) with 𝑘= (5.102)
√ 2𝐷
From substitution in the continuity equation, Eq. 5.96, we get:
𝑎
𝑢(𝑥, 𝑡) = √𝜔𝐷𝑒 −𝑘𝑥 cos(𝜔𝑡 − 𝑘𝑥 + 𝜋/4) (5.103)
ℎ
The amplitudes of the horizontal and vertical tide are progressively damped with dis-
tance 𝑥 from the inlet. Furthermore, the velocity leads the surface elevation by 45°.
The phase speed 𝑐 = 𝜔/𝑘 is given by 𝑐 = 𝜔/𝑘 = √2𝜔𝐷. Note that this differs from the
phase speed of a linear progressive wave 𝑐 = 𝜔/𝑘 = √𝑔ℎ.
Note that with a quadratic friction law (instead of the simple linear approximation)
the tidal diffusion coefficient is dependent on the flow magnitude and therefore varies
strongly over the tidal period. This non-linear character implies that the basin deforms
the tidal wave during diffusion and it loses its sinusoidal character. The symmetry
between ebb and flood is broken.

Around slack water the flow will not be described correctly by the tidal diffusion equa-
tion; during the slack waters friction is not dominant. The tide then propagates ap-
proximately as an undamped wave (see the progressive wave solution of the classical
wave equation of Intermezzo 3.5).
The Dutch former Zuiderzee is an example of a tidal basin of such length that the tidal
wave at the end of the basin was nearly damped out. The tidal wave entering through
the Texel Inlet or Marsdiep and the Vlie Inlet was largely damped out at Amsterdam
and the Veluwe banks. In tidal rivers (where the river flow velocity and the tidal flow
velocity are of similar order) we also find slowly-damping propagating tidal waves,
e.g. the Lek and the Waal in the Netherlands and the St. Lawrence river in Canada. In
these tidal rivers, the reflected tidal wave is negligibly small, so that the tide has the
character of an inland-diffusing wave. See Fig. 5.62.
estuary tidal river river

Qr
Qr Qr Qr Qr
t t t t
Figure 5.62: In a tidal river, the river and tidal discharge are of the same order of magnitude.
Inertia-dominated flow
Let us now consider the (mostly unrealistic) case of negligible friction. Eq. 5.100 re-
duces to the classical second-order wave equation. We have already seen this equation
when discussing tidal propagation in seas and ocean basins (Eq. 3.35). This equation
allows wave propagation in two directions (in positive and negative direction along
the channel axis). Since the incident tidal wave is not damped by friction, it is reflec-
ted at the basin end 𝑥 = 𝐿𝑏 . The surface elevation at the inlet 𝑥 = 0 must be equal to

a sinusoidal open sea or ocean tide 𝜂(0, 𝑡) = 𝑎 cos 𝜔𝑡. This results in a standing tidal
wave:
1 𝑎
𝜂(𝑥, 𝑡) = [ cos (𝜔𝑡 − 𝑘 (𝑥 − 𝐿𝑏 )) + ⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟
cos (𝜔𝑡 + 𝑘 (𝑥 − 𝐿𝑏 ))] =
2 cos 𝑘𝐿𝑏 ⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟
incident wave reflected wave
cos (𝑘 (𝐿𝑏 − 𝑥))

=𝑎 cos 𝜔𝑡
cos 𝑘𝐿𝑏
(5.104a)
1 𝑎𝑐
𝑢(𝑥, 𝑡) = [cos (𝜔𝑡 − 𝑘 (𝑥 − 𝐿𝑏 )) − cos (𝑘 (𝑥 − 𝐿𝑏 ) + 𝜔𝑡)] =
2 ℎ cos 𝑘𝐿𝑏
𝑎𝑐 sin (𝑘 (𝐿𝑏 − 𝑥))

=− sin 𝜔𝑡
ℎ cos 𝑘𝐿𝑏
(5.104b)
In Eq. 5.104b, 𝑐 is the phase speed and 𝑐 = √𝑔ℎ. Water level and flow velocity are 90°
out of phase, as can be seen from the terms cos 𝜔𝑡 and sin 𝜔𝑡 in Eqs. 5.104a and 5.104b
respectively. This is an important characteristic of a standing wave pattern.
The amplitudes along a basin are sketched in Fig. 5.63. The amplitude of the tidal
elevation along the estuary varies according to cos(𝑘(𝐿𝑏 −𝑥)) = 0. Hence, the standing
wave has a maximum amplitude (or antinode) in the surface elevation for 𝑘(𝐿𝑏 −𝑥) = 0,
i.e. at the landward end 𝑥 = 𝐿𝑏 . The tidal amplitude at the landward end is therefore
larger than that at the seaward end. The antinode in the surface elevation corresponds
to a minimum (or node) in the velocity amplitude; at the landward end the velocity is
zero. This can be seen from the term sin(𝑘(𝐿𝑏 − 𝑥)) in Eq. 5.104b.
node antinode
antinode node
basin
sea
Figure 5.63: Standing wave pattern in a tidal basin.

If the basin is long enough, a node in the elevation (zero amplitude) and an antinode
in the velocity (maximum amplitude) will occur for 𝑘(𝐿𝑏 − 𝑥) = 𝜋/2. The longer the
basin, the more nodes and antinodes it can accommodate.
The amplitudes of both the vertical and horizontal tide are inversely proportional to
cos 𝑘𝐿𝑏 . If 𝑘𝐿𝑏 = 𝜋/2, 3𝜋/2, …, then cos 𝑘𝐿𝑏 = 0 which means that the amplitudes
of 𝜂 and 𝑢 become infinitely large. Note that 𝑘𝐿𝑏 = 𝜋/2 is equivalent to 𝐿𝑏 = 1/4𝐿.
So, resonance occurs for basins with a basin length equal to a quarter of the tidal
wavelength 𝐿 or an uneven multiple of a quarter wavelength.
Resonance occurs because for 𝑘𝐿𝑏 = 𝜋/2 the incident and reflected wave cancel each
other at the mouth, or, in other words, we have a node of the surface elevation at the
mouth. The amplitude of the vertical tide at the mouth is therefore zero relative to
the amplitude in the basin. At the same time we had the condition that the tide at the
mouth was equal to the ocean tide 𝜂(0, 𝑡) = 𝑎 cos 𝜔𝑡. Since this must be zero relative
to the tide in the basin, the amplitude in the basin must go to infinity. In reality, this
would not occur, since bed friction would dampen the amplitudes.
So far, we considered a sinusoidal tidal motion and thus a single tidal frequency. In
reality, more tidal constituents are present. If the natural period of a basin is close
to the period of one of the tidal constituents, that constituent will be amplified by
resonance more than others.
Combination of friction and inertia

In most tidal basins both friction and inertia play a role. The tidal wave can be de-
scribed as the superposition of an incoming and a reflected damped wave. Without
friction a standing wave pattern will result. With friction, the incoming (and reflected)
wave is partly damped, so that the result is a total surface elevation pattern that has a
partly propagating and partly standing character.
Figure 5.64 shows the ratio of the amplitude of a harmonic tidal component at the
end of the basin over the amplitude at the mouth. Depicted is the solution of the 1D
tidal propagation in a prismatic basin for various values of 𝑠1 = 𝑟/ℎ𝜔. Check from
Eq. 5.100 that 𝑠1 is the ratio between the friction term and the local acceleration term.
For 𝑠1 = 𝑟/ℎ𝜔 = 0, we recognise the frictionless situation with resonance for 𝐿𝑏 = 1/4𝐿
and 3/4𝐿. For large values of 𝑠1 , the damping of the tide due to friction can be recognised.
The values in between are a combination of both effects.
Width convergence and shoaling

So far we have only considered prismatic basins. In such basins, dissipation of energy
due to friction inevitably dampens the tidal amplitude.
In the case of a funnel-shaped estuary, the cross-section is largest at the inlet and then
gets progressively smaller. This gives a convergence of tidal energy that increases the
tidal amplitude along the channel axis. The balance between the counteracting effects

10
9 s1 = 0 s1 = 0
7
s1=0.2
6
s1 = 0.3
4
s1 = 0 s1 = 0
3
s1 = 0.5
s1 = 2
2 s1 = 0.7
s1 = 5 s1 = 0.2
s1 = 1
1 s = 0.3
s11 = 0.5
s1 = 10 s1 = 0.7
0 s1 = 1
0 1 2 3 4 5
π/2 s2 3π/2
Figure 5.64: Tidal propagation in a prismatic basin (linear solution). The variable on the hori-
zontal axis is the basin length in terms of 𝑠2 = 𝜔𝐿𝑏 /√𝑔ℎ = 𝑘𝑟=0 𝐿𝑏 with 𝑘𝑟=0 the wavenumber for
the frictionless situation. The amplification factor on the vertical axis is the tidal amplitude at
the basin end over the tidal amplitude at the mouth. The amplification factors are shown for
different values of 𝑠1 = 𝑟/(ℎ𝜔). Note that 𝑠2 = 𝜋/2 and 𝑠2 = 3𝜋/2 are equivalent to 𝐿𝑏 = 1/4𝐿𝑟=0
and 3/4𝐿𝑟=0 , respectively, with 𝐿𝑟=0 the tidal wavelength for the frictionless situation. These are
the conditions for resonance found in the previous paragraph on inertia-dominated flow.
of energy loss due to friction and energy convergence due to width restriction determ-
ines whether the tidal amplitudes increase or decrease along the channel axis. Note
that a rapid convergence of energy for the incident wave implies a rapid divergence
of energy for a reflected wave.
Progressive shallowing of a basin gives a similar concentration of wave energy coun-
teracting the dampening by friction. The reason is that the propagation speed becomes
smaller in smaller water depths. This is identical to the phenomenon of shoaling in
wind waves (see Sect. 5.2.2).
In the case of gradual changes in width and depth (no reflection on the sides or on
a sill) and, in the absence of friction, the energy is conserved along the channel axis.
Analogous to Eq. 5.9 we have:
1/4 1/2
𝜂̂2 ℎ 𝑏
𝐸𝑛𝑐𝑏𝑠 = constant → 𝜂̂2 √𝑔ℎ𝑏𝑠 = constant → ∝ ( 1 ) ( 𝑠1 ) (5.105)
𝜂̂1 ℎ2 𝑏𝑠2

According to this equation, width reduction has a stronger effect on the tidal amplitude
than depth reduction.
Storage considerations and short basins

A short basin has a length that is short relative to the tidal wavelength: 𝐿𝑏 ≪ 1/4𝐿, say
𝐿𝑏 < 1/10𝐿 or 𝐿𝑏 < 1/20𝐿. For such a short basin we can find a solution for the tidal
elevation and velocity based on the continuity equation alone. Let us first look at the
general continuity equation Eq. 5.95. It can be written as:
𝜕𝑄 𝜕𝜂
− =𝑏 (5.106)
𝜕𝑥 𝜕𝑡
Integration along the channel axis from a location 𝑥 to the end of the basin gives:
𝐿𝑏
𝜕𝜂
𝑄(𝑡, 𝑥) = ∫ 𝑏 𝑑𝑥 (5.107)
𝑥 𝜕𝑡
This equation shows that the tidal discharge in a certain cross-section depends on the
amount of water needed to fill the basin landward of the cross-section under consid-
eration. This volume of water (excluding any fresh water) that has to flow in and
out through the inlet during one tidal cycle is called the tidal prism. It thus makes
sense that the tidal prism has empirically been found to determine the equilibrium, or
minimal stable cross-sectional channel area of a cross-section (Sect. 9.5.2).
In a short basin this situation can be simplified, since we can expect the water level
in the basin to immediately follow the water level in the sea. There is no variation
in the tidal amplitude along the channel axis: 𝜕𝜂/𝜕𝑥 = 0. For a sinusoidal tide at the
seaward boundary, the entire basin thus oscillates in exactly the same way as the first
harmonic. Eq. 5.107 then yields:
𝜕𝜂0 𝐿𝑏 𝜕𝜂0
𝑄(𝑡, 𝑥) = 𝐴𝑠 𝑢𝑠 = ∫ 𝑏 𝑑𝑥 = 𝐴 ⇔ (5.108)
𝜕𝑡 𝑥 𝜕𝑡 𝑏
𝜕𝜂0 𝐴𝑏
𝑢𝑠 (𝑥, 𝑡) = (5.109)
𝜕𝑡 𝐴𝑠
in which 𝐴𝑏 is the basin surface area upstream of 𝑥 and 𝐴𝑠 is the channel cross-
sectional area at 𝑥. At the time of high water in the sea, everywhere in the basin
the water levels are maximum. Low water also occurs everywhere at the same time.
Therefore the velocities are zero at high and low water. At the time of mean water
level, the ebb and flood currents are maximum. This corresponds with a velocity that
leads the surface elevation (by 𝜑 = −𝜋/2 = −90°). The velocities in a cross-section
are larger if a larger upstream surface area needs to be filled. Due to the uniformly
oscillating water level, this situation is referred to as pumping mode.

In the case of a narrow gorge to the small basin, friction can reduce the amplitude
inside the short basin and change the phase relationships.
5.7.4. Tidal asymmetry

In Sect. 5.3 it was described how wind waves propagating towards the shore become
more and more asymmetric until the point of wave-breaking. We had seen that this
shoaling process was characterised by:
• increase in amplitudes;
• peaking of the wave crest and a flattening of the trough, and;
• relative steepening of the face resulting in a pitched-forward wave shape.
Comparable effects occur for tidal propagation in basins. This is illustrated in Fig. 5.65
for a number of tidal basins in the Netherlands.
2 Eems-Dollard Texel Inlet 1
current velocity [m/s]

1 0.5
tide level [m]
0 0
-1 -0.5
-2 -1
75
2 Eastern Scheldt Western Scheldt 50

discharge [103 m3/s]
1 25
tide level [m]
0 0
-1 -25
-2 -50
-75
0 4 8 12 16 0 4 8 12 16
t [hours] t [hours]
Figure 5.65: Tidal curves for water level, flow velocity and discharge for several Dutch basins.
(Data from Dronkers, 1986).

As the tide propagates up the estuary, the water depth and basin width change. The
shoaling and narrowing of the estuary slows the progress of the tidal wave, increas-
ing its amplitude along the channel axis. Friction, on the other hand, will reduce the
amplitudes. The net effect depends on the situation. Furthermore, tidal asymmetry
develops. As a result, in most tidal basins the tidal curves for water level and flow
velocity (or discharge) strongly differ from a sinusoidal curve.
First consider the situation of a prismatic tidal river or basin without intertidal storage
areas. We also neglect reflection from the head. In the absence of friction, the propaga-
tion velocity is given by 𝑐 = √𝑔ℎ = √𝑔(ℎ0 + 𝜂). We can also write the propagation
velocity as 𝑐 = √𝑔𝐴𝑠 /𝑏. Without intertidal storage areas, the high and low tide use the
same channels and the propagation velocity of the high tide 𝑐 = √𝑔(ℎ0 + 𝑎) is larger
than that of the low tide 𝑐 = √𝑔(ℎ0 − 𝑎). We can thus expect the rising period to be
shorter than the falling period (see Fig. 5.66). Hence, the tidal wave becomes distorted,
with a steep vertical face.
HW
bore
rising
period
HW
falling
LW period
HW
0
sea
Figure 5.66: Tidal propagation along a basin (𝑥-axis) in the case of a faster propagation of High
Water (HW) with respect to the propagation of Low Water (LW). The rising period becomes in-
creasingly smaller than the falling period. A bore develops when the rising period is reduced to
zero (or when high water ‘catches up’ with low water).
In extreme cases, as a result of certain bathymetric conditions, the steep vertical face
can take on the form of a wall of water travelling up the basin. This causes an almost
instantaneous rise of the water level as the water wall passes (see Fig. 5.66). This
phenomenon is called a tidal bore (see Intermezzo 5.7). In the modelling of the surface
in breaking waves, an analogy with a bore is often made.
Friction gives an additional slowing down of the low tide with respect to the high tide,
since the low tide ‘feels’ the bottom more. In tidal rivers this effect is even larger; since
the river flow velocity and the tidal flow velocity are of similar order, the flow velocity
and hence the friction is much larger at maximum ebb than at maximum flood.
Friction also causes damping of the amplitude that is stronger during ebb around LW
than during flood around HW. The preferential damping of the LW is depicted in

Fig. 5.67. It results in an asymmetry of the surface elevation about the horizontal axis
(a positively skewed signal).
amplitude
HW
LW
Figure 5.67: Amplitudes of HW and LW for tidal propagation along a basin (𝑥-axis). In this case,
HW is damped less by friction than LW, and, consequently, HW propagates faster than LW. This
is particularly true if, due to river discharge, the flow velocity at ebb is much higher than at flood.
Intermezzo 5.7 Tidal bore
Some rivers located at the landward end of an estuary experience an extreme tide-
dependent condition - a tidal bore, an abrupt and migrating rise in the water level
(Figs. 5.68 and 5.69). Tidal bores can cause great inconvenience to ship traffic. In
those cases port authorities try and eradicate the bore by changing the shape and
depth of the estuary.
Figure 5.68: Spectators watching the tidal bore in the Qiantang Estuary, China. Photo by
Marcel J.F. Stive (‘Credits’ on page 575).

Figure 5.69: Tidal bore on the Nith River, Scotland. Photo by Richard P. Long (‘Credits’ on
page 575).
Bores are uncommon, forming only in special circumstances that depend on tidal
conditions and the morphology of the estuary. The bore in the Salmon river near
Truro at the head of the Bay of Fundy is typically only about half a metre high. In
the Bay of St. Malo on the northern coast of France, a bay with the world’s second
largest tidal range, the bore rarely exceeds a metre in height. Large tidal bores
occur in the Araguari River in the Amazon system and in the Qiantang estuary in
China. The bore reaches 5 m in the Araguari River and nearly the same height in
the Qiantang. In the Brazilian Amazon, tidal bores are known as pororoca.
Flood and ebb dominance

In the frictionless case and without reflection from the head, the vertical and horizontal
tides are in phase. As a result, the horizontal tide displays similar sawtooth asymmetry
as the vertical tide. The flood duration is identical to the ebb duration and the ebb
velocities are equal to the flood velocities.
Due to the effect of friction, we can expect a phase shift between vertical and hori-
zontal tide that increases inland. In the extreme case of the horizontal tide leading
the vertical tide by a phase difference of 𝜋/2 (see Fig. 5.54), a shorter rising period
directly corresponds to a shorter flood duration than ebb duration. In the absence of
other non-tidal forcing or river discharge, the net tide-averaged discharge should be
zero, even though the tidal current is asymmetric. Therefore, a shorter flood duration
means that the maximum flood velocities are higher than the maximum ebb velocities.
Systems in which the maximum flood velocities are higher than the maximum ebb
velocities are called flood-dominant. If the maximum ebb velocities are higher, we
speak about ebb dominance. Flood dominance can be expected for a large tide (or

more precisely a large ratio of tidal amplitude over water depth 𝑎/ℎ). In that case,
the propagation of high water is faster than of low water and thus the rising period is
shorter than the falling period. The result is floods with a higher velocity and a shorter
duration. This effect increases for longer basins.
Flood dominance can be counteracted by the presence of a river flow that increases
the seaward-directed velocities (see Fig. 5.72). For negligible river flow, the presence of
intertidal storage areas counteracts flood dominance. In many tidal basins, intertidal
flats are present, which fall dry and are flooded again during the tidal cycle. The small
water depths on these intertidal marshes and flats cause the high tide to propagate
slower than the low tide. This is sometimes indicated by the ratio of intertidal storage
volume over channel volume 𝑉𝑠 /𝑉𝑐 (Friedrichs & Aubrey, 1988). Dronkers (1986, 2005)
uses another indicator, viz. the wet surface area at HW over the wet surface area at
LW: 𝐴𝑏,HW /𝐴𝑏,LW (see Fig. 9.34).
Ebb dominance is further enhanced by the fact that the water level averaged over the
flood period is generally higher than that averaged over the ebb period. Since during
ebb the discharge is similar to the discharge during flood, the ebb velocities must on
average be larger because of the smaller cross-section for the flow (for the same flood
and ebb durations).
Summarizing, ebb dominance means that the ebb has higher maximum velocities
and a shorter duration than the flood, whereas flood dominance refers to floods
with a higher velocity and shorter duration. A large tidal amplitude and shallow
channels enhance flood dominance. A large intertidal storage volume (as com-
pared to channel volume) enhances ebb dominance. This shows that the basin
geometry controls the tidal distortion.
Sawtooth asymmetry of the horizontal tide

So far we mainly mentioned skewness (flood or ebb dominance) of the horizontal tide.
The horizontal tide can also display sawtooth asymmetry. In case of a short basin
(length much smaller than a quarter of the tidal length, with the amplitudes of the
incoming and reflected tidal wave of the same order), this will be the dominant form
of asymmetry.
In a short basin, the tidal phase difference between seaward and landward end is neg-
ligible. Further, in a short basin 𝜂 and 𝑢 are out of phase; LWS (flow reversal from ebb
to flood) occurs at LW, and HWS (flow reversal from flood to ebb) occurs at HW. At
slack water, 𝑑𝜂/𝑑𝑡 = 0. The rate of change of the velocity 𝑑𝑢/𝑑𝑡 around slack water is
proportional to the landward basin area 𝐴𝑏 and inversely proportional to the channel
cross-section 𝐴𝑠 (this can be found using Eq. 5.108).
Consider a short basin with shallow channels and little storage-offering flats. The
channel cross-section 𝐴𝑠 is much larger around HW than at LW. As a consequence,

the rate of velocity change at HWS is smaller; the HWS duration is longer than the
LWS duration.
On the other hand, in a basin with vast storage-offering tidal flats and deep channels,
𝐴𝑏 is much larger around HW than around LW. As a result the variation of the flow
velocity around HWS is much larger than around LWS; the HWS duration is shorter
than the LWS duration.
Impact of the tide at sea

The asymmetry of the velocity variation in a tidal basin will strongly depend on the
water level variation of the tide at sea. This holds especially for short basins, and to
a lesser degree also for longer basins. A faster water level rise than fall at sea causes
higher flood flow velocities than ebb flow velocities in the basin and vice versa. A long
duration of HW at sea causes a long HWS duration and a short duration of HW at sea
a short HWS duration. This holds in the same way for low water.
Deformation of the tidal wave along the coast has an impact on the ebb-flood asym-
metry in the adjacent tidal basins and hence on the net inward or seaward sediment
transport and the morphological evolution. Fig. 5.70 illustrates the tidal variation for
the Dutch coast and the adjacent tidal basins; this shows how the sea-borne tidal asym-
metry is transferred into the basins.
Relevance to sediment transport

The asymmetries of the horizontal tide are of extreme importance for the net sediment
transport. The first type of asymmetry, skewness or asymmetry about the horizontal,
is important for residual (net) sediment transport in tide-dominated areas (mainly tidal
inlet systems). For instance, if the maximum flood velocity exceeds the maximum ebb
velocity, a residual sediment transport in the flood direction is likely to happen, since
sediment transport responds non-linearly to the velocity. Hence, flood dominance
implies a residual transport in the flood direction. For medium to coarse sediment
this is the dominant effect. Note that in the same way, short-wave skewness in the
nearshore (see Fig. 5.14) leads to a net transport in the wave propagation direction.
Chapters 6, 7 and 9 will discuss transport due to skewness of the velocity signal in
both wave- and tide-dominated environments in more detail.
Flood-dominant systems tend to import sediment (assuming the transport is a power
function of the velocity, see Sect. 9.7.2). As a consequence, they tend to become shal-
lower and their channels tend to infill. Ebb dominance, on the other hand, results in a
net seaward transport, which tends to keep channels deep. Ebb-dominant behaviour is
an important driver for the formation of a tidal basin in coastal lowlands after a barrier
breach. This is further treated in Sect. 9.7.2.
The second type of asymmetry (around the vertical) affects the residual transport for
fine sediment in inlets and basins. This is because fines need time to settle; they are

Figure 5.70: Tidal variation of the water levels along the Dutch coast and in the adjacent basins
(data from https://waterinfo.rws.nl/). The reference time for each station corresponds to the
first LW on 3 October 2019. Note that a shorter rising period than falling period at sea causes
higher flood than ebb velocities in the basins. A shorter falling period has the opposite effect.
allowed to deposit if the slack duration is long enough. For asymmetries around the
vertical, slack durations are not equal; if for instance the duration of HWS (before ebb)
is longer than the duration of LWS (before flood), stronger sedimentation occurs just
before flow reversal to ebb. We will see in Sect. 9.7.3 that this leads to a net transport
of fines in the flood direction.
5.7.5. Overtides
Mathematically, tidal distortion and asymmetry can be described by the inclusion of
higher harmonics, that is, tidal periods that do not originate from a tidal forcing period
but are integer fractions (1/2, 1/3, etc.) of the period of the basic astronomical constitu-
ents generated by the attraction forces of earth, moon and sun. For that reason they
are named overtides. Since these higher harmonics are generated as a result of non-
linear effects in shallowing coastal waters and tidal basins (compare with wind waves,

Sect. 5.5), they are also called shallow-water tides. The phase relationships between
the tidal constituents determine whether the tidal elevation and velocity curves are
asymmetric about the horizontal or about the vertical axis.
Two important sources for non-linearity in the tidal propagation equations are bottom
friction and continuity. The continuity equation Eq. 5.96 or Eq. 5.95 leads to 𝑐 = √𝑔ℎ or
𝑐 = √𝑔𝐴𝑠 /𝑏. Different celerities of the high and low tide were seen to introduce tidal
asymmetry. The resulting distortion of the surface elevation profile can be approxim-
ated with a second harmonic: a wave with twice the frequency of the basic harmonic.
Equivalently, non-linear tidal propagation in shallow water thus generates a M4 tide
from the M2 tide with a period that is 1/2 the M2 period. In the same way, additional
higher harmonics can be generated (M8, etc.). The higher harmonics to S2 are S4, S8,
etc. If the different tidal components interact, interaction tides can be generated, e.g.,
M2 and S2 can generate MS4 (see Table 3.6). This explains why Table 3.6 is so much
longer than Table 3.5!
The second source of higher harmonics is the quadratic (and hence non-linear) bottom
friction term 𝜏𝑏 = 𝜌𝑐𝑓 𝑢|𝑢| ∝ cos 𝜔𝑡| cos 𝜔𝑡|. A Fourier expansion of this term can be
shown to give tidal constituents with a frequency that is three times as high as the basic
frequency. Friction thus generates an M6 tide from the M2 tide, with a period that is
1/3 of the M2 period. Similarly, from S2 the S6 tide is generated. In the denominator
of the friction term in Eq. 5.98 we have the water depth ℎ, which varies with the tidal
stage. This will introduce second harmonics (M4 etc.).
Figure 5.71 demonstrates the effect of the summation of M2 (semi-diurnal, 12.42 h
period) and M4 (quarter-diurnal, 6.21 h period) with amplitudes of 𝑎M2 = 1 m and
𝑎M4 = 0.2 m respectively. The surface elevation is given by:
𝜂(𝑡) = 𝑎M2 cos (𝜔M2 𝑡 − 𝜑M2 ) + 𝑎M4 cos (𝜔M4 𝑡 − 𝜑M4 ) (5.110)
Since 𝜔M4 = 2𝜔M2 and with 𝑡 ′ = 𝑡 − 𝜑M2 /𝜔M2 , this can be written as:
𝜂(𝑡) = 𝑎M2 cos (𝜔M2 𝑡 ′ ) + 𝑎M4 cos (2𝜔M2 𝑡 ′ − (𝜑M4 − 2𝜑M2 )) (5.111)
On the 𝑥-axis of Fig. 5.71 we have 𝜔M2 𝑡 ′ in radians. The phase lag between M2 and
M4 is 𝜑M4 − 2𝜑M2 . If the phase lag is 0° or ±180°, the result is an asymmetry about the
horizontal (either a positive or negative skewness). For 𝜑M4 − 2𝜑M2 = 0° the signal is
positively skewed (left panel), for 𝜑M4 − 2𝜑M2 = ±180° it is negatively skewed (right
panel). If 𝜑M4 − 2𝜑M2 = ±90°, the result is an asymmetry about the vertical axis (a
longer falling period than rising period or vice versa). Shown in the figure is the
situation that the falling period is longer than the rising period (middle panel). For
relative phase differences other than the special cases of 0°, ±90° and ±180°, the two
types of asymmetry will be combined.

M2 M4 M2+M4
1.5
1 φM4−2φM2=0 φM4−2φM2=−π/2 φM4−2φM2=±π

elevation [m]
0.5
−0.5
−1
−1.5
0 π/2 π 3π/2 2π 0 π/2 π 3π/2 2π 0 π/2 π 3π/2 2π
ωM2t′
Figure 5.71: Influence of the relative phase difference between the M2 and M4 tide on the asym-
metry about the horizontal axis and vertical axis. The combined signal becomes positively
skewed when the two are in phase (left panel), asymmetric (w.r.t. the vertical axis, i.e., lean-
ing forward in space, hence backward in time) when the phase shift 𝜑M4 − 2𝜑M2 = −𝜋/2 (middle
panel), and negatively skewed when the two are out of phase (i.e., 𝜑M4 − 2𝜑M2 = ±𝜋, right panel).
Note that not depicted is 𝜑M4 − 2𝜑M2 = 𝜋/2, which would result in a backward-leaning signal in
space and forward-leaning signal in time.
Sediment transport as a result of asymmetry of the horizontal tide is explained in

Sect. 9.7 in connection with the overtides.
5.7.6. Residual currents

River discharge is the most obvious source of a net tide-averaged flow velocity and
discharge (see Fig. 5.72).
outflow
constant river
discharge
0
inflow
Figure 5.72: Combined effect of tidal flow and river discharge.
In the absence of river discharge or other nontidal forcing (e.g. wind), the net tide-
averaged discharge should be zero. Nevertheless, vertical or horizontal circulation
currents and residual currents may exist. For instance, an irregular coastline or a
headland can induce residual currents, in pretty much the same way as the harbour
moles in Sect. 5.7.2. Other processes and mechanisms are treated below.

The relevance of these residual currents and secondary flows to the sediment transport
and the bed topography is described in Ch. 9.
Tidal residual current pattern in an inlet gorge

An important physical mechanism to bear in mind when interpreting tidal currents is
inertia. The tidal flow in an inlet gorge has so much momentum that it cannot spread
out fast enough when leaving the gorge: it forms a tidal jet. The length of such a tidal
jet is a few hundred times the water depth, so a few kilometres. On the other hand,
the water flowing into the gorge has to accelerate and is therefore inertia-dominated,
i.e. close to potential flow. This means that the highest velocities coincide with the
shortest path through the gorge, i.e. just around the tips of the islands (Fig. 5.73).
The tidal residual current pattern in this highly schematised situation boils down to a
quadruplet of gyres, two at either side of the inlet (Fig. 5.74).
sea sea
basin basin
(a) (b)
Figure 5.73: Outgoing tidal jet (a) and ingoing tidal jet (b).
sea
basin
Figure 5.74: Schematised tidal residual currents around an inlet.
Stokes drift
As for wind waves (Sect. 5.5.1), a Stokes drift or mass flux is associated with tidal
propagation of an at least partly progressive tidal wave. A Stokes drift in the direction

of tidal propagation will occur if more than half of the flood tide coincides with wa-
ter levels above the tidal cycle mean water level. The corresponding inflow of water
must, for a closed basin, be compensated by an opposing outgoing flow, much like the
undertow (Sect. 5.5.1).
The magnitude is dependent on the phase coupling between the horizontal and the
vertical tide. Suppose 𝑢 is the tidal current velocity in a 1D channel, ℎ the mean water
depth and 𝜂 the water surface elevation above mean sea level, and let 𝑢 and 𝜂 be given
by:
𝜂 = 𝑎 sin(𝜔𝑡 − 𝑘𝑥) and 𝑢 = 𝑢̂ sin(𝜔𝑡 − 𝑘𝑥 − 𝜑) (5.112)
The tidal residual flux of water is then given by (compare with Eq. 5.40):
𝑇
1 1
𝑞𝑟𝑒𝑠 = 𝑢𝑎̂ ∫ sin(𝜔𝑡 − 𝑘𝑥 − 𝜑) sin(𝜔𝑡 − 𝑘𝑥) 𝑑𝑡 = 𝑢𝑎̂ cos 𝜑 (5.113)
𝑇 0 2
Thus, if the horizontal tide (𝑢) and the vertical tide (𝜂) are 90° out of phase (𝜑 = −𝜋/2),
there is no residual flux. But if they are more or less in phase (𝜑 ≈ 0°), there can be a
considerable residual current.
Bathymetry-induced residual current

If the cross-section of a channel consists of deeper and shallower sections, the residual
flow in the deeper channels is usually in the ebb direction, whereas in the shallower
parts it is in the flood direction. This leads to a net (averaged over the tidal cycle)
flood current in the shallower section and a net ebb current in the deeper sections, as
indicated schematically in Fig. 5.75.
ide
sea s
HW
LW
-
flood ed
i n a t
dom
-
flood ed
at
do m i n ebb- d
om inate
d
Figure 5.75: Bathymetry-induced circulation.

Coriolis-induced residual current and secondary flow

Due to Coriolis, the flood and ebb flow in a basin tend to be concentrated along op-
posite banks: the flood flow more along the right bank, viewed from the flood flow
direction, and the ebb flow more along the right bank, viewed from the ebb flow dir-
ection (in a Northern Hemisphere basin). Averaged over the tide, this may give rise
to residual flood and ebb currents. In a confined channel that prevents a deviation
of the course, the Coriolis force is balanced, in a depth-averaged sense, by a water
level (pressure) gradient (see Intermezzo 3.6). Compare this with the cross-shore force
balance between a water level gradient due to set-up and the wave forces in the surf
zone (Sect. 5.5.4). As in the cross-shore profile, the channel prevents a depth-mean
transversal velocity. The depth distribution of the Coriolis force is determined by the
velocity profile (e.g. logarithmic) along the channel axis (see Eq. 5.114). Thus, the
depth distribution of the Coriolis and pressure forces are different (compare Fig. 5.43),
resulting in a secondary circulation in the channel cross-section (hence in the vertical
plane). Since the Coriolis force changes sign at ebb and flood, the effect is neutral in a
tide-averaged sense.
Curvature-induced and Coriolis-induced secondary flow

In river bends, a secondary circulation occurs perpendicular to the main flow direction.
Closer to the bottom it is directed towards the inner bend and closer to the surface
towards the outer bend. In tidal channels a similar curvature-induced circulation can
be found. The curvature-induced secondary flow pattern is the result of different depth
distributions of pressure and curvature-induced centrifugal forces, analogous to the
Coriolis-induced circulation described above. The curvature-induced flow, however,
does not change sign as the tide turns. In the following, we will combine both effects.
A simple description is obtained when the assumption is adopted that the flow (both
ebb and flood) has adapted to the channel bend under the influence of bottom fric-
tion and mean surface slope (pressure gradient) and that both the channel and flow
structure are angle-independent in a cylindrical coordinate system. The transversal
(perpendicular to the mean flow) force balance then reads:
𝑢2 1 𝜕𝑝 𝜕 𝜕𝑣
− + 𝑓𝑢 + − 𝜈 =0 (5.114)
⏟ 𝑅 ⏟ 𝜌⏟ 𝜕𝑦 𝜕𝑧 𝑇 𝜕𝑧
⏟⏟⏟⏟⏟⏟⏟⏟⏟
centrifugal force Coriolis pressure gradient turbulent viscosity
where:

𝑅 radius of the channel bend (assumed to be much larger than the m

channel width)
𝑢 local velocity along the channel axis m/s
𝑧 vertical coordinate; bottom: 𝑧 = –ℎ, surface 𝑧 = 𝜂 m
𝑣 transversal velocity m/s
𝑦 transversal coordinate m
𝜈𝑇 eddy viscosity m2 /s
If stratification (density differences) is neglected, the transversal pressure gradient

𝜕𝑝/𝜕𝑦 = 𝜌𝑔 𝜕𝜂/𝜕𝑦 is due to the water level slope, counteracting the centrifugal and
Coriolis forces. The equation describes the evolution of a transversal circulation cur-
rent 𝑣 due to the vertical shear in the lateral channel flow 𝑢. The bend radius 𝑅 is
an important factor in determining the relative importance of the centrifugal force
and the Coriolis force. The Coriolis contribution changes sign at ebb and flood; the
centrifugal force exerts its influence at ebb and flood in the same direction.
In a depth-averaged sense, and neglecting the bed shear stress due to the transversal
circulation, the centrifugal forces and Coriolis force are balanced by a water level gradi-
ent. As above, a circulation current arises, since the depth distribution of the various
forces is different (see the first two terms in Eq. 5.114 that are determined by the velo-
city profile along the channel axis).
By twice integrating the force balance Eq. 5.114 over depth, it is possible to derive
a formula for the transversal circulation current. Near the bottom this circulation
is towards the inner bend (tide-averaged, see Fig. 5.76). The Coriolis acceleration
strengthens the transversal circulation, when from the perspective of the main flow
the channel outer bend is a right bank (i.e. at ebb and flood this implies opposite
bends, and this holds for the Northern Hemisphere). The transversal circulation gives
the tidal flow a spiralling character.
V u or u
V
Figure 5.76: Cross-section of a channel bend showing tide-averaged transversal circulation. The
Coriolis contribution changes sign at ebb and flood; the centrifugal force exerts its influence at
ebb and flood in the same direction. Near the bed, the transversal circulation in a tide-averaged
sense therefore runs from the outer bend to the inner bend. The water level is higher in the outer
bend.

5.8. Long-wave phenomena in coastal waters 269
5.8. Long-wave phenomena in coastal waters

5.8.1. Seiches
Seiches are free oscillations that occur in basins of moderate size (harbour, lake, bay or
even sea). They are standing waves with a frequency equal to the resonance frequency
of the basin in which they occur. The oscillations may be caused by sudden changes
in wind conditions. After generation, the water sloshes back and forth until the wave
motion is dampened out by friction. Seiches generally have half-lives of only a few
periods, but may be frequently regenerated. The difference between seiches and the
tidal resonance phenomena in basins as discussed above is that the latter are not free
oscillations, but forced at a certain tidal frequency.
Seiches can have periods ranging from a few minutes up to several hours. They can
cause havoc in a harbour by setting up reversing currents at the entrance or by rocking
ships free of their moorings. They can also abruptly surge onto piers and beaches and
sweep people away. The Great Lakes of North America and some of the large lakes
in Switzerland are especially prone to seiches, because they are enclosed basins with
large fetches and strong winds.
As seiches are a resonance phenomenon, it is obvious that the basin size in relation to
the wavelength is an important factor. Therefore, measures against the generation of
seiches are usually based on size restrictions of a harbour or other basins, and on the
use of irregularly shaped basins.
Seiches can oscillate in a semi-enclosed (Fig. 5.77, right) and closed mode (Fig. 5.77,
left) or in a combination of these two modes if the open end is somewhat restricted.
In simple cases the wavelength is twice or four times the basin length (or a certain
fraction thereof).
node antinode
Figure 5.77: Standing wave in a closed body of water and in a semi-enclosed body of water.
Basin length 𝐿𝑏 and wavelength 𝐿
Usually, the vertical amplitude of a seiche, even at an antinode, is small. However,

especially at a node, the horizontal displacement of the water can be significant. This

can cause mooring difficulties for ships. Another related influence on large ships is
the effect of the water surface slope.
In the port of Rotterdam, a seiche was observed in the morning of 1 March 1990. It
appeared first as a minor fluctuation of around 10 cm at light island Goeree (an ob-
servation post some kilometres offshore) at 0:00 hours. Then, at about 01:30 hours, it
appeared as a huge standing wave of 1.75 m at Rozenburg lock, a navigation lock some
15 km inland. Figure 5.78 shows a smaller seiche at the same locations observed on 27
August 2001.
light island Goeree

Rozenburg lock
Δz = 0.11 m
Δz = 0.36 m
Figure 5.78: Seiche in the port of Rotterdam (data from https://waterinfo.rws.nl/). At the light
island Goeree, the seiche can be observed as a disturbance of 0.11 m (grey line) and, as it propag-
ates up the Rotterdam waterway, the disturbance grows to 0.36 m by the time it reaches the
Rozenburg lock (black line) after approximately one hour.
5.8.2. Bound long waves and surf beat

The concept of wave radiation stresses and wave-induced forces considered in this
chapter did not include temporal variations created by individual waves or wave groups,
because radiation stress was calculated as the resultant integrated over the wave period
and any wave groups. As we have seen in Sect. 3.5.3, waves generally travel in groups
of higher and smaller waves. Due to the wave height variation in the groups, the ra-
diation stresses vary as well, being highest under the highest waves. This results in
a time-varying set-down in the shoaling zone, with the largest depression under the
highest waves. The effect of this is a long wave motion on the wave group scale. The
long wave is forced; it has the length and frequency of the group and the phase speed
is not that of a free gravity wave, but it travels with the group at the wave group speed.
This phenomenon is referred to as the bound long wave associated with the group.
For a perfect bound long wave, the phase shift between the long wave and the short-
wave envelope (enveloping the maxima and minima of the individual waves) equals

5.8. Long-wave phenomena in coastal waters 271
−𝜋 (it is 180° out of phase with the group, meaning that the wave trough of the bound
long waves coincides with the maximum of the wave envelope, see Fig. 5.79).
short waves bound long wave

1.5
0.5
elevation [m]
-0.5
-1
-1.5
0 200 400 600 800 1000
distance [m]
Figure 5.79: A bound long wave, perfectly out of phase (𝜋 rad) with the wave group. Note that
the troughs of the bound long wave coincide with the maxima of the wave group envelope. The
bichromatic wave consists of two waves: 𝑎1 = 0.5 m, 𝑇1 = 5.75 s and 𝑎2 = 1.0 m, 𝑇2 = 6.15 s,
travelling in a water depth of 30 m. The group length 𝐿𝑔 = 420 m.
In reality, however, it appears that the correlation between the long-wave motion and
the group is smaller (but still negative as long as we stay offshore from the surf zone),
and that it changes into a positive correlation as we enter the surf zone. The change in
the correlation indicates that the long waves are no longer moving at the speed of the
group. It is said that the bound long wave is released from the group after breaking.
As we will see in Ch. 7, the correlation is important for the magnitude and direction
of sediment transport by the long wave.
Due to the variations of wave height in a group, wave-breaking can occur intermit-
tently within a group for waves that exceed a certain height. This generates a time-
varying set-up and therefore oscillations of the water level near the shore. A general
name for low-frequency water level oscillation near the shore is surf beat. The exact
generation mechanisms of surf beat are still uncertain and, besides the present explan-
ation, many alternative and sometimes conflicting explanations have been proposed
in the literature.
Surf beats occurring near port entrances can have a negative impact on mooring con-
ditions in harbour basins.

273
6
Sediment transport
6.1. Introduction
Changes in coastal morphology are the consequence of spatial gradients in net sedi-
ment transport rates (compare Sect. 1.4.3 and Eq. 1.1). Therefore, this chapter discusses
sediment transport mechanisms and prediction methods. Unfortunately, the interac-
tion between hydrodynamics and sediment is very complex and poorly understood. A
famous anecdote says that when Albert Einstein’s son started researching sediment
transport, his father warned him strongly of the difficulties in dealing with sediment
transport processes (see e.g. Vollmers (1989)). Whether this actually happened is ques-
tionable (Gyr & Hoyer, 2006), but the anecdote illustrates the frustration that coastal
engineers and scientists must sometimes feel when trying to understand and model
sediment transport. In view of the above, the modelling of sediment transport is largely
based on empiricism.
Sediment transport can be defined as the movement of sediment particles through a
well-defined plane over a certain period of time. The movement of sediment particles
depends on the characteristics of the transported material (grain size, fall velocity).
These sediment properties are discussed in Sect. 6.2. Next, in Sect. 6.3, initiation of
motion is considered. It is shown that sediment particles will start moving when a
so-called critical velocity (or critical shear stress) is exceeded. This bed shear stress
is the result of the combined wave-current motion. Section 6.4 discusses the various
transport regimes and modes. Generally, two transport modes are distinguished: bed
load and suspended load. If the particles are moved in bed load mode, they roll, shift or
make small jumps over the seabed, but stay close to the bed. Transport in suspended
load mode, on the other hand, implies that grains are lifted from the seabed at flows
above the critical flow velocity and transported in suspension by the (moving) water.
Below a certain flow velocity, the grains settle down again. General bed load and
suspended load transport formulations are treated in Sects. 6.5 and 6.6 respectively. In

274 6. Sediment transport
Sect. 6.7 the so-called energetics approach to bed load and suspended load transport is
discussed. This approach gives us an easy method to unravel sediment transport into
the separate contributions of waves and currents. Section 6.8 deals with some of the
additional complications that can be expected when dealing with fine sediment (mud
and finer sand fractions). Section 6.9 discusses the choice between various transport
models and schematisations that can be made for specific situations.
6.2. Sediment properties

6.2.1. General
In the coastal zone we find sediments like quartz (SiO2 ), carbonates and clay minerals
(sheets of silicates, see Intermezzo 2.3). Depending on the particle size we distinguish
silt and clay, sand, gravel, and cobbles (see Sect. 2.6.2). Clay particles are very small
with a large surface area compared to their volumes. This surface area is chemically
active, which, especially when wet, leads to the typical cohesive characteristics of its
bulk form. Quartz and carbonate sands on the other hand are non-cohesive; the grains
do not stick together. A handful of pure sand cannot be picked up by hand in the way
a piece of clay can be picked up. In these lecture notes the main focus is on sand.
6.2.2. Grain size, density and bulk properties

Two important parameters for sediment transport are the median particle diameter
𝐷50 and the grading, for example, 𝐷90 /𝐷10 . 𝐷𝑥 is defined as the sediment particle
diameter (in metres) for which 𝑥% by weight is finer. In American literature, a 𝜙 scale
is often used to identify the particle dimensions:
𝜙 = − log2 𝐷 (6.1)
where:
𝐷 sand grain diameter mm
For every 𝜙, the diameter 𝐷 [mm] can be found with 𝐷 = 2−𝜙 . Classifications of
sediments on the basis of size are found in Table 2.2. Sediment is called well-sorted if
𝐷90 /𝐷10 is small (say < 1.5, although there is no formal classification); for large values
of 𝐷90 /𝐷10 (for instance > 3) we speak of poorly sorted or well-graded sediment.
Besides grain size, other properties of either the grains or the bulk material are import-
ant for sand transport, such as grain shape (the grains are not perfect spheres), grain
density, fall velocity, angle of repose, porosity and sediment concentration:
• The grain density 𝜌𝑠 depends on the mineral composition of the sand. Most of
the world’s beach sands consist of quartz (feldspar is the second most common

6.2. Sediment properties 275
mineral), with a mass density of 2650 kg/m3 . Other minerals are often referred to
as heavy minerals, since their mass density is usually greater than 2700 kg/m3 ;
• Relative density 𝑠 is defined as the ratio of sediment density over water density
𝜌𝑠 /𝜌. For natural sediments, 𝑠 will be normally around 2.65;
• The fall velocity depends on the grain characteristics as well as on the fluid
characteristics (such as water density and viscosity). This is further discussed
in Sect. 6.2.3. Roughly speaking, the fall velocity of medium-sized sand particles
(0.1 mm < 𝐷 < 0.5 mm) in water varies from 0.01 m/s to 0.05 m/s;
• When (dry) sand is poured onto a flat surface, it will form a mound. The surface
of the mound has a slope tan 𝜑𝑟 (with 𝜑𝑟 is the angle of repose) which depends
mainly on the grain size;
• The porosity 𝑝 is defined as the ratio of pore space (voids) to the whole sediment
volume. Natural sands have porosities in the range of 0.25 to 0.50; a frequently
applied figure is 0.40 (or 40 %);
• The sediment concentration 𝑐 can be defined in two ways: mass concentration
and volume concentration. The mass concentration is the mass of the solid
particles per volume (𝑐 in kg/m3 or equivalently g/L) and is often used when
measuring sediment concentrations. Volume concentration is defined as the ra-
tio of the volume of solid particles to the whole volume (𝑐 in m3 /m3 or in %).
Sediment in a sediment bed has a volume concentration of 𝑛 = 1 − 𝑝. For sed-
iment in suspension, the volume concentration 𝑐 indicates the volume of sedi-
ment per volume of the mixture. If the sediment in such a mixture settles to
the bed, 1 m3 of solid particles will occupy 1/(1 − 𝑝) (𝑝: porosity) m3 at the bed.
Volume concentrations are obtained from mass concentrations by multiplication
by 1/𝜌𝑠 ;
• Contrary to the grain density (𝜌𝑠 ) as mentioned above, bulk density is the mass
of a unit volume of e.g. a mixture of particles and air or water. The dry bulk
density is defined as 𝑛𝜌𝑠 . If the whole pore volume is filled with water (density
𝜌), then we obtain the saturated bulk density, which is defined as 𝑛𝜌𝑠 + 𝑝𝜌. A
typical figure for the dry bulk density is 1600 kg/m3 , and for the saturated bulk
density 2000 kg/m3 .
6.2.3. Fall velocity

When a particle falls in still and clear water, it accelerates until it reaches a constant
vertical velocity that is called fall velocity or settling velocity. This velocity can be
assessed from the balance between the downward-directed gravity force 𝐹𝐺 (minus
the effect of buoyancy) and the upward-directed (retarding) drag force 𝐹𝐷 as indicated
in Fig. 6.1.

FD
FG
Figure 6.1: Forces on a ‘sphere’ in clear water.
Basic equation
The downward-directed gravity force 𝐹𝐺 on a sphere combined with the upward buoy-
ancy effect is given by the so-called underwater weight of the sphere, viz. by the
weight of the sphere minus the weight of the displaced volume of water. With weight
equal to mass times acceleration due to gravity, we have for a perfect sphere:
𝜋
𝐹𝐺 = (𝜌𝑠 − 𝜌) 𝑔 ( 𝐷 3 ) (6.2)
6
where:
𝜌𝑠 mass density of the particle kg/m3

𝜌 mass density of the surrounding fluid kg/m3
𝐷 particle diameter m
𝑔 acceleration of gravity m/s2
In this equation, the second term between brackets is the volume of the sphere.
The upward-directed force is equal to the so-called drag force denoted by:
1 𝜋
𝐹𝐷 = 𝐶𝐷 𝜌𝑤𝑠2 ( 𝐷 2 ) (6.3)
2 4
where:
𝐶𝐷 drag coefficient −
𝑤𝑠 particle fall velocity m/s
The second term between brackets refers to the cross-section of the sphere. The drag
force is non-zero only if 𝑤𝑠 > 0.
In equilibrium, both forces are in balance and the fall velocity 𝑤𝑠 (in m/s) is given by:
4(𝑠 − 1)𝑔𝐷
𝑤𝑠 = (6.4)
√ 3𝐶𝐷

6.2. Sediment properties 277
in which 𝑠 is the relative density (see Sect. 6.2.2).

A particle’s fall velocity depends on its size, its density and the magnitude of the drag
coefficient 𝐶𝐷 . This drag coefficient depends on the shape of the particle and its rough-
ness, but mainly on the grain’s Reynolds number:
Re = 𝑤𝑠 𝐷/𝜈 (6.5)
where:
𝜈 kinematic viscosity coefficient m2 /s
The kinematic viscosity is defined as the dynamic viscosity divided by the water dens-
ity: 𝜈 = 𝜇/𝜌. A characteristic value for 𝜈 is 10−6 m2 /s. The dynamic viscosity repres-
ents the fluid’s internal resistance to flow (‘thickness’) and is a function of the temper-
ature and to a smaller extent of the density.
Dependence on Reynolds number

For low grain Reynolds numbers (Re < 0.1 to 0.5) in the so-called Stokes range, the
drag coefficient can be described by (see Fig. 6.2):
𝐶𝐷 = 24/Re (6.6)
yielding:
(𝑠 − 1)𝑔𝐷 2
𝑤𝑠 = (6.7)
18𝜈
In this range, the fall velocity depends on the square of the grain diameter, the relative
density and the kinematic viscosity coefficient.
For high grain Reynolds numbers (400 < 𝑅𝑒 < 2 × 105 ), in the so-called Newton range,
the drag coefficient becomes a constant (𝐶𝐷 ≈ 0.5). In that case it follows from Eq. 6.4:
𝑤𝑠 = 1.6√𝑔𝐷(𝑠 − 1) (6.8)
In this range, the fall velocity depends on the square root of the grain diameter and
the relative density and is independent of the kinematic viscosity coefficient. This is
also the case for extremely high Reynolds numbers (𝑅𝑒 > 2 × 105 ), where the drag
coefficient is (constant) around 0.2.
For quartz spheres falling in still water, a Reynolds number of 0.5 corresponds roughly
to a particle diameter of 0.08 mm, while a Reynolds number of 400 corresponds to a

10 4
spheres
discs
10 3
[-]
10 2
drag coefficient
10 1
Stokes
= 24/Re discs
(Eq. 6.6) spheres
10 0
Newton
= 0.5
10 -1
10 -3 10 -2 10 -1 10 0 10 1 10 2 10 3 10 4 10 5 10 6
Reynolds number Re [-]
Figure 6.2: Drag coefficient as a function of Reynolds Number (Vanoni, 1975).
diameter of about 1.9 mm. Since most beach sediments fall in the range of 0.08 mm
to 1.9 mm, in most practical cases neither the Stokes approximation nor the Newton
approximation can be used.
For very small particles (silt, clay) the fall velocity is proportional to 𝐷 2 ; for gravel
size particles the fall velocity is proportional to √𝐷. For sand, the fall velocity falls
in the transition range between a 𝐷 2 dependency and a √𝐷 dependency. By lack of a
simple expression for the drag coefficient in this range, it is common practice to use
empirical formulas for the fall velocity. In literature, many formulas can be found (e.g.
Van Rijn (1984b) and Ahrens (2000)). As an example, in Fig. 6.3, for two formulas the
fall velocity as a function of the grain diameter is plotted.
Hindered settling
In high-concentration mixtures, the fall velocity of a single particle is reduced due to
the presence of other particles. This can be explained as follows: with each downward
grain movement, a similar fluid volume must flow upward; this upward flow slows
down the other grains. In order to account for this hindered settling effect, the fall
velocity in a fluid-sediment mixture should be determined as a function of the sediment
concentration 𝑐 and the particle fall velocity 𝑤𝑠 . An often-used formula for the effective
fall velocity in a mixture is:
𝑤𝑒 = (1 − 𝑐)𝛼 𝑤𝑠 (6.9)
where:

6.3. Initiation of motion 279
Van Rijn (1984b)

Ahrens (2000)
ws [m/s]
D [μm]
Figure 6.3: Sediment fall velocity as a function of the grain size in fresh water with a temperature
of 18 °C (𝜌𝑤 = 1000 kg/m3 , 𝜌𝑠 = 2650 kg/m3 and 𝜈 = 1.05 × 10−6 m2 /s). Comparison between Van
Rijn (1984b) and Ahrens (2000).
𝑤𝑒 effective fall velocity m/s

𝑤𝑠 fall velocity of one grain in clear water in rest m/s
𝑐 sediment concentration (volumetric %, volume of solid sediment −
particles in volume of water-sediment mixture)
𝛼 coefficient (ranging from 4.6 at low Reynolds numbers to 2.3 at −
high Reynolds numbers)
In many engineering cases related to sediment transport, the sediment concentration

can be as high as approximately 1 vol.% (2.65 kg/m3 ), meaning that the effect of sed-
iment concentration on the fall velocity is small. This is not the case in sheet flow
conditions (see Sect. 6.4.2) and in very silty environments, where sediment concentra-
tions can be significantly higher.
6.3. Initiation of motion

6.3.1. Forces on a single grain
Sediment can only be transported if the water movement exerts a large enough shear
stress 𝜏𝑏 on the grains. The so-called critical shear stress 𝜏𝑏,𝑐𝑟 describes the point of
initiation of motion. If this condition is exceeded, grains move, roll or are brought
into suspension. Note that instead of the term bed shear stress one may find the term
shear velocity (𝑢∗ ), which is related to the bed shear stress through 𝜏𝑏 = 𝜌𝑢∗ |𝑢∗ | .
In order to assess the critical condition of initiation of motion, the various forces acting
on an individual grain have to be taken into account (see Fig. 6.4). These forces can
be divided into forces which tend to move the grain – the drag force 𝐹𝐷 (Fig. 6.4a) and

the lift force 𝐹𝐿 (Fig. 6.4b) – and a force which tries to keep the grain in its place; the
gravity force 𝐹𝐺 (Fig. 6.4c).
skin friction
sub pressure
current FD
(a) drag force (in the direction of the current)

FL
ion
act
current c o ntr
(b) lift force
FG
(c) gravity force
Figure 6.4: Forces on an individual grain in a stationary situation.
The drag force is a combination of skin friction acting on the surface of the grain and
a pressure difference on the up- and downstream sides of the grain because of flow
separation at the downstream end of the particle. Analogous to Eq. 6.3, the drag force
is proportional to:
• the square 𝑢 2 of a typical upstream horizontal flow velocity;
• the particle’s surface area and hence for spheres to 𝐷 2 ;
• the water density 𝜌.
The lift force results from the flow separation, as well as from the flow contraction
above the grain. A higher local flow velocity results in a lower local pressure (Bernoulli
law). The difference in vertical pressure causes an upward-directed lift force. Similarly
to the drag force, the lift force is proportional to the particle’s surface area (and thus
to 𝐷 2 in the case of a sphere) and to 𝑢 2 .
The total driving force (drag and lift combined) is therefore proportional to 𝜌𝑢 2 𝐷 2 . The
resisting gravity force is proportional to (𝜌𝑠 −𝜌)𝑔𝐷 3 (see Eq. 6.2). Equilibrium of forces,
whether horizontal, vertical or rotational, therefore, is expressed through a formula of
the following type:
(𝜌𝑠 − 𝜌) 𝑔𝐷 3 ∝ 𝜌𝑢𝑐𝑟2 𝐷 2 (6.10)

6.3. Initiation of motion 281
in which 𝑢𝑐𝑟 is the critical velocity of the water at which grains start moving. Since
the bed shear stress is proportional to the velocity squared times the water density, we
could also write:
(𝜌𝑠 − 𝜌) 𝑔𝐷 3 ∝ 𝜌𝜏𝑏,𝑐𝑟 𝐷 2 (6.11)
2 is the critical bottom shear stress (critical in the sense that higher
Here, 𝜏𝑏,𝑐𝑟 = 𝜌𝑢∗,𝑐𝑟
bottom shear stresses lead to the initiation of motion).
From the proportionality Eq. 6.11, the so-called critical Shields parameter 𝜃𝑐𝑟 can be
deduced:
𝜏𝑏,𝑐𝑟
𝜃𝑐𝑟 = =𝐶 (6.12)
(𝜌𝑠 − 𝜌) 𝑔𝐷
The constant 𝐶 has to be determined experimentally. The experiments of Shields, per-

formed on a flat bed, are the most widely used. He defined the critical bed shear stress
as the bed shear stress at which the (extrapolated) measured transport rates were just
zero. For sand placed smoothly on this flat bed, 𝐶 was found to be around 0.05. Ap-
pendix D points out the similarities between Eq. 6.12 and stone stability and structural
damage approaches for slopes of loose rock and of breakwater elements.
6.3.2. Shields curve

Shields found experimentally that the ‘constant’ 𝐶 ≈ 0.05 is a weak function of the
grain Reynolds number defined as:
𝑢∗ 𝐷
Re∗ = (6.13)
𝜈
where:
𝑢∗ shear velocity m/s

𝐷 diameter of grains m
𝜈 kinematic viscosity coefficient m2 /s
The subscript ∗ is used to indicate that the Reynolds number is based on 𝑢∗ .

Figure 6.5 shows measured values of 𝐶 as a function of Re∗ . The shaded band separates
two zones: movement of sediment particles was observed in the zone above this shaded
band, whereas no movement was observed in the zone underneath the shaded band.
The shaded band therefore indicates initiation of motion. Sometimes, the shaded band
is represented by a single line, which is then referred to as the Shields curve. The
average value can be seen to be approximately 0.05.

mm
mm
m
mm
m
m
m
m
m
50
62
25
m
8m
2m
1m
4m
0.0
0. 1
0.2
0.5
16
D=
D=
D=
D=
D=
D=
D=
D=
D= u u u u
* = * = * = * =
0. 0. 0. 0.
u 02 05 1 2
* = 5 m m m
0. m /s /s /s
01 /s
25 movement
m
/s
rest
Figure 6.5: Shields curve (Shields, 1936). Note that the axes are drawn on a log-log scale. The
lines of constant 𝐷 and 𝑢∗ do not originate from Shields and are valid for constant density 𝜌𝑠 =
2650 kg/m3 and kinematic viscosity 𝜈 = 1.25 × 10−6 m2 /s at a water temperature of 12 °C.
Unfortunately, reality is more complex. Reasons for this include:

• The Shields curve is valid for uniform flow on a flat bed. The effect of bed
ripples and the effect of the combination of unidirectional and oscillatory flow
on initiation of motion are largely unknown;
• Gradation of the bed material may play a role, especially for poorly sorted sed-
iment (𝐷90 /𝐷10 > 3). In these cases, the smaller particles will be hidden in the
voids between the larger particles, while the larger particles are more exposed.
After exposed smaller particles are washed out, a top layer of coarser particles
(with higher critical flow velocities) remains and prevents movement of the un-
derlying smaller particles. This is called bed armouring;
• For a sloping bed in the flow direction, it can be argued that the critical flow
velocity will be somewhat smaller for downward-sloping beds and somewhat
higher for upward-sloping beds;
• Cohesive forces between the grains – due to the presence of cohesive sediment in
the bed – may drastically increase resistance against erosion (see also Sect. 6.8).
Biological activity and consolidation may be important in this respect as well.
Notwithstanding these complications, many practical sediment transport formulas use
the critical Shields parameter 𝜃𝑐𝑟 (Eq. 6.12) and the Shields curve to define initiation of
motion. Amongst others, Van Rijn (1984a) represented the Shields curve as a function
of a non-dimensional grain size 𝐷∗ :
1/3
𝑔(𝑠 − 1)
𝐷∗ = 𝐷so ( ) (6.14)
𝜈2

6.4. Basic principles of transport modelling 283
The Shields curve can be represented in terms of 𝐷∗ , since every grain diameter has
a corresponding 𝑢∗,𝑐𝑟 as can be seen in Fig. 6.5. With 𝜃𝑐𝑟 = 𝑓 (𝐷∗ ), no iteration is
necessary to obtain the critical shear stress, as would be the case when applying the
Shields curve. In addition, the threshold parameter can be corrected to account for the
effect of the bed slope tan 𝛼 on the threshold of motion. This formulation results in an
increase of the critical shear stress for upslope movement and a decrease of the critical
shear stress for downslope movement. We then have a formula of the type:
𝜃𝑐𝑟 = 𝑓 (𝐷∗ , tan 𝛼) (6.15)
6.4. Basic principles of transport modelling

6.4.1. Definitions
Sediment transport can be defined as the movement of sediment particles over a cer-
tain period of time through a well-defined plane. The vertical extent of such a plane
is generally from the bed to the water level. In the horizontal direction the plane
may extend from the edge of the surf zone to the water line. In that case the total
wave-induced longshore transport integrated over the surf zone is considered. Often,
however, the transport rates are expressed per metre width.
Since coastal engineers are interested in volumes of accretion and erosion, it is com-
mon practice to express the sediment transport rates 𝑆 in m3 /s/m (volumes of sand
per second per metre width). These sediment volumes may be expressed in terms of
volumes of solid grains. Other sediment transport formulas directly yield the deposited
volumes of sand (per second and metre width). Deposited volumes of sand include the
pores between the grains and are a factor 1/(1 − 𝑝) larger than the volumes of solid
grains (see Sect. 6.2.2). In the mass balance Eq. 1.1, it was assumed that the volumet-
ric sediment transport rates include the pores. If transport rates of solid grains are
considered instead, the mass balance reads (with zero local gains or losses):
𝜕𝑧𝑏 𝜕𝑆𝑥 𝜕𝑆𝑦

(1 − 𝑝) + + =0 (6.16)
𝜕𝑡 𝜕𝑥 𝜕𝑦
where:
𝑧𝑏 (𝑥, 𝑡) bed level above a certain horizontal datum m

𝑆𝑥 (𝑥, 𝑦, 𝑡) sediment transport rate in the 𝑥-direction (volume of m3 /m/s
solid grains per second and per m width)
𝑆𝑦 (𝑥, 𝑦, 𝑡) transport rate in the 𝑦-direction m3 /m/s
𝑝 porosity −

Instead of volumetric transport rates, sometimes mass transport rates are given. This
generally does not concern the dry mass of the sediment (with density 𝜌𝑠 ), but the
immersed (underwater) mass of the sediment. The relation between the volumetric
transport rate 𝑆 (of deposited material, hence including pores) and the immersed mass
transport rate 𝐼𝑚 is:
𝐼𝑚 = (𝜌𝑠 − 𝜌) (1 − 𝑝)𝑆 (6.17)
where:
𝐼𝑚 immersed mass transport rate kg/m/s

𝑆 volume transport rates of deposited material m3 /m/s
Other transport formulas are expressed in terms of immersed weight 𝐼 = 𝑔𝐼𝑚 .

The above-mentioned transport rates can be instantaneous transport rates, but also
transport rates averaged over various hydrodynamic conditions (a wave group, a storm,
a year). The latter is of most interest to coastal engineers.
Bed load versus suspended load

Different transport modes can be distinguished:
Bed load transport the transport of sediment particles in a thin layer close to the bed.
The particles are in more or less continuous contact with the bed. Bed load
transport at low shear stresses is shown in Fig. 6.6a. At higher shear stresses, an
entire layer of sediment is moving on a plane bed (Fig. 6.6b). This is called sheet
flow and is often considered bed load, since grain-grain interactions play a role.
Suspended load transport the transport of particles suspended in the water without
any contact with the bed. The particles are supported by turbulent diffusive
forces. Figure 6.6c shows this transport mode.
(a) (b) (c)
Figure 6.6: Different modes of sediment transport:(a) bed load at small shear stresses; (b) sheet
flow (often considered as bed load at higher shear stresses); (c) suspended load.
The sum of bed load and suspended load is called total load. In addition, a third cat-
egory exists that is named wash load. Wash load consists of very fine particles that

will only settle in still water and are not found in the bed. Since these particles do not
contribute to bed level changes, wash load is not taken into account into the total load
transport.
Bed load transport at low shear stresses

As soon as the bed shear stress exceeds a critical value (Shields parameter between
0.03 and 0.06, see Sect. 6.3), sediment particles start rolling or sliding over the bed. If
the bed shear stress increases further, the sediment particles move across the bed by
making small jumps, which are called saltations. As long as the jump lengths of the
saltations are limited to, say, a few times the particle diameter, this type of motion is
considered part of the bed load transport. Close to initiation of motion, the bed remains
flat, but for somewhat larger bed shear stresses the bed load occurs primarily via the
migration of small-scale bed forms. When the jumps become larger, the particles lose
contact with the bottom and become suspended.
Sheet flow transport

At higher shear stresses (Shields parameters higher than about 0.8–1.0), the particles
closer to the bed start moving in multiple layers, instead of rolling and jumping in
a single layer. In a wave tunnel with reversing flow (see Fig. 6.8), this process can
be observed very clearly; the top layer of the bed moves back and forth as a sheet
of sand over a flat immobile bed. If the thickness of the sheet flow layer is defined
as the distance between the non-moving bed and the level where the time-averaged
sediment concentration becomes lower than a volume concentration of 8 vol.% (see
Dohmen-Janssen et al., 2001), the thickness of the sheet flow layer is in the order of
centimetres. For 8 vol.% the distance between particles is on average equal to the grain
diameter. This means that intergranular forces and grain-water interactions are im-
portant. Bagnold (1956) defined bed load as that part of the total load that is supported
by intergranular forces. According to this definition, sheet flow transport can be con-
sidered bed load transport at high shear stresses.
Suspended load transport

Above the bed load layer or sheet flow layer, sediment may be in suspension. Particles
in suspension do not immediately return to the bed under the influence of their settling
velocity, but are kept in suspension by fluid turbulence. Suspended particles present
in a certain vertical plane can be assumed to move horizontally across the plane with
the water particles and thus at the same speed as the water particles. The particles
are suspended in the flow at relatively low concentrations (less than 1 vol.%), so that
intergranular forces are not important.

As mentioned above, in the sheet flow regime (Shields parameters above 0.8–1.0) an
entire layer of sediment is moving on a plane bed. Because the bed is plane, the as-
sumption seems reasonable that the suspended load above the sheet flow layer is sup-
ported by fluid turbulence (or in other words: that the vertical transport of sediment
is governed by turbulent diffusive processes). Suspended transport supported by fluid
turbulence is further treated in Sect. 6.6.
At lower Shields parameters (below 0.8–1.0), however, the bed will not remain plane.
Instead smaller and larger bed forms occur under the influence of currents and the
wave orbital motion. Orbital ripples have a length in the order of the free stream
orbital velocity amplitude 𝑢̂0 (see Sect. 5.4.3 and Eq. 5.27), whereas anorbital ripples
are much smaller and scale with the grain size. In the case of a rippled bed, the bed
roughness is related to the ripple geometry rather than the grain diameter, as was the
case for a plane bed. Furthermore, the flow separates behind the ripple crest and an
organised pattern of vortices is formed near the bed (see Fig. 6.7).
plane bed rippled bed
onshore
orbital motion
flow reversal
offshore
orbital velocity
Figure 6.7: Velocity field near a rippled bed in oscillatory flow. The sediment-laden vortices
formed by the larger onshore orbital motion are transported in offshore direction by the offshore
orbital motion, resulting in a wave-averaged offshore transport.
These vortices are capable of bringing large amounts of sediment into suspension. The
timing of these suspension events within the wave cycle is crucial for the magnitude
(and the direction, see Sect. 6.6.1 and Intermezzo 6.1) of the resulting wave-averaged
sediment transport. Whereas in the plane bed case, it is normally assumed that tur-
bulent diffusive forces keep the sediments in suspension; for a rippled bed, a different
approach should be taken, because the upward transport of sediment is now the result
of an organised motion. This complicates the computation of 𝑐(𝑧, 𝑡) as well as of 𝜈(𝑧, 𝑡)
considerably and is not explicitly considered in this book.

Intermezzo 6.1 Intra-wave sediment concentration
In a coastal environment, both the velocity and the concentration exhibit large
variations within a wave period. Unfortunately, our understanding of 𝑐(𝑧, 𝑡) espe-
cially is very poor. Even in a relatively simple laboratory case of a wave tunnel
(Fig. 6.8), the relation between flow kinematics and sediment concentration is far
from unambiguous. This can be seen from Fig. 6.9 where nearly a hundred differ-
ent records of 𝑐(𝑡) are shown, all measured at a constant near-bed elevation and
under identical regular wave conditions. The scatter in the measuring results is
very large. Nevertheless, we can see the following qualitative behaviour:
• A peak in the sediment concentration just after the maximum onshore and
offshore velocity (𝑡/𝑇 = 0.25);
• Two secondary peaks in sediment concentration just after flow reversal (as-
sociated with vortices developing over a rippled bed).
The behaviour of 𝑐(𝑧, 𝑡) in irregular progressive and breaking waves is even less
well understood and very difficult to model with reasonable accuracy. This has
serious implications for the accuracy of calculations of the intra-wave sediment
flux 𝑐(𝑧, 𝑡) ⋅ 𝑢(𝑧, 𝑡). Besides, due to the oscillatory behaviour of 𝑢(𝑧, 𝑡), its average
value is close to zero, making the computations very sensitive to errors.
Measurements in wave flumes (as opposed to tunnels) show the presence of sus-
pended sediment particles from the bed up to the (instantaneous) water surface.
The largest concentrations are found close to the bed, where the diffusivity is large
due to ripple-generated eddies. Further away from the bed, the sediment concen-
trations decrease rapidly because the eddies dissolve rather rapidly travelling up-
wards.

open cylindrical riser
servo cylinder 0.3 m
cross-section
piston 1.1 m
cylindrical riser
test
section
sand trap 15 m sand trap
Figure 6.8: Schematic representation of the Large Oscillating Water Tunnel (LOWT) used in
the Netherlands to study intra-wave sediment transport phenomena under controlled simu-
lated wave conditions at full scale. The system basically consists of a vertical U-tube with
one open leg. The other leg is provided with a piston. At the bottom of the test section,
a sediment bed may be installed. The generated oscillatory flow is purely horizontal; as
opposed to the case of progressive waves, vertical velocities and horizontal gradients, and
thus wave-induced streaming, are absent.
80
60
c [kg/m3]
40
20
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
t/T
Figure 6.9: Sediment concentrations as a function of time (99 individual records) (Bosman,
1982). On the 𝑥-axis: time 𝑡 relative to wave period 𝑇 . On the 𝑦-axis: sediment concentration
in kg/m3 (= g/l). Sinusoidal water motion 𝑢̂0 = 0.3 m/s and 𝑇 = 1 s, measured at ripple crest
position.

6.4.2. Practical modelling of sediment transport

The mechanisms behind bed load and suspended load are quite different. It is therefore
common practice to use separate transport formulas for the two modes of transport.
Bed load transport is almost exclusively determined by the bed shear stress acting on
the sediment particles that roll, slide and jump along the bed. Hence, bed load formu-
las are often expressed in terms of bed shear stress due to currents and waves, often
supplemented with a criterion that describes initiation of motion. These formulas are
based on either a time-averaged or an instantaneous bed shear stress (Sect. 6.5). In
the latter case the sediment is normally assumed to respond instantaneously (without
delay) to the intra-wave (within the wave period) shear stress variation. This is a so-
called quasi-steady approach to bed load transport. It assumes that inertia (resistance
to accelerations and decelerations) plays a minor role only. This is probably a valid
assumption for particles smaller than a few millimetres. Generally, the coefficients
in the formulas are empirically derived. In practical applications, bed load transport
formulas are assumed to predict sheet flow transport rates when used for large Shields
parameters.
Suspended load transport takes place above the bed load layer. This is depicted in
Fig. 6.10. The suspended sediment flux at a certain height above the bed is often mod-
elled as the product of the sediment concentration 𝑐 and the horizontal velocity 𝑢 of
the water that is transporting the sediment. The suspended sediment transport can
be computed by integrating the suspended sediment flux 𝑢𝑐 from the top of the bed
load layer to the water level. In order to compute the sediment concentration 𝑐, it is
generally assumed that turbulent diffusive forces are responsible for transporting the
sediment upwards in the water column, against the downward movement with the fall
velocity. This approach is treated in detail in Sect. 6.6.
Since sediment particles are allowed time to settle, suspended load does not respond
instantaneously to hydrodynamic conditions. At the top of the bed load layer, a bound-
ary condition needs to be supplied, for instance a prescribed concentration at 𝑧 = 𝑎
(called reference concentration) or a prescribed pick-up rate from the bed. Since the
bed load transport is determined by the excess shear stress (above a critical value),
the reference concentration should be a function of this as well. The exact distinction
between the bed load and suspended load is quite arbitrary.
In the coastal zone, where waves play an important role in water motion, both the wa-
ter velocity 𝑢 and the sediment concentration 𝑐 usually vary strongly as a function of
time, on a scale comparable to the wave period. However, in engineering applications
often only the wave-averaged velocity distribution 𝑈 (𝑧) and wave-averaged concen-
tration 𝐶(𝑧) are used (as shown in Fig. 6.10). The validity of this approach is discussed
later.
As mentioned before, the diffusion approach to suspended load transport only makes
real sense in case of a plane bed. In the case of a rippled bed, vortices shed off the

velocity concentration suspended load
bed load
Figure 6.10: Bed load transport takes place in a thin bed load layer close to the bed (between
𝑧 = 0 and 𝑧 = 𝑎). The suspended load transport takes place in the upper layer. Note that the
𝑧-coordinate is still vertically upward but now the bed is 𝑧 = 0 and the mean water level at 𝑧 = ℎ.
rippled bed bring sediment into suspension. This is an organised rather than a turbu-
lent motion. In practical situations, the presence of ripples can be taken into account
by a larger pick-up rate and diffusivity of sediment (Sect. 6.6).
An entirely different approach to bed load and suspended sediment transport is the
energetics approach (Bagnold, 1962, 1963, 1966; Bailard, 1981; Bailard & Inman, 1981;
Bowen, 1980). It assumes that a certain portion of the fluid energy is expended to
keep the sediment, in both bed load and suspended load, in motion. This results in
quasi-steady formulas that relate the bed load and suspended load transport to the
instantaneous intra-wave velocity at a certain height above the bed. The energetics
approach will be treated in Sect. 6.7.
The sum of the bed load transport rate 𝑆𝑏 and suspended transport rate 𝑆𝑠 equals the
total transport rate 𝑆𝑡 :
𝑆𝑡 = 𝑆 𝑏 + 𝑆 𝑠 (6.18)
Here 𝑆𝑏 and 𝑆𝑡 can denote the instantaneous or averaged (over sufficient wave periods)
sediment transport. Instead of separately modelling bed load and suspended load trans-
port, sometimes a total load formula is used that is assumed to predict the total trans-
port. Such a total load formula could for instance be a function of the time-averaged
bed shear stress.
6.5. Bed load based on the Shields parameter

6.5.1. Importance of the Shields parameter
Bed load transport occurs when the bed shear stress or the shear velocity 𝑢∗ = √𝜏𝑏 /𝜌
exceeds a critical value (initiation of motion, Sect. 6.3). In the bed load layer, turbulent
mixing is often assumed to be still small (due to the presence of the bed), so that it only

6.5. Bed load based on the Shields parameter 291
slightly influences the motion of sediment particles. Gravity limits vertical particle
movement. We can then assume that the bed load transport responds instantaneously
(without delay) to the bed shear stresses.
Many approaches for bed load transport are based on this reasoning and take the sed-
iment transport to be a direct function of the shear stress on the grains. In such
formulas the dimensionless sediment transport is invariably a function of a Shields
parameter 𝜃 (dimensionless shear stress). Many formulas used in coastal engineering
practice have been based on already existing formulas used in river engineering. In
Fig. 6.11 a comparison is made between various bed load transport formulas developed
for rivers.
Although the formulas seem quite different at first glance, Fig. 6.11 demonstrates that
they all represent dimensionless transport as a function of a Shields parameter. Further
note that the predicted transport rates for a certain value of the Shields parameter vary
by up to an order of magnitude. This is (unfortunately) quite common for sediment
transport predictions and underlines the fact that calibration of the transport formulas
for the locations and conditions under consideration is crucial.
6.5.2. Including waves

For nearshore applications the influence of waves needs to be included into the Shields
parameter. This can be done in two ways, either by using the time-averaged (wave-
averaged) bed shear stress for the combined wave-current motion or by using the in-
stantaneous bed shear stress (varying during the wave motion). Both approaches will
be explained below.
The instantaneous bed load transport vector 𝑆𝑏 for waves and currents combined can
be written in a dimensionless form as:
𝑆𝑏 (𝑡)
Φ𝑏 (𝑡) = (6.19)
3
√(𝑠 − 1)𝑔𝐷50
This dimensionless formulation was proposed by Einstein (1942, 1950) based on theor-
etical considerations. The denominator is the square root of a parameter representing
the specific underwater weight of sand grains. Further, 𝑆𝑏 is the bed load transport
rate in volume per unit time and width. In applying a certain formula, one should
check whether 𝑆𝑏 is defined including pores or excluding pores.
Based on the above, we may expect that instantaneous dimensionless bed load trans-
port Φ𝑏 (𝑡) responds quasi-steadily to instantaneous bed shear stress (above the threshold
for motion, in other words, the critical level of bed shear stress). Hence,
Φ𝑏 (𝑡) = 𝑓 (𝜃 ′ (𝑡), 𝜃𝑐𝑟 ) (6.20)

in which 𝑓 is an algebraic operator and 𝜃𝑐𝑟 is defined by Eq. 6.12. The instantaneous
dimensionless effective shear stress 𝜃 ′ (𝑡) due to currents and waves is given by:
𝜏𝑏′
𝜃 ′ (𝑡) = (6.21)
(𝜌𝑠 − 𝜌) 𝑔𝐷50
The Shields parameter 𝜃 ′ is a measure of the forcing on the sediment grains (drag and
lift) relative to the resisting force (see Sect. 6.3.1). The effective bed shear stress 𝜏𝑏′ is
that part of the total bed shear stress which is transferred directly to the grains in the
bed as skin friction. The form drag induced by bed forms is not effective in relation
to bed load transportation. One way to compute skin friction is to use a roughness
height related to the grain size, not to the bed form size. Also, so-called efficiency or
ripple factors 𝜇 can be used that represent the fraction of the total bed shear stress that
can be attributed to skin friction. The latter approach was followed in Fig. 6.11.
Kalinske
Meyer-Peter/MÜller
Einstein
Frijlink
Rottner
Ackers-White
dimensionless transport Φ [-]
skin friction θ' (see Eq. 6.21)
Figure 6.11: Comparison of various bed-load transport formulas developed for rivers. On the
𝑦-axis is the dimensionless transport Φ in the case of steady flow (cf. Eq. 6.19), on the 𝑥-axis a
Shields parameter based on skin friction (cf. Eq. 6.21). Note that Δ = (𝜌𝑠 −𝜌)/𝜌 and the subscript
0 to the shear stress refers to the bed shear stress. Adapted from Breusers (1983).

The non-dimensional critical shear stress parameter 𝜃𝑐𝑟 represents the threshold of
motion of sand grains. It can be computed using the classical Shields curve (Fig. 6.5)
or an explicit approximation of the form Eq. 6.15.
Time-averaging Eq. 6.20 gives a time-averaged bed load sediment transport:
⟨Φ𝑏 (𝑡)⟩ = ⟨𝑓 (𝜃 ′ (𝑡), 𝜃𝑐𝑟 )⟩ (6.22)
The brackets ⟨ ⟩ denote time-averaging.

Some time-averaged formulas for bed load transport relate the time-averaged bed load
transport directly to a time-averaged bed shear stress magnitude (averaged over the
wave motion) rather than an instantaneous bed shear stress:
⟨Φ𝑏 (𝑡)⟩ = 𝑓 (⟨|𝜃 ′ (𝑡)|⟩ , 𝜃𝑐𝑟 ) (6.23)
Note the differences between Eqs. 6.22 and 6.23.

By correlation of the non-dimensional parameters in Eqs. 6.22 and 6.23, using a range
of data sets for sediment transport, a general bed load transport formula can be ob-
tained. Many of these exist, based on different data sets and correlating different para-
meters.
6.5.3. Instantaneous bed load transport

As an example of a bed load transport formula of the form of Eq. 6.20, we mention the
quasi-steady bed load transport formula according to Ribberink (1998). The author
assumed that the sand transport rate is a function of the difference between the actual
time-dependent non-dimensional bed shear stress and the critical bed shear stress:
𝑆𝑏 (𝑡) 𝛽𝑠 ′ 1.8 𝜃 ′ (𝑡)

Φ𝑏 (𝑡) = = 9.1 {|𝜃 (𝑡)| − 𝜃𝑐𝑟 } (6.24)
3 (1 − 𝑝) |𝜃 ′ (𝑡)|
√(𝑠 − 1)𝑔𝐷50
The values of the coefficients in this equation were derived using various datasets of
sediment transport in oscillatory flow over horizontal beds. Therefore, the formulation
is supplemented with a correction 𝛽𝑠 to account for slope effects on the transport (see
Intermezzo 6.2). The modulus is used to obtain the correct direction of transport, i.e.
in the direction of the instantaneous bed shear stress. The transport magnitude is
dependent on the shear stress magnitude to the power 1.8. The transport rate 𝑆𝑏 (𝑡) is
given in volumes, per unit time and width, of deposited material (see the factor (1 − 𝑝)
in Eq. 6.24).

Intermezzo 6.2 Slope effect
An often-neglected transport component is the one due to gravitation along a slop-

ing bed. In case of a sloping bed, not only the effects of slope on the initiation of
motion (see Eq. 6.15) have to be taken into account, but also the transport directly
induced by gravity when the grains have been set in motion. The bed load formu-
las are mostly derived from data sets for a horizontal bed and do not automatic-
ally include the effect of slope. Nevertheless, we expect moving grains to rather
go downhill than uphill. This gives an additional transport component which is
directed downhill. Therefore, sometimes a slope correction parameter is intro-
duced (after Bagnold, see also Sect. 6.7.2) which increases the transport rates for
downslope transport and decreases the transport rates for upslope transport.
The downhill gravitational transport component has a smoothing effect on the
bed topography. This effect is of great importance for the morphological stability
of the bed and for the equilibrium state to which the bed topography tends (but
which it probably never reaches, since this state is a function of the ever-changing
input conditions).
To calculate 𝑆𝑏 (𝑡), we need to compute the instantaneous dimensionless bed shear

stress (Eq. 6.21). As already indicated in Ch. 5, the computation of time-averaged –
let alone instantaneous – bed shear stress under a combined wave-current motion is
not straightforward at all. Without detailed modelling of the vertical velocity structure
and turbulence, the computation of 𝜃 ′ (𝑡) is most easily done using a quadratic friction
law (see Sects. 5.4.3 and 5.5.5).
Grant and Madsen (1979) suggested expressing bed shear stress as a quadratic func-
tion of the combined wave-current velocity at some height 𝑧 above the bed. With the
velocity 𝑢0 (𝑡) at the top of the wave boundary layer (𝑧 = 𝛿), we have:
′ |𝑢 (𝑡)| 𝑢 (𝑡)
𝜏𝑏 (𝑡) = 1/2𝜌𝑓𝑐𝑤 (6.25)
0 0
in which 𝑓𝑐𝑤′ is a (skin) friction factor for the combined wave-current motion and 𝑢
0
is the time-dependent (intra-wave, that is: within the wave period) near-bottom hori-
zontal velocity vector of the combined wave-current motion. Compare Eq. 6.25 with
Eqs. 5.30 and 5.76. In principle the problem is 2DH, since waves and currents may
interact under an arbitrary angle. The bed shear stress 𝜏𝑏 and near-bed velocity 𝑢0 are
vectors in the same direction with varying magnitudes and varying directions during
the wave cycle.
The velocity 𝑢0 should be representative for irregular wave groups and therefore con-
tain contributions due to wave skewness and asymmetry, wave group-related amp-
litude modulation and bound long waves. These contributions to 𝑢0 (𝑡) were described

in detail in Ch. 5 and follow from an appropriate wave theory. In addition, the mean
flow at the top of the wave boundary layer must be taken into account, for instance
by solving the mean flow in the vertical. Wave-induced contributions to the near-bed
mean velocity are Longuet-Higgins streaming, undertow and longshore current, as
already discussed in Ch. 5. The near-bed velocity vector is then computed as the vec-
tor addition of the near-bed oscillatory velocity signal and the time-averaged velocity
at the same height.
We then have for the time-dependent 𝜃 ′ (𝑡):
1/2𝜌𝑓 ′ |𝑢0 (𝑡)| 𝑢0 (𝑡)

𝑐𝑤
𝜃 ′ (𝑡) = (6.26)
(𝜌𝑠 − 𝜌) 𝑔𝐷50
We have now reduced the problem to the determination of the skin friction factor 𝑓𝑐𝑤 ′ .
This is the major unknown and therefore the bottleneck in many transport computa-
tions. It is dependent on amongst others the bed roughness, which is highly variable
in nature and not easily measured in practical applications. This is why results from
laboratory experiments are still widely used to find a relationship between the friction
factor and the bed roughness.
′ is a skin friction factor (related to grains only
Note that, since the friction factor 𝑓𝑐𝑤
and not to bed forms), we need to use roughness heights related to the grain size and
not the height of bed forms.
Grant and Madsen determine the friction coefficient 𝑓𝑐𝑤 ′ for currents and waves in com-
bination by first computing the (skin) friction factors for ‘waves alone’ and ‘currents in
the presence of waves’. This can be done using formulas as presented in Sect. 5.4.3 (but
with 𝑓𝑐′ evaluated using the mean velocity at the top of wave boundary layer). Note
furthermore that currents in the presence of waves experience an increased roughness
due to the wave boundary layer. Next, 𝑓𝑐𝑤 ′ is calculated by weighting 𝑓 ′ and 𝑓 ′ lin-
𝑤 𝑐
early with the relative strength of the near-bed net current and oscillatory velocity
amplitude. Others propose different models, resulting in different relative contribu-
tions of waves and currents to bed shear stress.
The net wave-averaged bed load transport rate can be obtained by averaging of the
time-dependent transport vector 𝑆𝑏 (𝑡) over the duration of the imposed near-bottom
velocity time series. It includes the transport by the mean current as well as the net
transport as a result of the oscillatory wave motion. Due to the non-linear relation
between velocity and shear stress, the latter contribution will be zero only for a com-
pletely symmetric velocity signal.
At small shear stresses, this bed load transport formulation represents the transport oc-
curring as individual particles moving over a rippled bed, while at higher shear stresses
the formulation represents the sheet flow phenomenon where particles move as bed
load in several layers (sheets) over a plane bed.

6.5.4. Bed load transport based on time-averaged shear stress

Considerably simpler are the bed load formulas in which, instead of the instantaneous
bed shear stress, the time-averaged bed shear stress is used. Bijker (1967) was the first
to present such a model (of the type of Eq. 6.23) for the combination of waves and cur-
rents. As a starting point he used the Kalinske-Frijlink bed load transport formula for
currents only. This empirical formula had been used extensively for river applications.
Bijker adjusted this formula for the combined-wave current situation, by assuming
that the effect of waves is to enhance the stirring of sediment, which is consequently
transported by the mean current. Later we will explain in some detail that this as-
sumption is especially valid for longshore sediment transport calculations (oscillatory
movement approximately perpendicular to the current). The Bijker formula (1967) is
given here as an example of a transport formula for time-averaged bed load transport:
𝑈 −0.27(𝑠 − 1)𝐷50 𝜌𝑔
𝑆𝑏 = 𝐵𝐷50 √ 𝑔 exp [ ] (6.27)
𝐶 𝜇 ⟨|𝜏 𝑐𝑤 |⟩
⏟⏟⏟⏟⏟⏟⏟⏟⏟ ⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟
current only wave-current shear stress
transports stirs up the sediment
the sediment
where:
𝐵 Bijker coefficient (= 5) −
𝐷50 representative particle diameter m
𝑈 current velocity (e.g. longshore current) m/s
𝑠 relative density 𝜌𝑠 /𝜌 −
m ⁄2 /s
1
𝐶 Chézy coefficient (see Intermezzo 5.2)
𝜇 ripple coefficient: part of the total bed shear that is available −
for transporting material
⟨|𝜏𝑐𝑤 |⟩ time-averaged shear stress magnitude for the combined N/m2
wave-current motion
The value of the Bijker coefficient 𝐵 has been subject to much discussion. It has been
suggested that the value of 𝐵 should vary from 2 well outside the breaker zone to 5 in-
side the breaker zone. The main difference with the original Kalinske-Frijlink formula
is the use of the time-averaged wave-current shear stress magnitude ⟨|𝜏𝑐𝑤 |⟩ instead
of the time-averaged current-only bed shear stress. We can interpret Eq. 6.27 as the
product of the transporting mean velocity times the sediment load stirred up by the
combination of waves and currents: 𝑆 = 𝑈 × ‘sediment load’. In shallow water the
contribution of the wave motion to the bed shear stress magnitude (and therefore to
the stirring of sediment) is often more important than the contribution by the mean
current. Therefore, it is often said that waves stir up the sediment, while currents
transport it.

Bijker reasoned that at every moment in time, the bed shear magnitude determines the
stirring of sediment, irrespective of the direction of this shear stress. Therefore, the bed
load transport is dependent on ⟨|𝜏𝑐𝑤 |⟩ instead of, for instance, the time-averaged bed
shear stress ⟨𝜏𝑐𝑤 ⟩. The difference between the two becomes clear when considering
a sinusoidal wave only, for which the bed shear oscillates symmetrically as well. The
time-averaged bed shear stress ⟨𝜏𝑐𝑤 ⟩ is zero. Nevertheless, the waves are able to stir
up sediment when the absolute value of the bed shear |𝜏𝑐𝑤 | is larger than zero (no
threshold of motion criterion is considered). Similarly, for the combination of waves
and currents, the sediment load is determined by the shear stress magnitudes.
Bijker developed generalised formulas for the bed shear stress 𝜏𝑐𝑤 for waves at an arbit-
rary angle with the current. In Sect. 5.5.5 we have already seen that we cannot simply
add up the shear stress due to currents and waves; due to the non-linear relationship
between velocity and shear stress, we need to add the velocities instead (see Fig. 6.14).
Bijker started with a time series of (linear) waves plus current at a certain height 𝑧𝑡
above the bed, from which he derived – using an analytical turbulence model – approx-
imate formulas for not only the mean bed shear stress in the direction of the current
𝜏𝑚 but also for the mean shear stress magnitude ⟨|𝜏𝑐𝑤 |⟩. Bijker’s formula for the mean
value of the bed shear stress magnitude under combined waves and currents reads:
2
1 𝑢̂0
⟨|𝜏𝑐𝑤 |⟩ = 𝜏𝑐 [1 + {𝜉 } ] (6.28)
2 𝑈
where:
𝑢̂0 maximum orbital velocity at top of wave boundary layer m/s

𝑈 depth-averaged velocity m/s
1 𝑓𝑤 𝑓𝑤
𝜉 combination of various parameters 𝜉 = =𝐶 −
√ 2 𝑓𝑐 √ 2𝑔
This formula is dependent on the waves-only friction factor 𝑓𝑤 (Eq. 5.33) and the
current-only friction factor 𝑓𝑐 (Eq. 5.32). It represents the average length of the vector
⟨|𝜏𝑐𝑤 |⟩ of Fig. 6.14 and is independent of the angle between waves and current.
For a typical surf zone situation with approximately normally incident waves (perpen-
𝑢̂
dicular to the longshore current velocity 𝑉 ) and under the assumption of 𝜉 𝑉0 ≫ 1,
Bijker’s formula for the mean bed shear stress 𝜏𝑚 in the current direction reduces to a
formula identical to Eq. 5.77.
Bijker’s model for wave-current shear stress is just one of many wave-current in-
teraction models that can be used in morphodynamic models. Soulsby et al. (1993)
compared various wave-current interaction models, ranging from the Bijker model to
sophisticated wave boundary layer models, and analysed how waves and currents in
a combined wave-current motion contribute to bed shear stress (see Intermezzo 6.3).

Intermezzo 6.3 Parameterisation of wave-current interaction models
The two parameters that were considered in the intercomparison were the time-
averaged bed shear stress in the direction of the current (denoted 𝜏𝑚 ) and the max-
imum bed shear stress (denoted 𝜏max ). See also Fig. 6.12. The choice for these
parameters is based on the fact that 𝜏𝑚 is generally required for the computation
of the mean current profile in wave-current motion, whereas the parameter 𝜏max is
important to determine the threshold of motion and the entrainment of sediment.
Both the maximum shear stress during a wave cycle and the mean shear stress in
the current directionare related to shear stresses that would occur if the wave- and
current-only shear stresses could be summed linearly.
Figure 6.12: Schematic for bed shear stresses with wave-current interaction. The current-
alone stress (𝜏𝑐 , left), and the wave-alone stress (with amplitude 𝜏𝑤 , 𝑚𝑖𝑑𝑑𝑙𝑒) combine non-
linearly to give stresses having mean 𝜏𝑚 and maximum 𝜏max (right). After Soulsby et al.
(1993). The orbital velocity makes an angle 𝛾 with the mean current.
Figure 6.13 shows a result of this intercomparison. The lower figure shows that
mean bed shear stress goes to zero for waves only and to the current-only shear
stress for currents only. For the combination of waves and currents, all models
show the expected non-linear enhancement of the bed shear stress when currents
and waves are combined; the dotted line in both the upper and lower figure shows
the result for a linear addition of the effect of waves and current (shear stress due
to waves and currents simply added up). The models differ in the relative effect
of waves and currents, with the Bijker model overestimating the effect of waves.
This is caused by the fact that the Bijker model does not take into account that
the mean flow near the bed is reduced (and the shear stress enlarged) due to the
extra resistance introduced by the wave boundary layer. In Fig. 6.12 it was already
demonstrated that the angle between waves and current is important. Soulsby et
al. (1993) showed that the non-linear enhancement decreased for angles 𝛾 going
from 0° (waves and current aligned) to 90° (waves and current perpendicular).
The majority of the compared models require extensive computations for the pre-
diction of the time-averaged bed shear stress. Soulsby et al. (1993) parameterised
the models, so that they can be used in a computationally efficient way in morpho-
dynamic models. The models are derived for monochromatic waves (a single har-
monic). When applying the models to irregular waves, the orbital excursion amp-
litude should be based on the 𝐻𝑟𝑚𝑠 (the root-mean-square wave height, Sect. 3.4.2).

1
2
3
4
5
6
7
8
without nonlinear
enhancement
3
6
7
2
1
8
without nonlinear 5
enhancement 4
1. Bijker 5. Huynh-Thanh & Temperville

2. van Kesteren & Bakker 6. Myrhaug & Slaattelid
3. Grant & Madsen 7. Christoffersen & Jonsson
4. Fredsøe 8. Davies, Soulsby & King
Figure 6.13: Intercomparison of various wave-current interaction models for an angle 𝛾 of

0° between waves and currents and for given values of 𝑧0 /ℎ and 𝜉0̂ /𝑧0 where 𝜉0̂ = 𝑢̂0 /𝜔 is
the particle excursion amplitude (see Eq. 5.24) close to the bed and 𝑧0 = 𝑘𝑠 /30 with 𝑘𝑠 is the
Nikuradse roughness height. Adapted from Soulsby et al. (1993).
6.5.5. Summary and concluding remarks

Some theoretical and many (semi-)empirical bed load formulas have been proposed in
the literature. Many formulas were originally developed for rivers, and later applied to
coastal environments by using a bed shear stress due to waves and currents in the ori-
ginal formulas (Meyer-Peter Müller, Kalinske et cetera). These formulas can normally
be written in the form Eq. 6.23.
Unfortunately, there is a massive difference in the results of the different (bed load)
formulas, making the uncertainty in sediment transport computations rather large.
Transport formulas can only be used with enough confidence if they have been prop-
erly calibrated, preferably with data for the considered site and for representative hy-
drodynamic conditions.
A common attribute in bed load formulas is that the shear stress is raised to a certain
power, say 1.5 to 2. In Eq. 6.24, for instance, at any moment in time, the transport is in

Figure 6.14: Left: plan view of velocity components 𝑢𝑡 and 𝑉𝑡 at a certain elevation 𝑧𝑡 above the
bed. Top right: the resultant velocity is 𝑉𝑟 and varies in magnitude and direction during the wave
cycle (between 𝑉𝑟,min and 𝑉𝑟,max ). Bottom right: bed shear stress components during the wave
cycle. Since 𝜏𝑐𝑤 ∝ |𝑉𝑟 | 𝑉𝑟 , the larger velocities contribute relatively more to the bed shear stress.
the direction of the intra-wave velocity and has a magnitude proportional to the shear
stress magnitude to the power 1.8. Since the shear stress itself is related quadratically
to the velocity signal, we may simply state that the magnitude of the instantaneous bed
load transport depends on the modulus of the near-bed velocity raised to a power 𝑛 = 3
to 4. As expected, the velocity direction does not influence the sediment load. Since the
𝑛
transport is in the direction of the instantaneous velocity, we have 𝑆(𝑡) ∝ sign(𝑢) |𝑢|
𝑛−1
or equivalently 𝑆(𝑡) ∝ 𝑢 |𝑢| . The latter function may be interpreted as the product
of a transporting velocity 𝑢 and the sediment load stirred by waves and currents pro-
𝑛−1
portional to |𝑢| . This concept we already encountered in Sect. 6.5.4 for the more
specific case of transport dominated by a mean current. Averaged over time, for in-
𝑛−1
stance over the short-wave period or the tidal period, we have ⟨𝑆⟩ ∝ ⟨𝑢 |𝑢| ⟩, with
brackets denoting time-averaging. Intermezzo 6.4 demonstrates that even a purely
oscillatory motion may give rise to a net (time-averaged) sediment transport.
Intermezzo 6.4 Net sediment transport due to an oscillatory motion?
An oscillatory velocity signal may, under certain circumstances, result in a wave-

averaged sediment transport. First imagine a velocity signal 𝑢(𝑡) that is purely
symmetric about the horizontal axis (a sine or sawtooth wave). In that case 𝑢|𝑢|𝑛−1
is also symmetric and it follows that the wave-averaged transport ⟨𝑆⟩ = 0 (check
this by sketching 𝑢, |𝑢|2 and 𝑢|𝑢|2 as a function of time). The symmetrical orbital
motion simply moves an amount of sediment back and forth without a net wave-
averaged transport. Now consider a positively skewed velocity signal, typical for
shoaling waves (Sect. 5.3) or a flood-dominant tide (Sect. 5.7.4), with larger peak

6.6. Diffusion approach for suspended transport 301
velocities in the wave propagation direction than in the opposite direction. Even
though ⟨𝑢⟩ = 0, we now find ⟨𝑆⟩ ≠ 0. This is because the sediment load responds
non-linearly to the velocity, such that more sediment is stirred up during the part
of the wave cycle with velocities in the propagation direction (again draw 𝑢|𝑢|2 ).
The result is a net (bed load) transport in the propagation direction. Similarly, a net
current superimposed on a sinusoidal velocity signal introduces asymmetry about
the horizontal axis, leading to a net sediment transport (in the current direction)
that is larger than the transport for the current alone situation. Verify this!
6.6. Diffusion approach for suspended transport

6.6.1. General formulation
When the actual bed shear stress is (much) larger than the critical bed shear stress,
the particles will be lifted from the bed. If this lift is beyond a certain level, then the
turbulent upward forces may be larger than the submerged weight of the particles. In
that case, the particles go into suspension, which means that they lose contact with
the bottom for some time.
For not too high sediment concentrations and not too heavy particles (so that they fol-
low the water motion), we can assume that at every height the particles move through
a vertical plane with the horizontal water velocity. The sediment flux can then be com-
puted from the vertical distribution of fluid velocities and sediment concentrations, as
follows:
𝑞(𝑧, 𝑡) = 𝑐(𝑧, 𝑡)𝑢(𝑧, 𝑡) (6.29)
where:
𝑞 sediment flux m3 /s/m2

𝑐 local instantaneous sediment concentration at height 𝑧 above m3 /m3
bed
𝑢 local instantaneous fluid velocity at height 𝑧 above bed m/s
The instantaneous suspended transport rate 𝑆𝑠 is found by integrating the sediment

flux (Eq. 6.29) from the top of the bed load layer (𝑧 = 𝑎, Fig. 6.10) to the instantaneous
water level ℎ = ℎ0 + 𝜂:
ℎ
𝑆𝑠 (𝑡) = ∫ 𝑐(𝑧, 𝑡)𝑢(𝑧, 𝑡) 𝑑𝑧 (6.30)
𝑧=𝑎

The suspended sediment transport is in the direction of the water velocity. Time-
averaging over a representative period gives the time-averaged suspended sediment
transport (exclusive of pores).
The instantaneous velocity and concentration at a certain height consists of a mean
part and an oscillatory part, which is fluctuating on the wave scale but has a zero time
mean. Hence:
𝑢 = 𝑈 + 𝑢̃ and 𝑐 = 𝐶 + 𝑐̃ (6.31)
in which:
𝑈 time-averaged fluid velocity at height 𝑧 m/s

𝐶 time-averaged concentration at height 𝑧 m3 /m3
𝑢̃ oscillating fluid component m/s
𝑐̃ oscillating concentration component m3 /m3
Substituting Eq. 6.31 into Eq. 6.30 and time-averaging yields:
ℎ ℎ
⟨𝑆𝑠 ⟩ = ∫ 𝑈 𝐶 𝑑𝑧 + ∫ 𝑢̃ 𝑐 ̃ 𝑑𝑧 (6.32)
⏟⏟⏟⏟⏟⏟⏟⏟⏟ ⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟
𝑎 ⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟
𝑎
time-averaged current-related wave-related
sediment transport rate part part
Time-averaging is indicated by the brackets and the overbar. In order to express the
transport rate in terms of deposited volumes, we need to multiply Eqs. 6.30 and 6.32
with 1/(1 − 𝑝), with 𝑝 the porosity of the deposited material.
The first part of the right-hand side of Eq. 6.32 is the so-called current-related suspended
sediment transport: the transport of sediment particles by the time-averaged current
velocities. In the coastal zone, the transporting currents are often wave-induced, for
instance rip currents and longshore currents. Furthermore, secondary currents can
give rise to a net current-related sediment transport, even though the depth-averaged
mean current velocity is zero, see Intermezzo 6.5. An example is the undertow, below
the wave trough level, which is responsible for an offshore-directed suspended sed-
iment transport under breaking waves. The time-averaged sediment concentrations
are affected by the wave motion. Considerable amounts of sediment are brought into
suspension by turbulence, generated at the surface under breaking waves and in the
wave boundary layer near the bed.
The second part of the right-hand side is the wave-related suspended sediment trans-
port: the transport of sediment particles by the oscillatory water motion. As for the
bed load transport, skewness of the oscillatory velocity signal can result in a net wave-
related suspended sediment transport (see Intermezzo 6.4). It is however complex and

(computer) time-consuming to compute the wave-related suspended sediment trans-

port by solving the intra-wave velocities and concentrations. Moreover, very little is
known about the intra-wave concentrations and such a computation would be very
uncertain. The exact phase relationship between velocity and concentration at every
height above the bed is crucial in determining the magnitude and direction of the wave-
related suspended sediment transport. In the case of a plane bed, one may expect a
wave-related suspended sediment transport in the wave propagation direction. But in
the case of a rippled bed (where the concentrations are largest around flow reversal,
see Intermezzo 6.1) this may lead to an offshore-directed wave-related suspended load
transport! The latter could be the case if more sediment is brought into suspension at
flow reversal from onshore to offshore.
The wave-related part of the suspended sediment transport is often assumed to be
smaller than the current-related part. This simplifies the modelling significantly. In
practical models it is therefore often assumed that the suspended load transport is
dominated by the transport by the mean current; the suspended sediment flux is com-
puted as the product of the wave-averaged current and concentration profiles (using
e.g. formulations according to Van Rijn (1989, 1993, 2000) and Soulsby and Van Rijn
(Soulsby, 1997).
ℎ
⟨𝑆𝑠 ⟩ ≃ ∫ 𝑈 𝐶 𝑑𝑧 (6.33)
⏟⏟⏟⏟⏟⏟⏟⏟⏟ ⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟
𝑎
time-averaged current-related
sediment transport rate part
The contribution of the oscillatory velocity to the sediment transport is then thought
to be taken into account in bed load transport computations. The mean (wave-induced)
current profile can be computed as explained in Sect. 5.5.6. In order to compute the
time-averaged concentration, an advection-diffusion equation is often used. This will
be treated below.
Note that even with a zero depth-averaged velocity – as in the case of a circulation
current – a net current-related sediment transport can occur (see Intermezzo 6.5).
By using a bed load transport formula of the form as in Eq. 6.22 and a suspended load
transport according to Eq. 6.33, it is possible that bed load and suspended load transport
are in opposite directions. Let us first consider cross-shore sediment transport. In the
surf zone, the bed load transport will generally be directed onshore (due to short-wave
skewness) and the suspended transport offshore as a result of the undertow. This is fur-
ther discussed in Sect. 7.5. In the alongshore direction, further simplifications may be
made. Since, due to refraction, the oscillatory wave motion is almost perpendicular to
the coast, the transport in the alongshore direction is governed by the slowly varying
longshore current (wave- or tide-induced). The role of waves in longshore transport is
merely to stir up sediment that is consequently predominantly transported along the

coast by the current (cf. Sect. 6.5.4). Specific longshore sediment transport formulas
are treated in Sect. 8.2.
Intermezzo 6.5
Net suspended sediment transport due to secondary flows
The combination of suspended load transport with secondary flows can give a
net sediment transport, even though the depth-averaged velocity is zero. This
is the result of the non-uniform distribution of the sediment concentration over
the vertical: most of the sediment is concentrated in the lower part of the water
column. An example is the cross-shore secondary flow under breaking waves
discussed above, that gives a net transport component against the direction of
wave propagation; the sediment concentrations are largest in the lower layers of
the water column, where the velocity is offshore-directed (undertow). Under non-
breaking waves, wave-induced streaming close to the bed may result in an onshore
directed transport. In the same way, secondary flows in a tidal basin can have
a net effect on the transport direction. The curvature-induced secondary flow
in a channel bend (Sect. 5.7.6), for instance, will give a net transport component
towards the centre of curvature (from the outer bend to the inner bend). Another
potentially important contribution of this type is associated with the deformation
of the velocity profile as the flow moves up or down a steep slope (see Fig. 6.15).
When averaged over the tide, this gives a net upslope transport. This may explain
why the steep banks of tidal channels can be stable.
ebb flood residual
Figure 6.15: Residual tide-averaged velocity profile resulting in a net upslope transport.
6.6.2. Sediment continuity

In order to obtain the sediment concentration, a mass balance equation for the sedi-
ment needs to be solved. The general conservation statement for the sediment reads:
𝜕𝑐 𝜕𝑢𝑐 𝜕𝑣𝑐 𝜕𝑤𝑐 𝜕𝑤 𝑐

+ + + − 𝑠 = 0 (6.34)
𝜕𝑡
⏟ 𝜕𝑥 𝜕𝑦
⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟ 𝜕𝑧
⏟⏟⏟⏟⏟⏟⏟ 𝜕𝑧
⏟⏟⏟⏟⏟⏟⏟⏟⏟
change in net import of sediment net upward transport net downward transport
sediment by the horizontal of sediment by the with the fall velocity
concentration fluid velocity vertical fluid velocity

or using the continuity equation for the fluid:
𝜕𝑐 𝜕𝑐 𝜕𝑐 𝜕𝑐 𝜕𝑤 𝑐
+𝑢 +𝑣 +𝑤 − 𝑠 =0 (6.35)
𝜕𝑡 𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑧
Please note that in Eqs. 6.34 and 6.35 the velocity and the concentration are the total
signals consisting of a mean, an oscillatory and a turbulent part.
The horizontal advective terms are often (but not always) smaller than the vertical
advective terms (see Intermezzo 6.6). Let us for simplicity neglect the horizontal ad-
vective terms of Eq. 6.34, so we are left with:
𝜕𝑐 𝜕𝑤𝑐 𝜕𝑤𝑠 𝑐
+ − =0 (6.36)
𝜕𝑡 𝜕𝑧 𝜕𝑧
Intermezzo 6.6 Drift of suspended sediment
Similarly to the drift of water (Stokes drift, Sect. 5.5.1), there is also a drift of sus-
pended sediment. Under the wave crest the suspended sediment concentration
is stretched out, whereas under the trough it is compressed. Stokes drift was ex-
plained by the fact that water levels are higher during the forward orbital motion
than during the backward motion. Analogously, a suspended sediment drift oc-
curs. This drift is only taken into account when the horizontal advective terms
are included in the advection-diffusion equation.
As said previously, the velocity and the concentration consist of a mean, an oscillatory
and a turbulent part. Hence, 𝑤 = 𝑊 + 𝑤̃ + 𝑤 ′ and 𝑐 = 𝐶 + 𝑐 ̃ + 𝑐 ′ . We are certainly not
going to resolve the turbulent motion and would like to average over that motion. This
is called Reynolds averaging. If we average Eq. 6.36 over the turbulence scale, most
terms with turbulent fluctuations average out, except for one term, viz. 𝜕⟨𝑐 ′ 𝑤 ′ ⟩/𝜕𝑧.
Here the brackets denote averaging over the turbulence timescale and ⟨𝑐 ′ 𝑤 ′ ⟩ is the
(upward) sediment flux by turbulence. Further, the Reynolds averaged vertical water
velocity can be assumed to be negligible compared to the fall velocity of the sediment.
With the simplification of a constant fall velocity (in reality the fall velocity depends
on the concentration, see Sect. 6.2.3), we end up with the following often-used form of
the advection-diffusion equation:
𝜕𝑐 𝜕𝑐 𝜕⟨𝑐 ′ 𝑤 ′ ⟩
− 𝑤𝑠 + =0 (6.37)
𝜕𝑡 𝜕𝑧
⏟⏟⏟⏟⏟⏟⏟⏟⏟ 𝜕𝑧
⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟
sediment net going downward sediment net going upward
with its fall velocity with fluid turbulence
Please note that in this equation the concentration 𝑐 now denotes the turbulence-
averaged concentration: 𝑐 = 𝐶 + 𝑐.̃

In order to model the sediment flux due to turbulence, we make a similar assumption
that upward transport of sediment is due to turbulent diffusion, as we did for the fluid
(see for instance Eqs. 5.85 and 5.88):
𝛿𝑐
−⟨𝑐 ′ 𝑤 ′ ⟩ = 𝜈𝑡,𝑠 (6.38)
𝛿𝑧
in which 𝜈𝑡,𝑠 is the turbulent diffusivity of sediment mass in m2 /s and 𝑐 is now defined
as volume concentration (see Sect. 6.2.2). This upward transport by turbulent diffusion
can be understood as follows. Turbulent exchange makes sure that a sediment-laden
fluid parcel goes upward to a level with a lower sediment concentration. A sediment
parcel going downward contains less sediment than the average parcel at the level
where it arrives. Since more sediment particles are carried upward than downward,
the net effect is upward transport.
Sometimes the turbulent diffusivity 𝜈𝑡,𝑠 of sediment mass is taken to be equal to the
turbulent viscosity 𝜈𝑡 of the water. However, it can also be argued that the mixing of
water and sediment are two different things. Whatever approach is taken, normally
the damping of turbulence due to high sediment concentrations is taken into account.
This refers to the influence that sediment particles have on the turbulence structure of
the fluid. This effect becomes increasingly important for high sediment concentrations
that result in stratification and hence damping of turbulence. This affects both the
water motion and the sediment distribution. Empirical formulations are sometimes
used, which reduce the eddy viscosity dependent on the sediment concentration.
The non-steady advection-diffusion equation now reads:
𝜕𝑐 𝜕𝑐 𝜕 𝛿𝑐
− 𝑤𝑠 − 𝜈𝑡,𝑠 = 0 (6.39)
𝜕𝑡 𝜕𝑧 𝜕𝑧 𝛿𝑧
As a bottom boundary condition, a sediment concentration at a certain level near the

bed is prescribed, the so-called reference concentration. This concentration is pre-
scribed at a certain reference level, for instance 𝑧𝑎 = 2𝐷50 , also depending on the bed
load formula. The reference concentration is often assumed to be a function of the
bed shear stress (much like the bed load transport formulas). Hence, at the bed the
response is quasi-steady, whereas higher in the vertical, the sediment concentration
lags behind the shear stress at the bed. An alternative boundary condition is a pick-up
function that prescribes the vertical concentration gradient instead of the concentra-
tion.
Detailed models that also resolve the wave boundary layer (Sect. 5.4.3) resolve the
time-dependent concentration in the vertical, in order to model sheet flow transport
and suspended load transport. The time-dependent suspended sediment concentration
can be determined with an unsteady advection-diffusion equation such as Eq. 6.39.

This approach involves the modelling of the turbulence, which increases and decreases
during a wave period; this is difficult and time-consuming (see also Intermezzo 6.1).
6.6.3. Time-averaged concentration distribution

In a steady situation 𝜕𝑐/𝜕𝑡 = 0 and 𝑐 = 𝐶. A balance must exist between the upward
transport by turbulence and the downward transport with the fall velocity. Integration
of Eq. 6.39 over the depth (with a zero vertical flux at the water surface) leads to:
𝑑𝐶(𝑧)
𝑤𝑠 𝐶(𝑧) + 𝜈𝑡,𝑠 (𝑧) =0 (6.40)
𝑑𝑧
This equation indicates an equilibrium between the downward transport with the fall
velocity 𝑤𝑠 𝐶(𝑧) and the net upward movement of grains by turbulence.
As mentioned above, local turbulence in the water column leads to an upward and a
downward exchange of sediment-laden fluid parcels. Since the sediment concentration
is larger close to the bed, more grains will be transported in the upward direction than
in the downward direction. This leads to a turbulent transport of sediment, which
depends on the gradient in concentration over the vertical. This is why it is referred to
as a gradient-type transport. On average, turbulence transports sediment from levels
of high concentration to levels of lower concentration, i.e. from lower levels in the
water column to higher levels.
Generally the suspended sediment is somewhat finer than the bed material. Hence, in
order to compute the constant fall velocity, a smaller grain diameter is often used than
in the computation of the bed load transport.
The assumption of upward transport due to turbulent diffusion seems reasonable for
a plane bed. In case of a rippled bed however, it can be expected that upward trans-
port is by eddies generated by the ripples (see Intermezzo 6.1). These are coherent
fluid motions that bring the sediment upward by convection. In that case, the whole
concept of upward transport proportional to a concentration gradient does not hold
anymore. Sometimes, the diffusion concept is stretched a bit by assuming that the
sediment diffusivity accounts for the convection processes also.
It makes sense to relate the turbulent diffusivity 𝜈𝑡,𝑠 of sediment mass to the eddy
viscosity of the fluid through a factor 𝛽: 𝜈𝑡,𝑠 = 𝛽𝜈𝑡 . A value of 𝛽 < 1 reflects that
the sediment particles cannot respond fully to the turbulent fluid velocity fluctuations
(inertia). On the other hand 𝛽 > 1 (supported by laboratory experiments) indicates
a more effective mixing for sediment particles than for water particles. It has been
argued that this is the result of larger centrifugal forces on sediment particles in eddies
than on fluid particles (because of their higher density). In the following 𝛽 = 1 is used.

With an appropriate distribution for 𝜈𝑡,𝑠 (for which many options are discussed in liter-
ature!) and a bottom boundary condition, Eq. 6.40 can be integrated either numerically
or analytically. Analytical solutions are possible for simple diffusivity distributions.
The integration is performed from a near-bed reference level 𝑎 to the water surface. At
the reference level 𝑎, a concentration-type boundary condition can be used. Since in
principle 𝑧 = 𝑎 corresponds to the top of the bed load layer, the reference concentration
must be somehow related to the bed load transport rate, for instance via:
𝑆𝑏
𝐶𝑎 = (6.41)
𝑢𝑎
where:
𝑆𝑏 bed load transport m3 /m/s

𝑢 average fluid velocity in the bottom layer m/s
𝑎 thickness of the bottom layer (order of magnitude of the bottom m
roughness 𝑟)
The reference concentration is often formulated as a function of the time-averaged bed

shear stress and should take the enhanced stirring by waves into account.
The most general solution of Eq. 6.40 is:
𝑧
𝑤𝑠
𝐶(𝑧) = 𝐶𝑎 exp [− ∫ 𝑑𝑧] (6.42)
𝑧=𝑎 𝜈𝑡,𝑠 (𝑧)
where:
𝐶𝑎 time-averaged concentration at level 𝑧 = 𝑎 −

𝑤𝑠 fall velocity m/s
𝑧 height above the bed m
Einstein (1942, 1950) and Rouse (1937) suggested a parabolic distribution of the diffu-
sion coefficient over the water depth:
𝑧
𝜈𝑡,𝑠 (𝑧) = 𝜅𝑢∗ (ℎ − 𝑧) (6.43)
ℎ
where:
𝜅 Von Karman constant = 0.4 −

A parabolic sediment diffusivity is in accordance with the generally assumed parabolic

eddy viscosity in the water column (increasing from zero at the ‘walls’ to a maximum
away from the walls, see Eq. 5.88 and the text below it).
This results in the following concentration distribution (see Fig. 6.16):
𝑧∗
ℎ−𝑧 𝑎
𝐶(𝑧) = 𝐶𝑎 [ ] (6.44)
𝑧 ℎ−𝑎
where:
𝑎 thickness of the bottom layer m

𝐶𝑎 concentration at reference level 𝑧 = 𝑎 −
𝑤
𝑧∗ Rouse number defined as 𝑧∗ = 𝜅𝑢𝑠 −
∗
z
(linear scale)
Ca
Cz [kg/m³]
0.1 1 10 100 (log scale)
Figure 6.16: Sediment concentration distribution over water depth (parabolic mixing coeffi-
cient). Concentration magnitude and distribution over depth are representative for wave or wave-
current situations, as for instance in the surf zone.
The Rouse number is a non-dimensional number that not only defines the sediment
concentration profile, but also determines the mode of transport. It is the ratio between
the downward sediment fall velocity and the upwards velocity on the grain represented
by 𝜅𝑢∗ . For Rouse numbers larger than, say, 2.5 (or 𝑤𝑠 /𝑢∗ > 1) all transport is bed
load transport. For a Rouse number smaller than, say, 0.8, we have only wash load.
Between these extremes, sediment suspension occurs. If 𝑤𝑠 /𝑢∗ > 1 we may say that
the sediment is coarse and responds quasi-steadily to the flow field. For 𝑤𝑠 /𝑢∗ < 1
(‘fine’ sand in suspension) non-instantaneous sediment responses start playing a role.

6.7. Energetics approach

6.7.1. Introduction
The energetics approach to bed load and suspended load transport is a more integrated
approach to bed load and suspended load transport and was first developed by Bagnold
(1963, 1966) for rivers. The underlying idea is that a certain amount of energy is needed
to keep the bed load moving and the suspended load at a certain height above the bed.
The sediment transport formula of Bagnold is therefore proportional to the rate of
energy dissipation of the stream. It also includes an efficiency factor for bed load and
suspended load that determines how efficiently the energy is used for the sediment
transport. Besides, the effect of downward transport by gravity is included. In this
section the energetics transport model is discussed in general terms. In Sect. 7.5, we
apply the model to unravel the various contributions to cross-shore sediment transport
as well as treat model applications to obtain equilibrium beach profiles.
6.7.2. Energetics approach for combination of waves and currents

Bagnold derives formulas for the immersed weight suspended load 𝐼𝑠 and bed load
transport 𝐼𝑏 (see text below Eq. 6.17 for the definition of immersed weight transport
rates) for uni-directional flow. In his derivations, it was hypothesised that the fluid
acts as a machine expending energy at a prescribed efficiency rate to offset the work
done in transporting sediment. His reasoning is briefly demonstrated here for bed load
transport.
The immersed weight of a mass of bed load 𝑚𝑏 is given by 𝑊 = (𝜌𝑠 − 𝜌) /𝜌𝑠 𝑔𝑚𝑏 . The
component of this perpendicular to the sloping bed – with bed slope tan 𝛼 – is 𝑊 cos 𝛼,
whereas the component parallel to the bed is 𝑊 sin 𝛼. Hence, the frictional resistance
for downslope transport is equal to 𝑊 (𝜇 cos 𝛼 − sin 𝛼). Here, the friction coefficient
𝜇 = tan 𝜑𝑟 with 𝜑𝑟 is the angle of repose (Sect. 6.2). ‘Work’ is defined by the exerted
force times the distance travelled in the direction of the force. Therefore, the work
done (per unit time) in maintaining bed load in motion is equal to the force required
to overcome frictional resistance times the mean speed 𝑈𝑏 at which the grains travel.
Further, since 𝑈𝑏 cos 𝛼 is the horizontal component of the grain velocity in a vertical
section, the immersed weight transport rate through a vertical section is defined as
𝐼𝑏 = 𝑊 𝑈𝑏 cos 𝛼. We can now write:
work done per unit time = 𝑊 (𝜇 cos 𝛼 − sin 𝛼) 𝑈𝑏 = 𝐼𝑏 (tan 𝜑𝑟 − tan 𝛼) (6.45)
Some fraction 𝜀𝑏 of the dissipated fluid power 𝜔 is expended to offset the work done
in maintaining the bed load.

6.7. Energetics approach 311
Hence:
𝜀𝑏 𝜔
𝐼𝑏 = (6.46)
(tan 𝜑𝑟 − tan 𝛼)
A similar derivation for suspended load leads to:
𝜀𝑠 𝜔
𝐼𝑠 = (6.47)
(𝑤𝑠 /𝑈𝑠 − tan 𝛼)
Eqs. 6.46 and 6.47 apply to uni-directional flow along a downsloping bed. Bowen (1980)
rewrites these equations for a cross-shore situation with normally incident waves. The
formulas for the instantaneous transport rates then read:
𝜀𝑏 𝜌𝑐𝑓 𝑢 3
𝐼𝑏 = (6.48a)
tan 𝜑𝑟 − 𝑢/|𝑢| tan 𝛼
𝜀𝑠 𝑐𝑓 𝜌𝑢 3 |𝑢|
𝐼𝑠 = (6.48b)
𝑤𝑠 − 𝑢 tan 𝛼
where:
𝜀𝑏 , 𝜀𝑠 efficiencies for bed and suspended load −

𝑐𝑓 friction coefficient −
𝑤𝑠 sediment fall velocity m/s
tan 𝜑𝑟 tangent of angle of repose of the sediment −
tan 𝛼 bed slope −
The time-dependent velocity 𝑢 is defined as positive seawards, the direction of 𝑥 pos-

itive. The bed load and suspended load velocity of the grains (𝑈𝑏 and 𝑈𝑠 ) are repres-
ented by 𝑢. Further, the dissipated fluid power 𝜔 (work done per unit time) is com-
puted as the energy dissipation rate 𝐷𝑓 , due to bottom friction. 𝐷𝑓 is given by the
product of bed shear stress 𝜏 and the near-bottom time-varying flow 𝑢. The bed shear
stress is assumed to be described by a quadratic friction law 𝜏𝑏 = 𝜌𝑐𝑓 |𝑢|𝑢, so that
𝜔 = 𝐷𝑓 = 𝜌𝑐𝑓 𝑢 2 |𝑢|. Note that in this way only the dissipation in the wave boundary
layer is taken into account. For applications in the surf zone, Roelvink and Stive (1989)
add to this the energy dissipation due to turbulence near the bottom induced by wave
breaking. The modulus signs in Eqs. 6.48a and 6.48b are chosen such as to properly ac-
count for the direction of transport in terms of velocity and slope. Hence, for seaward
transport (𝑢 positive) the denominators are reduced; material is therefore more easily
transported downslope.

The bed slope is limited by the following two conditions:
tan 𝛼 → tan 𝜑𝑟 giving avalanching or slumping (6.49a)
tan 𝛼 → 𝑤𝑠 /𝑢 giving autosuspension (6.49b)
Autosuspension refers to the fact that for a certain bed slope (and hence a certain
amount of gravitational energy) the sediment ‘suspends itself’. Bailard (1981) argued
that the condition in Eq. 6.49b needs to be supplied with an efficiency factor as well
(such that 𝜀𝑠 𝑢 instead of 𝑢 should be used).
Note that the model lacks an initiation of transport condition like the critical Shields
parameter and that both the bed load and the suspended sediment transport respond
instantaneously to the flow field (as opposed to the diffusion approach of Sect. 6.6).
In a parallel development, Bailard (Bailard, 1981, 1982; Bailard & Inman, 1981) gen-
eralised the Bagnold model to a total load model of time-varying sediment transport
over a plane, sloping bed. In the Bailard formulation both the bed load and suspended
load transport rate vectors are composed of a velocity-induced component directed
parallel to the instantaneous velocity vector and a gravity-induced component direc-
ted downslope. Now alongshore and onshore-offshore currents as well as oscillatory
wave-induced orbital motion with a local angle of incidence are considered.
Here we will not go into the details of the derivation and the final formulas, but follow
Roelvink and Stive (1989), who write the Bailard formula in general terms (for the
simplified case of the velocity aligned with the local bed slope):
𝑛−1 𝑚
𝑆(𝑡) = 𝐶 1 𝑢(𝑡) |𝑢(𝑡)|
⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟ + 𝐶 2 |𝑢(𝑡)| tan 𝛼
⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟ (6.50)
quasi-steady response to time-varying flow response to downslope gravity force
Here tan 𝛼 is the local bed slope and 𝑢(𝑡) is the near-bottom time-varying cross-shore
flow. For bed load, the powers are 𝑛 = 𝑚 = 3, and for suspended load the powers are
𝑛 = 4 and 𝑚 = 5. After time-averaging (indicated by brackets) we find that:
• the bed load transport ⟨𝑆𝑏 ⟩ is proportional to the odd moment ⟨𝑢|𝑢|2 ⟩ and the
even moment ⟨|𝑢|3 ⟩; and
• the suspended load transport ⟨𝑆𝑠 ⟩ is proportional to the odd moment ⟨𝑢|𝑢|3 ⟩ and
the even moment ⟨|𝑢|5 ⟩.
Note that for the even moments, the terms within brackets (i.e. before time-averaging)
are positive for every moment in time, whereas for the odd moments these terms have
the same sign as the instantaneous velocity. The transport terms containing the odd
moments ⟨𝑢|𝑢|2 ⟩ and ⟨𝑢|𝑢|3 ⟩ reflect the quasi-steady bed load and suspended sediment

6.8. Some aspects of (very) fine sediment transport 313
load transport respectively, due to the time-varying flow. The terms containing the
even moments ⟨|𝑢|3 ⟩ and ⟨|𝑢|5 ⟩ reflect the downslope-directed, gravity-driven trans-
port and are usually an order of magnitude smaller than the terms containing the odd
velocity moments.
We could therefore make the following approximate statement:
⟨𝑆𝑏 ⟩ ∝ ⟨𝑢|𝑢|2 ⟩ (6.51a)
⟨𝑆𝑠 ⟩ ∝ ⟨𝑢|𝑢|3 ⟩ (6.51b)
Note that in Sect. 6.5.5, we found similar dependencies for bed load transport, viz.
⟨𝑆𝑏 ⟩ ∝ ⟨𝑢|𝑢|𝑛−1 ⟩ with 𝑛 = 3 to 4. There we also found that net sediment transport is
either due to net currents or due to skewed oscillatory velocity signals. This distinc-
tion is comparable to the distinction between current-related and wave-related sus-
pended sediment transport, as discussed in Sect. 6.6.1. We will further explore these
different transport contributions in Sect. 7.5 (for cross-shore sediment transport) and
in Sect. 9.7.2 (for tide-induced sediment transport).
As said before, a criterion for initiation of motion was not taken into account in the
energetics approach discussed above. This can be expected to increase the asymmetry
in sediment stirring. Another simplification is the assumed quasi-steady approach for
suspended load transport. Coarse sediment responds more or less instantaneously to
the flow velocity, but for finer material this may not be the case. Fine material in this
respect is defined as having a diameter such that 𝑢∗ /𝑤𝑠 > 1, where 𝑢∗ is the shear
velocity and 𝑤𝑠 is the fall velocity (see also Sect. 6.6.3). The larger this parameter, the
more time the suspended particles need to settle. As a result of these time-lags in the
vertical sediment distribution, the transport rates may be reduced (since the maximum
concentrations no longer coincide with the maximum velocities at every height above
the bed). An example of a sediment transport formula in which such time-lag effects
are taken into account is given by Dibajnia and Watanabe (1993).
6.8. Some aspects of (very) fine sediment transport

6.8.1. Memory effects
Suspended load transport is a quite common condition in the coastal environment,
maybe even more common than pure bed load. From a morphological modelling point
of view, we must distinguish two types of suspended load: one which is determined
entirely by the hydrodynamic conditions and the sediment properties at the point of
consideration, and one which includes a ‘memory effect’ and responds to the condi-
tions in all points it has come through in the past. The former type can be modelled
with a sediment transport formula (e.g. the Bijker or the Bailard formula), or with an

intra-wave model which describes the suspension process during a wave cycle (see
Sect. 6.6).
The depth-averaged sediment concentration associated with the second type of suspen-
ded load transport is described by an advection/diffusion equation of the type Eq. 6.52,
as given by, amongst others, Katopodi and Ribberink (1992) for tidal currents, and
Wang and Ribberink (1986) for nearshore applications:
𝜕𝑐 𝑢 𝜕𝑐 𝜈 𝜕𝑐
𝑇𝐴 + 𝐿𝐴 [ + ] = 𝑐𝑒𝑞 − 𝑐 (6.52)
𝜕𝑡 𝑢tot 𝜕𝑥 𝑢tot 𝜕𝑦
Eq. 6.52 describes the adjustment of the depth-averaged concentration 𝑐 to its equilib-
rium value 𝑐𝑒𝑞 according to a relaxation process. The parameters 𝑇𝐴 and 𝐿𝐴 represent
the characteristic scales of this adjustment process. The adaptation time 𝑇𝐴 is a times-
cale of the order of magnitude ℎ/𝑤𝑠 and the adaptation length 𝐿𝐴 is a length scale of the
order of magnitude 𝑢tot ℎ/𝑤𝑠 . The adaptation time- and length scales increase for finer
sediment (smaller settling velocity). The equilibrium concentration, 𝑐𝑒𝑞 , corresponds
to the spatially uniform situation and is usually derived from a sediment transport
formula, or from a simpler model which applies to uniform situations.
Note that Eq. 6.52 can also be considered a decay equation of the type:
𝐷𝑐 𝑐 𝑐𝑒𝑞
+ = (6.53)
𝐷𝑡 𝑇𝐴 𝑇𝐴
in which 𝐷/𝐷𝑡 stands for the material derivative, i.e. moving along with the sediment
flow. If the concentration at 𝑡 = 0 is given, the general solution can be written as:
𝑡
1
𝑐(𝑡) = 𝑐(0)𝑒 −𝑡/𝑇𝐴 +∫ 𝑐 (𝜏 )𝑒 −(𝑡−𝜏 )/𝑇𝐴 𝑑𝜏 (6.54)
0 𝑇𝐴 𝑒𝑞
in which 𝜏 is a formal time variable which runs from 0 to the actual time 𝑡. Apparently,
all values of the equilibrium condition encountered in the time interval between 0 and
𝑡 contribute to the forcing term (i.e. the second term in the right-hand part of the
equation), but their contributions become smaller as they occurred longer ago. This
shows that the memory of the system is not infinite, but has a certain timescale 𝑇𝐴 .
The memory effect of the suspended load concentration can be rather important in
a tidal inlet system, with its strong spatial variations of the bed topography and the
hydrodynamic conditions (see Sect. 9.7.3).
Note that for constant 𝑐𝑒𝑞 , Eq. 6.54 reduces to 𝑐(𝑡) − 𝑐𝑒𝑞 = (𝑐(0) − 𝑐𝑒𝑞 )𝑒 −𝑡/𝑇𝐴 . This can
be rewritten as 𝑐(𝑡) = 𝑐(0) + (𝑐𝑒𝑞 − 𝑐(0))(1 − 𝑒 −𝑡/𝑇𝐴 ), which is similar in form to Eq. 1.2.

6.8.2. Critical shear stress and settling velocity

Sand, silt and clay particles are continuously entrained, transported and deposited in
the coastal system by currents and waves. Due to the difference in grain size and
cohesiveness, the sediment transport characteristics of sand and mud are different. In
general, sand is transported as bed load (at the bed-water interface) and as suspended
load (higher in the water column). For low sediment concentrations, silt and clay are
only transported in suspension. For high concentrations, silt and clay can also be
transported as fluid mud in a layer near the bed. In that case, the fluid mud will act as
a viscous layer on top of the bed, blocking the exchange of sand between the bed and
the water column.
We have seen that the fall velocity is an important parameter determining the ver-
tical distribution of suspended sediment (see for instance Eq. 6.40). The fall velocity
depends on the size, shape and specific density of the particles and of the water tem-
perature and the sediment concentration. As a guideline, an individual silt particle of
10 µm has a fall velocity in still water of about 0.1 mm/s. For sand, the particle prop-
erties are relatively independent of time (because the quartz crystals are chemically
hard and stable, it takes years to change the size of a sand grain). But, as we have seen
in Sect. 2.6.2, the size and shape of mud can change quite easily. As a result of floc-
culation, the size of individual mud particles increases. With this increase, particles
become heavier, which will speed up the settling of the particles. On the other hand,
due to the increase in size, the drag force – the resistance the particle ‘feels’ when
settling – will also increase. This will slow down the particle and can eventually cause
deflocculation.
Another factor influencing the sediment transport is the exchange between the bottom
layer and the water column. The upward sediment flux from the bed into the water
column strongly depends on the bed composition. From laboratory experiments it was
concluded that two regimes can be distinguished: a non-cohesive and a cohesive re-
gime (Van Ledden & Wang, 2001). The clay content (the weight percentage of particles
with a grain size smaller than 4 µm) of the bed material is the governing parameter for
the transition between both regimes. For clay content of less than 5 % to 10 % the bed
behaves more or less non-cohesively. Sand and mud are eroded more or less independ-
ently, so the ‘normal’ formulations for the bed erosion can be used. For clay content
exceeding 5 % to 10 % the bed behaves cohesively. The sand and mud particles are
eroded simultaneously. Figure 6.17 gives the critical velocities as a function of mud
(clay and silt) content in the bottom material.
In sediment transport modelling, the exchange processes between water column and
bottom layer are often assumed to be governed by the critical bed shear stress of the
bottom material and the actual occurring bed shear stresses (Sect. 6.3).

3.5
Kaolinite
3.0 Montmorillonite
critical erosion shear stress [N/m2]

2.5
2.0
1.5
1.0
0.5
0
0 5 10 15 20 25 30
mud content [%]
Figure 6.17: Effect of mud content on the critical erosion shear stress for two different sand-mud
mixtures (data from Torfs, 1995). Mud content on the horizontal axis represents the percentage
of fines smaller than 63 µm, hence including silt and clay fractions.
For actual bed shear stresses larger than the critical bed shear stress of the bed material,
the grains can be eroded, whereas for actual bed shear stresses smaller than the critical
bed shear stress of the bed material, the grains can settle:
𝜏𝑏,actual > 𝜏𝑏,𝑐𝑟 → erosion (6.55a)
𝜏𝑏,actual < 𝜏𝑏,𝑐𝑟 → sedimentation (6.55b)
For non-cohesive material, the critical bed shear stress 𝜏𝑏,𝑐𝑟 for erosion is equal to the
critical bed shear stress for sedimentation. For cohesive material, where there is a
strong binding force between the particles, this is not the case. The critical bed shear
stress for erosion is larger than the critical bed shear stress for sedimentation. This
implies that for a range of actual bed shear stresses, there is no exchange with the
bottom layer in the case of cohesive sediments:
𝜏𝑏,actual > 𝜏𝑏,𝑐𝑟,𝑒𝑟 → erosion (6.56a)
𝜏𝑏,𝑐𝑟,𝑠𝑒𝑑 < 𝜏𝑏,actual < 𝜏𝑏,𝑐𝑟,𝑒𝑟 → no exchange with bottom layer (6.56b)
𝜏𝑏,actual < 𝜏𝑏,𝑐𝑟,𝑠𝑒𝑑 → sedimentation (6.56c)
The exact formulations for sediment transport of mixtures of sand and mud are still a
subject of research.

6.8.3. Environmental issues

Due to the chemical constitution of mud and especially clay, the particles are easily
bound together, as we have seen in Sect. 2.6.2. Organic matter is also easily bound to
mud particles. This is one of the main reasons why delta areas are so fertile and thus
one of the main reasons why people settle in these low-lying, rather dangerous areas.
The fertile coastal areas are also good habitats for flora and fauna.
Unfortunately, contaminants are also easily bound by mud particles. At certain low
concentrations, naturally occurring elements (i.e. copper and zinc) may have benefi-
cial effects. But with increasing concentration, contaminants can become toxic and
cause damage to the natural environment. An increase in concentration can occur
due to fluvial inputs in the coastal system, but also as a result of a disturbance of a
contaminated bed caused by, for example, dredging.
During dredging operations, extremely dangerous contaminants such as heavy metals
and PCBs may be part of the dredged material. Highly contaminated mud has to
be treated as chemical waste. This can have serious consequences for the costs of
a dredging project in a harbour. The dredged material must be transported to special
depots, for example the ‘Slufter’ near the Port of Rotterdam (see Fig. 6.18) and isolated
from the environment.
Figure 6.18: The Slufter after the Maasvlakte Extension 1 in 1972 (Rotterdam, the Netherlands).
From Rijkswaterstaat (‘Credits’ on page 575).
For the construction of large land reclamation projects, huge volumes of sand are re-
quired. These volumes are often dredged at the sea bottom as close as possible to the
construction site, in a very vulnerable area, close to the coast. Let us assume that for a
very large project 500 million m3 of sand is required. The sand can be dredged in the
open sea, but the bottom contains 2 % silt (a quite normal content for natural seabeds).
During the dredging operation with hopper dredges, the bed material is pumped into

the hopper. The sand remains in the hopper; the fines (silt) return to the sea with
the overflow. Dredging 500 million m3 for the land reclamation means that about 10
million m3 of silt is mobilised in a relatively short period of time. This might result
in a heavy additional silt burden for the receiving sea system. Detrimental effects are
likely.
6.9. Discussion
6.9.1. Choice of models
The quantitative description of sediment transport is still largely empirical (e.g. Fred-
søe and Deigaard (1992) and Van Rijn (1989)). There is a variety of transport formulas
and models, each with its own strengths and weaknesses. Note that all the models
discussed in this chapter are geared towards describing sub-aqueous (‘underwater’)
transport. Aeolian transport (by wind) and transport in the swash zone around the
water line are not considered. They can, however, be important for the sediment bal-
ance.
Results of sediment transport computations often show large discrepancies compared
with actual measurements. A factor 2 to 5 too large or too small is certainly not excep-
tional. Many researchers are trying to improve the results, for instance by proposing
better and more reliable descriptions of the various elements in a formula. For practical
problems in coastal engineering, however, the simpler models – e.g. Bijker (Sect. 6.5.4)
and Bailard (Sect. 6.7.2) – are still the ones to beat.
A. G. Davies et al. (2002) present a very interesting intercomparison study of various
research and practical sand transport models. They summarise their study as follows:
“A series of model intercomparisons, and model comparisons with field

data, was carried out as part of the EU MASTIII SEDMOC Project (1998–
2001). Initially, seven ‘research’ models were intercompared over a wide
range of wave and current conditions, corresponding to both plane and
rippled sandbeds. These models included both one-dimensional vertical
(1DV) formulations, varying in complexity from eddy viscosity and mix-
ing length models to a full two-phase flow formulation, and two-dimen-
sional vertical (2DV) formulations capable of representing vortex shed-
ding above sand ripples. The model results showed greatest convergence
for cases involving plane beds, with predicted sand transport rates agree-
ing to well within an order of magnitude, and greatest divergence for cases
involving rippled beds. A similar intercomparison involving (mainly) prac-
tical sand transport models, carried out over wide wave and current para-
meter ranges, also showed greatest variability for cases involving rippled
beds. Finally, (mainly) practical models were compared with field data

6.9. Discussion 319
obtained at five contrasting field sites. The results showed that suspen-
ded sand concentrations in the bottom metre of the flow were predicted
within a factor of 2 of the measured values in 13 % to 48 % of the cases con-
sidered, and within a factor of 10 in 70 % to 83 % of the cases, depending
upon the model used. Estimates of the measured alongshore component
of suspended sand transport yielded agreement to within a factor of 2 in
22 % to 66 % of cases, and within a factor of 10 in 77 % to 100 % of cases. The
results suggest that, at the present stage of research, considerable uncertainty
should be expected if untuned models are used to make absolute predictions
for field conditions. The availability of some measurements on site still ap-
pears to be a necessary requirement for high-accuracy sand transport pre-
dictions. However, for morphological modellers, the results may be viewed
as more encouraging, since many of the present models exhibit agreement
in their relative behaviour over wide ranges of wave and current conditions,
which is a prerequisite to obtaining correct morphodynamic predictions.”
6.9.2. Specific situations

In this chapter we have discussed the topic of sediment transport in quite general terms.
In the next three chapters, sediment transport is considered for specific situations, as
is the resulting morphodynamic response of the system.
Chapters 7 and 8 are devoted to wave-dominated coastal systems: uninterrupted coastal
stretches, pocket beaches and wave-dominated deltaic coastlines. We make a distinc-
tion between cross-shore (Ch. 7) and longshore (Ch. 8) sediment transport (Fig. 6.19).
cross-shore
transport
longshore
current longshore
transport
Figure 6.19: Distinction between longshore transport parallel to the shoreline and cross-shore
transport transverse to the shoreline.
Such a distinction may seem artificial at first glance, but is often made because of the
following differences:
• Long-term changes in the coastline are often the result of gradients in longshore
transport. They are easily observed near disturbances, such as breakwaters or
river mouths. Cross-shore transport is responsible for short-term variations,

such as changes in the position and size of breaker bars and dune erosion during
storms. Long-term changes due to cross-shore transport (for instance a long-
term loss to deeper water) may also occur, but are harder to detect (and may be
difficult to distinguish from long-term changes due to alongshore processes);
• Where coastline change as a result of human-induced changes is concerned,
alongshore and cross-shore effects are either of equal importance (on low wave
energy coasts, e.g. the Mediterranean) or alongshore effects dominate (on high-
wave energy coasts, e.g. the Dutch North Sea coast). This may explain why
cross-shore impacting structures – such as offshore breakwaters and perched
beaches (see Ch. 10) – perform better on low-energy coasts than on high-energy
coasts;
• The wave orbital motion is very important in transporting material in the cross-
shore direction, but not in the alongshore direction. The wave orbital motion
is approximately cross-shore directed in the nearshore. Since every wave in
principle moves sand back and forth, gross cross-shore transports (per m width)
are large and much larger than longshore transport rates (per m width). For
wave-dominated coasts not influenced or interrupted by coastal inlets, wave-
induced surf zone longshore flow is the main driving agent. The direction and
magnitude of the wave-induced longshore transport is determined by the wave
conditions (wave height, period and direction). Net cross-shore transport rates,
however, are generally an order of magnitude smaller than longshore transport
rates.
In Ch. 9, the specifics of tide-dominated systems (tidal basins) and tide- and wave-
dominated coastal inlets are considered. The sediment exchange between coast and
basins is also discussed. Dynamics of fine sediment transport are very important in
tidal basins; coarser sediment (quartz and carbonate sands) is predominantly found in
seaside regions, while the finer sediment (silt and clay) settles in the more protected
landward regions.

321
7
Cross-shore transport and profile
development
7.1. Introduction
In this chapter we consider the hydrodynamic, sediment transport and morphody-
namic processes in the cross-shore direction that determine the coastal profile shape.
In the first instance, we will disregard any structural losses both due to longer-term
cross-shore effects like sea level rise, shoreface feeding and aeolian losses (discussed
later in this chapter), and due to longshore sediment transport gradients (discussed in
Ch. 8) that lead to structural changes in the mean shoreline position.
Let us first define what the coastal profile is. On a long timescale, say decades to mil-
lennia, the coastal profile or shoreface extends from the shelf to the sub-aerial (‘under
the air’, i.e. exposed to the air) beach and dune system (see Figs. 7.1 and 7.2). In this
zone we observe generally parallel or nearly parallel depth contours, while the profile
slopes are generally steeper than 1 in 1000. We conjecture that this depth contour con-
figuration is mainly caused by cross-shore processes and that these processes are dom-
inated by wave action. On the shelf (see Fig. 7.1) we may observe three-dimensional
morphologies (sand banks, sand ridges and shoreface-connected sand ridges) that are
commonly a result of tidal action and generally have only secondary impacts on the
coastal profile. Shelf slopes are generally smaller than 1 in 1000.
The response of the coastal profile to wave action is extremely depth-dependent, i.e.
the shallower the depth, the faster the response. We therefore introduce a rough zon-
ing, viz. the lower shoreface, the upper shoreface (Fig. 1.14) and the backshore (the
coastal plain in Figs. 1.15 and 7.4). As stated above, at this moment we ignore longer-
term cross-shore and alongshore effects and concentrate on shorter-term cross-shore
effects. We assume that the backshore is simply a dune or cliff (as in Fig. 7.2), high
enough to prevent overwash and without alongshore gradients. In this particular case

322 7. Cross-shore transport and profile development
the lower shoreface (large depths) responds slowly to wave action, while the upper
shoreface and the dune or cliff backshore (shallow depths) respond fast.
Figure 7.1: The shoreface (horizontal lines) and the shelf (white) in the southern part of the North
Sea with linear sand banks or ridges on the shelf and along the central Dutch coast or Holland
coast connected to the shoreface (dashed box) (adapted from Van de Meene & Van Rijn, 2000).
In front of the Southwest Delta and Wadden coasts, the bed slope flattens from ∼ 1 ∶ 100 to
∼ 1 ∶ 1000 at water depths of about 20 m, whereas in front of the Holland coast the transition
occurs at a water depth of about 16 m (see also Intermezzo 7.1).
We define the upper shoreface as consisting of the surf zone, the beach and the first
dune row or cliff face. This zone responds nearly instantaneously to wave action,
which we may notice on the shoreline if disturbed by human interference, such as a
sand castle. Also in the surf zone we observe that surf zone bars – if present – respond
on the timescale of events, i.e. storm events may move surf zone bars offshore, while
more moderate wave action may move them onshore (Sect. 7.5). Under more extreme
conditions, when the water level rises due to a storm surge, the upper beach or even
the dune or cliff face will respond. The upper beach may develop a scarp (a nearly

7.1. Introduction 323
20
Ameland
Egmond
15
Noordwijk
elevation [m +NAP]
10
-5
-10
-15
-500 0 500 1000 1500 2000 2500
cross-shore distance w.r.t. dune foot [m]
Figure 7.2: Three instantaneous cross-shore profiles along the Dutch coast at the three locations
indicated in the map of Fig. 7.3. The cross-shore distance is relative to the dune foot, here
defined as MSL +3 m. Note the flat coastal profile of Ameland as compared to the steeper
Holland or central Dutch coast (i.e. Egmond and Noordwijk). (Data from JARKUS, n.d.)
Ameland
3001440
Egmond
7004125
Noordwijk
8007900
Figure 7.3: The location of the three instantaneous cross-shore profiles along the Dutch coast
shown in Fig. 7.2. The JARKUS transect numbers refer to the number of the coastal section
or, in Dutch, kustvak (first digit), and the last digits refer to the alongshore distance with regard
to a reference transect which differs for each coastal section. For example, the transect near
Egmond is in coastal section 7, 41.25 km south of Den Helder (which is the first transect in
coastal section 7).
vertical slope along the beach as a result of erosion) under moderate surge, while the
dune or cliff face under a high surge level reaching the dune or cliff face may undergo
surge erosion. We will treat these extreme events at a later stage.
On longer timescales, say many decades to millennia, the whole shoreface profile is
morphodynamically active (cf. (Stive & De Vriend, 1995)). However, on shorter times-
cales, say hours to a few decades (engineering scales), the lower shoreface shows neg-
ligible activity compared to the upper shoreface. In the field of coastal engineering it
is therefore often assumed that the morphologically active zone extends from the first

Figure 7.4: The shoreface is the zone between the shore and the continental shelf. On larger
timescales the entire shoreface is influenced by wave action and morphodynamically active.
The shelf extends to the shelf break.
dune or cliff face to just a little offshore the surf zone at the upper part of the lower
shoreface at a depth (relative to MSL) of about twice the wave height extreme, which
is exceeded, for example, for twelve hours per year (see also Sect. 7.2.3). A second
coastal engineering assumption is that – although the active zone may display pro-
file variability in response to instantaneous, episodic or seasonal forcing – the active
profile shape remains at a dynamic equilibrium when averaged over time (say years
to decades) and alongshore space (say 100 m to 1000 m). Unless alongshore sediment
transport gradients exist, it is a third common coastal engineering assumption that the
amount of sediment in the active zone remains unchanged, which implies that there
is no structural loss. The above three assumptions form the basis for the modelling of
the year- to decade-averaged shoreline changes treated in Ch. 8.
However, there exist a number of arguments to also pay attention to the morphody-
namic behaviour of the upper shoreface on timescales shorter than a year. Amongst
these arguments is the necessity to gain insight into the profile and plan form variation
of the upper shoreface for infrastructure design and beach use purposes.
With the above in mind, we will therefore treat the following topics. In Sect. 7.2 we dis-
cuss the concept of a dynamic equilibrium shoreface profile, which ignores profile and
plan form dynamics on scales shorter than years. This concept is instrumental in the
modelling of year- to decade-averaged shoreline changes and in the design of upper
shoreface nourishments. Subsequently, in Sect. 7.3, we treat the dynamic variations

7.2. Equilibrium shoreface profile 325
of the upper shoreface profile and plan form on timescales shorter than a year, includ-
ing surge-induced dune erosion. This gives insight into the upper shoreface profile
variations on shorter timescales, which is relevant for the safety of nearshore infra-
structure and for recreational beach use. Section 7.4 then considers structural changes
to the position of the upper shoreface profile on longer – say decades to millennia
– timescales, when the assumption that the amount of sediment in the active zone
remains unchanged is violated. Finally, in Sect. 7.5 we discuss cross-shore sediment
transport processes that are relevant for both the dynamic equilibrium profile and the
inter-annual variations therein.
Intermezzo 7.1 The Dutch shoreface
The shoreface is the zone between the shore and the continental shelf. It is morpho-
dynamically active on decadal to millennial timescales. On annual to decadal
timescales, morphodynamic activity involves mainly the upper shoreface. The up-
per shoreface is the zone with regular and dominant wave action, whereas on the
lower shoreface wave action only occurs during larger storm events. The Dutch
upper shoreface is the beach and surf zone with breaking waves and breaker bars
between the waterline and approximately the NAP −8 m depth contour, with mean
bed slopes varying between 1:50 to 1:200. The Dutch lower shoreface is the zone
between approximately the NAP −8 m and NAP −20 m depth contours with typical
bed slopes between 1:200 and 1:1000, where sand ridges may be present. Offshore
the shoreface merges with the continental shelf where the slope is generally less
than 1:1000; tidal sand waves and sand banks may be present here.
7.2. Equilibrium shoreface profile

7.2.1. The concept
Consider a coastal stretch that has no alongshore sediment transport gradients. If
we adopt the coastal engineering assumption that the amount of sediment in the up-
per shoreface remains constant, the shoreline position will remain the same when
averaged over several years. Figure 7.5 shows the Ameland, Egmond and Noordwijk
shoreface profiles (Fig. 7.3) from 1965–2019 (from the JARKUS (n.d.) database with
yearly measurements along the Dutch coast). The cross-shore distance is relative to
the relevant local beach pole or Rijksstrandpaal (RSP) in Dutch. In order to cancel any
structural gains or losses, the profiles have been shifted to the same duneface location
(Fig. 7.6).
We observe that the various profiles show a dynamic variation that is clearly related
to the position of the bar(s). We will discuss bars further on. We also observe that the
profile variations remain in an envelope that seems stable over the years. One may

15
Ameland
10
elevation [m +NAP] 5
MHW
0
MLW
-5
-10
20 Egmond
15
elevation [m +NAP]
10
0
MHW
MLW
-5
-10
-15
20
Noordwijk
15
elevation [m +NAP]
10
0
MHW
MLW
-5
-10
-15
-500 0 500 1000 1500 2000 2500
Figure 7.5: Upper shoreface profile variation for Ameland (upper panel), Egmond (middle panel)
and Noordwijk (lower panel) over the period 1965–2019. The cross-shore distance is relative to
the relevant local beach poles or Rijksstrandpaal (RSP). The average HW and LW lines are also
indicated. Note the significantly stronger variability of Ameland’s dune and nearshore profile as
compared to the other two locations. (Data from JARKUS, n.d.).
therefore speak of the existence of a dynamic equilibrium profile (e.g. as follows from
a best fit through the profile data).
It is commonly assumed that if one exposes, e.g. in the laboratory, a beach profile to
constant wave forcing at a fixed water level, a stable equilibrium profile (cross-shore
profile of constant shape) will be reached after a sufficiently long time. Is there a
relation between this stable equilibrium profile and the dynamic equilibrium profile
as found in the field? The answer is partly yes and partly no. We may expect that a
shore profile in the field will respond to a constant forcing condition and that it will also

15
Ameland
10
elevation [m +NAP] 5
MHW
0
MLW
-5
-10
20 Egmond
15
elevation [m +NAP]
10
0 MHW
MLW
-5
-10
-15
20
Noordwijk
15
elevation [m +NAP]
10
0 MHW
MLW
-5
-10
-15
-500 0 500 1000 1500 2000 2500
cross-shore distance w.r.t. dune foot [m]
Figure 7.6: Upper shoreface profile variation for Ameland (upper panel), Egmond (middle panel)
and Noordwijk (lower panel) over the period 1965–2019. The profiles have been shifted to the
same duneface location (MSL +3 m), so any structural gains or losses are cancelled out. The
average HW and LW lines are also indicated. (Data from JARKUS, n.d.).
search for an equilibrium form to that forcing. However, in nature the forcing (waves
and water levels) is far from constant and varies so rapidly that a stable equilibrium is
never reached. It is for this reason that the shoreface profile continuously oscillates in
response to the varying forcing. However, interestingly, these oscillations are confined
to a steady envelope, the mean position of which we define as a dynamic equilibrium
profile.
In the following sections we will first describe empirical and semi-empirical approaches
to generalise the best profile fit for arbitrary upper shoreface profiles (Sect. 7.2.2).

Secondly, in Sect. 7.2.3, we will discuss the engineering applications of the dynamic
equilibrium concept.
7.2.2. (Semi-)empirical derivations

Bruun
Bruun (1954) was one of the first coastal engineers to introduce the existence of a
dynamic equilibrium profile. He proposed an empirical equation for the equilibrium
beach profile based on an analysis of beach profiles along the Danish North Sea coast
and the California coast. The equation consists of a simple power law relating the
water depth ℎ to the offshore distance 𝑥 ′ (where 𝑥 ′ = 0 is around the mean waterline
and is positive in the offshore direction), as follows:
ℎ = 𝐴(𝑥 ′ )𝑚 (7.1)
where:
𝑚 exponent equal to 𝑚 = 2/3 −

ℎ water depth m
Interestingly, the value of 𝑚 = 2/3 corresponds to the theoretical derivation of Bowen

(1980), see Eq. 7.24. The dimensional constant 𝐴 (the dimension of which depends on
the magnitude of the exponent 𝑚; with 𝑚 = 2/3 the dimension of 𝐴 is 𝑚1/3 ) is a so-called
shape factor that depends on the “stability characteristics of the bed material” (what
this actually means is somewhat unclear). Bruun found that 𝐴 = 0.135𝑚1/3 provided
the best correlation for North Sea beaches in the Thyborøn area in Denmark. Hughes
and Chiu (1978) show that 𝐴 = 0.10𝑚1/3 provides the best correlation for beaches
along the coast of Florida. The larger 𝐴 is, the steeper the profile. So, apparently
North Sea beaches are somewhat steeper than Florida beaches. Whether this can be
explained physically will be discussed next.
Dean
The Bruun equation – a simple power law – was supported by Dean (1977) on semi-
empirical grounds by reasoning that for a certain grain size, nature strives towards a
uniform energy dissipation 𝜀 (𝐷50 ) = 𝐷/ℎ (loss in wave power) per unit volume of
water across the surf zone (in W/m3 ). The reasoning behind this is that the energy
dissipation per unit volume is a measure for the “destructive forces” (causing offshore

sediment transport) acting on a sediment particle. With Dean’s equilibrium assump-

tion and for waves normally incident to an alongshore uniform coast 𝜃 = 𝜑 = 0, the
energy balance Eq. 5.2 can be written as:
𝑑 (𝐸𝑐𝑔 )
= −𝐷 = −ℎ𝜀 (𝐷50 ) (7.2)
𝑑𝑥
where:
𝐸 wave energy per unit sea area J/m2

𝑐𝑔 group velocity m/s
We again use the simple dissipation model for the surf zone 𝐻 = 𝛾 ℎ (see Ch. 5). If we
further assume shallow water (𝑐𝑔 = 𝑐 = √𝑔ℎ), Eq. 7.2 leads to:
𝑑 1
( /8𝜌𝑔𝛾 2 ℎ2 √𝑔ℎ) = −ℎ𝜀 (𝐷50 ) ⇒
𝑑𝑥
1 𝑑 1
( /8𝜌𝑔𝛾 2 ℎ /2 √𝑔) =
5
𝜀 (𝐷50 ) = − (7.3)
ℎ 𝑑𝑥
5 3/2 2 𝑑ℎ
= 𝜌𝑔 𝛾 √ℎ ′
16 𝑑𝑥
Note that in the last step of this equation we have reverted to the cross-shore coordin-
ate 𝑥 ′ which is positive offshore. The depth is the only parameter that varies with 𝑥 ′ .
If we integrate Eq. 7.3 we get:
2/3
24𝜀 (𝐷50 ) 2/3
ℎ (𝑥 ′ ) =( ) (𝑥 ′ ) (7.4)
5𝜌𝑔 𝛾 2
3/2
Apparently, based on monochromatic waves and a constant breaker index across the
surf zone 𝛾 = 𝐻 /ℎ, the magnitude of the exponent 𝑚 in Eq. 7.1 can be derived and is
found to be 2⁄3, in agreement with Bruun’s suggestion. Furthermore, the dimensional
shape factor 𝐴 (in m1/3 ) is given by:
2/3
24𝜀 (𝐷50 )
𝐴=( ) (7.5)
5𝜌𝑔 /2 𝛾 2
3
in which the equilibrium energy dissipation rate 𝜀 (𝐷50 ), in W/m3 , depends on the
particle diameter. Hence, in Dean’s approach the shape factor 𝐴 is some function of
the particle diameter. Assuming that the breaker index does not vary with the wave
conditions, 𝐴 is independent of the deep-water wave conditions. The magnitude of 𝐴
is seen to vary from 0.079 m1/3 to 0.398 m1/3 (Dean, 1977).

The shape parameter 𝐴 was empirically related to the median grain size by Moore
(1982), showing that a coarser grain size implies a larger value of 𝐴, and thus a steeper
cross-shore profile. Dean (1987) showed that this relation could be transformed to a
relation using the fall velocity 𝑤𝑠 (in cm/s) as a parameter, viz.:
𝐴 = 0.067𝑤𝑠0.44 (7.6)
Or for 𝑤𝑠 in m/s:
𝐴 = 0.5𝑤𝑠0.44 (7.7)
If this result is accurate and universally valid, we have a powerful tool. Once we know
the fall velocity of the sediment (which depends largely on the grain size, Sect. 6.2.3),
we can predict the equilibrium profile after for instance nourishment with borrow
material of a different grain size than the native material (see also Sect. 10.7.2). Qualit-
atively, this result is in line with the finding that coarser beaches are generally steeper
than finer beaches.
7.2.3. Engineering applications

Before discussing the importance of the dynamic equilibrium concept for engineering
applications, we need to know how accurate the predictive ability is. We therefore
apply a parabolic fit to the three typical profiles along the Dutch coast described above
(Fig. 7.5), knowing that these profiles differ considerably (see Figs. 7.7 and 7.8). Two
profiles are from the Holland coast, one in the southern section and one in the northern
section and one from the Wadden Sea coast (in all cases 𝐷50 = approximately 200 µm).
The fits were made to match the average profile over the 55 yearly measurements.
𝑚
Note that the formula ℎ = 𝐴 (𝑥) results in a vertical slope of the beach profile at
the waterline (𝑥 ′ = 0). This is not realistic. In practical applications it should be
considered to position this point at a level somewhere above the waterline (1 m in our
example). This results in more reasonable slopes at the waterline. Furthermore, we
need to define the offshore extent of the profile fit, for which one may adopt the active
zone concept as defined by Hallermeier (1978), who introduced the concept of depth
of closure (Sect. 1.5.1). Hallermeier’s inner closure depth (see Fig. 1.14) corresponds to
the surf zone width for extreme conditions which are exceeded for twelve hours per
year. It marks the seaward limit of the upper shoreface, where the majority of the bed
dynamics takes place. Since ℎ𝑏 = 𝐻𝑏 /𝛾 , the depth of closure is more than twice the
wave height extreme.
Even though the mean grain diameter is the same for all three locations, the optimal
values of 𝐴 with fixed power 𝑚 = 2/3 (see Fig. 7.7 and Table 7.1) differ by a factor of
almost 2! According to Eq. 7.7, the value of 𝐴 should be 0.10 (based on a fall velocity
𝑤𝑠 = 0.0256 m/s). These results indicate that the predictive ability of Eq. 7.1 is limited.

0
Ameland average
A = 0.055 parabolic fit
elevation [m +NAP]
-5
-10
-15
-20
0
Egmond average
elevation [m +NAP]
-5
-10
-15
-20
0
Noordwijk average
elevation [m +NAP]
-5
-10
-15
-20
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Figure 7.7: Parabolic fits (black lines) to the long-term-averaged (1965–2019) profiles (blue lines)
of Ameland (top panel), Egmond (middle panel) and Noordwijk (bottom panel) according to
Eq. 7.1 with 𝐴 as a free parameter and 𝑚 = 2/3. Note that each panel also shows the profiles of
the other two locations.
We hypothesise that the value of 𝐴 is not only dependent on the fall velocity, but also
on several other variables, like wave climate, tide and surge water level variations and
coastal currents. Hence, we must be hesitant to use Eq. 7.1 for engineering applications,
given its limited predictive ability. Obviously, a fit with two free parameters improves
the fit somewhat (Fig. 7.8), but the variation in the values of 𝐴 is even larger and the
theoretical foundation is lacking.
Nonetheless, the dynamic equilibrium concept is one of the few tools that a coastal
engineer has available to make a prediction of the expected profile of a new land re-
clamation or a newly constructed offshore island. A particularly useful application

0
Ameland average
m = 0.537
elevation [m +NAP]
-5
-10
-15
-20
0
Egmond average
m = 0.773
elevation [m +NAP]
-5
-10
-15
-20
0
Noordwijk average
m = 0.631
elevation [m +NAP]
-5
-10
-15
-20
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Figure 7.8: Parabolic fits (black lines) to the long-term (1965–2019) averaged profiles (blue lines)
of Ameland (top panel), Egmond (middle panel) and Noordwijk (bottom panel) according to
Eq. 7.1 with 𝐴 and 𝑚 as free parameters. Note that each panel also shows the profiles of the
other two locations.
concerns the expected changes in the dynamic equilibrium profile when a coast is
nourished with borrow sediment sizes that differ from the native sediment size. Dean
(2002) discusses this application at length.
Table 7.1: Optimal values for 𝐴 for Eq. 7.1 with fixed power 𝑚 = 2/3.
A (Eq. 7.7) A (Ameland) A (Egmond) A (Noordwijk)

0.10 0.055 0.093 0.070

7.3. Morphodynamics of the upper shoreface 333
7.3. Morphodynamics of the upper shoreface

7.3.1. Introduction
In the previous section we have focused on the dynamic equilibrium profile of the
upper shoreface, temporally averaged over years and alongshore spatially averaged
over about a kilometre. We assumed that a cross-shore profile will evolve towards a
(stable) equilibrium, corresponding to the concurrent forcing, and that disequilibrium
implies that not enough time has been available to achieve equilibrium. In such an
approach, the focus is on the order and regularity in the profile evolution.
On timescales shorter than a year and especially on the timescale of episodic events,
we may observe highly dynamic variations of the upper shoreface profile and plan
form. In this section, the emphasis is on variability and diversity instead of order and
regularity. We will first discuss the high dynamic variability of the intertidal beach and
the surf zone on the timescale of events (Sect. 7.3.2). Next, we introduce the typical
seasonal variations that are common on many beaches worldwide (Sect. 7.3.3). Cyclic
behaviour of bars on the timescale of years is discussed in Sect. 7.3.4. Last, we focus
on the impact of episodic events on a dune-backed shoreface (Sect. 7.3.5). These issues
are considered relevant for the safety of nearshore infrastructure and for recreational
beach use.
7.3.2. Beach states

Let us first investigate morphologic features that we may encounter on the upper
shoreface on the timescale of wave events (short-term variation). Remember that we
defined the upper shoreface to range from the surf zone at its furthest offshore reach
(i.e. the active depth as defined by Hallermeier) to the first dune or cliff face. Parts of
the upper shoreface are the beach (normally sub-aerial) and the intertidal zone, which
is the zone between low water and high water, at the transition between the surf zone
and the beach.
Looking at the intertidal zone, we observe strongly three-dimensional morphology,
probably the strongest in the upper shoreface. An example is shown in Fig. 7.9, where
we observe a ridge-runnel structure on the intertidal beach. Although it impacts beach
recreation and use, the importance of ridge-runnel structures in terms of shoreface
behaviour is limited. We expect that the surf zone morphological structure has a larger
impact on upper shoreface behaviour. On many beaches worldwide we may observe
one, two and sometimes more surf zone bars, while in the intertidal zone a bar may
also be present. In the examples presented on the Holland coast (a so-called dissipative
beach, see later on in this section) we observe one surf zone bar and an intertidal bar.
The bars determine the locations and rates of energy dissipation due to wave-breaking
and may thus dictate the morphological response. Further on, we will discuss the

behaviour of the surf zone bars averaged over many years, which determines their
cross-shore location (Sect. 7.3.4).
Figure 7.9: Ridge (intertidal bar) runnel (drainage channel) structure at low water at the north-
ern flank of the Sand Engine on 16 February 2016. Photo by Jurriaan Brobbel, Rijkswaterstaat
On the timescale of wave events, we may observe highly dynamic variations of the
upper shoreface profile and plan form. In the literature, the various morphodynamic
regimes (also called beach states) have been classified according to their overall ap-
pearance, which is often related to previous weather conditions on daily to monthly
timescales. Wright and Short (1984) distinguished a series of six beach states, ranked
hierarchically from the highest state, dissipative, to the lowest state, reflective. In
between the two end states, four intermediate beach states can be discerned. They
based their analysis on observed behaviour at Australian beaches. Dissipative and
reflective beaches are relatively two-dimensional, although reflective beaches usually
have pronounced arc-shaped shoreline formations that are called beach cusps. Due
to their small alongshore variability, the dissipative and reflective end states can be
characterised by their cross-shore profile. The cross-shore profiles of reflective and
dissipative beaches correspond roughly to ‘summer’ and ‘winter’ beaches, respectively
(see Fig. 1.3 and Sect. 7.3.3). They can be described as follows:
Reflective beaches are characterised by a relatively steep and narrow beachface with
a berm and a narrow surf zone without bars. Nearshore and beach slopes are
between 0.10 and 0.20. The sandy material is relatively coarse (see Sect. 7.2.2).
For these beaches a large Iribarren parameter (Eq. 5.20) can be expected (say
larger than 2). Collapsing or surging breakers are common on reflective beaches.
The corresponding waves have a low steepness (long, small-amplitude waves).
Thus, reflective beaches are the result of a period of mild wave conditions that
transport sediment onshore. Reflective beaches are often found in swell and

monsoon wave climates (see Sects. 4.3.2 and 4.3.3). Since in these climates the
conditions have a low variability, the resulting morphodynamic behaviour is
less dynamic than is the case for storm wave climates that are characterised by
a higher variability.
Dissipative beaches – at the other end of the spectrum – are characterised by a wide
and flat sandy coastal zone with one or multiple linear bars and with dunes back-
ing a wide beach. The nearshore slope is about 0.01 and the beach slope about
0.03. The sandy material is relatively fine. The corresponding Iribarren numbers
are small (around 0.2 to 0.3, corresponding to spilling breakers). A dissipative
beach is the result of high-energy waves that start breaking far offshore (in a
wide surf zone, say up to 500 m wide). These high-energy, short waves are typ-
ical for a storm wave climate and the associated variability results in a highly
dynamic coastal profile.
The four intermediate beach states are all strongly three-dimensional. Whereas rip
currents (Sect. 5.5.7) and corresponding rip channels are generally absent on pure dis-
sipative and reflective beaches, they are an important attribute of intermediate beach
morphologies. For intermediate beach states, plunging waves tend to occur.
Instead of the Iribarren number, another more or less equivalent parameter is often
used to indicate the beach state, viz. the dimensionless fall velocity:
𝐻𝑏
Ω= (7.8)
𝑤𝑠 𝑇
where 𝐻𝑏 is the wave height at breaking, 𝑇 is the wave period and 𝑤𝑠 is the fall velocity.
The reflective beaches at one end of the spectrum typically have Ω < 1, whereas for
dissipative beaches at the other end of the spectrum we have Ω > 6. For values in
between, intermediate beach states are found.
Klein et al. (2005) used the distinction between dissipative, intermediate and reflect-
ive beach states, as explained above, to classify southern Brazilian beaches (Fig. 7.10).
Their work is very illustrative in showing the dependency on grain size and beachface
slope.
The beach state at a certain location is not invariable; beaches can move through a
series of beach states. Wright and Short (1984) use the dimensionless fall velocity to
explain whether the response will be slow (low energy, reflective) or fast (high en-
ergy, dissipative) and whether erosion or accretion will occur. Mild wave conditions
slowly force a beach towards a reflective beach state through onshore sediment trans-
port, whereas storm waves are responsible for fast offshore movement of sediment,
resulting in a dissipative beach state. Hence, wave conditions which move a beach
toward a higher state (the dissipative state being the highest in rank) cause erosion;
wave conditions which move a beach downstate1 cause accretion.
1
Downstate implies consecutive beach states under gradually more moderate energy forcing.

dissipative intermediate reflective

1
0.8
7
mean diameter [mm]
8 coarse sand
2 5
6
0.6
1
9
0.4 17
medium sand
4
10 15 11
0.2 14 16
13 3 fine sand
12
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
beachface slope [degrees]
exposed beaches semi-exposed beaches sheltered beaches
1. Barra Velha 6. Taquaras 11. Piçarras 15. Armação
2. Itajuba 7. Estaleirinho 12. Balneário Camboriú 16. Laranjeiras
3. Navegantes 8. Estaleiro 13. Itapema 17. Zimbros
4. Brava 9. Ilhota 14. Bombas
5. Taquarinhas 10. Mariscal
Figure 7.10: Relationship between sedimentary grain size and slope of the beachface indicates
the importance of morphodynamic stages and energy level (data from Klein, 2003).
Based on ARGUS video imaging techniques, Lippmann and Holman (1990) and Ranas-
inghe et al. (2004) gained detailed insight into intermediate beach state dynamics at two
different single-barred beaches: Duck, North Carolina, USA, and Palm Beach, Sydney,
Australia, respectively. These studies showed that beaches continuously cycle through
the four intermediate beach states in response to changing wave conditions. The three-
dimensional structure of the morphology can be wiped out by an episodic event (a reset
event, a high-energy event that resets the three-dimensional character to an along-
shore uniform bar position typical for the highest beach state). In conclusion, we may
encounter high three-dimensional variability, which may turn into two-dimensional
variability (alongshore uniform morphology) after a reset event (Stive & Reniers, 2003),
see also Fig. 7.11). The progression from alongshore uniform morphology to the re-
flective beach takes place in weeks or months.
Depending on the beach state, rhythmicity can be observed in for instance beach
cusps and rip channels2 . Scientists have long tried to couple the spatial scales of these
rhythmic features to certain length scales in the hydrodynamic forcing or geological
constraints. However, in the last two decades or so, the focus has been more and more
on self-organisation (see e.g. Coco and Murray (2007)). Self-organisation means that
certain spatial patterns are associated with internal dynamics rather than with external
forcings. These two contrasting approaches are further explained in Intermezzo 7.2.
2
Rip channels are found in between shore-connected transverse bars or as cross-shore depressions in a
shore-parallel bar. The latter bars are also called surf zone crescentic bars systems in literature: surf
zone sand bars with alongshore undulations in height and cross-shore position of the bar crest.

1500 1500
rip channel breaker bar

l
breaker bar shoreline
1000 1000
–3
–11
–10 –9
y [m]
y [m]
–4
–3
–2
–7
500 500
–6
–4
–2
0 0
0 200 400 0 200 400
x [m] x [m]
Figure 7.11: ARGUS imagery is used to obtain photographic images of incident wave-breaking.
Waves break over shallow bars, where the foam of the breaking waves shows up as an area of
high light intensity. Left panel: Example of a time exposure of Palm Beach, Australia, and su-
perposed bottom contours displaying a complex pattern of shallow shoals (light areas corres-
ponding to intense wave-breaking) cut by deep rip channels (dark areas without wave-breaking).
The beach is located on the left. Right panel: Another example of a time exposure at the same
beach showing intense wave-breaking on a linear bar and additional breaking at the shoreline
(Argus images courtesy of Graham Symonds).
Obviously, we have discussed interesting notions about intertidal and surf zone vari-
ability due to antecedent and concurrent forcing conditions. With timescales of ante-
cedent forcing conditions of days to months, these observations are of importance
for the beach state and its suitability for recreation and beach use, but are of limited
importance on longer timescales.
Intermezzo 7.2
Spatial scales: governed by external forcing or internal dynamics?
Traditionally, it has been assumed that the spatial structures in the forcing (called
forcing templates) become imprinted in the sediment bed. Observed length scales
in the resulting morphology would then correspond directly to length scales in the
forcing. A complicating factor is that the morphodynamic response is not always
instantaneous. This implies that some time is required for the spatial scales of the

morphological patterns to evolve towards the scales of the forcing. Due to natural
variability in the forcing, not enough time may be available to get to that point.
Hence, the morphology of the system generally reflects the mix of the antecedent
and concurrent hydrodynamics (thus a history of wave height, period and direc-
tion combinations). Predictability of the morphological length scales then strongly
depends on knowledge about the entire history of the forcing and – if only the con-
current forcing is known – on the relative importance of antecedent and concur-
rent forcing. Note that due to the stochastic nature of the hydrodynamic forcing
conditions, future forcing is at best known in statistical terms!
In a self-organisation approach it is assumed that rhythmic features are initiated by
positive feedbacks between hydrodynamics and morphology (leading to divergent
behaviour) and are stabilised by negative feedbacks (leading to convergent beha-
viour), see Sect. 1.5.2. This implies that the observed length scales in the coastal
morphology do not necessarily correspond to the length scales of the initial for-
cing processes. Only after a feedback process between forcing and morphology
do certain length scales predominate. To differentiate it from forced behaviour,
self-organisation has also been termed free behaviour.
Note that although the length scales of the morphodynamic features do not match
those of the external forcing, self-organisation approaches assume the system to
respond to (changing) hydrodynamic forcing. This is in accordance with observa-
tions that when hydrodynamic conditions change, bars and other features respond.
The self-organisational processes, in which certain length scales prevail, need suf-
ficient time for the dominant length scales to appear (hence the concurrent forcing
should last sufficiently long).
Let us consider the example of rip-channel spacings in alongshore non-uniform
surf zone bars (generally in the order of 100s of metres). A once-popular explan-
ation of the fact that rip channel spacing is related to length scales in the forcing,
was forcing due to edge waves. Edge waves are waves trapped against a shoal-
ing beach. They are the wind-wave equivalent of coastally trapped tidal Kelvin
waves (Sect. 3.8.3). Just as with Kelvin waves, the amplitude of edge waves varies
sinusoidally along the shore and diminishes rapidly seawards from the shoreline.
Whereas Coriolis is the mechanism behind the Kelvin waves, the trapped-wave
phenomenon of edge waves has been attributed to wave refraction near the shore
as a result of the variable water depth. The standing edge-wave patterns are
presumed to become imprinted on the underwater sediment bed resulting in rip-
channel systems. Although it has been argued that rip-channel spacings are correl-
ated to length scales of edge waves (e.g. Bowen and Inman (1971) and Holman and
Bowen (1982)), observed patterns could not be correlated with concurrent or ante-
cedent (one week) hydrodynamic forcing (Holman et al., 2006; Turner et al., 2007).
Later (modelling) work could not verify these relationships either (e.g. Caballeria

et al., 2002; Reniers et al., 2004). Wave groups are another possible forcing mech-
anism for rip currents (e.g. Reniers et al. (2004)). This is discussed further in the
TU Delft course Coastal Dynamics II (CIE4309).
According to the self-organisation hypothesis, rip channels are thought to develop
as a response to small, initial disturbances of the seabed that grow due to pos-
itive feedback. Where the water depth is locally smaller due to seabed disturb-
ance, wave-breaking is enhanced, as is the resulting wave set-up and onshore
flow, leading to a 2D circulation pattern (cf. Sect. 5.5.7). Furthermore, since sed-
iment concentrations decrease from the breaking point to the shoreline, at loca-
tions where the water depth is smaller than average, the onshore sediment trans-
port decreases shoreward. This transport convergence leads to deposition that
further decreases the water depth. Where the water depth is locally larger, the
flow is offshore-directed and the offshore sediment transport increases in the sea-
ward direction (hence transport divergence). The resulting erosion leads to the
progressive carving out of rip channels.
The testing of self-organisation hypotheses has been undertaken using process-
based models as well as more abstract models. Such self-organisation approaches
have suggested that rip channel spacings increase with an increase in hydrodynamic
energy and depend on the antecedent morphology as well (for a review see Coco
and Murray (2007) and Smit (2010)). Field observations, however, do not always
convincingly support these relationships (e.g. Ranasinghe et al. (1999) and Turner
et al. (2007)).
What are the implications of self-organisation for modelling efforts? As stated
earlier, in the self-organisation hypothesis, small disturbances may grow to finite-
amplitude shapes as a result of a feedback process. Theoretically, a model in which
the governing internal dynamics are correctly represented can therefore be expec-
ted to properly represent the scale of the evolving morphodynamic features under
known forcing. However, in practice, numerical models have only shown very
limited success. Also, the exact positioning of the features (e.g. the rips) is very
sensitive to small variations in the initial conditionsa . Since the ‘true’ initial con-
ditions (the positioning of the disturbances in nature) are not known precisely, it
is not possible to predict the exact positioning.
a
This sensitivity to initial conditions is typical for chaotic dynamics and is often popularly referred
to as the butterfly effect.
7.3.3. Storm and seasonal changes

In the previous section we treated the beach state, which is a result of everyday mor-
phological response to the antecedent and concurrent forcing. Here, we look at the

response of the upper shoreface profile due to storm events with low surge heights and
– largely in line with this – to the summer-winter response of the upper shoreface.
The cross-shore profile responds differently to summer and winter conditions, leading
to a seasonal behaviour and profile characterisation as shown in Fig. 1.3. In response
to the milder summer wave conditions, the offshore bars of the ‘winter’ profile move
onshore and finally attach to the shore and rebuild the wider berm associated with
the ‘summer’ profile. Hence, in summer the beach is ‘rich’ in sediment (a high beach
profile; slope high) and in winter it is ‘poor’ in sediment (a low beach profile; slope
gentle). In the classification of Wright and Short (1994), typical summer profiles re-
semble reflective beaches and winter profiles dissipative beaches (Sect. 7.3.2).
The summer-winter seasonality is strongest for the Northern Hemisphere storm wave
climate, which exhibits a large seasonality (Sect. 4.3). Note that situations also exist in
which the seasonal behaviour is reversed. For instance, the beaches in the Rhone delta
on the Mediterranean coast show the reverse behaviour due to the summer mistral.
The summer-winter behaviour can also occur when a relatively calm period is inter-
rupted by a storm event (Fig. 7.12). This behaviour was recorded by List and Farris
(1999) along a stretch of the eastern USA coast (see Fig. 7.13).
20 20
15 15
elevation [m +NAP]
10 10
5 5 SSL
0
MHW 0 MSL
-5 -5
0 200 400 600 0 200 400 600
cross-shore distance w.r.t. RSP [m] cross-shore distance w.r.t. RSP [m]
Figure 7.12: Left: profile during a calm period with normal water level and wave conditions
(‘summer’ profile, see Fig. 1.13 for details). Right: profile as a result of storm wave height and
surge conditions (‘winter profile’). The post-storm profile is computed according to Vellinga, see
Fig. 7.18 for further info on the Vellinga profile and the storm conditions considered here.
Figure 7.13 shows that after a calm period a ‘summer’ beach is present which turns into
a ‘winter’ beach during a storm (note the relatively low storm surge heights). After
the storm the beach nearly completely returns to its summer state: the beach breathes,
so to speak. Interestingly, there is a large variation in the degree of breathing. The
average recession/accretion is 10 m but some locations show much less variation, while
others reach 15 m to 20 m. The latter locations are also known as ‘hotspots’, locations
with more erosion than their surroundings. The reasons for this are not well-known,
but it is likely that a variable offshore bathymetry may cause wave energy focusing
(see Sect. 5.2.3 and specifically Fig. 5.6) and that the position and height of dissipative
bars play a role.

20 Oct 20–25
shoreline change [m]
accretion
10
-10
erosion
-20
Oct 13–20 FRF
-20 -10 0 10 20 30 40
distance from FRF [km]
4
Hs [m]
0
Oct 13–20 Oct 20–25
depth [m]
9
8
7
1 5 10 15 20 25 30
day in October 1997
Figure 7.13: Upper figure: shoreline change between October 13 and 20 and between October 20
and 25 along a stretch of coast that is approximately 80 km long. FRF is the location of the Field
Research Facility of the Army Corps of Engineers. Lower figures: the significant wave height at
8 m depth and tidal variation in time. The three vertical lines in the lower figures indicate October
13, 20 and 25 respectively. Note that surge is included in the tidal level variation. (Adapted from
List et al., 2005; List & Farris, 1999).
7.3.4. Bar cycles over years

In Sect. 7.3.2 on beach states, we have discussed the three-dimensional behaviour of
bars under antecedent and concurrent wave forcing conditions. This behaviour is
either forced or self-organisational behaviour of the bar around its mean position of
that particular instant. However, that mean position varies over time, as we will dis-
cuss here.
If we study the behaviour of bars on longer timescales, i.e. years, we observe that bars
generally and on average move offshore under more energetic conditions and may
move a little back onshore during less energetic but skewed waves. The net offshore
movement generally shows cyclic behaviour; the initial bar formation is in the inter-
tidal zone, after which the bars move offshore and grow in size until a maximum is
reached somewhere around the initiation of the surf zone, after which they gradually
decrease in size and amplitude and finally disappear at the end of the active shoreface
profile. Along the Holland coast, this cycle varies between 4 to 5 years at the South
Holland coast and around 15 years at the North Holland coast. The cyclic behaviour

is indicated in Figs. 7.14 and 7.15 for the North Holland and South Holland coast, re-
spectively.
Figure 7.14: 3D image of the coast of North Holland from Den Helder (transect 7000150) to
IJmuiden (transect 7005450) in 1990 and 2019, as well as the time evolution in two transects
indicated by the transparent panels in the 3D images. (Data from JARKUS, n.d.).After completion
of the sandy reinforcement of the Hondsbossche and Pettemer sea defence in 2018, it was
renamed ‘Hondsbossche Dunes’ (‘Hondsbossche Duinen’ in Dutch).
The difference in cycle time between North Holland and South Holland is most prob-
ably explained by the fact that the shoreface slope is steeper and the bars are larger on
the NH coast (see Fig. 7.5), so that more energy is needed to induce sufficient sediment

Figure 7.15: 3D image of the coast of South Holland from IJmuiden (transect 7005600) to Hoek
van Holland (transect 9011900) in 1990 and 2019, as well as the time evolution in two transects
indicated by the transparent panels in the 3D images. (Data from JARKUS, n.d.).
transport. The importance of this behaviour is most probably that the bar’s position
will influence the energy dissipation that energetic waves will undergo. This may lead
to regional differences in shoreline response (see List and Farris results in Sect. 7.3.3).
The North Holland transect 7001827 between Callantsoog and Petten demonstrates
an undisturbed, stationary bar movement (Fig. 7.14, top right). On the contrary, near
Bergen aan Zee (transect 7003350) the initial 15-year bar cycle stagnated around 2000
(Fig. 7.14, bottom left). When comparing the 3D images of 1990 and 2019 (Fig. 7.14,

middle) the sandy reinforcement of the Hondsbossche and Pettemer sea defence – com-
pleted in 2018 – is clearly visible.
The South Holland transect 8006900, just south of Zandvoort, exhibits an initially
rapid offshore bar migration (4-year cycle) that gradually slowed down to a 5-year
cycle from 1990–2010 and a still longer cycle since 2010 (Fig. 7.15, top right). Near
Katwijk (transect 8008850) the initial rapid offshore bar migration (5-year cycle) stag-
nated around 2000 (Fig. 7.15, bottom left). The 3D images (Fig. 7.15, middle) clearly
show the nourishment intervention known as the Sand Engine, comprising an unpre-
cedented 21.5 Mm3 concentrated mega-nourishment, which was implemented in the
summer of 2011 (Sect. 10.7.6).
We believe that the changes to the autonomous bar cycle in both North Holland and
South Holland are the result of the Dutch coastline preservation strategy, which con-
sists of frequent sand nourishments on the beach and shoreface to maintain the coast-
line in its 1990 position.
7.3.5. Episodic changes (dune erosion)

Dune erosion in the Netherlands and more generally along the southwest North Sea
coasts takes place during storm surges, when the mean water level increases and waves
can reach the dune face and impact it (Fig. 7.16). Sand eroded from the dunes (by a kind
of episodic avalanching process) is transported offshore by a strong undertow. The
combination of a strong undertow and high concentrations of suspended sediment
in the proximity of the dunes, leads to a large offshore transport capacity. Further
seaward the transport capacity of the flow decreases and the sediment starts to settle
forming a new coastal profile that better fits the storm surge conditions. The newly
developed foreshore is more efficient in dissipating the energy associated with the
incoming waves and consequently dune erosion rates decrease as the storm progresses.
post-surge beach width

R
post-surge max.storm surge level
dune foot
pre-surge dune foot
MSL
pre-surge beach width
Figure 7.16: Typical storm surge dune erosion when the surge level reaches the dune face.
Existing design rules in the Netherlands allow for erosion rates of 80 m to 100 m of
dune retreat during design storm conditions (Fig. 7.17). Design conditions are usually
extreme conditions with a specified return period used in the design of coastal works.
During a severe storm surge, with design conditions of offshore water levels approx-
imately 5 m to 6 m above MSL and with severe wave conditions (wave heights 𝐻𝑠 ≈

7 m to 9 m and peak periods 𝑇𝑝 ≈ 12 s to 18 s), the dunes will erode quickly and fiercely.
Figure 7.17 presents retreat rates for lower design conditions as well.
80
60
retreat [m]
40
20
0
10 -1 10 -2 10 -3 10 -4 10 -5
exceedance frequency per year
Figure 7.17: Retreat rates as a function of different design conditions (frequency of exceedance).
Figure 7.16 suggests that a new equilibrium profile is formed that fits the extreme
conditions. Indeed this was the basis of the ‘erosion profile’ approach of Vellinga (1986),
who extended earlier work of Van de Graaff (1977). The difference with the dynamic
equilibrium profile methods of Bruun and Dean is that the timescale is different. This
erosion profile method is presently still used for the basic assessment of the safety of
the Dutch dunes (ENW, 2007).
For sediments with a 𝐷50 between 90 µm and 225 µm, Vellinga conducted a compre-
hensive set of experiments on different scales, and derived a scale relation showing
the effect of grain size on the erosion profile shape:
0.28
𝑛𝑙 𝑛ℎ
=( 2) (7.9)
𝑛ℎ 𝑛𝑤
where 𝑛𝑥 represents the ratio between two cases with different scales of a variable 𝑥.
If a small-scale physical model is made to represent reality (the ‘prototype’), this scale
relation can be used to derive the physical model dimensions and grain diameter. Note
that in a scale series, the prototype can be a larger-scale model.
The following scales are specified in Eq. 7.9:
𝑛𝑙 ratio between a horizontal distance in prototype and the equivalent −

distance in the model
𝑛ℎ ratio between water depth in prototype and the water depth in the −
model
𝑛𝑤 ratio between fall velocity of bottom material in prototype and fall −
velocity of bottom material in model

Using the scale relation Eq. 7.9, Vellinga derived for the shape of the erosion profile
(the post-storm profile):
ℎ = 𝐴(𝑥 ′ )0.78 (7.10)
This equation relates the water depth ℎ to the offshore distance 𝑥 ′ through a dimen-
sional shape factor 𝐴, which can also be derived from Eq. 7.9:
𝐴 = 0.39𝑤𝑠0.44 (7.11)
where 𝑤𝑠 is in m/s. The position of 𝑥 ′ = 0 in Eq. 7.10 is now defined by the storm surge
level instead of by the MSL. Based on a fall velocity of 𝑤𝑠 = 0.0256 m/s, we find that
the erosion profile is described by ℎ = 0.078(𝑥 ′ )0.78 . Clearly, the shape of the erosion
profile is affected by grain size, with coarser material resulting in steeper slopes.
Note the similarity between Eqs. 7.10 and 7.11 on the one hand and Eqs. 7.1 and 7.7 on
the other hand, but also note the different power and the different coefficient used in
determining the shape parameter.
An interesting empirical finding by Vellinga concerns the extent of the sedimentation
zone (see Fig. 7.18). This zone was found to reach a depth of approximately 75 % of
the offshore significant wave height relative to the surge level. Equations 7.10 and 7.11
thus describe the erosion profile down to a depth of 75 % of the significant wave height.
The horizontal position of the erosion profile is defined by the conservation of sedi-
ment in the cross-shore profile.
ini
tial p
rofile
1:1
storm surge level

eros
ion
pro
file h mean sea level
1:1
2.5
Figure 7.18: Extent of the sedimentation zone to a depth ℎ equal to approximately 75 % of the
offshore (significant) wave height 𝐻𝑠 = 8 m. The ‘erosion profile’ or post-storm profile (after
Vellinga, 1986) has a prescribed shape (Eq. 7.10 with 𝐴 = 0.078) and its location is determined by
balancing the volumes of erosion (higher in the profile) and accretion (lower in the profile). The
vertical and horizontal scale are not the same (the 1:1 slope is a 45° slope in reality). The initial
profile matches Fig. 1.13 and 𝑆𝑆𝐿 = 5 m.

We can now make a rough estimate of the dune retreat. For simplicity, we assume that
the erosion profile and the pre-storm equilibrium profile have the same shape, but that
the erosion profile is shifted upward to the elevation of the storm surge level instead of
the MSL. A landward translation (retreat) of the profile is required to ensure conserva-
tion of sediment. Since we have assumed a shape-invariant profile, the required dune
retreat distance can be estimated using Bruun’s rule (see Sect. 2.5.2 and Sect. 7.4). Fol-
lowing Bruun’s rule, the volume of sediment eroded from the upper profile balances
the deposited volume in deeper water for a retreat distance equal to (cf. Eq. 7.13):
𝑅SS = 𝑆𝑆𝐿(𝐿/𝑑) (7.12)
where 𝑆𝑆𝐿 is the storm surge level above MSL, 𝐿 the length over which the erosion
and sedimentation takes place and 𝑑) the corresponding height.
We assume that the dune height above MSL is 10 m. Let us take the same design
storm conditions as given before, i.e. a storm surge level of approximately 5 m to
6 m above MSL and wave heights of 𝐻𝑠 ≈ 7 m to 9 m. From Vellinga’s finding that
the sedimentation reaches a depth relative to the surge level of 75 % of the offshore
significant wave height, we find that this depth is of the same order of magnitude as
the storm surge level (viz. 5 m to 6 m). Hence, the region of sedimentation extends
seawards until the point where the seabed is approximately at MSL, so that the height
over which the sedimentation and erosion take place is equal to the dune height above
MSL and 𝑑 = 10 m in Eq. 7.12. The length 𝐿 over which the changes take place can
be estimated from the erosion profile, which we assume to be determined by Eqs. 7.10
and 7.11. Using 𝑤𝑠 = 0.0256 m/s, we have ℎ = 0.078(𝑥 ′ )0.78 or 0.75𝐻𝑠 ≈ 0.078𝐿0.78 and
thus 𝐿 = 220 m to 300 m. From Eq. 7.12, we obtain a dune retreat distance of 110 m to
180 m. This is (somewhat) larger than the design rule outcome of 80 m to 100 m.
We will get a significantly closer estimate using a dune erosion prediction model such
as DUROS-plus, which was largely developed from a prediction model that was pro-
posed by Vellinga, which in essence is not very different from our crude hand compu-
tation. As in the above, the depth to which the erosion profile reaches is computed as
75 % of the offshore significant wave height. The initial profile, however, is not neces-
sarily identical in shape to the erosion profile. Furthermore, the shape of the erosion
profile now is a function of both the sediment fall velocity and the offshore signific-
ant wave height and has a 1:1 slope above storm surge level. Again the position of
the erosion profile must be determined by a horizontal translation, until erosion and
sedimentation volumes are balanced.
More refined methods to assess dune erosion exist, especially as a function of the
specific storm surge conditions and duration, including variation over the tide. With
these, the evolution in time of the erosion and sedimentation can be mimicked, which
is necessary when the final ‘equilibrium erosion profile’ is not reached. These methods
are commonly used for advanced designs.

After a storm surge, the beach width has become substantially wider (Fig. 7.16) and
the coastal profile is not in equilibrium with the post-surge hydrodynamic conditions.
Waves, tide and wind reshape the foreshore and the dunes gain eroded sand back gradu-
ally. In a situation without alongshore sediment transport gradients, the dunes recover
to pre-storm volume. However, the timescale of dune recovery is considerably larger
than that of erosion.
7.4. Structural losses or gains

In the previous sections we have assumed that while the beach and surf zone may
breathe and may display three-dimensionality or simply two-dimensionality under
cross-shore wave forcing, the amount of sediment in the active zone remains constant.
On engineering scales (1 to 100 years) a number of processes exist that violate these as-
sumptions. A prime one is alongshore transport gradients, treated in Ch. 8. Shoreline
orientations and hence alongshore transport rates will vary, which will lead to either
shoreline advance or retreat, while the upper shoreface profile will remain in profile
equilibrium. On these engineering scales, aeolian (transport by wind) sand loss from
the beach to the dunes and nearshore offshore canyons may be important as well. They
can be accounted for as sinks.
A further ‘sink’, which is virtual because no sediment is lost, is due to relative sea level
rise. This effect was first described by Bruun (1962). As explained in Sect. 2.5.2, Bruun
argued that the response of the upper shoreface to an increased MSL is so fast that
the equilibrium upper shoreface profile will adjust to the same profile, but relative to
the new MSL (Figs. 2.22 and 7.19b). Hence, even though no sediment is lost from the
profile, the shoreline retreats.
The retreat distance 𝑅 follows from sediment continuity considerations. The new equi-
librium profile requires that the shoreline retreats to supply sediment to the lower surf
zone in order to maintain the equilibrium profile. Horizontal retreat 𝑅 leads to a sedi-
ment yield of 𝑅 × 𝑑, where 𝑑 is the height over which erosion and sedimentation takes
place (viz. the closure depth plus dune crest height). Relative Sea Level Rise (RSLR)
leads to a demand of sand of 𝐿 × RLSR, where 𝐿 is the length over which the erosion
and sedimentation takes place. Balancing the two (no net loss or gain) gives:
𝑅RSLR = RSLR(𝐿/𝑑) (7.13)
where RSLR is the rise of relative sea level, 𝐿 the length over which the erosion and
sedimentation takes place and 𝑑 the corresponding height. In practice the ratio 𝐿/𝑑 is
of the order 50 to 150 (Zhang et al., 2004).
Sediment supply may counteract the effect of shoreline retreat. Fig. 7.19c assumes a
horizontal shift of the profile in the case of a sediment source. If the rate of sediment
supply keeps up with the rate of sea level rise, the accommodation space created by the

7.4. Structural losses or gains 349
SLR
(a) (b)
SLR
sediment sediment
supply supply
(c) (d)
Figure 7.19: Equilibrium shoreface response to sea level rise and sediment supply: (a) the equi-
librium profile before sea level rise; (b) shoreline retreat and shoreface erosion due to sea level
rise (Bruun rule); (c) coastal progradation in the case of sediment supply; (d) balance between
sea level rise and sediment supply.
sea level rise is filled by incoming sediments (Fig. 7.19d) and the profile moves upward
only; the position of the shoreline remains unchanged.
Quite often the Bruun rule is used to assess possible future effects of sea level rise.
Although it gives qualitative insight into profile response to sea level changes, it is not
a valid model approach in general due to oversimplifications (see e.g. Ranasinghe and
Stive (2009)). One of the problems with the application of the Bruun rule – for instance
when considering shore nourishment to counteract the effects of sea level rise – is that
time is required for equilibrium to be established. Often this point is not considered by
coastal engineers. Rapid sea level rise induces rapid coastal response. However, the
timescale on which present sea level rise influences coastal evolution in general and
coastal erosion in particular, is relatively large. Present coastal evolution on smaller
time- and spatial scales is dominated by other causes.

7.5. Cross-shore sediment transport

7.5.1. Introduction
The research field of cross-shore hydrodynamics, sediment transport dynamics and
resulting bed profile dynamics is highly topical and very complex, and simple, ana-
lytical treatments do hardly exist. The reason is that over the whole of the shoreface
the constituent processes vary strongly. In the upper shoreface, say the surf zone, un-
der normal conditions we encounter a mix of bed and suspended load transport due
to undertow, bound and free long waves, short-wave skewness in combination with
breaking-induced turbulence (Bailard, 1981; Roelvink & Stive, 1989). In this zone, dur-
ing extreme conditions the undertow is dominant (Steetzel, 1993), although long-wave
effects (bound and free long waves) cannot be ignored (Van Thiel de Vries, 2009). On
the middle and lower shoreface, again a mix of bed and suspended load transport is
encountered, but here wave boundary layer streaming, bound long waves and short-
wave skewness are relevant. Also, upwelling and downwelling due to stratification
and Ekman currents may play a role here (but are outside the scope of this book).
For calculating bed profile dynamics of the upper shoreface (under storm surge and
non-storm surge conditions), process-based, numerical models have been developed.
These models describe sediment transport as a function of the wave evolution in the
surf zone, which requires a largely empirical description of the highly variable wave
energy dissipation and related flow variations. For the middle and lower shoreface,
analytical approaches are available due to Bowen (1980), who derived formulas for
equilibrium profiles by balancing onshore and offshore transport components.
In Sect. 7.5.2, the cross-shore sediment transport rates are decomposed into contribu-
tions due to undertow, short-wave skewness and bound and free long waves associated
with wave groups. This gives new insight into the relevance of the various flow com-
ponents to net sediment transport. In Sect. 7.5.3, we reproduce some of the analytical
approaches for the middle and lower shoreface as introduced by Bowen.
7.5.2. Decomposition of the transport rate

As explained in Ch. 5, the velocity 𝑢 close to the bed can be assumed to consist of a
wave group-averaged component 𝑢,̅ a short-wave-averaged oscillatory component 𝑢𝑙𝑜
and a short-wave component 𝑢ℎ𝑖 :
𝑢= 𝑢̅
⏟ + 𝑢⏟
𝑙𝑜 + 𝑢⏟
ℎ𝑖 (7.14)
time-averaged component low-frequency motion oscillatory motion
(streaming outside surf zone, at wave group scale at short-wave scale
undertow in surf zone)
We are now interested in the relative contributions of these components of the time-
varying flow to the net sediment transport (net meaning averaged over the wave

7.5. Cross-shore sediment transport 351
group). Let us for simplicity focus on the third odd velocity moment that determ-
ines the bed load transport (see Eq. 6.51). By assuming that 𝑢 ̅ ≪ 𝑢𝑙𝑜 ≪ 𝑢ℎ𝑖 and using a
Taylor expansion (see Intermezzo 7.3), Roelvink and Stive (1989) showed that the most
important contributions to the third odd velocity moment are given by:
2 2 2
⟨𝑢|𝑢|2 ⟩ = 3 ⟨𝑢 ̄ |𝑢ℎ𝑖 | ⟩ + ⟨𝑢ℎ𝑖 |𝑢ℎ𝑖 | ⟩ + 3 ⟨𝑢𝑙𝑜 |𝑢ℎ𝑖 | ⟩ + … (7.15)
⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟ ⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟ ⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟
1 2 3
The term |𝑢ℎ𝑖 |2 in all three terms reflects that the sediment load is stirred by short waves.
The first term on the right-hand side of Eq. 7.15 is related to the subsequent transport
by the mean current. The mean current could be the onshore-directed wave-induced
near bed streaming in non-breaking waves (Sect. 5.4.3) or the offshore-directed under-
tow in the surf zone (Sect. 5.5). The second and third term represent the contribution
to the velocity moment (and hence the net transport) of the oscillatory part of the in-
stantaneous velocity; the second term is related to the short-wave skewness, whereas
the third term is associated with the interaction between the long-wave velocity and
the short-wave velocity variance (the latter will change periodically on the timescale
of short-wave groups).
The second term is zero for a completely symmetrical oscillating velocity signal. How-
ever, in shoaling waves the velocity signal becomes asymmetric about the horizontal
axis (positively skewed, meaning ⟨𝑢 3 ⟩ > 0, see Sect. 5.3). As a result the second term
is non-zero and onshore-directed. In Intermezzo 6.5, we explained this transport con-
tribution as follows: the larger onshore peak velocities under the wave crests are more
effective in stirring up sediment than the smaller offshore velocities under the wave
troughs. The relation between the amount of sediment stirring and velocity is non-
linear; a twice as high velocity stirs up (in this case) four times more sediment. The
result is an onshore transport.
The third term is non-zero as long as there is a correlation between the slowly varying
short-wave velocity variance (the short-wave envelope) and the long wave forced by
the wave group. Outside the surf zone this correlation is negative (where the bound
long wave is not yet ‘released from the group’); the mean water surface decrease
(trough of the bound long wave) is found under the larger amplitude waves in the
group (Fig. 7.20).
As long as the trough of the bound long wave coincides with the highest velocities in
the group, most sediment is stirred up while the long-wave velocities are offshore-
directed. The net transport by the bound long wave motion is therefore offshore-
directed (Fig. 7.20 shows this for suspended sediment transport). If the phase rela-
tionship between the short-wave envelope and the long wave changes, this situation
may change. The main reason why the phase relationship may change is that bound
long waves become released during breaking of the short waves.

bound long wave short waves

cˉ
cg
surface elevation η
ηhi ηlo
motion induced
by long waves ulo
suspended sediment
concentration
Figure 7.20: Offshore (suspended) sediment transport under a bound long wave (outside the
surf zone). The sediment suspended by the highest waves is transported offshore by the out of
phase long-wave motion. The effect is a net offshore transport.
Roelvink and Stive (1989) analysed the decomposition of the velocity moments using
laboratory measurements in a large (full-scale) wave flume. Calculated and measured
total third and fourth odd flow moments (important for bed load and suspended load
transport respectively) are compared in Fig. 7.21. The figure clearly shows the onshore-
directed component due to short-wave skewness. It increases in shoaling waves and
decreases again in the surf zone (Sect. 5.3). The undertow component is offshore-
directed in the entire surf zone. The long-wave contribution is offshore-directed un-
til the correlation between the short-wave variance and long wave becomes positive
somewhere in the surf zone.
From Fig. 7.21 it also becomes clear that gross cross-shore transports are much higher
than net cross-shore transports. This makes accurate cross-shore transport predictions
very difficult.
As mentioned in Sect. 7.5.1, during extreme conditions the transport contribution
due to undertow is dominant. This explains that under higher and longer waves, as
present during storms, a net offshore transport is observed, leading to a ‘winter’ pro-
file (Sect. 7.3.3). By contrast, milder waves build up the profile to a ‘summer’ profile
via onshore transport, due to predominantly short-wave skewness.
A similar decomposition of the transport rates can be made for the situation in which
tidal flow dominates the velocity 𝑢. For this we refer to Sect. 9.7.2.
7.5.3. Analytical solutions for the middle and lower shoreface

Bowen (1980) derives formulas for the equilibrium shape of the middle and lower shore-
face based on balancing onshore and offshore transport terms. He discusses the con-
tribution of the main process of oscillatory orbital motion in combination with a mean

60
40
[m 3 s -3 10-4 ]
20
0
-20
-40
-60
-80
-100
20
[m 4 s -4 10-4 ]
-20
-40
-60
0
depth [m]
0.2
0.4
0.6
40 35 30 25 20 15 10
x [m]
total, calculated total, measured
short waves, calculated short waves, measured
long waves, calculated
return flow, calculated
Figure 7.21: Total third (𝑛 = 3) and fourth (𝑛 = 4) odd flow moments and their constituent com-
ponents (i.e. for the third odd moment, the terms 1, 2 and 3 on the right-hand side of Eq. 7.15);
wave flume measurements (symbols) and model predictions (lines). This situation – with a dom-
inant transport component due to undertow and a net offshore transport – represents storm
wave conditions. For milder conditions the onshore transport components are relatively more
important. From Roelvink and Stive (1989).
flow, a higher harmonic orbital motion (asymmetry3 ) and a long wave motion, for in-
stance a bound long wave. Because of the non-linear coupling of these contributions
in sediment transport, their interactions are important. His analytical approach gives
a good insight into these process interactions, and we note that the model efforts of
Bailard (1981), Bailard and Inman (1981), Roelvink and Stive (1989) and Stive and De
Vriend (1995) are very much inspired by this.
In this section, we present an abstracted form of certain parts of Bowen’s landmark
paper.
Bowen starts by stating that:
3
In this section – as opposed to elsewhere in the book – we follow the wording of Bowen and use the
word asymmetry for a skewed wave that is asymmetric about the vertical axis.

“Any understanding of the relationship between the incident waves and

the topography of a beach is greatly complicated by the beach being rarely,
if ever, in equilibrium with the existing wave field. The morphology de-
pends on some complex integral of past wave conditions, an integral heav-
ily weighted towards periods of high waves. (…) The real situation is per-
haps too complex for parameters to be developed without some guidance
as to the relative importance of various possible processes. Simple theoret-
ical models are particularly useful in defining these possibilities. (…) The
purpose of the present paper is to develop a consistent model for onshore-
offshore sediment transport under the influence of waves, currents and
gravity. (…) The model is, however, neither unique nor necessarily cor-
rect (…), but a doubtful model is clearly preferable to no model. (…) The
most important aspect of the present approach is the rigorous develop-
ment of a theory, starting from any given model of sediment transport.”
Note that he speaks of a beach, but in what follows he concentrates rather on the
middle and lower shoreface, not on the surf zone. The probable reason is that we lack
straightforward analytical descriptions of the interactions between undertow, asym-
metric oscillatory flow motion, breaking-induced turbulence and the bound and free
long waves on the upper shoreface.
He explains the existence of a beach as follows:
“If a beach was exposed to waves having an exactly symmetrical, orbital

velocity, all the sediment would slide down the slope and out to sea. The
existence of the beach depends on small departures from symmetry in
the velocity field, balancing this tendency for gravity to move material
offshore. (…) The development of the theory assumes as a basic hypothesis
that the orbital motion of the incoming waves is the dominant motion.”
Hence, the basic idea is that the velocity 𝑢 consists of two interacting components, the
symmetrical orbital velocity 𝑈0 = 𝑢0 cos(𝜔𝑡) and a perturbation 𝑈1 , where:
𝑢 = 𝑈 0 + 𝑈1 , with generally 𝑈0 ≫ 𝑈1 (7.16)
The perturbation 𝑈1 of the primary symmetrical harmonic 𝑈0 can take many forms;
cases of particular interest are:
1. A constant, steady current: 𝑈1 = 𝑢1 ;
2. The velocity field associated with a higher harmonic of the incoming wave: 𝑈1 =
𝑢𝑚 cos(𝑚𝜔𝑡 + 𝜃𝑚 ), with 𝑚 = 2, 3, 4, …;
3. A perturbation due to a wave with a frequency 𝜔𝑡 unrelated to 𝜔: 𝑈1 = 𝑢𝑡 cos(𝜔𝑡 𝑡).
Note that in principle the steady current can be any current. Especially relevant for the
shoreface is undertow in the surf zone and boundary layer streaming, further offshore.

The higher harmonics are relevant in the surf zone, with forward pitching or even
sawtooth waves and 𝜃𝑚 is non-zero, while further offshore we have only skewed or
peaked waves (symmetrical around the vertical), where 𝜃𝑚 is zero. The third item (a
wave with a frequency unrelated to the primary harmonic) is not discussed in this
book.
Bowen adopts Bagnold’s transport description for uniform flow and transforms it for
non-linear oscillatory motion, resulting in Eqs. 6.48a and 6.48b. According to Bowen,
Bagnold’s energetics model is the simplest available model that includes the basic prop-
erties that are necessary to describe cross-shore sediment transport. He especially
stresses the explicit inclusion of the gravitational effect of a sloping bed.
Bowen also discusses the limitations of the energetics model:
“First, and perhaps most serious, the transport in this model depends only
on the immediate flow conditions, adjusting instantaneously to changes
without any time lag. Second, the theory applies to fully developed flow
and does not describe initiation of movement; it does not apply to large
particles that may move only intermittently at peak flows.”
First Bowen looked at suspended sediment transport and then at bed load transport
both on the middle and lower shoreface. We will only repeat his suspended sediment
transport derivations, because Stive and De Vriend (1995) showed by evaluating the
transport terms for a typical shoreface situation that this transport mechanism is dom-
inant on the middle and lower shoreface.
From Eq. 6.48b, the suspended load transport can be expressed as:
𝜀𝑠 𝑐𝑓 𝜌 𝑢 3 |𝑢| tan 𝛼
𝐼𝑠 = with 𝛾= (7.17)
𝑤𝑠 (1 − 𝛾 𝑢) 𝑤𝑠
For normal transport conditions 𝛾 𝑢 < 1. If 𝛾 𝑢 approaches 1, 𝐼𝑠 goes to infinity and

autosuspension effects (avalanching, i.e. the slope is so steep that all particles go into
suspension) totally dominate the transport.
Bowen now expands Eq. 7.17, using the Taylor series as in Intermezzo 7.3, and takes a
time average (denoted by the overbar), arriving at a net transport ⟨𝐼𝑠 ⟩ over a number
of wave periods. If we consider normal transport conditions, the expansion involves
the small quantities 𝑈1 /𝑈0 and 𝛾 𝑈0 . To second order, Bowen’s transport formula reads:
𝜀𝑠 𝑐𝑓 𝜌
⟨𝐼𝑠 ⟩ = [𝑈03 |𝑈0 | + 4𝑈1 𝑈02 |𝑈0 | + 6𝑈12 𝑈0 |𝑈0 |+
𝑤𝑠
(7.18)
𝛾 (𝑈04 |𝑈0 | + 5𝑈1 𝑈03 |𝑈0 |) + 𝛾 2 𝑈05 |𝑈0 | + …]

Terms of the form 𝑈0𝑛 |𝑈0 | vanish if 𝑛 is odd, since 𝑈0 is assumed to be a symmetric
oscillation.
Now, to first order, the equation reduces to:
𝜀𝑠 𝑐𝑓 𝜌
⟨𝐼𝑠 ⟩ = [4𝑈1 𝑈02 |𝑈0 | + 𝛾 𝑈04 |𝑈0 | + …] (7.19)
𝑤𝑠
The first term describes the transport, onshore or offshore, due to a perturbation of
the flow field 𝑈1 . The second term involves the slope tan 𝛼 and is generally positive,
representing the tendency for downslope transport.
Bowen continues by evaluating the transport equations for the first form of the per-
turbation 𝑈1 , viz. 𝑈1 = 𝑢1 is a constant in time. Using Eq. 7.19, this yields to first
order:
𝜀𝑠 𝑐𝑓 𝜌 16 16 5
⟨𝐼𝑠 ⟩ = [ 𝑢1 𝑢03 + 𝛾𝑢 ] (7.20)
𝑤𝑠 3𝜋 15𝜋 0
Equation 7.20 is a general result for any distribution of a steady flow 𝑈1 (𝑥). One could
consider undertow, upwelling or downwelling but Bowen chooses to continue with
boundary layer streaming as the relevant effect, which we support for the middle and
lower shoreface.
An equilibrium profile, purely in suspended load, exists if the gravitational effects bal-
ance the influence of the steady current everywhere and ⟨𝐼𝑠 ⟩ vanishes. From Eq. 7.20,
when ⟨𝐼𝑠 ⟩ = 0:
tan 𝛼 5𝑢
𝛾= = − 21 (7.21)
𝑤𝑠 𝑢0
This equation essentially contains no free parameters, which is an attractive feature

of Bagnold’s model.
The second-order, Eulerian mean velocity due to boundary layer streaming is of the
order −𝑢02 /𝑐 (Eq. 5.29), where 𝑐 is the wave phase speed (Eq. 3.23), so that:
5𝑤𝑠 5𝑤𝑠 𝜔
tan 𝛼 ≃ = (7.22)
𝑐 𝑔 tanh 𝑘ℎ
where 𝑘 is the local wavenumber, ℎ the water depth. This yields a formula for the
equilibrium slope in terms of 𝑤𝑠 𝜔/𝑔, a dimensionless parameter also used by Dean
(1973).
If sediment of a given grain size is in equilibrium with the local slope, so that ⟨𝐼𝑠 ⟩
vanishes for this grain size, any coarser material with a larger fall velocity has a smaller
value of 𝛾 . The term involving gravity is reduced, the onshore term remains constant,

coarser material therefore moves onshore; similarly, finer material moves offshore as
observed. As can be observed from Eq. 7.22, profiles of coarser material are steeper,
and any material that finds itself on a beach that is ‘too steep’ moves seawards. This
leads to a new null hypothesis that is in far better agreement with observations than
classical models.
In shallow water 𝑐 tends to √𝑔ℎ and Eq. 7.22 is readily integrated:
𝑑ℎ 2
tan 𝛼 = ≃ 5𝑤𝑠 /√𝑔ℎ ⇒ ℎ3 ≃ (7.5𝑤𝑠 𝑥) /𝑔 (7.23)
𝑑𝑥
And thus the equilibrium profile is described by:
2/3
ℎ ≃ (7.5𝑤𝑠 ) /𝑔 1/3 𝑥 2/3 (7.24)
Intriguingly, this result accurately resembles the results earlier presented for the equi-
librium shoreface profiles of Bruun and Dean (Eq. 7.1). The profile shape factor 𝐴
according to Eq. 7.24 is for 𝑤𝑠 = 0.0256 m/s:
2/3
𝐴 = (7.5𝑤𝑠 ) /𝑔 1/3 = 0.16 (7.25)
This result is reasonably close to 𝐴 = 0.10 for the same fall velocity according to Eq. 7.7.
Next, Bowen considered the role of the higher harmonics associated with the incoming
wave field (thus, wave asymmetry). The perturbation 𝑈1 is now a higher harmonic of
𝑈0 = 𝑢0 cos(𝜔𝑡), viz. 𝑈1 = 𝑢𝑚 cos(𝑚𝜔𝑡 + 𝜃𝑚 ).
The effect of wave asymmetry on the transport is derived from the first-order equation
Eq. 7.19. The first-order term 4𝑈1 𝑈02 |𝑈0 | vanishes if 𝑚 is odd. We therefore only have
to take even harmonics into account, the second harmonic (𝑚 = 2) being the most
important. For the second harmonic we have:
16 3
4𝑈1 𝑈02 |𝑈0 | = − 𝑢 𝑢 cos 𝜃2 (7.26)
5𝜋 0 2
The term arising from Eq. 7.26 is not an alternative to the contribution due to the
Longuet-Higgins boundary layer streaming, but an additional factor. We therefore
add the term from Eq. 7.26 to Eq. 7.20 and find:
𝜀𝑠 𝑐𝑓 𝜌 16
⟨𝐼𝑠 ⟩ = 𝑢 3 [5𝑢1 − 3𝑢2 cos 𝜃2 + 𝛾 𝑢02 ] (7.27)
𝑤𝑠 15𝜋 0
Again, an equilibrium profile for suspended load is found for ⟨𝐼𝑠 ⟩ = 0:
tan 𝛼 5𝑢 3𝑢 cos 𝜃
𝛾= = − 21 + 2 2 2 (7.28)
𝑤𝑠 𝑢0 𝑢0

To evaluate this equation, Bowen assumed 𝜃2 = 0 and used the second-order Stokes
solution (see Intermezzo 5.1) to estimate 𝑢2 . Herewith he derived that:
9 𝑤𝑠
tan 𝛼 ≃ [2 + (sinh 𝑘ℎ)−2 ] (7.29)
4 𝑐
The term in brackets is shown in Fig. 7.22. In deep water, the effect of wave asymmetry
is negligible compared to that of the wave boundary layer streaming. As the waves
shoal, the wave harmonics become increasingly important. Since in very shallow wa-
ter the Stokes solution is not necessarily a good approximation, the trend is indicated
by a dashed line. At some point, the waves break and the form of 𝑢2 is only known
from very limited empirical data.
10 2
f (kh)
10 1
Stokes wave
streaming
0
10
10 -2 10 -1 10 0 10 1
k0 h
Figure 7.22: The term 𝑓 (𝑘ℎ) = 2 + (sinh 𝑘ℎ)−2 (see Eq. 7.29), plotted as a function of the nondi-
mensional depth 𝑘0 ℎ, where 𝑘0 = 𝜔 2 /𝑔 is the deep-water wavenumber.
According to Eq. 7.29, wave asymmetry becomes as important as drift when:
sinh 𝑘ℎ ∼ 2− /2
1
(7.30a)
ℎ ∼ 0.01𝑔𝑇 2 (7.30b)
where 𝑇 is the wave period. Equation 7.30 suggests that the extent of the nearshore
area in which wave asymmetry is the dominant effect is strongly dependent on the
period of the incoming waves. On the west coast of North America, which is generally
exposed to waves of much longer periods, wave asymmetry should have a much more
significant effect than on the east coast.

As 𝑘ℎ becomes small, then from Eq. 7.29 a formula for the equilibrium profile due to
asymmetry can be derived, comparable to Eq. 7.24:
5.7𝑤𝑠 𝑥 2
ℎ5 ∼ ( ) 𝑔 (7.31)
𝜔2
Bowen’s landmark paper contains many more interesting analytical derivations and
again we note that these analytical ideas have been at the basis of many numerical
process-based model efforts. It is a paper worth studying if one is interested in solving
the cross-shore morphology problem.
Intermezzo 7.3 Taylor series expansion of velocity moments
Bowen (1980) outlined a Taylor series expansion of the velocity moment 𝑢 𝑛 |𝑢|. He
assumed the velocity 𝑢 to consist of two interacting components, viz. 𝑈0 and a
perturbation 𝑈1 , hence 𝑢 = 𝑈0 + 𝑈1 . Under the assumption that 𝑈0 is much larger
than 𝑈1 , the Taylor series expansion reads:
𝑛(𝑛 + 1) 2 𝑛−2
𝑢 𝑛 |𝑢| = 𝑈0𝑛 |𝑈0 | + (𝑛 + 1)𝑈1 𝑈0𝑛−1 |𝑈0 | + 𝑈1 𝑈0 |𝑈0 | + … (7.32)
2
In practice, 𝑈0 is the wave orbital velocity 𝑈0 = 𝑢0 cos(𝜔𝑡) which vanishes during
the orbital cycle, and the assumption that 𝑈1 is small compared to 𝑈0 is not justified
during the stages of the orbital velocity when 𝑈0 is small. The approximation is
really that 𝑈1 ≪ 𝑢0 , the maximum orbital velocity, and that the significant periods
for transport are when 𝑈0 is large. This is probably a reasonable assumption, as
the transport is proportional to the third or fourth power of the total velocity.

361
8
Longshore transport and coastline
changes
8.1. Introduction
The purpose of this chapter is to illustrate how alongshore sediment transport con-
tributes to shaping our coasts. The insight gained in the previous chapters, describing
the water and sediment movements along coasts, will now be combined to provide the
necessary explanations.
The main principle behind coastal change has already been indicated in Chs. 1 and 7:
coastal change occurs where there are spatial sediment transport gradients and/or sedi-
ment sinks or sources. In this chapter we specifically look at gradients in mainly wave-
driven longshore transport or, in other words, littoral transport. Consider an infinitely
long, straight sandy coast with parallel depth contours, as sketched in Fig. 5.4. If waves
approach this coast at a uniform angle along its entire length, and there are no other
current-driving forces such as tides and wind, then there will be a constant, uniform
transport of sand along this coast. What, then, causes erosion or deposition? This is
caused by changes (gradients) in transport rates along a coast. This change may result
from changing any of the factors influencing the longshore sediment transport rate,
such as nearshore wave height and angle of wave incidence.
Various methods to compute longshore sediment transport are treated in Sect. 8.2. The
calculation of shoreline position is described in Sect. 8.3. Section 8.4 discusses the vari-
ous coastal shapes and characteristics that can be explained from longshore sediment
transport gradients.
In Sect. 6.9.2 it was mentioned that coastline change is dominated by alongshore effects
in the case of human-induced changes on high-wave energy coasts (e.g. the Dutch
North Sea coast). In Sects. 8.3 and 8.4, we therefore mainly focus on typical time
and spatial scales for engineering or coastal maintenance (years to decades and 1 km

362 8. Longshore transport and coastline changes
to 100 km, see Sect. 1.5.3) and implicitly assume dominance of alongshore transport
processes.
8.2. Longshore transport formulas

8.2.1. General transport formulas
Longshore sediment transport is the net movement of sediment particles through a
fixed vertical plane perpendicular to the shoreline. The direction of this transport is
parallel to the shoreline and the depth contour lines. Without further specifications
it is often meant to be the total or bulk transport in the alongshore direction (e.g. the
total transport along a coast in the entire active zone).
Wave-driven longshore sediment transport depends, amongst other factors, on the
hydrodynamics in the breaker zone (see Ch. 5) and on the sediment properties (see
Sect. 6.2). However, regardless of the strength of the transport-generating hydraulic
forces, sediment transport will only occur if moveable sediment is available in a cer-
tain area, either in the bed or in the water column through supply from an adjacent
area. If the seabed is fixed (for instance by bottom protection, biochemical processes or
vegetation or in the case of absence of sediment), erosion is prevented and the actual
transport may be smaller than the local transport capacity based on the hydrodynam-
ics. The sediment transport formulas presented in this section all reflect the longshore
transport capacity assuming that the material is actually present and available for trans-
port.
The computation of the sediment transport parallel to the coast can, in principle, be
handled with any of the general sediment transport models introduced in Ch. 6. When
these models are applied to compute longshore sediment transport rates, we can sim-
plify the problem by assuming that transport in the alongshore direction is by a time-
invariant longshore current.
Let us, for instance, consider the (suspended) sediment transport description based on
velocity multiplied by concentration, Eq. 6.30. In principle the current velocity in the
alongshore direction 𝑣(𝑧, 𝑡) at a certain position in the surf zone is a function of 𝑧 and 𝑡.
It consists of a mean and oscillatory component: 𝑣(𝑧, 𝑡) = 𝑉 (𝑧)+𝑣(𝑧,̃ 𝑡). The short-wave
motion acts mainly in the cross-shore direction (𝜑 is small due to refraction) and thus
nearly perpendicular to the sediment transport direction. In practice it is therefore
often assumed that the alongshore oscillatory velocity 𝑣(𝑧, ̃ 𝑡) is small and that during
a period with no significant change in the offshore wave conditions, a relatively time-
invariant current velocity 𝑣(𝑧, 𝑡) ≈ 𝑉 (𝑧) is present. This assumption implies that the
mean water motion is the main contributor to longshore sediment transport. Note
that the sediment concentration, in principle, fluctuates on the timescale of the waves
and can – just as the velocity – be seen as the sum of a mean component and an
oscillatory component, viz. 𝑐(𝑧, 𝑡) = 𝐶(𝑧) + 𝑐(𝑧,̃ 𝑡). The purely oscillatory component

8.2. Longshore transport formulas 363
of the sediment concentration, however, will not give rise to a net transport since
𝑉 (𝑧) ⋅ 𝑐(𝑧,
̃ 𝑡) is purely oscillatory and therefore has a time average equal to zero. Hence,
(suspended load) transport models based on solving the velocity and concentration
distribution can be reduced to the current-related part only (cf. Eqs. 6.32 and 6.33):
ℎ
⟨𝑠𝑦 ⟩ = ∫ 𝑉 (𝑧) ⋅ 𝐶(𝑧)𝑑𝑧 (8.1)
𝑎
where:
𝑠𝑦 net longshore sediment transport excl. pores m3 /m/s

𝑉 (𝑧) longshore current velocity at height 𝑧 above the bottom m/s
𝐶(𝑧) time-averaged sediment concentration at height 𝑧 m3 /m3
𝑎 thickness of bed load layer m
ℎ local (still) water depth m
Note that such a computation yields the longshore (suspended) sediment transport at
a certain cross-shore location. If we like to quantify the total transport in e.g. the surf
zone, we must integrate the outcome of the equation over the width of the surf zone.
The longshore current velocity 𝑉 (𝑧) can have many different driving forces, but for
most beaches this current is driven predominantly by breaking waves which approach
the coast at an angle. The longshore current is concentrated more or less in the surf
zone and occurs regardless of whether there is sediment transport or not. Methods to
compute this longshore current were outlined in Sect. 5.5.5.
We have concluded that the role of the oscillatory wave motion in transporting the sed-
iment in the alongshore direction is limited. What then is the role of waves – besides
generating the longshore current – in longshore sediment transport? Their role is to
enhance the amount of sediment in suspension 𝐶(𝑧). As a result of turbulence, gener-
ated in the wave boundary layer and at the surface under breaking waves, considerable
amounts of sediment are brought into suspension:
1. Due to the orbital motion (in mainly the cross-shore direction), the magnitude
of the bed shear stress varies over the wave cycle and peaks twice every wave
cycle. Especially during these peaks a lot of sediment is mobilised and entrained;
2. Breaking waves strongly increase the turbulence in the water column and there-
fore relatively easily bring suspended sediments into the upper part of the flow.
Summarizing, since the wave motion in the breaker zone is nearly perpendicular to
the resulting current, the major influence of waves is to stir more material loose from
the beach and keep it in suspension, thereby increasing the sediment concentration. It
is the (wave-induced) longshore current (and not the oscillatory wave motion) that is
mainly responsible for the net movement of material along the coast.

We can also demonstrate these findings using a transport formula of the form Eq. 6.24
or Eq. 6.51. With the orbital motion approximately cross-shore and the bed shear stress
according to Eq. 5.76, we have:
(𝑛−1)/2
⟨𝑠𝑦 ⟩ = 𝑚1 ⟨|⃗𝜏𝑏 | 𝑢 |𝑛−1 ⟩ 𝑉
⟩ 𝑉 = 𝑚2 ⟨|⃗ (8.2)
where:
𝑠𝑦 longshore sediment transport (per unit width) m3 /m/s

𝑉 longshore current velocity at some level above the bed (e.g. m/s
depth mean velocity)
|𝜏⃗𝑏 | magnitude of the bed shear stress vector due to combined N/m2
wave-current motion
𝑢|
|⃗ magnitude of the combined wave-current velocity vector m/s
𝑚1 , 𝑚 2 (dimensional) coefficients −
𝑛 power (typically 3 to 5) −
(𝑛−1)/2 𝑛−1
In Eq. 8.2, 𝑚1 |𝜏⃗𝑏 | = 𝑚2 |⃗
𝑢| describes the sediment load that is transported by
the longshore current velocity 𝑉 . The sediment load is stirred up by the combination of
longshore current and wave orbital motion (see Fig. 6.13 for the non-linear enhance-
ment of the bed shear stress in the combined wave-current motion). Especially the
wave orbital motion is important because of the small boundary layer thickness and
𝑢 | = √𝑢 2 + 𝑉 2 (see Fig. 5.38),
hence large shear stresses (see Sect. 5.4.3). Substitution of|⃗
with 𝑢 = 𝑢̂ cos 𝜔𝑡 the cross-shore time-varying orbital motion, yields for 𝑛 = 3:
1
⟨𝑠𝑦 ⟩ = 𝑚 2
2 ⟨𝑢 + 𝑉 ⟩
2
⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟ 𝑉⏟ = 𝑚2 𝑢̂ 2 𝑉 + 𝑚2 𝑉 3 (8.3)
2
sediment load stirred by longshore current
wave-current motion responsible for transport
Assuming the stirring due to the short-wave motion is dominant, the longshore trans-
port is approximately proportional to ⟨𝑠𝑦 ⟩ ∝ 𝑢̂ 2 𝑉 . This again reflects that the effect of
short waves is to mobilise the material which is consequently transported by the long-
shore current. In Sect. 8.2.3, specifically in Intermezzo 8.1, we will see that commonly
used bulk longshore transport formulas can be interpreted in a similar way.
8.2.2. Cross-shore distribution of longshore transport

The cross-shore distribution of longshore sediment transport is, of course, strongly de-
termined by the cross-shore distribution of the longshore current. The longshore cur-
rent distribution on a barred beach for different wave conditions (Fig. 5.42) is repeated
in Fig. 8.1a. Furthermore, Fig. 8.1 shows the longshore transport rates as calculated
with the longshore sediment transport model Unibest-CL+. Comparing the upper and

middle panel of Fig. 8.1a while keeping Eqs. 8.1 and 8.3 in mind, we may conclude that
most of the sediment stirring takes place in a relatively narrow zone on the seaward
flanks of the breaker bars, where most of the wave-breaking takes place. Also note
the non-linear dependency of the transport on the velocity. Figure 8.1b demonstrates
that the outcomes of uncalibrated transport formulas may differ substantially.
Bayram et al. (2001) analyse the cross-shore distribution of longshore sediment trans-
port according to a few well-known predictive formulas and field measurements at
Duck, North Carolina. Measured hydrodynamics were used as much as possible as
input for the transport models. The transport models were used with standard coef-
ficient values without further calibration. Figure 8.2 gives the result for one specific
condition (𝐻𝑟𝑚𝑠 = 3.18 m, 𝑇𝑝 = 12.8 s) representative for a large transport on a barred
sandy profile during a storm. The mean longshore current velocity in the surf zone was
0.6 m/s. The water depth at the most offshore measurement point was 8.6 m. The peak
in the transport rate was observed some distance shoreward of the bar crest, whereas
the formulas predicted the peak to occur more seaward (i.e., close to the bar crest).
For this specific run, most formulas overpredict the sediment transport. Furthermore,
the differences between the various models are rather large. The formulas used differ
particularly in the way the influence of the waves is taken into account. Besides, sens-
itivities to certain input parameters vary between the formulas. This could be seen
from runs under different conditions that showed a different relative behaviour of the
various formulas. Total transports between the various transport formulas can easily
differ by up to a factor ten! This illustrates the need for calibration of the formulas
before results from computer computations can be used in specific coastal engineering
cases. The data used for calibration should be representative for the specific site and
hydrodynamics conditions. Examples of such calibration data are observed coastline
changes following a certain ‘event’ (such as the construction of a new harbour along
a coastline, or the damming of a river and the subsequent erosion of the shoreline of
the river delta).
8.2.3. Bulk longshore transport formulas

The general transport formulas, which were compared in the previous section, were
developed from the early seventies onwards (Bijker was the first to include the effect of
waves in a general transport formula, see Eq. 6.27). Before that time only bulk formulas
were available. These bulk longshore transport formulas did not give the distribution
over the surf zone, but only the total transport over the entire width of the littoral
zone. Note however that, although general formulas may resolve the cross-shore dis-
tribution over the surf zone, they are not necessarily better in predicting bulk transport
rates, considering the large uncertainties involved in transport computations. Clear
advantages of bulk transport formulas are that they are robust and easy to calibrate
and apply.

Hs = 1.0 m Hs = 2.0 m Hs = 3.0 m

Tp = 5.0 s Tp = 7.0 s Tp = 9.0 s
1
V [m/s]
0.5
0
3
Sy = 0.01 m 3/s
sy [10 -3 m 3 /m/s]
2
Sy = 0.12 m 3/s
Sy = 0.46 m 3/s
1
0
5
2012
0
z [m]
-5
-10
-1500 -1000 -500 0

x [m]
(a) longshore current velocity, sediment transport rates (with the Bijker transport formula
in deposited volumes) and coastal profile using three different deep-water wave condi-
tions: 𝐻𝑠 = 3 m and 𝑇𝑝 = 9 s, 𝐻𝑠 = 2 m and 𝑇𝑝 = 7 s and 𝐻𝑠 = 1 m and 𝑇𝑝 = 5 s and 𝜑 = 30° in
deep water
5
Sy = 0.27 m 3/s (CERC)
4 Sy = 0.12 m 3/s (BIJK)
[10 -3 m 3 /m/s]
Sy = 0.10 m 3/s (VR)

3
Sy = 0.03 m 3/s (SOULVR)
2
sy
0
-1500 -1000 -500 0
x [m]
(b) sediment transport rates (deposited volumes) for 𝐻𝑠 = 2 m and 𝑇𝑝 = 7 s using four
different transport formulas (CERC, Bijker (BIJK), Van Rijn (VR) and Soulsby Van Rijn
(SOULVR)) with default settings (uncalibrated)
Figure 8.1: Example of model results (longshore current velocity 𝑉 (𝑥), longshore transport
rates 𝑠𝑦 (𝑥) and transport integrated over the cross-shore 𝑆𝑦 ) computed with Unibest-CL+ (https:
//www.deltares.nl/en/software/unibest-cl) for a profile measured in 2012 near Egmond (tran-
sect 7003850 from JARKUS, n.d.) using 𝐷50 = 200 µm and 𝐷90 = 300 µm. The cross-shore distri-
bution of the CERC transport is determined by assuming that the transport is proportional to the
3
third power of the longshore current velocity 𝑉 (𝑥) .

70
SandyDuck 98/02/04
60
transport rate [kg/m/s]

50
longshore sediment
40
B VR EH W Bl AW
30
20
10
0
150 200 250 300 350 400 450 500 550
Figure 8.2: Comparison between calculated and measured cross-shore distribution of longshore
sediment transport rate for one run of the SandyDuck experiment (98/02/04). Adapted from
Bayram et al. (2001). The letters indicate different models, amongst them B: Bijker (Eq. 6.27),
BI: Bailard-Inman (Sect. 6.7.2), VR: Van Rijn (Eq. 6.33 for suspended load plus separate bed load
formula).
In this section, three bulk formulas are presented: 1) the Coastal Engineering Research
Council (CERC) formula; 2) a formula according to Kamphuis (1991); and 3) a formula
according to Bayram et al. (2007).
The CERC formula

Although one of the oldest longshore transport formulas, the CERC formula is still
widely used. It was developed by the Coastal Engineering Research Council (CERC)
of the American Society of Civil Engineers (ASCE). The development of this formula
took place in the late 1940s, well before longshore current theory was developed. The
formula was calibrated using a large number of prototype and laboratory measure-
ments.
The CERC formula gives the bulk longshore sediment transport – the total longshore
sediment transport over the breaker zone – due to the action of waves approaching
the coast at an angle. Hence, only the effect of the wave-generated longshore currents
is included; tidal currents or other alongshore currents are not considered. If the long-
shore current is exclusively driven by waves, one can imagine that both the sediment
concentration and longshore current velocity can be related in some way to the incid-
ent wave conditions. This is reflected in the CERC formula which reads in its most
general form:
𝐼 𝐾
𝑆= = (𝐸𝑛𝑐)𝑏 cos 𝜑𝑏 sin 𝜑𝑏 (8.4)
𝜌𝑔(𝑠 − 1)(1 − 𝑝) 𝜌𝑔(𝑠 − 1)(1 − 𝑝)
where:

𝐼 the immersed (underwater) weight of sediment transported, cf. N/s

Eq. 6.17
𝑆 the deposited volume of sediment transported m3 /s
𝜌 density of the water kg/m3
𝑠 the relative density of the sediment 𝜌𝑠 /𝜌 −
𝑝 porosity −
𝑔 gravitational acceleration m/s2
𝐾 coefficient −
𝐸 wave energy J/m2
𝑐 the wave phase velocity m/s
𝑛 the ratio between group and phase velocity −
𝜑 the wave angle of incidence −
𝑏 subscript referring to conditions determined at the outer edge of −
the breaker zone
Equation 8.4 is a dimensionally correct form of the CERC formula as presented by

Komar and Inman (1970). The original CERC formula had a dimensional coefficient
(such that the coefficient was dependent on the units used) and was formulated in
terms of US Customary Units. The Komar and Inman form does not have this problem.
According to Eq. 8.4 the bulk longshore sediment transport rate 𝑆 is a function of
the product of the energy flux (𝐸𝑛𝑐)𝑏 at the point of breaking (see also Eq. 5.4) and
cos 𝜑𝑏 sin 𝜑𝑏 . This term 𝑃 = (𝐸𝑛𝑐)𝑏 cos 𝜑𝑏 sin 𝜑𝑏 was named ‘longshore component of
wave power’ and its physical interpretation has been the topic of many discussions
over the years. It was not until 1972 – years after the original presentation of the
CERC value – that Longuet-Higgins related 𝑃 to the then new concept of radiation
stresses (Sect. 5.5.2). From Eq. 5.53c and Fig. 5.30 it becomes clear that 𝑃 is the shear
component of the radiation stress at the breaker point, multiplied by 𝑐 at the breaker
point: 𝑃 = 𝑆𝑦𝑥,𝑏 𝑐𝑏 . We know that the cross-shore gradient in radiation shear stress is
responsible for driving the longshore current (Eq. 5.72). The radiation shear stress at
the point of breaking can therefore be seen as this driving force integrated over the
surf zone. But that still does not explain how the product of 𝑆𝑦𝑥,𝑏 and 𝑐𝑏 should be
interpreted. The CERC formula was probably rather intuitively derived. Inman and
Bagnold (1963) derived a bulk longshore transport formula in a more fundamental way,
namely based on an energetics approach, and found a formula similar in form to the
CERC equation. This is demonstrated in Intermezzo 8.1.
Often the CERC equation is rewritten in terms of wave height parameters by substi-
tuting 𝐸 = 18 𝜌𝑔𝐻𝑏2 . Shallow-water wave-breaking is assumed, such that 𝑛𝑏 can be
taken equal to 1 and 𝑐𝑏 ≈ √𝑔ℎ𝑏 in which ℎ𝑏 is replaced with 𝐻𝑏 /𝛾 . For the breaker

index 𝛾 , a value of 0.78 is often assumed. (Sect. 5.2.5). Using the double-angle formula
2 cos 𝜑𝑏 sin 𝜑𝑏 = sin 2𝜑𝑏 , the CERC formula Eq. 8.4 can be expressed as:
𝐾 𝑔
𝑆= sin 2𝜑𝑏 𝐻𝑏2.5 (8.5)
16(𝑠 − 1)(1 − 𝑝) √ 𝛾
For irregular waves, the value of the coefficient 𝐾 depends on whether, for the wave
height at breaking 𝐻𝑏 , the root-mean-square wave height 𝐻𝑟𝑚𝑠 or the significant wave
height 𝐻𝑠 is used. A value of 𝐾for 𝐻𝑟𝑚𝑠 = 0.77 was derived from a field study by Komar
and Inman (1970) corresponding to 𝐻𝑟𝑚𝑠 . This is the value commonly seen in long-
shore transport rate computations. In the Shore Protection Manual (now CEM), the
US Army Corps of Engineers mentions a value of 𝐾for 𝐻𝑟𝑚𝑠 = 0.92. In more recent stud-
ies, Schoonees and Theron (1993, 1996) suggested a significantly lower value of the
coefficient, namely about 0.5, based on a re-examination of available field data. For a
specific project, it is best to determine the coefficient by calibration. Engineers often
prefer to use, for 𝐻𝑏 , the significant wave height 𝐻𝑠 at breaking. In that case the value
5/2
1
of the coefficient 𝐾for 𝐻𝑠 is smaller, viz. 𝐾for 𝐻𝑠 = ( ) 𝐾for 𝐻𝑟𝑚𝑠 ≈ 0.4𝐾for 𝐻𝑟𝑚𝑠 (since
√2
for a Rayleigh distribution: 𝐻𝑠 = √2𝐻𝑟𝑚𝑠 , see Table 3.1).
Intermezzo 8.1
Bulk longshore transport based on the energetics approach
Inman and Bagnold (1963) applied the energetics concept of Bagnold (1963), as
explained in Sect. 6.7.2, to the littoral zone. In the surf zone, the wave oscillatory
motion, with angle of incidence 𝜑, is thought to set an amount of sediment into
motion without resulting in a net transport. The sediment supported by the wave
action (with a total immersed weight 𝑊 ) is passively advected with a representat-
ive longshore current velocity 𝑉 , such that the bulk immersed weight longshore
transport rate is given by 𝐼 = 𝑊 𝑉 .
Per unit crest width, the wave power (rate of transport of energy) that is available
for dissipation in the entire surf zone is equal to (𝐸𝑐𝑔 )𝑏 , see Eq. 5.3, with 𝑏 referring
to the values at the breaker line. The crucial assumption now is that a proportion
𝜀 of the wave power is dissipated by means of bottom friction and used in trans-
porting sediment as bed load. The mean frictional force applied to the whole of
the sediment bed per unit crest width is proportional to (𝐸𝑐𝑔 )𝑏 /𝑢̂𝑏 , where 𝑢̂𝑏 is
mean frictional velocity relative to the bed within the surf zone and is assumed to
be proportional to the orbital velocity near the bottom just before wave-breaking.
Per unit shoreline, the frictional force is proportional to (𝐸𝑐𝑔 )𝑏 cos 𝜑𝑏 /𝑢̂𝑏 .
The frictional resistance of the sediment is equal to 𝜇𝑊 , in which the friction
coefficient 𝜇 = tan 𝜑𝑟 is the tangent of the internal angle of repose indicated by 𝜑𝑟 .

Equating the frictional force on the bed and the resisting force leads to 𝑊 =
𝜀(𝐸𝑐𝑔 )𝑏 cos 𝜑𝑏 /(𝜇 𝑢̂𝑏 ). Hence, according to Inman and Bagnold:
(𝐸𝑐𝑔 )𝑏 cos 𝜑𝑏
𝐼 = 𝑊𝑉 = 𝜀 × 𝑉 (8.6)
𝜇 𝑢̂𝑏
Let us now rewrite the analytical result for the longshore current velocity (Eqs. 5.82
and 5.83) using the shallow-water approximation Eq. 5.23 for the orbital velocity.
This gives for the maximum longshore current velocity at the breaker line (a rep-
resentative longshore current velocity 𝑉 at a mid-surf zone location would have
half this maximum value):
5 tan 𝛽
𝑉𝑏 = 𝜋 𝑢̂ sin 𝜑𝑏 (8.7)
8 𝑐𝑓 𝑏
Substitution in Eq. 8.6 yields:
tan 𝛽
𝐼 = constant × (𝐸𝑛𝑐)𝑏 cos 𝜑𝑏 sin 𝜑𝑏 (8.8)
𝑐𝑓 𝜇
Neglecting dependencies on the beach slope and the physical properties of the
sand yields:
𝐼 = constant × (𝐸𝑛𝑐)𝑏 cos 𝜑𝑏 sin 𝜑𝑏 (8.9)
This is identical in form to the CERC equation Eq. 8.4. We have however not been
totally consistent, since in the derivation of Eq. 8.7 we assumed cos 𝜑𝑏 ≈ 1.
Equation 8.5 is a very practical form of the CERC equation. It shows that, as long as
the other parameters are constant, the transport magnitude increases with increasing
wave angle at the breaker point until a maximum is reached at 𝜑𝑏 = ±45° (in practice
𝜑𝑏 will be in the range −20° to 20° due to refraction). Interesting is also that the long-
shore transport is proportional to the wave height to the power of 2.5. This can be
interpreted, using Intermezzo 8.1, as a sediment load proportional to 𝐸𝑏 and hence to
𝐻𝑏2 , that is transported with a velocity proportional to 𝑢̂ = 21 √𝑔𝛾 𝐻𝑏 and hence to √𝐻𝑏 .
The breaking parameter is often taken as a constant (0.78). However, from Sect. 5.2.5
it is known that the breaking parameter increases with increasing Iribarren parameter
(Eq. 5.20), hence with increasing relative steepness of the slope. A larger relative bot-
tom slope (larger bottom slope or longer wave) would therefore slightly decrease the
sediment transport, at least according to the CERC formula. This is, however, only a

weak dependence in the CERC equation, and for all practical purposes one may con-
sider the formula to be independent of wave period or bottom slope.
In some applications it is more practical to use deep-water wave parameters in the
equations. The conversion to deep-water parameters is quite straightforward, when
using an important finding of Sect. 5.5.5. There, it was explained that if the water depth
contours in the area are all parallel, 𝑆𝑦𝑥 does not vary outside the breaker zone. Hence:
𝑃 = 𝑆𝑦𝑥,𝑏 𝑐𝑏 = 𝑆𝑦𝑥,0 𝑐𝑏 . We can thus evaluate 𝑆𝑦𝑥 using deep-water wave characteristics,
which means that all wave parameters can be taken as deep-water parameters, except
𝑐𝑏 . Remember that in deep water we need to take 𝑛0 = 12 . This yields for straight,
parallel depth contours:
𝐾
𝑆= 𝑐𝑏 sin 2𝜑0 𝐻02 (8.10)
32(𝑠 − 1)(1 − 𝑝)
There are many equivalent forms of the CERC equation and all have at least one wave
parameter at the breaker line. Due to the different representations, any application
of the CERC formula must be made carefully (like, of course, all sediment transport
formulas). This holds in particular for the value of the coefficient, the choice of 𝐻𝑟𝑚𝑠 or
𝐻𝑠 and the choice between deep-water wave parameters and parameters at the breaker
line.
Limitations of the CERC formula

Due to its simplicity the CERC formula can be helpful in understanding and solving
many practical problems. However, the simplicity of the CERC formula has the fol-
lowing limitations (some of which have already been mentioned):
1. Only the wave-induced longshore current is taken into account; all other along-
shore current driving forces, such as tidal currents, are ignored. In order to take
the latter into account, more general transport formulas need to be applied. In a
formula of the form Eq. 8.6, in principle, also tidal and wind-driven alongshore
currents could be used;
2. The sand transport is independent of sand properties such as grain size. Also, the
beach slope and hence the type of breakers is ignored (although the breaker
index may be assumed to be dependent on the breaker type). Eq. 8.8 already
suggested a dependency on beach slope and grain properties. To include these
variables, the formula of Kamphuis (below) can be used;
3. Only the total sediment transport in the breaker zone is given. It is often of prac-
tical importance to know how this transport is distributed over the width of
the breaker zone, for instance if bars are present in the coastal profile, or if
coastal structures are considered that do not entirely cover the breaker zone
(such as groynes: see Ch. 10). However, this distribution could be estimated
from a distribution for the longshore current velocity and wave-stirring capacity.
For example, Unibest-CL+ determines the cross-shore distribution of the CERC

transport by assuming that the transport is proportional to the third power of

the longshore current velocity 𝑉 (𝑥)3 . This procedure is followed for the CERC
transport in Figs. 8.1b and 8.5b.
Grain size and beach slope

Dean et al. (1982) suggested that the coefficient 𝐾 in Eq. 8.5 should be dependent on
the grain size, see Fig. 8.3.
1.5
K [-]
0.5
0
0 0.2 0.4 0.6 0.8 1
diameter D [mm]
Figure 8.3: Variation in 𝐾 with grain size (denoted by 𝐷) according to Dean et al. (1982). Note
that the dependency of 𝐾 on 𝐷 is convincing only for larger grain sizes.
Such a relation is not unexpected since the CERC formula only has the wave character-
istics and one can expect the sediment transport to depend on the sediment properties
too, with larger transport for smaller grain sizes. It is often believed that the transport
should be inversely proportional to the grain size to a power of about three, hence
𝑆 = 𝑓 (𝐷 −3 ). The CERC formula was originally derived for beaches with uniform sand
ranging between 175 µm to 1000 µm and thus represents an average over this range of
conditions.
Kamphuis (1991) made an analysis of field and lab data and suggested an alternative
bulk longshore transport formula that includes the effects of not only grain size but
beach slope and wave steepness as well.
His formula for the immersed mass of transported sediment 𝐼𝑚 = 𝜌(𝑠 − 1)(1 − 𝑝)𝑆 (see
Eq. 6.17) reads:
2 𝑇 1.5 (tan 𝛼 ) 0.75 0.6

𝐼𝑚 = 2.27𝐻𝑠,𝑏 𝑝 𝑏 𝐷 −0.25 (sin 2𝜑𝑏 ) (8.11)
with tan 𝛼𝑏 the beach slope at the breaker point.

The wave period, grain size and beach slope (absent in the CERC equations) now in-
fluence the longshore transport rate. Note that the dependency on the wave angle

seems relatively weak, compared to the CERC formula. The grain size proportional-
ity with 𝐷 −0.25 also seems rather weak. However, it should be noted that the vari-
ables in the empirically derived Eq. 8.11 are interdependent. For example: given con-
stant hydraulic conditions, coarser beach materials tend to form steeper slopes1 ; gravel
beaches are far steeper than sand beaches. On the other hand, if the beach material is
of a given diameter, then higher waves tend to result in flatter beach slopes. Similarly,
the wave parameters 𝑇𝑝 , 𝜑𝑏 and 𝐻𝑠,𝑏 are interrelated.
Recent bulk longshore transport formula

Bayram et al. (2007) presented a new bulk formula based on the energetics concept, sim-
ilar to Inman and Bagnold’s model in Intermezzo 8.1. By assuming that the dissipated
fluid power maintaining the sediment load is dissipated by bottom friction, Inman and
Bagnold considered bed load sediment transport. By contrast, Bayram et al. assume
that suspended sediment transport is the dominant mode of transport in the surf zone,
as a result of turbulence induced by breaking waves. They assume that the dissipated
fluid power per unit shore length used in suspending sediment is 𝜀 (𝐸𝑐𝑔 )𝑏 cos 𝜑𝑏 (sim-
ilar to Intermezzo 8.1). In contrast with Intermezzo 8.1, the work done (per unit time),
in maintaining the suspended load at a certain level, is the product of 𝑊 times the fall
velocity 𝑤𝑠 . Hence: 𝑊 = 𝜀 (𝐸𝑐𝑔 )𝑏 cos 𝜑𝑏 /𝑤𝑠 .
Instead of Eq. 8.6 we now get:
𝜀 (𝐸𝑐𝑔 )𝑏 cos 𝜑𝑏
𝐼 = 𝑊𝑉 = 𝑉 (8.12)
𝑤𝑠
For 𝑉 an analytical solution is used for the mean longshore current velocity that is de-
rived using a Bruun-Dean’s equilibrium profile (Eq. 7.1) instead of a constant bed slope.
Based on comparison with field and laboratory data, Bayram et al. (2007) conclude that
the predictive capability of their formula is higher than of the CERC, Inman-Bagnold
and Kamphuis formulas. A recent attempt to improve the accuracies of these bulk
longshore transport formulas is Mil-Homens et al. (2013).
8.2.4. The (S,𝜑)-curve

2.5 sin 2𝜑 .
According to Eq. 8.5, bulk longshore sediment transport 𝑆 is proportional to 𝐻𝑠,𝑏 𝑏
The power 2.5 was explained in Intermezzo 8.1 as a sediment load proportional to 𝐻𝑠,𝑏 2
transported by a wave-induced longshore current velocity proportional to √𝐻𝑠,𝑏 . Both

𝐻𝑠,𝑏 and 𝜑𝑏 depend on the offshore wave height, wave period and angle of incidence
𝜑0 .
1
Kamphuis (2000) mentions this as the probable reason for the absence of the grain size and beach slope
dependency in the CERC formula. In addition, for grain sizes in the order of 200 µm, the grain size
dependency is not so clear according to Fig. 8.3.

The angle of incidence is relative to the local depth contours, which, in the nearshore
zone, are often approximately parallel to the shoreline. Given a relatively constant
wave climate along a stretch of coast, alongshore variations in coastline orientation
cause gradients in longshore sediment transport and hence cause the coastline orient-
ation to change over time. The relationship between sediment transport 𝑆 and angle
of wave incidence 𝜑 is therefore a central concept in shoreline modelling (see also
Sect. 8.3). In order to examine the effect of the angle of incidence on longshore sed-
iment transport, we consider the situation of straight depth contour lines parallel to
the shoreline and compute the sediment transport rate 𝑆 from Eq. 8.10 for a constant
offshore wave height and period and for a range of angles of attack 𝜑0 (Example 8.1).
Example 8.1
Bulk longshore sediment transport rate 𝑆 for different values of 𝜑0
Input parameters
Offshore wave height: 𝐻𝑟𝑚𝑠,0 = 2 m
Wave period: 𝑇 = 7s
Required Longshore sediment transport rate 𝑆 for different values of 𝜑0
Method We consider the simple situation of straight depth contour lines parallel
to the shoreline. In that case we can use the relation between 𝑆 and the deep-water
wave angle 𝜑0 (Eq. 8.10). For 𝐻0 we use the deep-water root-mean-square wave
height 𝐻𝑟𝑚𝑠,0 and a corresponding value for 𝐾 of 𝐾𝑟𝑚𝑠 ≈ 0.7. With a value for the
porosity 𝑝 = 0.4 and the relative density 𝑠 = 2.65, this equation can be written as:
2
𝑆 ≈ 0.02𝑐𝑏 𝐻𝑟𝑚𝑠,0 sin 2𝜑0 (8.13)
We compute ℎ𝑏 as 𝐻𝑟𝑚𝑠,𝑏 /𝛾 with 𝛾 = 0.8. This is a rather large value for the breaker
index based on 𝐻𝑟𝑚𝑠 , but consistent with the determination of the coefficient value
for 𝐾 .
First choose 𝜑0 . Then calculate 𝑐𝑏 using for instance Table A.3 (alternatively, the
dispersion relation can be solved using a small computer routine or an explicit
approximation). The computation of ℎ𝑏 requires an iteration (choose ℎ𝑏 , calculate
𝐻𝑟𝑚𝑠,𝑏 according to Table A.3, check whether 𝐻𝑟𝑚𝑠,𝑏 /ℎ𝑏 = 𝛾 ; if not, choose new ℎ𝑏 ,
repeat process). The sediment transport rate can now be computed using Eq. 8.13.
The results can be found in Table 8.1 and Fig. 8.4.

Table 8.1: Bulk longshore transport rates as a function of the deep-water wave angle 𝜑0 . The
maximum transport is found for 𝜑0 = 43°. Note that 𝐻𝑏 = 𝛾 ℎ𝑏 .
𝜑0 [°] ℎ𝑏 [m] 𝜑𝑏 [°] 𝑐𝑏 [m/s] 𝑆 [m3 /s]

5 2.72 2.27 4.97 0.069
10 2.71 4.52 4.96 0.136
15 2.69 6.73 4.95 0.198
20 2.66 8.87 4.93 0.253
25 2.63 10.91 4.90 0.300
30 2.59 12.85 4.86 0.337
35 2.54 14.64 4.82 0.362
40 2.48 16.26 4.76 0.375
43 2.42 17.14 4.72 0.377
45 2.40 17.68 4.69 0.376
50 2.31 18.87 4.61 0.364
55 2.21 19.79 4.52 0.340
60 2.09 20.42 4.40 0.305
65 1.95 20.69 4.26 0.261
70 1.79 20.58 4.09 0.210
75 1.59 19.99 3.87 0.155
80 1.35 18.77 3.57 0.098
85 1.01 16.46 3.11 0.043
maximum
φ0 = 43˚
Sy [m3/s]
φ0 [˚]
Figure 8.4: Bulk longshore transport rates 𝑆 as a function of the deep-water wave angle 𝜑0 .
The dotted line indicates 𝜑0 = 43°. The maximum transport occurs for deep-water wave
angles somewhat smaller than 45° (43° in this case). The angle 𝜑 is positive for a transport
𝑆 in the direction of a positive alongshore coordinate axis and vice versa.

Figure 8.4 shows the longshore sediment transport against the wave angle in deep
water. This is a so-called (𝑆, 𝜑)-curve; it gives the transport as a function of the wave
angle for a given set of wave conditions. The (𝑆, 𝜑)-curve indicates that the maximum
longshore transport occurs for an angle somewhat smaller than 𝜑0 = 45°. The term
sin(2𝜑0 ) in Eq. 8.10 suggests a maximum of S at exactly 𝜑0 = 45°. However, the wave
propagation speed at the breaker line (as a function of the local depth) also depends
on the deep-water wave angle, as it depends on the wave refraction from deep water
to the breaker line: for larger values of 𝜑0 , wave-breaking takes place at smaller water
depths, reducing 𝑐𝑏 (Sect. 5.2.3). This slightly affects the angle at which the maximum
occurs. The longshore transport reduces to zero if 𝜑0 goes to 0° or 90°. An angle 𝜑0 = 0°
implies normally incident waves at the breaker line and hence only wave stirring and
no longshore current. Note that for small angles, 𝜑0 up to approximately 20° to 30°, the
longshore transport varies almost linearly with the wave angle. Table 8.1 furthermore
shows that, due to refraction, the angle of incidence at the breaker line is smaller than
about 20° for all values of 𝜑0 .
8.2.5. Yearly-averaged sediment transport

In Sect. 8.2.4, we discussed the longshore transport rate for a single wave state (repres-
ented by its root-mean-square wave height and wave period) and examined the effect
of the angle of incidence on the transport rates. The maximum computed transport
rate (for a deep-water wave angle of 43°) amounts to 0.38 m3 /s. In the (unrealistic) situ-
ation that this wave condition is present all year round, we would find a net longshore
transport of the order 10 × 106 m3 per year.
In order to obtain a realistic estimate of the long-term average annual longshore sedi-
ment transport rate, we need to take the variability of the wave climate into account.
Wave heights, periods and angles are weather-dependent. Different wave heights and
periods result in different transport magnitudes. Changes in the wave angle may lead
to transports that not only have different magnitudes, but opposite directions as well.
The wave climate variability can be taken into account by schematization of the wave
climate in several classes (see Sect. 3.6). This requires that the wave climate is divided
into sectors of for instance 30°, 0.5 s 𝑇𝑝 and 0.5 m 𝐻𝑠 with a certain percentage of occur-
rence. For morphological computations, such a full wave climate is then often reduced
to a limited number of wave conditions. For the reduced wave climate to be represent-
ative of the full climate, it must reflect the dominant wave directions and also represent
a morphological criterion of interest. Such a criterion could for instance be that the net
longshore sediment transport at some coastal sections for the reduced climate must be
identical to the net transport rate for the full climate. A reduced wave climate typic-
ally has 10 conditions, but sometimes a single representative wave condition is chosen
that reproduces the net longshore transport rate of the full climate. The net longshore
transport rate is defined as the residual transport rate as a result of all conditions and
is generally much smaller than the gross longshore transport rates up and down the

coast. Of course, for a single wave condition, the gross and net longshore transport
rates are identical.
As an example we consider the wave climate for the Dutch coast, which is charac-
terised by relatively short-period waves (of about 5 s). Waves offshore exceed 2 m
approximately 10 % of the time and 3 m approximately 2 % of the time. Most waves
arrive arrive from directions between southwest and northwest, so that the average
wave incidence is almost normal to the coast. The highest waves come from the north-
west direction, because of the longer fetches in this sector. Swell is also predominantly
from the northwest direction. The orientation of the central Dutch coast is approxim-
ately N-S (e.g. 7°N at Egmond, or 277°N for the shore-normal direction, and 15°N at
IJmuiden). Along this coast, storms arriving from the northwest generate longshore
currents and a southwards longshore transport. Conversely, waves from the southw-
est result in a northward-directed transport.
Figure 8.5 shows the cross-shore distribution of the longshore transport around Noord-
wijk according to model computations with Unibest-CL+. A distinction is made between
northward and southward transport. For engineering purposes, we are interested in
the gross transport rates as well as in the net transport rate (which is directed north-
ward for most of the central Dutch coast).
Gradients in net transport rates determine the long-term evolution of a stretch of coast.
Gross transport rates are important to determine the coastal response near breakwa-
ters and groynes. This is because these structures provide shadow zones on either side
for part of the wave conditions (see Sect. 8.4.2). Especially when the gross longshore
transport rates along a coast are quite similar in magnitude, the magnitude (and dir-
ection!) of the net longshore transport is very difficult to determine with sufficient
accuracy. This is because two large and inaccurate numbers are subtracted. Note fur-
ther that net longshore transport rates (per metre width) are generally much higher
than net cross-shore transport rates (per metre width).
A strongly reduced wave climate for the Dutch coast, taking into account the two
dominant directions, may for instance consist of the following two wave conditions:
1. From southwesterly directions (say 225°N), occurring 20 % of the time and with
𝐻𝑠 = 1.7 m and 𝑇𝑝 = 5 s;
2. From northwesterly directions (say 315°N), occurring 8 % of the time and with
𝐻𝑠 = 1.8 m, 𝑇𝑝 = 6 s.
For a coastline orientation of about 15°N and hence a shore-normal of 285°N (IJmuiden),
this means angles of wave incidence in deep water of 60° and −30° for condition 1 and
2 respectively. This would yield (using Eq. 8.13 as in Example 8.1) a net northward
directed sediment transport of 250 000 m3 /yr as a result of:
1. Northward transport of ±530 000 m3 /yr;
2. Southward transport of ±280 000 m3 /yr.

2
[10 3 m 3 /m/yr]
1
-1 Sy = 901 000 m 3/yr (northward)

Sy = 375 000 m 3/yr (net)
sy
-2
Sy = -526 000 m 3/yr (southward)
-3
2
0
-2
z [m]
-4
-6
-8
-1500 -1000 -500 0
x [m]
(a) net transport rates 𝑠𝑦 (𝑥) and 𝑆𝑦 (integrated over the cross-shore) using the Bijker trans-
port formula and separated into gross (northward and southward) components
6
Sy = 641 000 m3/yr (CERC)
5
Sy = 375 000 m3/yr (BIJK)
[10 3 m 3 /m/yr]
4 Sy = 225 000 m3/yr (VR)

Sy = 45 000 m 3/yr (SOULVR)
3
2
sy
0
-1500 -1000 -500 0
x [m]
(b) net transport rates 𝑠𝑦 (𝑥) and 𝑆𝑦 (integrated over the cross-shore) using the transport
formulas according to CERC, Bijker (BIJK), Van Rijn (VR) and Soulsby Van Rijn (SOULVR)
without calibration (i.e. default settings are used)
Figure 8.5: Cross-shore distribution of longshore transport at Noordwijk according to

model computations with Unibest-CL+ (https://www.deltares.nl/en/software/unibest-cl) using
a yearly wave climate from a wave look-up table developed by Deltares for the Building with
Nature project HK 3.2. The cross-shore distribution of the CERC transport is determined by as-
suming that the transport is proportional to the third power of the longshore current velocity
3
𝑉 (𝑥) .
Note that we have chosen the two conditions such that: 1) they reflect the prevailing
wave conditions; and 2) the net transport magnitude and direction correspond with es-
timates based on sediment balance considerations. As a reference: the net yearly trans-
port in the central part of the Dutch coast (from Wassenaar to Zandvoort) is thought
to be around 200 000 m3 /yr northwards. We could further finetune our wave condi-
tions to arrive at the figure of 200 000 m3 /yr rather than 250 000 m3 /yr. This is the

8.3. Calculation of coastline position 379
process of defining a morphologically representative reduced wave climate. Note that

in practice we would generally take more conditions into account.
8.3. Calculation of coastline position

8.3.1. Introduction
On high-energy coasts, long-term (years to decades) shoreline changes are predom-
inantly due to human-induced longshore effects. Cross-shore movement of sediment
typically occurs on short timescales (days) and has little direct influence on the longer-
term changes in beach position, unless material is permanently lost from or introduced
to the system. Possible sediment sinks are offshore canyons from where any depos-
ited sediment cannot return, the hinterland (through aeolian transport) or sand min-
ing from the beaches (for instance for construction purposes). Examples of sediment
sources are dredge disposal or sediment input from rivers. Often however, longshore
effects are the most important cause of shoreline changes on the longer timescales.
Coastal change due to longshore processes will be the focus in the following.
Please note that the coastline is aligned with the 𝑥-axis in the remainder of this
chapter. This is different from the other parts of this book.
The long-term coastline changes are governed by net, yearly-averaged longshore sedi-
ment transport rates, averaged over several wave conditions with varying magnitudes
and directions. In computations these are taken into account by a schematised wave
climate with a limited number of conditions (as discussed in Sect. 8.2.5). The very pres-
ence of sediment transport does not lead to either erosion or deposition. Indeed, if one
considers a portion of beach as sketched in plan in Fig. 8.6, the coastline will remain
stable as long as 𝑆in is equal to 𝑆out . On the other hand, if 𝑆out is greater than 𝑆in (𝑆
𝑑𝑆
increases as we move along the beach, 𝑑𝑥 > 0, or sediment transport divergence), ma-
terial must be eroded in the area under consideration in order to maintain a sediment
mass balance. The coastal profile, including the shoreline, will recede. The shoreline
will move seawards (accretion) if the longshore sediment transport decreases in the
𝑑𝑆
transport direction ( 𝑑𝑥 < 0 or sediment transport convergence). A sediment mass bal-
ance relates morphological changes to sediment transport gradients, and to sediment
sinks and sources, see e.g. Eq. 1.1.
There are different ways to solve coastal changes from a sediment balance (Eq. 1.1):
• In complex morphological computer models (such as Delft3D), the sediment bal-
ance is computed for each cell in a fine grid that covers the area of interest. At
every time step, a field of sediment transport rates is computed, based on the
relevant hydrodynamic parameters. Subsequently, the bed levels are updated
for every grid cell, based on sediment continuity. Such detailed morphological

shoreline at t = 0 Sin Sout
shoreline at t > 0 if Sout > Sin
Figure 8.6: Longshore transport continuity.
models are generally used in complex areas or in complex applications in which

cross-shore and longshore transports are important. They are relatively expens-
ive to run and it is complex to interpret the computational results. At Delft Uni-
versity of Technology, students will be taught to use these models in Coastal
Dynamics II (CIE4309);
• If wave-induced longshore transport is the dominant mechanism, we can sim-
plify the problem. The coarsest schematization is made by the single-line or one-
line theory, in which the behaviour of the coast is mapped onto a single line, the
coastline. In this approach it is assumed that the shape of the coastal profile itself
does not change over the considered period of time or along the coast. The coast
can thus be schematised as one single line, which moves seaward (accretion) or
landward (erosion), depending on the sediment balance. This is what we (impli-
citly) assumed in Fig. 8.6. The single-line theory is discussed in Sect. 8.3.2 and
analytical solutions in Sect. 8.3.3. It gives a valuable insight into the principles
of coastline dynamics;
• Multiple (two and more) line theories (Sect. 8.3.4) are based on the same prin-
ciple, but now the coastal (cross-shore) profile is schematized into a number of
sections (depth zones), which can each be represented by one line in the along-
shore direction. This is a method to account for the fact that different depth
zones respond differently to changing sediment budgets.
8.3.2. Single-line theory

Beach profile schematization
The basic assumption is that the shape of the cross-shore profile does not change in
time, which implies an equilibrium profile of arbitrary shape. As the coast erodes
or accretes, the entire profile moves seaward or landward with a horizontal distance
𝑎. Figure 8.7 shows that this assumption also requires the assumption that there is a
more or less horizontal part in the underwater profile; otherwise, unrealisticly large
quantities of sand would be required for a small seaward movement of the shoreline.
In practical applications, the closure depth (see Sects. 1.5.1 and 7.2.3) is taken as the
lower limit of the coastal profile. Depth changes seaward of this closure depth are
assumed to not directly contribute to shoreline dynamics. The closure depth is gov-
erned by the highest waves that may occur in a certain period of time (normally several

Figure 8.7: Beach profile schematization in single line or one-line theory.
years). For the Holland coast (the central Dutch coast) the order of magnitude of the
closure depth is MSL−6 m to MSL−12 m.
The upper limit of the coastal profile that needs to be taken into account depends –
on shorter timescales – on whether the coast is eroding or not. For an eroding coast,
the upper limit should be the dune height, allowing volume changes of the dunes. In
the case of accretion, the upper limit is determined by the representative wave run-up
added to the high-water level. (It is thus assumed that the build-up of dunes by aeolian
transport is a very slow process). This is usually about 1 m to 3 m above MSL. In the
discussion below, we simply assume that the active profile has a height 𝑑.
The wave characteristics in the assumed horizontal part are needed to compute changes
in the schematised beach. Note that these characteristics are not necessarily equal to
the deep-water characteristics.
Governing equations
Consider a stretch of a beach that is changing – either eroding or accreting. The 𝑥-axis
is roughly parallel to the coastline, the 𝑦-axis is normal to the coast, see Fig. 8.8. The
position of the coastline is at 𝑦 = 𝑌 .
z ΔY
Δx d
ΔY
y
Y
x
Figure 8.8: Longshore transport 𝑆𝑥 and coastal change Δ𝑌 in a coastline model.

If during a short time interval Δ𝑡 the shoreline moves forward with Δ𝑌 , the amount of
accumulated volume over a distance Δ𝑥 is:
Δ𝑥Δ𝑌 𝑑 (8.14)
The net volume of sediment entering the segment with length Δ𝑥 during time Δ𝑡 is:
𝜕𝑆
− Δ𝑥Δ𝑡 (8.15)
𝜕𝑥
Equating the accumulation of sediment and the net inflow of sediment yields the fol-
lowing continuity equation:
𝜕𝑌 1 𝜕𝑆𝑥
+ =0 (8.16)
𝜕𝑡 𝑑 𝜕𝑥
In Sect. 8.2.3 we have seen that the angle of wave attack relative to the coastline 𝜑 is
an important variable in determining the sediment transport 𝑆𝑥 . Via the chain rule we
can therefore write Eq. 8.16 as:
𝜕𝑌 1 𝜕𝑆𝑥 𝜕𝜑
+ =0 (8.17)
𝜕𝑡 𝑑 𝜕𝜑 𝜕𝑥
The term 𝜕𝑆𝑥 /𝜕𝜑 follows from any longshore transport formula that we choose to use.
In Sect. 8.3.3 we will do this analytically, by assuming that the sediment transport is a
linear function of the angle of incidence. With a more complicated sediment transport
formula, we can empirically determine 𝜕𝑆𝑥 /𝜕𝜑 by varying the angle slightly in the
formula and calculating the transport rates. The term 𝜕𝜑/𝜕𝑥 is directly related to the
shoreline position that we are trying to solve for. This can be understood from Fig. 8.9,
where an originally straight coastline is shown together with the coastline after some
time.
y
wave direction at depth h
−∂Y/∂x
r some time
coastline afte
x
original coastline
Figure 8.9: Shoreline position 𝑌 in original position and after some time. In this figure 𝜑 ′ is the
original angle of wave incidence. After the change in shoreline orientation, this angle has been
reduced with 𝜕𝑌 /𝜕𝑥.

The shoreline has rotated with an angle 𝜕𝑌 /𝜕𝑥. This directly implies a reduction of
the angle of wave attack 𝜑 with 𝜕𝑌 /𝜕𝑥, since the wave angle is relative to the coastline.
Hence, 𝜑 = 𝜑 ′ − 𝜕𝑌 /𝜕𝑥 and 𝜕𝜑 = −𝜕𝑌 /𝜕𝑥. Substituting this in Eq. 8.17 gives:
𝜕𝑌 1 𝜕𝑆𝑥 𝜕 2 𝑌
− =0 (8.18)
𝜕𝑡 𝑑 𝜕𝜑 𝜕𝑥 2
Equation 8.18 is a parabolic partial differential equation for the coastline position 𝑌
and must generally be solved numerically. This equation is solved in coastline devel-
opment models (such as Unibest-CL+ or Genesis, see Szmytkiewicz et al. (2000)). In
this equation the angle 𝜑 is the angle of incidence at the closure depth.
In Sect. 8.2.4, the (𝑆, 𝜑)-curve was introduced. We determined this curve by changing
the angle of wave attack relative to a fixed coast. Since 𝜑 is the angle of wave attack
relative to the coastline, we could just as well determine the curve by changing the
coastline orientation relative to the waves, which is what is done in coastline model-
ling. The longshore sediment transport rate as a function of the changes in coastline
orientation must be determined for the stretch of coast and the wave climate under
consideration. This is done by calculating the longshore transport rates for the ini-
tial coastline orientation. Subsequently, the coastline orientation is changed a little bit
and the longshore transport calculation is carried out again for the entire wave climate.
This is equivalent to constructing a (𝑆, 𝜑)-curve. An example of such a procedure is
given in Intermezzo 8.2.
Intermezzo 8.2 Transport curve as computed in a coastline model
A transport curve in a coastline model describes the relation between the yearly
averaged longshore transport 𝑆𝑥 and the coastline orientation. To determine this
curve, the full local wave climate needs to be taken into account, generally schem-
atised as a limited set of individual conditions (see Sect. 8.2.5) that are each de-
scribed by a number of parameters, namely:
• The significant wave height 𝐻𝑠 (e.g. at the deep-water boundary);
• The peak wave period 𝑇𝑝 ;
• The angle of wave approach 𝜑 (e.g. at the deep-water boundary);
• The storm-related set-up ℎ𝑠 ;
• The percentage of occurrence.
Each of the conditions contributes to the yearly-averaged transport, weighted with
their percentage of occurrence. The sum of the percentages of occurrence equals
100 % (or 365 days per year, although one of the conditions may be a zero to small
wave height).

Figure 8.10 shows such a transport curve as computed with the coastline model
Unibest-CL+ (https://www.deltares.nl/en/software/unibest-cl). For this computa-
tion a bottom slope of 1 ∶ 100 was assumed from MSL to MSL−20 m. The sediment
grain size 𝐷50 = 200 µm. The transport zone was assumed to extend to MSL−10 m.
The wave input at MSL−20 m is given in Table 8.2 (wave angles are defined such
that positive wave angles give positive sediment transports). A total of 16 wave
conditions were defined in four wave height classes and four directional sectors.
From the table it can be seen that the number of days per year with hardly any
waves is 365 − 181 = 184.
0.5 490 000 m3/year

[106 m 3 /year]
15°
S
-0.5
-1
-40 -30 -20 -10 0 10 20 30 40

coastline orientation [°]
Figure 8.10: Yearly-averaged transport as a function of changes in coastline orientation.

The reference (initial) coastline orientation is 0°.
Table 8.2: Wave input: number of days with certain wave conditions (combination of wave
height and direction).
direction with respect to shore normal

𝐻𝑠 −45° −15° 15° 45°
0.5 10 10 25 8
1.0 8 15 35 16
2.0 4 10 21 11
4.0 1 1 4 2
For the reference coastline orientation the yearly-averaged longshore transport is

490 000 m3 /yr. If the coastline would rotate with 15°, the total longshore transport
becomes zero. The angle for which 𝑆 = 0 is sometimes called the equilibrium ori-
entation, which is somewhat confusing since equilibrium requires zero gradients
and not necessarily zero transports.

More insight in how to apply Unibest-CL+ can be obtained through e.g. Luijendijk
et al. (2011) and Van der Salm (2013).
Note that the transport curve computed in this way as a function of changing coastline
orientation is not necessarily constant along the coast, for instance where local wave
conditions are affected by wave sheltering. As an example, at either side of a shore-
normal breakwater, waves from certain directions are affected by diffraction, so that
wave heights are reduced and the directions of the wave rays are more shore-normal.
The alongshore variation of the yearly wave climate can be taken into account by
relating a specific wave climate to a specific alongshore position and computing trans-
port curves for various alongshore positions. In areas where the wave conditions are
affected by non-uniform bottom topographies, by natural coastal features like head-
lands, or by man-made structures like breakwaters, the shape of the transport curve
should be allowed to change alongshore.
8.3.3. Analytical solution for accretion near breakwater or jetty

The first person to describe the single line theory was Pelnard-Considère (1956). He
came up with an analytical solution that can be used for a quick assessment of the
effects of a structure, for instance a breakwater, on a coastline.
In order to find analytical solutions, we need to simplify Eq. 8.18. If we restrict ourselves
to small changes in the angle of wave attack, we can assume 𝜕𝑆𝑥 /𝜕𝜑 to be constant.
Such an assumption implies that a segment of the (𝑆, 𝜑)-curve has been replaced by a
straight line, which is reasonable for relatively small changes in 𝜑. We can now write:
𝜕𝑆𝑥
=𝑠 (8.19)
𝜕𝜑
in which 𝑠 is a coastal constant. For −20° < 𝜑 < 20°, 𝑆 is a linear function of the wave
angle, such that 𝑠 = 𝑆/𝜑 ′ (see Sect. 8.2.5). Note that 𝑆 is the transport for the original
angle of incidence 𝜑 ′ .
Substituting Eq. 8.19 in Eq. 8.18 changes the parabolic differential equation to:
𝜕𝑌 𝑠 𝜕 2 𝑌
− =0 (8.20)
𝜕𝑡 𝑑 𝜕𝑥 2
This equation shows that the curvature in the coastline, scaled by the coastal constant,
determines the coastal change. To solve Eq. 8.20, we need one initial condition and
two boundary conditions. Very often these conditions are the position of the coastline
at 𝑡 = 0 (the initial situation) and the sediment transport on both borders of the coastal
area as a function of time.

Consider the construction of a breakwater or a jetty running perpendicular to an ini-

tially straight shoreline (Fig. 8.11) and assume, for simplicity, that the construction
time of the structure is negligible. If the longshore transport is from left to right, then
the coast to the left of the structure is called the updrift coast, whereas the coast to the
right of the breakwater is called the downdrift coast. The breakwater is assumed to
extend into the sea until at least the depth of closure.
H, T Sx = 0
Sx = S x
x = −∞ x=0
Figure 8.11: Accretion of the shore near a breakwater at different times. The wave conditions
given in the sketch refer to the conditions at the assumed horizontal part in the coastal area.
The initial condition is the shape of the coastline at time 𝑡 = 0:
𝑌 = 0, for 𝑡=0 and for −∞<𝑥 <0 (8.21)
One boundary condition is that at a great distance from the breakwater, 𝑥 = −∞, the
sand transport remains constant and equal to its value before breakwater construction:
𝑆𝑥 = 𝑆, for 𝑥 = −∞ and for all 𝑡 (8.22)
The second boundary condition is imposed at the breakwater; it is impermeable to sand

(no sand particles slipping through the pores of the breakwater). This means that the
local longshore transport at the breakwater must be zero (100 % blocking):
𝑆𝑥 = 0, for 𝑥=0 and for all 𝑡 (8.23)
This boundary condition is only valid as long as the breakwater blocks the entire trans-
port, viz. as long as the active zone has not moved past the tip of the breakwater. Note
that a zero transport means that the wave angle with respect to the orientation of the
coastline must be zero as well. Since the offshore wave direction does not change, this
can only be accomplished by a shoreline rotation, such that the wave angle relative to
this rotated coastline becomes zero (Fig. 8.11). The boundary condition at 𝑥 = 0 can
therefore also be written as:
𝜕𝑌 /𝜕𝑥 = 𝜑 ′ , for 𝑥=0 and for all 𝑡 (8.24)

In other words, the beach accretion progresses seaward, always making an angle 𝜑 ′
with respect to the 𝑥-axis at the breakwater. Hence, at the breakwater, the coastline
and the depth contours tend to become parallel to the approaching waves.
With the help of Eq. 8.20, applying the reference system of Fig. 8.11 and using the
initial and boundary conditions mentioned above, the shape of the coastline 𝑌 (𝑥, 𝑡)
on the updrift side of the breakwater can be solved as a function of time. Here, we
only give some characteristics of the resulting solution (see also Fig. 8.12). First, the
outward growth of the coastline at the breakwater 𝐿(𝑡) at 𝑥 = 0 is found to be:
𝜑 ′ 𝑆𝑡 4𝑎𝑡
𝐿(𝑡) = 2 = 𝜑′ (8.25)
√ 𝜋𝑑 √ 𝜋
where:
𝑡 time yr
𝑑 profile height m
𝑎 = 𝑑𝑠 = 𝜑𝑆′ 𝑑 3
(m /yr)/rad/m
B
one
breaker z L
S
x
A C O
L/φ′
Figure 8.12: Accretion geometry. Note that line CB is tangential to the coastline at the breakwa-
ter.
Secondly, it can be shown that the influence of the breakwater (in terms of accretion
of the shoreline relative to the accretion at the breakwater) is negligible at a distance
5√𝑎𝑡 = 2.5√𝜋𝐿/𝜑 ′ from the breakwater. So, the influenced shoreline is dependent
on the steepness 𝑠 of the considered segment of the (𝑆, 𝜑)-curve and the height of the
active zone 𝑑 (both of which are strongly dependent on the wave height). Both the
horizontal extent of the influenced zone and the seaward growth at the breakwater
vary with √𝑡. This means that the progress of the shoreline gradually slows down;
suppose that after a time 𝑡1 the accretion length at the breakwater is 𝐿1 . At 𝑡2 = 2𝑡1
the accretion length has not doubled, but only increased to 𝐿2 = √2𝐿1 . Equation 8.25

furthermore shows that, of course, the accretion speed increases with increasing sedi-
ment transport 𝑆. The surface area OCB in Fig. 8.12 is equal to:
1 𝐿2 2 𝑆𝑡
surface OCB = ′ = (8.26)
2𝜑 𝜋 𝑑
Also, from continuity, the total surface area AOB is:
𝑆𝑡
surface OAB = (8.27)
𝑑
Therefore, 64 % (viz. 2/𝜋) of the total accretion is stored in the shaded area OCB.
For the coastal changes on the lee side of the breakwater (downdrift), an analogous ana-
lytical solution can be derived. In principle, the coastline changes on the updrift side
are mirrored on the downdrift side and the total volume of updrift-accreted sediment
equals the total volume of downdrift-eroded material. This is illustrated in Fig. 8.13.
H, T Sx = 0
Sx= S x
x = –∞
x=0
accretion erosion
dSx / dx < 0 dSx / dx > 0
Sx
initial transport at t = 0
transport at some time
Figure 8.13: Shoreline development on the lee side of a breakwater. As a first approximation we
may say that at the breakwater (𝑥 = 0) the lee-side shoreline orientation reflects the angle of
wave attack (since 𝑆𝑥 = 0), just as was the case for the updrift side. Thus, the shoreline at the
lee side (or the downdrift shoreline) is the mirror image of the updrift shoreline.
In reality, however, there is a difference in the wave attack between the updrift and
downdrift side. At the downstream side, a part of the shore is sheltered from wave
attack by the breakwater. Waves propagating past the end of the breakwater diffract
into the shadow zone (see Sect. 5.2.4). As a result, the wave heights are significantly
lower than in the undisturbed region (see Fig. 8.14). Furthermore, the wave angles are
significantly altered, since the wave rays turn somewhat towards the breakwater. At

the breakwater, the rays run parallel to the breakwater. In principle, at the breakwater
(zero transport), the coastline reorients itself normal to the local waves (hence parallel
to the original coastline). There is, however, an additional shadow zone effect due to
alongshore set-up differences. Due to the alongshore differences in wave height, the
wave set-up in the surf zone changes along the shore (Sect. 5.5.4). The lower wave
heights just behind the breakwater result in a lower set-up at the shoreline close to
the breakwater, than elsewhere along the coast. The alongshore water level gradients
result in a secondary current pattern that – close to the coast – is directed towards the
breakwater (Fig. 5.45) and induce sediment transport towards the breakwater.
H=
50%
H = 100%
H < 50%
Figure 8.14: Wave height reduction and shoreline development on the lee side of the breakwater
with, at the wave ray that separates the shadow zone from the wave zone, a wave height of 50 %
of the original wave height according to linear theory. The dashed line shows the mirrored image
of the accretion. The solid line is the equilibrium orientation, including shadow zone effects.
The resulting downdrift coastline development depends on the combination of all ef-
fects. The transport must increase from 0 at the breakwater to 𝑆 in the undisturbed
region. However, shadow zone effects will reduce the transport magnitude directed
away from the breakwater or may even cause a net transport towards the breakwater.
This leads to a coastline development as shown schematically in Fig. 8.14 (solid line).
Analytical solutions also exist for different boundary conditions at the breakwater and
for different breakwater layouts. For example, after some time a beach can build up
on the updrift side to such an extent that sand will be transported around the break-
water tip. This bypassing of sediment can be taken into account. Analytical solutions
make hand computations of coastline changes possible and hence facilitate a quick
assessment of the impact of certain coastal structures on the shoreline development.
Neither tidal influences (see Fig. 5.60 for tidal currents passing a breakwater) nor vari-
ations along the coast of wave height, wave direction or sediment characteristics can
be taken into account. Besides, shoreline evolution is calculated under the assumption
of steady wave conditions. This means that, as in the above example, only a single
wave condition is taken into account. This limits the applicability of these analytical
solutions, especially in the vicinity of structures where sheltering effects are important
and vary strongly with the wave direction. In Sect. 8.4.2 this is further explained.

Numerical methods based on Eq. 8.18 often include the above-mentioned processes.
These methods still suffer from limitations as a result of the basic assumption that
coastline models are based on, namely the assumed equilibrium coastal profile. This
assumption does not allow for profile changes over time or along the coast and as-
sumes an instantaneous cross-shore redistribution of sediment when erosion of accre-
tion occurs. This may be appropriate in the case of for instance a perturbation by a
nourishment (Sect. 8.4.4). However, many situations are truly 2D, which implies that
cross-shore and alongshore processes are not physically independent and the assump-
tion of an instantaneous equilibrium cross-shore profile is invalid. Two examples of
two-dimensional effects are the lee-side circulation patterns, as mentioned before, and
‘outbreaking’ longshore currents as they approach the breakwater. The latter refers to
the fact that the water flowing towards the breakwater on the updrift side has no other
option than to go offshore, since the longshore current decreases to zero at the break-
water (Fig. 8.15). This contributes to the redistribution of material in the cross-shore
direction.
V
V
V
e
breaker zone coastlin
accreted
Figure 8.15: Longshore current approaching a breakwater. The current is diverted offshore while
decreasing in strength.
8.3.4. Multiple-line theory

If the single-line shoreline approach is considered to be over-schematised, it is possible
to use multiple-line theory. The principles of multiple- and single-line theory are the
same. The most significant difference is that cross-shore exchange of sediment is taken
into account in multiple-line theory, which allows a non-equilibrium cross-section to
be considered. Furthermore, the shore (cross-section) can be schematised in two or
more zones of the profile.
Bakker (1968) proposed a two-line approach in order to study the dynamics of a coast
with a groyne system that only partially blocked the longshore transport (since the
groynes were shorter than the width of the breaker zone). He recognised the two
distinct transport zones shoreward and seaward of the tip of structure and made a
schematization into two zones, as shown in Fig. 8.16. In some special cases he found
analytical solutions to the two-line approach.

For each zone in a two- (or more-) line model, a computation of the development of one
depth contour can be made, similar to the computation of the single line in the single-
line coastline models, but with the addition of an extra term in the continuity equation,
viz. the cross-shore sediment transport component. In a schematization as in Fig. 8.16
each longshore transport component 𝑆𝑥𝑖 is directed parallel to the coast and describes
the littoral sand movement in its zone. The horizontal planes, the boundaries between
the zones, should in principle be selected at elevations which correspond more or less
to flat portions of the total profile. If structures, such as groynes, are present, the
boundaries between the zones can be chosen such that they correspond to the limits
of a transport zone. The longshore transport components can be determined with
any longshore transport formula, except for bulk longshore transport formulas. The
cross-shore sediment transport exchange can be evaluated based on deviations from
an equilibrium profile (see Sect. 1.5.4 and Sect. 7.2).
cross-shore profile
breakwater
Sy1 Sx1
Sx2
zone 2 zone 1
plan view
breaker zone
Sy1
Sx2 Sx1
Figure 8.16: Shore plan and profile (two-line schematization).
In addition to the analytical models, numerical models have been developed based on
an arbitrary number of lines connected by cross-shore sediment transport. As with the
single-line theory, a multi-line model assumes that the cross-shore and alongshore
physics are decoupled. It has proven to be difficult to specify realistic formulas for
both the cross-shore sediment transport and the cross-shore distribution of the long-
shore sediment transport. Hence, for strongly 2D coastlines, a more physics-based
area model should be applied instead.

8.4. Coastal features and coastal change due to long-

shore transport
8.4.1. Introduction
The knowledge acquired in the previous sections will be used in this section to explain
a variety of coastal features and coastline changes as observed in nature. The basis of
the explanations lies in the gradients in littoral transport rates. The littoral transport
rates are dependent on the nearshore wave exposure and wave incidence angle, as
explained in Sect. 8.2. Hence, for a given offshore wave climate, coastal changes are
determined by changes in the depth contours and shoreline orientation and the degree
to which waves refract, shoal and diffract along the shoreline.
Assuming the offshore wave conditions do not vary along the shore (which is a reas-
onable assumption on a scale of the order of 10 km), a curved coastline will always be
subject to a longshore sediment transport gradient. This is because:
1. The changes in coastline orientation are reflected in changes in the angle of wave
incidence. Changes in shallow-water depth contours and coastline orientation
are generally more important than directional changes in offshore wave climate
along the coast;
2. Due to refraction, convergence and divergence of wave energy occurs (Fig. 5.6).
As long as the wave angles are relatively small, viz. a deep-water wave angle with
respect to the shoreline smaller than about 45°, the transport rates increase with in-
creasing wave angle (Fig. 8.4). This implies that convex depth contours result in sedi-
ment transport divergence along the coast and thus in erosion, whereas concave depth
contours lead to sedimentation (see Fig. 8.17). Note that convex and concave are here
defined as seen from the sea.
As a consequence, a shoreline has a tendency to flatten out bumps and dents until
the straight, equilibrium coastline is restored (hence a negative feedback process, see
Sect. 1.5.2). The specific case of high-angle waves is discussed in Sect. 8.4.4.
In this section, we consider the structural coastal response to large disturbances or
interruptions (natural and human-induced) that can be explained based solely on long-
shore transport gradients. The onshore and offshore movements of sediment that take
place on short timescales, such as storm-induced erosion and seasonal variations, are
assumed to cancel out on the longer timescales of shoreline change. Any net gains or
losses of sediment in the cross-shore direction (e.g. due to offshore canyons or inland
aeolian transport) can be considered as sinks or sources. Natural disturbances are for
instance coastline interruptions by tidal basins or rivers, the presence of an offshore
island and river sediment supply. Human-induced disturbances are for instance har-
bour moles, shore protection structures, river regulation works, nourishment schemes
and maintenance dredging.

8.4. Coastal features and coastal change due to longshore transport 393
φb
surf zone
MWL
beach
dune
S
0 + – + 0 – + – 0
Figure 8.17: Longshore sediment transport along a coast for small, deep-water wave angles.
Zones with positive transport gradients (transport divergence, erosion) are indicated by ‘−’ and
zones with negative transport gradients (transport convergence, accretion) by ‘+’. The ‘bump’ in
the shoreline erodes and the ‘dent’ accretes.
8.4.2. Blockage of longshore transport by shore-normal structures

In this section, we discuss the effect of (partial) blockage of longshore transport by
groynes and breakwaters. As in Sect. 8.3.3, we expect updrift accretion and downdrift
erosion. Figure 8.18 (in Example 8.2) serves as a further illustration. Whereas in
Sect. 8.3.3 we considered steady wave conditions, here we consider a full wave climate
as well a schematization into primary and secondary waves. Primary or prevailing
waves are waves that give a gross transport in the net transport direction, averaged
over all wave conditions. By contrast, secondary waves give a gross transport smaller
in magnitude and in the opposite direction.
Example 8.2 Coastline development in lee of jettied entrance
The development in the lee side of a jettied entrance is illustrated in Fig. 8.18. Let
us consider irregular wave conditions from one primary direction. The transport
away from the downdrift jetty by the alongshore current is 𝑆𝑥,1 . It is reduced in the
lee side of the jetty due to the reduced wave heights (changes in wave angle are
neglected). The shoreline development in response to 𝑆𝑥,1 only is shown at the
top left of the figure. The secondary transport component due to set-up-driven
currents is 𝑆𝑥,2 and is directed towards the downdrift jetty. The total transport is
𝑆𝑥,3 = 𝑆𝑥,1 + 𝑆𝑥,2 . In the zone directly downdrift of the jetty the transport gradi-
ent 𝑑𝑆𝑥,3 /𝑑𝑥 is negative, resulting in sedimentation (see top right of the figure).
Erosion occurs where 𝑑𝑆𝑥,3 /𝑑𝑥 is positive. Further away from the jetty, a con-
stant 𝑑𝑆𝑥,3 /𝑑𝑥 implies no coastline change.

steepest slope
in S curve and
maximum erosion
H S
Sx3
H70%
steepest slope
S S
Sx1 Sx2
maximum erosion
at steepest slope
Figure 8.18: Shoreline development on the lee side of a jettied entrance. Left: Assuming
irregular waves, the wave height 𝐻 , at the wave ray that separates the shadow zone from the
wave zone, is order 70 % of the original wave height. Since the transport depends nonlinearly
on the wave height, the transport 𝑆𝑥,1 away from the jetty is reduced even more, for example
by 50 %. The shoreline development in response to 𝑆𝑥,1 is shown at the top. Right: due to
set-up differences, a transport 𝑆𝑥,2 towards the breakwater occurs, yielding a total transport
𝑆𝑥,3 = 𝑆𝑥,1 + 𝑆𝑥,2 . The shoreline development in response to 𝑆𝑥,3 is shown at the top.
Let us – as in Sect. 8.3.3 – consider a long breakwater that extends far beyond the surf
zone, such that at least initially no appreciable transport will take place around the
tip of the breakwater. The wave conditions are as given in Intermezzo 8.2 and hence
Fig. 8.10 shows the transport curve. This curve is valid for the undisturbed coastline
far from the breakwater, that is reached by waves from all directions. On the updrift
side of the breakwater, the secondary waves with negative angles of attack are initially
blocked by the long breakwater. This locally increases the yearly-averaged transport
rate (i.e. only the gross positive transport should be considered) and hence not only
the transport for the unchanged coastline orientation, but the coastline orientation for
zero transport as well. This means that the equilibrium coastline orientation, to be
expected on the updrift side of the breakwater at 𝑥 = 0, is initially larger than 15°.
Similarly, on the downdrift side, the waves with positive angles of attack are blocked
by the breakwater. Waves with negative wave angles will reach the breakwater and

result in a transport towards the breakwater, such that local accretion can be expected
at the breakwater. Note that the term downdrift refers to the net transport direction
along the undisturbed coast averaged over all conditions. The angle of the accretion
at the breakwater is determined by the condition of zero transport for the waves that
can reach the ray just on the lee side of the breakwater. This angle follows from a
transport curve (for the conditions giving a negative transport only), that intersects
the 𝑦-axis at a negative transport rate and the 𝑥-axis at a negative angle. Note that the
effects of diffraction of the waves with positive wave angles into the sheltered area (see
Fig. 8.14) and corresponding set-up-induced currents are not yet taken into account in
the discussion above about the downdrift equilibrium angle at the breakwater.
The expected development of the coastline for a comparable case (from Mangor, 2004)
is shown in Fig. 8.19. The wave climate is schematised into prevailing waves from the
NW and secondary waves from the NE. The ‘representative wave condition’ shown in
Fig. 8.19 is a single condition that would reproduce the net transport of the full climate
as well as the corresponding shoreline equilibrium angle. The wave angles are smaller
than 45°. The figure shows the longshore transport rates along the coast, both initially
after construction of the harbour and after bypass has started. The transport rates are
expressed in units LDR (Longshore Drift Rate). The prevailing and secondary waves
give transports of 10 LDR eastward (E) and 5 LDR westward (W) respectively. Hence,
the net transport far from the breakwater is 5 LDR eastward.
Initially, the transport on the updrift side increases from 5 LDR E (outside the area of
influence of the breakwater) to 10 LDR E closer to the breakwater, where the secondary
waves do not penetrate. At the breakwater, of course, the transport is zero. In the be-
ginning therefore, the zone on the updrift side that is sheltered from secondary waves,
shows accretion close to the breakwater (where the transport gradients are negative)
and a smaller (only initial) erosion a bit further away from the breakwater (in the zone
with positive transport gradients). The updrift side as a whole has a sediment surplus
of 5 LDR and hence experiences accretion.
Downdrift of the breakwater the expected initial transport rates are 5 LDR E further
from the breakwater. Very close to the breakwater – in the shadow zone for the pre-
vailing waves – the transport is 5 LDR W (resulting in a small accumulation at the
breakwater). The lee side as a whole has a sediment deficit of 5 LDR and therefore
experiences erosion. From sediment continuity considerations, it follows that the de-
posited sediment volumes in the entire affected area must match the eroded volumes.
At any given time the coastline orientation at the breakwater (updrift) is normal to the
direction of the waves that can reach the advanced coastline. The angle can therefore
be seen to gradually change from normal to the direction of the prevailing waves to
normal to the direction of the ‘resulting’ (representative for the net transport) waves.
After the bypass starts, a bar will build up in the harbour entrance and the updrift
coastline will try to re-orientate itself towards the original direction (but will never

representative waves
prevailing or
primary waves secondary waves
buildup of bar
bypass in front of entrance
ion bypass shoal

ccumulat
sand a littoral
port zone
leeside erosion
initial transport
10
S [LDR]
5
0
-5
transport after start bypass

10
S [LDR]
5
0
-5
Figure 8.19: Schematic shoreline development, morphological development and net littoral drift
budgets (in unspecified units of Longshore Drift Rate (LDR)) for a port at a coast with a slightly
oblique resulting wave attack (after Mangor, 2004). Given the fact that continuity requires equal
accreted and eroded volumes, we expect that the lee-side erosion must stretch over a much
longer length.
succeed). Although the downdrift coastline will continue to experience erosion, the
erosion will diminish after the coastline receives sediment via the migrating bypass
shoal.
8.4.3. Shadow effects due to obstacles away from the shoreline

An obstacle in front of a coast, such as a rocky outcrop, an offshore breakwater or even
a shipwreck, will reduce the wave activity in the zone of wave shadow between the
object and the shore. Since the reduced wave activity in the shadow zone will result
in a reduced sediment transport capacity, material being carried along the shore will
be deposited in the shadow zone, forming a tombolo. Initially only a shoal will form,
which we call a salient. This can, however, develop into a point of land connecting the
original shoreline to the obstacle (see Figs. 8.20 and 8.21). We now speak of tombolo. A
tombolo completely blocks longshore transport in the zone landward of the obstacle.
These shadow zone effects can be used to stimulate and preserve a recreational beach
(Sect. 10.5.4). Have a look at Fig. 10.41, which shows salient and tombolo formation
behind a series of emerged (i.e. having their crests above MSL) offshore breakwaters.

Figure 8.20: The holocenic tombolo of Mariscal formed behind an island in Bombinhas, Brazil.
The photo was taken from a hill in the original island (see black arrow in map) which is now
connected to the Porto Belo peninsula by the tombolo. Photo by Carolina Piccoli, map Open-
StreetMap (‘Credits’ on page 575).
Figure 8.21: Tombolo formed behind a detached breakwater in Alicante, Spain. Photo by Leonid
Andronov (‘Credits’ on page 575).

The development of a tombolo, as explained above, depends upon a transport of ma-

terial parallel to the coast; the reasoning is that in the shadow zones behind obstacles,
the breaking wave heights and thus longshore transport capacity are reduced and sed-
iment is deposited. Note that due to diffraction of waves around the ends of the break-
water, still some wave action will be present behind the breakwater (the wave heights
at the end of the breakwater are in the order of magnitude of 50 % of the original in-
coming wave height, see Sect. 5.2.4). Example 8.3 discusses the shoreline development
resulting from an emerged breakwater (partly) blocking the longshore sediment trans-
port by obliquely incident waves.
Besides the shoreline development, as a result of breaking wave height reduction in the
shadow zone and subsequent gradients in longshore sediment transport (illustrated in
Example 8.3), two subtler effects play a role, namely: 1) changes in wave angles in the
shadow zone; and 2) secondary current patterns as a result of set-up differences. These
additional or secondary effects enhance the pattern of deposition behind the obstacle
for obliquely incident waves and are the reason that even with normally incident waves
a salient or tombolo can be formed (Fig. 8.23). This can be explained as follows. Due to
diffraction into the shadow zone, not only the wave heights, but also the wave angles
are affected. Hence, even in the case of normally incident waves, the diffracted waves
behind the breakwater approach the original coastline at some angle. This results in an
initial sediment transport pattern with transports from both breakwater ends towards
the shadow zone. Furthermore, as shown in Fig. 5.46, nearshore currents run towards
the shadow zone from both sides of an emerged obstacle. This secondary current
pattern also enhances transport towards the shadow zone.
Example 8.3 Coastline development near an emerged detached

breakwater as a result of longshore transport gradients
Consider an emerged detached breakwater along a coast with initially straight

and parallel depth contours (Fig. 8.22). Waves are incident at an angle of 15°.
The natural surf zone width is smaller than the distance between the breakwa-
ter and the shore. The length of the breakwater is equal to its distance from the
shore. As in Fig. 8.17, Fig. 8.22 qualitatively shows the sediment transport as a
function of the distance alongshore, for the initially straight shoreline, estimated
based on a hypothesised variation of the wave height behind the breakwater. Next,
the areas where the expected shoreline response is accretive (negative transport
gradients, sediment transport convergence, indicated with ‘+’) and erosive (pos-
itive transport gradients, sediment transport divergence, indicated with ‘−’) can
be identified. Based on the initial sediment transport variations alongshore, the
expected shoreline development can be sketched. The result is a wave-like initial
pattern of updrift accretion (the shoreline bulges seaward) and downdrift erosion
(the shoreline curves landward) as in Fig. 8.22.

detached breakwater
shadow zone breaker line
accretion zone erosion zone
Sx
0 + − 0
S
Figure 8.22: Shoreline development behind an emerged breakwater for obliquely incident
waves from one direction. Only the development due to the longshore transport gradients
is considered; other mechanisms are ignored. Note that continuity requires that accreted
volumes are equal to eroded volumes.
Figure 8.23 shows the resulting symmetrical shoreline development, for normally in-
cident waves, with erosion on either side of the breakwater and accretion behind the
structure. The initial shoreline development takes the form of a salient (lower panel),
which may or may not develop over time into a tombolo (upper panel). A tombolo is
more likely to develop if the breakwater is located in or just outside the surf zone (since
the responsible currents are generated in the surf zone) and if its length is relatively
large. If the length is larger than, say, twice the distance from the coast, the diffracted
wave heights reduce to zero in the centreline of the structure (Fig. 8.24, left figure).
In that case the sediment stirring and transporting capacity behind the breakwater is
not enough to keep a passage open. For a relatively short breakwater, the equilibrium
shoreline will take the form of a salient (Fig. 8.24, right figure).
In the case of submerged obstacles, tombolos or salients may also develop. However,
in those cases, any accretional tendency may be counteracted by a (possibly stronger)
erosional effect. This is related to the wave-induced current pattern as shown in
Fig. 5.49. The onshore-directed wave forces, as a result of waves breaking on a shoal or
submerged breakwater, result in a large on-shore flow of water. This is compensated
for by rip currents returning seaward at either side of the shoal or breakwater. These
currents can carry large amounts of sediments from the shadow zone seaward and
hence create large levels of erosion. Some consequences for submerged breakwater
design are discussed in Sect. 10.5.4.

8.4.4. Shoreline perturbation

This section discusses the evolution of a single perturbation (a ‘bump’) in the shoreline.
Such a bump in the shoreline can have various origins, for instance a nourishment
scheme or a significant supply of sediment by rivers, leading to delta formation (see
Sect. 8.4.6). In Example 8.4, we consider an initial disturbance of the system by a
nourishment; the same development occurs for any other perturbation of the coastline.
breaker lines
0 – 0 + 0 – 0
S
salient
breaker lines
tombolo
breaker lines
Figure 8.23: Shoreline development behind an emerged breakwater for normally incident waves.
In this situation, the approach angles of the diffracted waves and the current patterns induced
by the set-up gradients determine the coastline development. The figure also shows the initial
sediment transport variations. Note that the coastline development is symmetrical (as opposed
to the development for oblique incidence as sketched in Fig. 8.22) and that the accretion behind
the obstacle necessarily means erosion on both sides of the obstacle.

L L
C
D=½L D>½L
undisturbed wave undisturbed wave

reaches shoreline reaches shoreline
Figure 8.24: Wave diffraction behind a detached emerged breakwater with length 𝐿 and distance
to the shore 𝐷. Left: wave heights reduce to zero in the centreline of the structure. Right: wave
energy penetrates in the entire zone between the emerged breakwater and the shoreline.
Example 8.4 Coastline change caused by a perturbation of the shoreline
Assume a beach nourishment has recently been executed on an initially straight

stretch of coast. Some time after the nourishment, the sediment has become re-
distributed over the cross-shore and the shoreline is given by the dotted line in
Fig. 8.25. As a result of the nourishment, the shoreline has advanced, after cross-
shore redistribution, over an alongshore distance of 1000 m. The maximum shoreline
advance is 50 m. The angle between the original and nourished coastline is every-
where smaller than 10°.
after nourishment
original shoreline
Figure 8.25: Shoreline some time after a beach nourishment, over a limited alongshore
distance.
Questions
1. The waves are normally incident to the original shoreline. Sketch how
the coastline will qualitatively develop over time. In order to substantiate
your answer, draw the initial longshore transport rates as a function of the
distance along the coast (as in the middle panel of Fig. 8.19) and indicate,
based on the longshore transport gradients, where erosion and accretion
take place. What are the locations with the largest shoreline change?
2. As 1. but now the waves have an angle of incidence in deep water of 20°
with respect to the original shoreline;
3. As 1. but now the waves have a deep-water angle of incidence of 70° with
respect to the original shoreline.

Figure 8.26, based upon concepts by Ashton and Murray (2006), contains some of the
answers to the questions in Example 8.4. Figure 8.26b shows the transport as a function
of the deep-water wave angle relative to the shore (denoted 𝜑0 in this book). It can be
seen that for low-angle waves the transport increases with larger relative angles (from
1 to 2 to 3 in Fig. 8.26b). By contrast, for high-angle waves the transport decreases with
increasing angle (from 4 to 5 to 6 in Fig. 8.26b).
Figure 8.26c shows the shoreline response to low-angle waves. At the ‘horizontal’
parts of the shoreline the transport magnitude is equal and given by point 2 in Fig. 8.26.
However, at the ‘left’ and the ‘right’ flank of the ‘bump’, the relative wave angles and
thus the transport magnitudes are smaller (point 1) and larger (point 3) respectively.
The crest of the bump erodes, since there the transport increases (diverges) in the
transport direction (positive transport gradient). A decreasing (converging) transport
in the transport direction (negative transport gradient) causes deposition. The end
result is a flattening of the shape towards a straight coastline, for which no further
changes occur. For high-angle waves (greater than 45°) the pattern of erosion and
deposition is reversed, leading to growth of the bump (Fig. 8.26d).
The shoreline response to low-angle waves is a clear example of negative feedbacks
stabilising the shoreline by eroding the perturbations (see Sect. 1.5.2). For high-angle
waves, positive feedbacks promote the growth of perturbations, leading to self-organised
patterns (see also Intermezzo 7.2).
8.4.5. Interrupted coasts: spits

A spit is a pointed tongue extending into the sea. Spits develop where the longshore
transport capacity is diminished due to coastline interruptions (river, estuary, end of
island). The direction of a spit usually is a continuation of the shoreline from which
sediment is supplied. An example of a spit is shown in Fig. 8.27. Waves coming pre-
dominantly from the southwest cause a sand transport toward the north along the
western slope of the island. As the water becomes deeper at the north end of the is-
land, the waves no longer break, the sediment transport decreases (hence a negative
transport gradient is created), the sand settles and the spit gradually builds out as an
extension of the coastline.
The spit of Block Island is a good example of a spit formed at the end of a beach, where
the longshore current loses its transport capacity. For the same reasons, spits can also
form near the entrances of harbours or estuaries or where a river mouth interrupts an
otherwise straight coast. Let us consider the situation of a modest river flowing into
the open sea (see Fig. 8.28). The undisturbed longshore transport rate is 𝑆 m3 /yr. If the
river mouth is sufficiently wide and deep to strongly reduce the longshore transport
rate (hence a negative transport gradient), coastal material is deposited on the updrift
side of the entrance, narrowing the mouth. At the downdrift side the waves regain
their capacity to transport sediment (capacity is 𝑆 m3 /yr), whereas only a fraction of

(a) (b)
refraction
Sx
φ0 3 4
φb
2 5
1 6
θ
0 ≈45 90
relative wave angle (φ0-θ) [°]
(c) (d)
S S
accretion erosion accretion erosion accretion erosion

2 1 2 3 2 5 4 5 6 5
φ0 – θ [°] φ0 – θ [°]
90 90
45 45
0 0
Sx Sx
– + + – + – – +
+ – – + shoreline – + + –
change
Figure 8.26: Response of a perturbation (a ‘bump’) in the shoreline, based on concepts by Ashton
and Murray (2006) with (a) depiction of the terms and axes; (b) the transport curve as a func-
tion of the relative deep-water wave angle (𝜑0 − 𝜃) showing a maximum for an angle of around
45°; (c) response to low-angle waves (smaller than 45°) for which the transport increases with
larger relative angles resulting in flattening of the shape and (d) response to high-angle waves
(greater than 45°) for which the transport decreases for increasing angle, resulting in growth
of the bump. The bottom figures represent the initial variation of the wave angle and sediment
transport along the shore and indicate the zones where sedimentation and erosion can be ex-
pected. The numbered transport magnitudes in (b) correspond to the numbers in (c) and (d).

Figure 8.27: Spit at north end of Block Island, USA. High Resolution Orthophotos obtained in
April 2014. From U.S. Geological Survey (‘Credits’ on page 575).
this bypasses the river mouth. The positive transport gradient results in erosion on the
downdrift side. The result is a slow displacement of the river mouth in the direction
of the longshore transport; a spit develops and grows over time.
accretion erosion
S
growing spit
Figure 8.28: Spit and river mouth.
Landward of the spit, the river flows more or less parallel to the coast for some distance.
Because of the growing of the spit, the length of the river increases, and as a result
the water levels in the river behind the spit rise, eventually forcing a breakthrough
somewhere on the updrift side of the spit. This is a periodic process for a fully natural
river mouth with breakthroughs of the slender spits occurring during periods of large
river discharges (e.g. during a wet season).
Spits may also develop where coastlines with sufficient longshore sediment transport
are interrupted by lagoons or bays instead of by a river. If the longshore transport rate

is large enough, a spit can also develop in combination with a delta (see for instance
Fig. 8.32). Spits may eventually close off a lagoon or bay, acting as a ‘bay barrier’
(Fig. 8.29a). Figure 8.29b shows a barrier spit that separates a tidal basin from the
marine environment; tidal action keeps the entrance open. Breaches in long barriers
can lead to barrier island formation.
(a) Willapa Bay, Washington, USA (b) Keta Lagoon, Ghana
Figure 8.29: Barrier spit development as a result of longshore transport.
The coastline interruptions may not only locally diminish wave-induced longshore
transport capacity, but act as a source (e.g. a river, see Sect. 8.4.6) or sink (e.g. a
tidal basin) of sediment as well. Tidal basins (bays, lagoons, estuaries) are known to
act as a sink for longshore sediment transport, thereby depriving adjacent coastlines
of sediment. Tidal basins and their interaction with adjacent coasts are discussed in
Ch. 9.
8.4.6. Deltaic coastlines

Built primarily from river-borne sediments, deltas form when the amount of sediment
delivered at the mouth of a river exceeds the amount removed by waves and tidal
currents (see Ch. 2). The main effect of waves and tides is to redistribute the sediment
along the coast. Section 2.7.3 has discussed the classification of deltas based on the
relative river, wave and tidal influence.
Figure 8.30 shows the schematised development of a wave-dominated deltaic coastline
(cf. Fig. 1.7). An equilibrium coastline position requires that all of the sediment sup-
plied by the river is transported to the coastlines on either side of the river mouth by
the wave-driven longshore currents. From the wave characteristics and the river sedi-
ment supply, the equilibrium coastline orientation (see Fig. 8.30) can be calculated; at
the delta mouth the angle 𝜑 with the waves, which are normally incident to the initial
coastline, will be such that the longshore transport rate is equal to 12 𝑆𝑟 .

0.5 Sr 0.5 Sr
Qr [m3/s]
Sr [m3/y]
Figure 8.30: Development of deltaic coastline with waves that are normally incident to the ori-
ginal coastline and without tidal influence. The river discharge is 𝑄𝑟 and the sediment discharge
𝑆𝑟 . The angle 𝜙 is determined by the sediment discharge and wave conditions.
Man-induced changes in the characteristics of the river (e.g. sand mining in the river
bed or damming of the river for irrigation or hydro-power purposes) might change the
natural accreting tendency of the coast to an eroding tendency in the vicinity of the
river outlet (see Fig. 8.31).
Sr [m3/y]
Figure 8.31: Delta erosion after the sediment supply is cut off. Due to wave-induced longshore
transport gradients, sediments are redistributed along the coast, resulting in erosion of the delta
(cf. the development of a shoreline perturbation in Sect. 8.4.4).
For obliquely incident waves, the redistribution of the river sediment along the shore
is asymmetric. Also, a sand spit can develop in these cases due to a strong longshore
sediment transport, as already described in Sect. 8.4.5. Section 2.7.3 gives examples
of various wave-dominated deltas. The Ebro delta (Fig. 2.49) and the Senegal delta
(Fig. 8.32) are examples of spit formation in a deltaic environment.

Seneg al
5 km 2003-2011
water depth > 5 m flood plain beach/dune
Figure 8.32: Senegal river delta.

409
9
Coastal inlets and tidal basins
9.1. Introduction
In the previous two chapters, we mainly focused on wave-dominated coastal systems.
These coastal systems, with well-developed beaches, are typically shaped by waves
and wave-generated currents. The moderate to high wave energy (see Sect. 4.3.1)
overwhelms any tidal energy present, for which reason these coasts are called wave-
dominated. The shoreline in wave-dominated environments is characterised by elong-
ated sediment (mainly sand) bodies. These include spits (Ch. 8), alongshore bars (Ch. 7)
and beaches. Besides straight coastlines – whether or not they are interrupted by
breakwaters, groynes, rivers and lagoon entrances – pocket beaches may also fall into
the wave-dominated category.
So far we have only encountered tidal influence in deeper water outside the surf zone,
around long structures and at low wave-energy coastlines. In general, tidal conditions
dominate where wave energy is relatively low. The word ‘relative’ is crucial here,
since it is the relative influence of waves and tides that determines the morphology
(Fig. 4.13). Tide-dominated environments may occur due to restricted fetch or where
incident wave energy is trapped or reflected. Such environments include tidal basins.
As discussed in Ch. 2, tidal basins are the result of breakthroughs and flooding of low-
lying areas due to the global rise of the post-glacial sea level. Processes contributing to
basin formation include tectonic subsidence, fluvial erosion and glacial action. Bottom
subsidence of the coastal plains due to human interference (peat harvesting, impolder-
ing and water, oil and gas extraction) can be an additional cause for the creation of
tidal basins. Finally, basins can also evolve due to the formation of barriers enclosing
a body of water.
A few examples of tidal basins are Chesapeake Bay, San Francisco Bay, Waddenzee and
Baie d’Arcachon. All of these were created by flooding of low-lying coastal plains on

410 9. Coastal inlets and tidal basins
coasts with a strong tidal energy and with little sediment discharge from rivers. The
emphasis in this chapter is on understanding the relevant physics in order to make
predictions about changes, which are expected to occur on the scale of the tidal basin
itself. This does not only concern changes in water motion, but also changes in the
physical structure of the basin, i.e. the morphology. Relevant practical questions are
amongst others:
• What is the response of a tidal basin to sea level rise (does the basin floor follow
the sea level rise by sedimentation)?
• What are the consequences of dredging for navigation or sand extraction (does
sedimentation increase)?
• What is the effect of a dike or breaching of a beach barrier (does a new tidal
basin form)?
• What are the impacts of land reclamation (does the inlet close, does sedimenta-
tion or erosion occur)?
• How do basin changes influence the adjacent coast (will the adjacent coast erode
or accrete)?
The answers to a few of these practical questions are discussed in the final section of
this chapter, Sect. 9.8, where the knowledge acquired in the preceding sections is ap-
plied. We start this chapter with a discussion of general basin and inlet types (Sect. 9.2)
and a description of the main morphological units of tidal inlet systems (Sect. 9.3). The
subsequent three sections focus on the three important morphological units, viz. the
ebb-tidal delta (Sect. 9.4), the entrance or inlet channel (Sect. 9.5) and the inner basin
(Sect. 9.6). Besides relevant processes, these sections discuss how the stability of these
morphological units can be described with empirical relations, relating the geometric
properties to hydraulic boundary conditions. In Sect. 9.7 the mechanisms responsible
for net sediment import into and export from basins are discussed.
9.2. Basin and inlet types

9.2.1. Bays, lagoons and estuaries
Three distinct types of tidal basins can be discerned (Carter (1988), see Table 9.1 for a
summary):
• Tidal lagoons are basins that are enclosed by wave-shaped coastal barrier islands
or barrier spits. Almost 12 % of the world’s coastline is made up of barriers, many
of them enclosing lagoons. The Dutch Wadden Sea is a good example. Due to
the presence of the barriers, the penetration of waves into the lagoons is limited.
Water flows into the lagoon with the flood and out during the ebb through passes
or inlets between the barrier islands. These narrow waterways are also called
throats or gorges. In most cases the tidal fluxes in and out of the lagoon are
balanced over a tidal period. Sometimes it is a sub-surface rather than a surface

9.2. Basin and inlet types 411
connection with the open sea that allows the water levels in the basin to be
modulated by the tide. Where the basins proper are tide-dominated, the inlets
to the basins experience both wave and tidal influence. Sediment tends to be
finer in the more protected regions away from the entrances. The freshwater
run-off is typically limited;
• In the absence of barrier islands, tidal bays are basins that are more open to the
deep water of the sea or ocean. Bays have a limited freshwater run-off. The Baie
de St. Michel is a good example. Waves can enter unhindered, but generally lose
their energy not far from the entrance (mouth) before the shorelines in the bay
are reached, due to depth-limited breaking (Sect. 5.2.5) and bottom friction;
• Estuaries are different from bays in that they experience a (strong) fresh-water
run-off. Therefore, the seawater in these basins is measurably diluted by fresh-
water. But unlike river mouths, estuaries are typically tide-dominated; the water
motion in estuaries is controlled more by the tides than by the river discharge.
Also, sedimentation is primarily controlled by import from the adjacent coastal
region. The coarser sediment (sand) settles predominantly in the seaward re-
gions, while the finer sediment (silt) settles in the more protected landward re-
gions. The entrances of some estuaries may be constricted by the development
of spits, shoals or barriers across their mouths, as a result of wave effects.
Table 9.1: Distinctive attributes of tidal environments (according to Carter, 1988)
Environment Distinctive attributes

Tidal bays
• Baie de St. Michel, France • High levels of wave dissipation
(Normandy/Brittany) • Little freshwater runoff
• West coast of South Korea
• Hangzhou Bay, China
Tidal lagoons
• Wadden Sea, • Waves excluded by barriers
Netherlands/Germany/Denmark • Tidal flows via passes
• Laguna Madre, Texas • Infilling wetlands
• Little freshwater runoff
Estuaries
• Bay of Fundy, Nova Scotia • Waves possibly excluded by barriers or
• Bristol Channel sand shoals
• High freshwater runoff
The three types of basins mainly differ in terms of the characteristics of the entrance
and the importance of freshwater run-off. The basins are either interlinked with the

adjacent coast or barrier islands through tidal inlets, openings in the shoreline, such
as between two barrier islands, or wider entrances such as estuary mouths. Freshwa-
ter run-off and hence the interaction between salt- and freshwater is a fundamental
characteristic of estuaries (see Intermezzo 9.1). Based on the degree of mixing of the
two water masses, estuaries can be classified as stratified, partially mixed/stratified
and mixed or homogeneous. One of the phenomena related to fresh-salt water mixing
is the turbidity maximum. In this book we mostly neglect the mixing between fresh-
water and saltwater, a topic which would require a rather extensive treatment. We
therefore mostly assume a negligible influence of freshwater run-off in this chapter.
Intermezzo 9.1 Estuarine circulation
In estuaries, there exists a transitional region between salt- and freshwater. In

the case of small and medium freshwater discharges the salt-freshwater transition
region is located near the estuary head (the location of the river mouth). Seaward
of the transition region the tidal water motion dominates and the water is mainly
saline. More landward, the river flow dominates and the water is mainly fresh.
The pressure gradient associated with the density difference between the saline (of
sea origin mainly) water and the upstream fresh river water drives a vertical flow
circulation, the so-called estuarine circulation (see Fig. 9.1). When the freshwater
meets the saline water, the less dense freshwater overrides the denser saline water.
Hence, along the bottom the net (i.e. tidally-averaged) flow is in the landward
direction, while it is seaward along the surface. Note that wide estuaries may
show little salinity variation with depth, but considerable variation with width; as
a result of Coriolis, the denser flood and less dense ebb currents are concentrated
along different banks.
river 10 20 30 river 10 20 30
salt strongly
wedge tide stratified tide
river 10 20 30 river 10 20 30
weakly well
stratified tide mixed tide
Figure 9.1: Isohalines (lines of equal salinity) and estuarine circulation for four different
salinity distributions. The basic flow pattern is a surface flow of less dense freshwater
towards the ocean and an opposite flow of salty seawater into the estuary along the bottom.

As suspended sediment concentrations are higher near the bed, estuarine circu-
lation promotes sediment import. A related phenomenon is the formation of a
turbidity maximum, which is the concentration of fine sediments not far from the
end of a saltwater wedge. It occurs since saltwater stimulates the flocculation of
clay particles being transported by the river.
When discussing lagoons, we assume an approximately equal influx and outflux of

water (as is the case for the Wadden Sea basins). This means that we exclude lagoons
for which the seawater inflow exceeds the outflow due to evaporation. These lagoons
are called sabkhas and are common on low-latitude arid coasts.
In this book we often use the two distinct types of Dutch coastal basins, the estuaries of
the Southwest Delta1 and the Wadden Sea barrier inlet systems, to illustrate the above
described differences. They are indicated in Fig. 9.2. Three of the estuaries in the
Southwest Delta were partly or entirely closed as part of the Delta Works (numbers 3,
4 and 5 in Fig. 9.2). The Delta Works are a series of dams, sluices, locks, dikes and storm
surge barriers built between 1950 and 1997 (Maeslant storm surge barrier) to protect a
large area of land around the Southwest Delta from the sea. In the Dutch Wadden Sea
system, two closure barriers have also been constructed in the 20th century (numbers
1 and 2 in Fig. 9.2). The morphological response to (semi-)closures of tidal basins is
further discussed in Sect. 9.8.
9.2.2. Hydrodynamical classification

Following our preferred process-based classification, we could make a slightly different
distinction between the different basin types than in Sect. 9.2.1.
Tidal basins (estuaries, bays, lagoons) are of course tide-dominated, see Sect. 2.7.1. The
influence of the tide is evident from tidal current ridges, extensive salt marshes and
tidal flats. In tidal basins with an extremely strong tide (amplitudes of the order of
10 m) the tidal wave is deformed so strongly that a tidal bore may develop (see Inter-
mezzo 5.7). Examples of such basins are the Amazon, the Qiantang and the Severn.
We can distinguish between tidal basins with and without a strong river influence:
• Examples of tide-dominated coastal systems with strong river influence are the
estuaries of the Western Scheldt, the Thames, the Elbe (all three in the winter
season), the Seine, the Yangtze, the Mekong, the Ganges-Brahmaputra, the Fly
River, Chesapeake Bay and the Rio de la Plata;
1
The southwest coast of the Netherlands is often called our Delta Coast. It is here that Rhine, Meuse and
Scheldt rivers exit into the North Sea. However, there is no real delta present and these rivers have
mainly indirectly contributed to the Holocene formation of this coastal area; marine feeding (with
reworked Pleistocene riverine sediments) was the major source of sediment.

Lauwerszee
1969
2
Zuiderzee
1932
1
Haringvliet
1969
3
4
5
Grevelingen
1971
Eastern Scheldt
1986
Figure 9.2: Closure of tidal basins in the Dutch coastal system, with the estuaries of the south-
western delta area and the northern Wadden Sea. The Western Scheldt (not numbered) is loc-
ated to the south of the Eastern Scheldt.
• Examples of tidal basins without strong river influence are the (eastern) Wadden
Sea, the (present) Eastern Scheldt, the Gulf of St. Lawrence, the Colorado River
and San Francisco Bay (the latter two in summer).
In the entrance area of the basins (the inlet, see Intermezzo 9.2), which is maintained
by the tide, wave and tidal influences are combined. The tidal range outside an inlet
depends primarily on the ocean tides and their interaction with the continental shelf.
Micro-tidal, meso-tidal and macro-tidal ranges can be distinguished (see Sect. 4.4.1).
The wave conditions are generated further seaward and thus independent of the inlet.
Wave energy can be classified as low, medium and high according to Sect. 4.3.1. Since
both wave energy and tidal energy in the entrance area of basins are independent of
the inlet system configuration, they are very suitable to be used for inlet classification.
Hayes (1979) and Davis Jr. and Hayes (1984) distinguish five hydrodynamical classes,
from wave-dominant to tide-dominant, based on the above-mentioned tidal range and
wave energy classification (see Sect. 4.4.2). Each class develops its own specific mor-
phologic features. For instance, in Sects. 9.3 and 9.4 we will see that the relative sizes
of the flood- and ebb-tidal deltas and the mechanism for sand-bypassing of entrances
are dependent on the relative wave/tide dominance. Barrier islands are typically wave-
built features and the entrance areas to barrier tidal inlet systems (see nr 7 in Fig. 2.42)
are typically mixed-energy or wave-dominant systems. As an example, Fig. 9.4 classi-
fies the Frisian Inlet (Friesche Zeegat in Dutch), one of the inlets of the Dutch Wadden

Sea as a mixed-energy environment. The Dutch Delta Coast can be classified as a

mixed-energy environment as well.
Intermezzo 9.2 Terminology
Note that the term ‘tidal inlet’ generally refers to an opening in the shore that
provides a connection between the ocean or sea and a basin, that is maintained by
tidal currents. Alternatively, the term ‘inlet’ is used in a broader sense to describe
the entire morphological system consisting of the entrance itself, the ebb-tidal
delta (outside) and the flood basin, with possibly a distinct flood-tidal delta. In this
book, we want to make a clear distinction between the two definitions and will
use the term inlet only to refer to the short, narrow waterway that lets the tide in
(the inlet, gorge or throat). The gorge is usually dominated by a main channel, but
some are split by a shoal or a small island (cf. the Frisian Inlet, between the isles
of Ameland and Schiermonnikoog in the western (Dutch) Wadden Sea). We use
the term tidal inlet system to refer to the combined system of inlet, ebb-tidal delta
and basin.
intertidal fluvial delta

deposits
dunes
beach
intertidal
deposits
ebb-tidal delta flood-tidal delta
Figure 9.3: An estuary along a wave-dominated coast with a barrier constricting the mouth.
On either side of the narrow inlet deltas are present, built by the flood and ebb tide respect-
ively (see Sect. 9.3.2). Loosely based on Dingle Bay and Danube delta.
Tidal inlets mostly exist at places where there are breaks in a barrier coast. The
term tidal inlet is generally not used for entrances to estuaries or bays with large
unconstricted mouths, such as the Western Scheldt, although the tide may enter
through inlet channels cutting through sand shoals. We will just speak of an es-
tuary entrance or mouth. Entrances to estuaries along wave-dominated coasts or
lagoons may also be constricted by wave-built barrier spits (Fig. 9.3 and Ch. 8). In
that case it makes sense to speak of an inlet.

low medium high wave energy

6
macrotidal
gh
5 hi
(
d
mean tidal range [m]
te
ina
w)
(lo
om
d )
4
te ted
e-d
ina
tid
om
a
in
e-d
om
mesotidal
tid
y(
e-d
3 rg
e
tid
en d)
2.3 ixe
d
om inate
m ve-d
2 gy (wa
ener
xed
mi
microtidal
1
-dominated
wave
0
0 0.5 0.7 1 1.5 2 2.5
mean wave height [m]
Figure 9.4: Hydrodynamical classification according to Hayes (1979) and Davis Jr. and Hayes
(1984) with, as an example, the classification of the Frisian Inlet (Fig. 9.5), one of the inlets of the
Dutch Wadden Sea. Every class covers a spectrum of tidal ranges and wave heights; the relative
effects of waves and tides rather than the absolute tidal ranges and wave height are important.
9.2.3. Hydraulic boundary conditions and geometric controls

The hydraulic boundary conditions are crucial for determining the morphology of tidal
basins and inlets. Not only is the relative dominance of waves or tides important,
but other hydraulic conditions also control the morphodynamic behaviour. At the
same time, the basin morphology has an influence on the hydraulic conditions. Due to
this feedback mechanism, the basin in essence determines its own evolution, although
external conditions such as sediment availability and storm surges also play a role.
The following geometric and hydraulic controls will be further explained in this chapter:
1. The surface area of the basin, in combination with the tidal range, determine in
principle the tidal prism (Intermezzo 9.3), the volume of water that has to flow
in and out through the inlet during one tidal cycle (excluding any freshwater).
For a short basin2 , this is usually estimated by multiplying the mean surface
area of the estuary by the mean tidal range in the estuary. This is only valid in
the case of a negligible river discharge, since a river flow will also contribute to
the filling of the basin. The tidal prism has empirically been found to determ-
ine the equilibrium, or minimal stable cross-sectional channel area of an inlet
(Sect. 9.5.2). The channel cross-sectional area and volume (Sect. 9.6.2) as well as
2
Try to answer the following question: why is this method to estimate the tidal prism only appropriate
for a short basin? See also Intermezzo 9.3 and App. E.

the sand volume stored in the ebb delta (Sect. 9.4.3) have also been empirically
related to the tidal prism;
2. The morphologically active part of basins consists of deeper areas, where the
flow is concentrated during lower tidal water levels and of tidal flats, that are
covered during higher water and exposed during lower water. Because of the
different geometry (surface area, water depths) during low tide and high tide,
the high tide propagates at a different phase speed (see Sect. 5.7.4) than the low
tide. This can either strengthen or weaken the magnitude of the maximum flood
flow, compared to the maximum ebb flow (and thus shorten or lengthen the flood
duration compared to the ebb duration). This leads to a net import or a net export
of sediment respectively and hence steers the morphological development of
the basin in time. The important controls for the tidal distortion are the surface
areas at low and high water and the mean water depth at high and low water,
both of which are determined by a combination of tidal range, channel depth
and intertidal storage areas or flats (see Sect. 9.7.2);
3. Tidal waves propagating into basins may be either progressive or standing, or
a mixture of the two, depending on, amongst other factors, the length of the
basin (see Sect. 5.7.3). The length of tidal basins is generally much shorter than
the tidal wavelength. In short tidal basins, the tidal wave is reflected and has
a standing character. The tidal range in the basin is of similar magnitude as in
the open sea. In longer basins, resonance can occur when the basin length is
approximately a quarter of the tidal wavelength (or a multiple of that). The tidal
amplitude will then increase. The longer the basin, the more the tidal wave is
dampened by friction. The resulting weakening of the reflected wave, assures
that the tidal wave has a stronger propagating character;
4. In very short basins (length much smaller than the tidal resonance length), the
combination of tidal range, channel depth and intertidal storage areas or flats
results in a different type of asymmetry than asymmetry between ebb and flood
duration (see under 2.). In the case of a short basin, the duration of the flow
change is different at HWS and LWS (see Sect. 5.7.4). As we will discuss later,
this asymmetry is of great importance for the net transport of fine sediment in
the basin (Sect. 9.7.3).
Intermezzo 9.3 Tidal prism
The tidal prism is defined as the volume of water entering a tidal basin during flood
tide and leaving the basin again during ebb tide. Neglecting freshwater outflow,
the tidal prism 𝑃 is equal to half the time integral of the inflow and outflow during
a tidal cycle:
𝑇
1
𝑃 = ∫ |𝑄(𝑡, 0)|𝑑𝑡 (9.1)
2 0

with 𝑄(𝑡, 0) the tidal discharge in the inlet of the basin.

For a short basin the size of the basin is small compared to the tidal wavelength,
such that the spatial variation of the water level in the basin can be neglected
(Sect. 5.7.3). From Eq. 5.108, we have 𝑄(𝑡, 0) = 𝜕𝜂0 /𝜕𝑡𝐴𝑏 with 𝐴𝑏 the surface area
of the basin. With this, Eq. 9.1 can be approximated as 𝑃 = 𝐴𝑏 𝐻 . Therefore, the
tidal prism is often estimated by multiplying the mean surface area of the basin by
the mean tidal range 𝐻 in the basin.
For longer basins, the spatial variation of the water level in the basin cannot be
neglected and the full Eq. 9.1 must be computed. Also, for quite long basins the
tidal wave may exhibit a (partly) standing wave pattern with nodes and antinodes
in the surface elevation and discharge (Fig. 5.63). A node in the discharge at a
particular location implies that no water is exchanged through that cross-section
during the tidal cycle. As a consequence, only the part of the basin from the inlet
to the antinode contributes to the tidal prism. An example is given in App. E.
Under calm conditions, a basin like the Wadden Sea can be thought to consist of
multiple basins – each with their ‘own’ tidal prism – separated by tidal divides or
watersheds (Fig. 9.5).
9.3. The main morphological elements

9.3.1. Introduction
The morphodynamic behaviour of tidal inlet systems is highly dynamic and strongly
determined by the tide. As an illustration of the complex tidal dynamics, Fig. 9.5 and
Table 9.2 give an overview of the Wadden Sea. Many of the Wadden Sea inlet systems
have distinct catchment areas, which in calm conditions are separated by a watershed
(or tidal divide or wantide – from the Dutch wantij). The locations of the tidal divides
in the Wadden Sea are not fixed, but can move due to human interference.
Table 9.2: Wadden Sea basin data (Data from Lodder et al., 2019).
Inlet 𝐴𝑏 [km2 ] 𝐻 [m]

Texel 655 1.65
Eierland 157.7 1.65
Vlie 715 1.9
Ameland 276.3 2.15
Pinke 49.6 2.15
Zoutkamperlaag 105 2.25

9.3. The main morphological elements 419
Figure 9.5: Wadden Sea tidal dynamics model, displaying the types of tidal passes and the wan-
tide zones. The movement of the tide on the North Sea is from left to right.
At watersheds the inflowing flood currents through different inlets meet and tidal cur-
rents are practically zero. The watersheds are conducive to fine-grain deposition. This
is because the low current velocities permit the fines to settle. Large tidal eddies are
common along the back barrier flanks, sometimes superimposed on a unidirectional
stream creating a spiralling current. Each catchment area exchanges large volumes of
water with the outside area (the tidal prism – dependent on the tidal range and the
surface area of the catchment area). This results in large inflowing flood and outflow-
ing ebb currents that keep the inlets from closing naturally. The basin area consists
of deeper areas where the flow is concentrated during lower tidal water levels and of
tidal flats that are covered during higher water and exposed during lower water. In
the Wadden Sea, the tidal flats are covered by salt marshes, but they may be covered
with mangroves in other climatic regions (Sect. 2.6.6).
Although the Wadden Sea can be conceptually divided into separate basins during
calm conditions, wind-driven currents and storm surges may lead to inter-basin flow
and transport over the tidal watersheds (Duran-Matute et al., 2016).
9.3.2. Tidal deltas

Because of the large tidal discharges and hence velocities in the inlet or estuary mouth,
there is a strong sediment exchange between the basin and the outside area. Especially
in constricted inlets, this leads to the formation of extensive sand deposits at either side
of the entrance, which are called tidal deltas. In the divergent ebb-tidal discharging

flow, the flow velocities decrease and sediment is deposited, thus creating the ebb-
tidal delta. Just as the ebb-tidal delta consists of material deposited by the diverging
ebb flow leaving the inlet, the flood-tidal delta is created by sediment deposited by
the inflowing flood current. Consequently, basins with large tidal ranges (and limited
wave influence) tend to have well-developed ebb- and flood-tidal deltas and very deep
inlet gorges.
The ebb- and flood-tidal deltas are bisected by channels followed by the tidal currents.
The flats and channels exhibit a highly dynamic behaviour; their locations are continu-
ously changing with time.
Figure 9.6 shows a sketch of a barrier inlet system with typical flood- and ebb-tidal
deltas. The boundaries of the ebb-tidal delta can be found via the no-inlet bathymetry,
where the differences in bottom height are nil between the actual and no-inlet bathy-
metry of the adjacent coast (Fig. 9.20).
A typical ebb-tidal delta includes:
Marginal flood channels The term marginal in the context of marginal flood channels
is used to indicate the positioning of the flood channels to the sides of the inlet.
It is not used to indicate that they are of marginal importance or marginal in
size; for instance, in the estuaries at the SW of the Netherlands the marginal
flood channels are quite distinct.
Main ebb channel The occurrence of separate ebb- and flood-dominated tidal chan-
nels can be explained as follows. In the beginning of the flood cycle, the water
in the main ebb channel continues to flow seaward as a result of inertia. As
a consequence, water initially enters the basin via the path of least resistance,
around the margin of the delta.
Channel margin linear bars These flank the ebb-tidal channel and are built up from
deposits as a result of the interaction of flood- and ebb-tidal currents with wave-
generated currents. In the Dutch inlet systems, these distinct bars are not present;
instead, there are wide flats.
Terminal lobe A rather steep seaward-sloping body of sand, which forms the outer
end of the ebb-tidal delta.
Swash platforms and bars The main ebb channel is flanked by swash platforms, which
are broad sheets of sand. On these swash platforms, isolated swash bars can be
recognised, built up by swash action of waves (the water that washes up on the
platforms after waves have broken). Marginal flood channels are usually found
between the the coast of the barrier islands and the swash platforms.
The flood-tidal delta may be fan- or horseshoe-shaped, with a flood ramp that slopes
upward ending in the ebb shield, that is, the most elevated outer edge of the flood-tidal
delta (Fig. 9.6). The ebb shield is the flood-tidal delta’s equivalent of the terminal lobe
and helps to divert ebb-currents along the margins of the flood-tidal delta. The flood
currents follow the shallowing flood ramp and the flood channels that continue from

basin ocean
t er m
6 1 6 1
ina
l lo
3 4 3 4
be
6 A’ 6
5 5
6 6
3 1 3 1
2 2
6
6
A
flood flow
section A-A’ ebb flow
flood-tidal delta tidal gorge ebb-tidal delta 1. marginal flood channel
2. ebb shield
tidal flat 3. ebb spit
4. swash bar
flood ebb ramp
ramp 5. channel margin linear bar
ebb shield terminal lobe 6. tidal flat
Figure 9.6: Sketch of the morphological elements of a tidal inlet system based on a photo from
an inlet near Sheep Island, North Carolina (2016 imagery from NC OneMap Geospatial Portal
(2020), ‘Credits’ on page 575). The lower part of the figure shows a cross-section of the depth
variation along the inlet system.
the flood ramp. The ebb currents (and their interaction with the flood currents) are
responsible for the ebb spits and the bar-like spillover lobes.
For relatively small basins with abundant sediment supply, such as the Wadden Sea,
the flood-tidal delta spans the entire basin area, viz. flood-tidal deposits are found
everywhere in the basin. For the Wadden Sea we therefore generally only speak of
the outer delta () and the basin; the term flood-tidal delta is not used as a separate
morphological unit. This is quite contrary to, for instance, the USA East Coast basins,
which have relatively small, distinct flood-tidal deltas (as in Fig. 9.6). In these cases,
the flood-tidal deltas can be regarded as the morphologically active parts of otherwise
morphologically inactive basins. Following our Dutch experience, where the entire
basin acts as a flood-tidal delta, we will focus on the ebb-tidal delta in Sect. 9.4. The
flood-tidal delta is not treated separately, but in Sect. 9.6 on basins.
The overall morphology of the tidal deltas, especially the ebb-tidal delta, depends on
the combined action of waves and tides. Wave action is generally considered to act
as a bulldozer on the ebb-tidal delta morphology (Hageman, 1969); it moves sediment
onshore and limits the area over which the ebb-tidal delta can spread out. Hence the
ebb-tidal delta morphology is generally determined by the (dynamic) balance between
a net offshore-directed sediment flux induced by the inlet currents (building up the
ebb-tidal delta) and a net onshore-directed sediment flux induced by offshore waves
(removing sediment from the ebb-tidal delta). It is the flood-tidal delta that benefits

from this onshore sediment flux. Therefore, for tide-dominated entrances, viz. with
limited wave action, ebb-tidal deltas tend to be large relative to the flood-tidal delta.
For wave-dominated entrances, the ebb-tidal deltas tend to be relatively small, whereas
the flood-tidal deltas are well developed, with flood shoals that can be emergent at low
tide.
9.3.3. Basin characteristics

The tides fill and empty the basin via channels that cut through (large) lower and
higher tidal sand and mud flats. The intertidal flats serve to accommodate the tidal
prism: at low water these flats fall dry, while at high water they are submerged. In-
tertidal flats are called platen, slikken or wadden in Dutch3 . Only the most landward,
higher parts remain dry at high water; these vegetated parts are named supratidal flats
or salt marshes (in Dutch: schorren or kwelders3 ).
The combined action of amongst others centrifugal forces, earth rotation (Coriolis)
and inertia, causes the existence of separate ebb and flood channels, channel sills and
channel bifurcations (Sect. 9.6). In general, ebb-dominant channels follow a meander-
ing course, whereas flood-dominant channels shoal landwards.
A very good description of the ebb and flood channel systems is found in Van Veen
et al. (2005). It was originally published in Dutch by Van Veen (1950), translated and
annotated by Van Veen et al. (2002) and republished by Van Veen et al. (2005). Along
barrier island coasts, the basins are often rectangular or nearly square, and the channel
structure is often more branched than braided (Fig. 9.7). The branching structure is
found to show fractal characteristics, viz. if you zoom in on a part, you find subsets
that look like the whole figure.
In the case of larger rivers discharging into tidal basins, the tidal basin is often funnel-
shaped, and the channel structure is not as branched, but is potentially braided (Figs. 9.8
and 9.9). In areas without width restriction (very large estuaries, e.g. Thames estuary),
the channels followed by the flood current are generally different from the channels
followed by the ebb current. They may connect due to meander action. If the width
is more restricted (e.g. Scheldt estuary), the flood and ebb currents partly follow the
same pathways.
Section 9.6.1 describes the complex geometry of ebb and flood channels in a bit more
detail.
3
The terms slikken and schorren are mostly used for the southwestern delta area, whereas the terms
wadden and kwelders are exclusively used for the Wadden Sea.

ebb-
f tida
ld
f elt
f
a
f f
Wadden Sea
Figure 9.7: Sketch of ebb and flood channels in a typical Wadden Sea basin with an inlet between
dune-islands, several flood channels coming in (indicated with f) and an ebb channel (e) and
delta with a tendency to turn to the left, due to the ebb tide leaving to the left, consistent with
the propagation direction of the tidal wave along the Dutch coast (see Fig. 3.31). After Van Veen
et al. (2005).
f f f f
f f
Figure 9.8: Sketch of ebb (e) and flood channels (f) in a wide estuary (for example, the Thames or
Wash). Meander action may bring the ebb channel in connection with any of the flood channels
(after Van Veen et al., 2005).
f f
e f f f
e
f
Figure 9.9: Sketch of an ideal system of ebb (e) and flood (f) channels (Scheldt estuary). It has
a meandering main (ebb) channel and flood channels starting in every bend. The latter have a
double function; viz. 1. filling the tidal sand flat in the inner bend of the main channel; 2. serving
the cut-off current of the bend. After Van Veen et al. (2005).

9.4. The ebb-tidal delta or outer delta

9.4.1. Waves and currents at the outer delta
Clearly, the morphology of the outer delta is highly complicated and variable. This
means that waves and currents encounter a very complex bed topography, with length
scales which are not much larger than the wavelength of wind waves or swell, for
instance. Via refraction, diffraction and reflection, this can lead to complex wave pat-
terns with a strong spatial variability.
The currents in the vicinity of a tidal inlet are partly tidal, partly wave-driven and
partly wind-driven. The tidal currents are primarily concentrated in the main channels,
the wave-driven currents in areas where waves are breaking. Wind-driven currents
occur mainly during storm events and are therefore rather episodic, but nonetheless
important (though often forgotten in modelling studies!).
Wave patterns
As to the overall pattern of the waves, there are a few points to consider, such as the
penetration of wave energy into the gorge and the sheltering from wave exposure in
various parts of the system. Let us consider a typical inlet between two barrier islands,
representative of one of the Wadden Sea inlets. In the case of normally incident waves,
the entire delta edge and the gorge are exposed to wave energy. Due to refraction, the
wave crests turn to run more and more parallel to the depth contours. Therefore, wave
energy is concentrated in the central front edge of the ebb-tidal delta, whereas the side
lobes of the delta will be less exposed than the central front part. Due to breaking on
the shoals of the outer delta, the gorge will also be less exposed (Fig. 9.10).
exposed
less
exposed
less
exposed
less
exposed
sheltered sheltered
Figure 9.10: Exposure of outer delta to normally incident waves.
In the case of obliquely incident waves, one side lobe will be exposed, while the other
will be sheltered. By implication, the wave climate at the sheltered side lobe will be
different from that offshore or at the front edge of the delta. The western lobe in
Fig. 9.11, for instance, will be fully exposed to westerly waves, less exposed (due to
refraction) to northerly waves and sheltered from easterly waves. As a consequence,

9.4. The ebb-tidal delta or outer delta 425
wave-driven longshore currents in the vicinity of this western lobe are predominantly
eastbound. Similarly, the currents near the eastern lobe are predominantly westbound.
exposed exposed
sheltered exposed
exposed sheltered
less less
exposed exposed
sheltered sheltered sheltered sheltered
Figure 9.11: Sheltering and exposure to obliquely incident waves.
Another aspect which deserves attention is wave penetration into the inlet system.
Clearly, the barrier islands provide considerable shelter to the basin, but wave energy
can penetrate through the gorge. The latter, however, is sheltered to a certain extent
from the open sea by the outer delta. Hence, the wave energy which reaches the back
of the gorge is much less than offshore. Subsequently, this energy usually radiates
into the basin, where the energy density and the wave height rapidly decay. As a con-
sequence, the wave energy which penetrates from the open sea into the basin is usually
rather small and is restricted to the area right behind the gorge. Wave energy further
into the basin is generally due to waves generated inside the basin, if the prevailing
wind meets a sufficiently long fetch.
A drastically different situation may arise if channelling of the wave energy occurs
(Fig. 9.12). This phenomenon is associated with trapping of the waves in the channel
if the wave ray direction is almost parallel to the banks. In such a case, areas which are
at first sight sheltered from sea waves can be exposed to relatively high wave energy
and thus to much more erosion than expected.
energy
input
exposed
Figure 9.12: Exposure of a sheltered area due to wave channelling.
The complexity of the bed topography and the mixture of sea/swell waves coming
from offshore, and locally generated waves of a much shorter period inside the basin,

make wave modelling for tidal inlets particularly difficult. A model for long-crested
monochromatic waves (i.e. wave fields with one direction, one period and one height
in each point of the model domain) will not work here. Instead, one would need a
wave model which allows for irregular, short-crested waves of various periods and
includes wave generation. Such (often fully spectral) wave models have been available
for some time for deep-water wave prediction, but this concept has been translated
only relatively recently to shallower water.
Tidal residual currents

In Sect. 5.7.6 we already discussed the tidal residual flow pattern as a result of the
accelerating flow through the inlet gorge. The tidal residual current pattern in the
highly schematised situation discussed there boils down to a quadruplet of gyres, two
at either side of the inlet (Fig. 5.74). In reality, the residual current picture is much more
complicated than this. There is usually a distinct ebb-dominated current (i.e. averaged
over the tide the flow is in the ebb direction) over the outer delta, and often there are
flood channels near the tips of the islands. In the Wadden Sea, well-developed flood
deltas are hardly found: the entire basin acts as a flood delta, and the corresponding
flood-residual current is difficult to distinguish. Many inlets on the East Coast of the
USA, however, do exhibit such a feature and have a well-developed residual circulation
inside the inlet.
Another type of residual current is Stokes drift (Sect. 5.7.6) which is dependent on
the phase-coupling between the horizontal and the vertical tide. It was shown that
if the horizontal tide and the vertical tide are more or less in phase (𝜑 ≈ 0), there
can be a considerable residual current. This becomes even more apparent for shoals
bordering a tidal channel which are flooding and drying during the tide. There, the
largest part of the flooded stage coincides with the flood tide, so there must be a flood-
dominated residual current (averaged over the tide the flow is in the flood direction).
By implication, the residual current in the channel must be ebb-dominated.
Secondary flow components can arise from, amongst others, the curvature of the tidal
current (Sect. 5.7.6). The curvature-induced secondary flow does not change sign as the
tide turns and is therefore, in the upper part of the water column, always directed away
from the centre of curvature of the flow; in the lower part always towards it. Thus
the curvature-induced secondary flow contributes, for instance, to the maintenance of
shoals (see Fig. 9.13).
Curvature-induced secondary flow on an outer delta has been hypothesised to play an
important role in the case of the groyne which was built in 1995 to protect the north
part of the coast of Texel, one of the barrier islands of the Dutch Wadden Sea (Fig. 9.14).
The idea was to let this groyne interrupt the longshore drift into the inlet called Eier-
land Inlet (Eierlandse Gat in Dutch). Studies with 2D depth-averaged models (thus
not resolving secondary flow patterns) revealed that this would lead to the ‘textbook’
pattern of updrift accretion and downdrift erosion, that is to say, to strong erosion

shoal
Figure 9.13: Curvature-induced secondary flow: near the bed the flow is always directed towards
the shoal.
north of the groyne. It was foreseen that the groyne would have to be connected by a
hard structure to the existing revetment which protects the tip of the island. Once the
groyne had been built, however, accretion occurred at either side of it (Fig. 9.14), to the
effect that it is now largely buried in sand (so the project turned out to be extremely
successful, although at the cost of other coastal elements). Based on model computa-
tions by Steijn et al. (1998), amongst others, this has been attributed to a large extent
to the curvature-induced secondary flow in the gully around the tip of the groyne,
which consistently brings sediment towards the groyne (see Fig. 9.14). However, new
computations suggest that the ebb-tidal currents during spring tides, rather than the
secondary flow, are responsible for the large amounts of sedimentation at the north
side of the dam (see Visser, 2014).
(a) 1990 (b) 1999
(c) 2005 (d) 2011
Figure 9.14: Observed morphological evolution around the Eierlandsedam, a long groyne at the
north point of Texel built in 1995. Photos from Rijkswaterstaat (see ‘Credits’ on page 575).

In summary, there is a variety of mechanisms which lead to tidal residual currents,

and hence to residual sediment transports and morphological changes. In shallow
areas such as a tidal inlet system, we should not think of the tide as a small-amplitude
wave.
Wave-induced currents
As stated before, an outer delta is nowhere near a straight prismatic coast. This means
that the way of thinking about wave-driven currents as used in Ch. 5 is hardly ap-
plicable here. In that chapter, the alongshore and cross-shore momentum equations
were separated, the former describing the momentum balance for a uniform longshore
current and the latter the set-up balance. In the case of a complex topography with in-
terrupted breaking, however, the situation is much more complicated; the momentum
equations cannot be separated and need to be resolved in two horizontal dimensions; in
principle, the wave-driven current in this type of situation is 3D. A relatively simple ex-
ample, a shoal on which waves are breaking, was already given in Sect. 5.5.7 (Fig. 5.49).
The general pattern was a net flow in the wave direction over the shoal, turning sea-
ward again via the channels between the shoals. In tidal inlets, the wave-driven cur-
rents around shoals on the outer delta can be so strong that they dominate the tidal
residual currents. This has strong implications for the sediment bypassing mechanism,
see Sect. 9.4.2.
Wind-induced currents
An often forgotten type of currents in tidal inlet systems are wind-driven currents,
either directly, via the wind-induced shear stress on the water surface, or indirectly, via
the set-up of the water level against the coast. This means that the wind shear stress
components need to be added to the depth-averaged momentum equations. These
wind stress terms have the water depth in the denominator, which means that the
wind tends to be more effective in driving a current when it acts on shallower water.
Due to the large variations in water depth which are inherent to a tidal inlet system,
the wind-driven current field will therefore strongly vary in space. In general, the
flow will tend to follow the wind in the shallower parts, and to oppose it in the deeper
parts, but this picture can be complicated greatly by spatial interactions via water level
gradients.
As the wind forcing acts at the water surface, there will also be an effect on the vertical
structure of the flow: the primary flow profile (e.g. logarithmic) will be disturbed by
a ‘secondary’ flow component which follows the wind in the upper part of the water
column and goes against it in the lower part. Note that this secondary flow has to
be superimposed on the primary flow (the depth-averaged circulation with a primary
flow profile). The result can be a complex 3D flow pattern.
Probably even more important than the direct wind-induced forcing is the effect of
the water level set-up during a severe storm (Sect. 5.6). Although the peak of the

wind speed usually does not last much more than a few hours, the water level set-up
can last much longer. However, it takes time for the backbarrier basin to follow the
water level in the open sea, as raising the water level of such a basin by a few metres
takes huge amounts of water, all of which has to be squeezed in through the inlets.
Hence it sometimes occurs that the ebb current is entirely suppressed and that there
is a flood current in the inlet for a day or more. Clearly, such events may influence
the inlet morphology considerably. It is also clear that new inlets and channels will
preferentially be created under such conditions, when there is a large head difference
between the sea and the basin.
3D combined current field

In summary, we have seen that the current field around an inlet is essentially more
complex than on a uniform straight coast, and that all constituents of this field are
essentially 3D. So when modelling currents around inlets, we have to think 3D, even if
we decide to use a 2D depth-averaged model. The interpretation and post-processing
of the results (e.g. bed shear stress to be put into the sediment transport model) is
much less straightforward here.
Wave-current interaction
The tidal and wave-driven current pattern on the outer delta is largely concentrated in
the deeper channels. Consequently, there can be strong currents which affect the wave
propagation via current refraction (see Sect. 5.2.3). This may even go as far as wave
blocking (for a wave with opposing current). When standing on the coast overlooking
a tidal inlet, one often observes a sharp distinction between areas with waves and areas
with a flat water surface. This is simply because the current is strong enough to prevent
the waves from entering this area. Moreover, where waves do occur, their pattern is
often quite irregular, again as a consequence of refraction on a strongly varying current
field. This form of wave-current interaction makes it particularly difficult to predict
the wave field on the outer delta. Such a prediction should be based on a combined
wave and current model, and both should be carefully calibrated in order to find the
right pattern.
A more straightforward form of wave-current interaction is the effect of waves on
bottom shear stress experienced by the current (see Sect. 5.5.5 and Sect. 6.5). The
mechanisms underlying this effect are not essentially different from those in the case
of a prismatic coast, though waves and near-bed currents can have arbitrary directions
now. The result is a complex 3D boundary layer with a strongly veering velocity vector
and a non-trivial direction of the bed shear stress. In large-scale models like the ones
we use for tidal inlet systems, however, the effect on the direction is usually ignored
and the shear stress is assumed to be opposite to the mean current. What we do have
to take into account, however, is the bed shear stress enhancement induced by the
waves. This may have major effects on the current pattern: especially the tidal flow

will tend to avoid shallow areas, where wave action and shear stress enhancement are
strongest.
9.4.2. Sediment transport patterns

Figure 9.15 shows the effect of wave-driven alongshore sediment transport gradients
on the shape of barrier islands. The combination of wave, current and sediment trans-
port mechanisms on the outer delta gives rise to a typical residual sediment circulation
pattern. Figure 9.16 shows a highly schematised picture. Reality is usually more irreg-
ular and diffuse (Fig. 9.17).
shoal
attachment
spit formation
predominant wave direction

predominant longshore transport direction
Figure 9.15: Wave-driven longshore sediment transport along the barrier islands on either side of
a tidal inlet. The transport gradients lead to spit formation (left) and a drumstick-shaped barrier
island (right).
inlet
wave induced throat
tide induced
Figure 9.16: Schematic diagram of the outer delta residual sediment transport (hypothesised
by De Vriend et al., 1994): 1) the wave-induced longshore sediment transport that bypasses the
inlet; 2) the ebb-dominated main channel that exports sediment and 3) the flood channels that
import sediment. In this case the tidal wave is assumed to propagate along the coast from west
(left) to east (right).
In general, the flood channels on the outer delta carry sediment from the adjacent
coasts to the inlet, mostly during episodic events. Depending on the demand4 of the
4
For the sand demand (as opposed to sand surplus) of a basin, the term ‘sand hunger’ is commonly used
in the Netherlands. Internationally more established is the geological term ‘accommodation space’,

-5
-10 tide-induced transport
wave-induced
Vlieland -5 longshore transport
-5
-5
-10
-5
Texel
0 1 2 3 km
Figure 9.17: Residual sediment transport in the Eierland Inlet before the Eierlandsedam was built,
composed from observations and various numerical modelling studies according to Ribberink
et al. (1992).
basin, this sediment is transported either into the basin or sometimes into the main
ebb channel and then towards the edge of the outer delta. There, part of it is picked up
by wave-driven and/or tidal longshore currents and transported along the delta edge,
ultimately towards the downdrift island, or towards the updrift one, depending on the
local tidal state and wave direction.
Another part of the sediment from the ebb channel, as well as part of what is brought
into the area by the longshore drift along the updrift island, ends up in the shoal system
on the outer delta. Due to the various hydrodynamic and sediment transport processes
around these shoals, there is a long-term residual transport and a slow migration of
the shoals in the direction of the longshore drift (Fig. 9.18). The time needed for a
shoal to cross the inlet may be tens of years (for instance, for the Borndiep, also called
Ameland Inlet, between Terschelling and Ameland, it is typically 40 years).
It can be concluded that part of the littoral drift continues its way over the ebb-tidal
delta to the downdrift coast, while another part is diverted into the tidal inlet by the
flood. During ebb, part or all of the sediment carried into the tidal inlet is returned
to the ebb-tidal delta, from where it can be transported to the downdrift coast again.
The ratio between the volumes of littoral drift that are bypassed directly and that are
bypassed via the ebb-tidal delta depends on the tidal prism and the magnitude of the
littoral drift.
which can be both positive and negative and could apply to a backbarrier area as well as to a shoreface
(see Sect. 7.4).

wave-driven currents (zigzag pattern)
tide
sediment drift (bypassing)
tide
Figure 9.18: Sediment bypassing via shoal migration on the outer delta. The zigzag pattern
of wave-driven currents results in a net sediment drift from the updrift to the downdrift island.
(concept by De Vriend et al., 1994; Ehlers, 1988).
Bruun and Gerritsen (1959) propose a parameter 𝑟 to indicate the type of bypassing.
Later Bruun and Gerritsen (1960) and Bruun (1978) converted the original formula for
𝑟 to:
𝑃
𝑟= (9.2)
𝑀tot
in which:
𝑃 the tidal prism m3

𝑀tot the total littoral drift m3 /yr
The tidal prism is the volume of water entering or leaving the basin per half tidal cycle.
Note that the parameter 𝑟 is a measure for the relative wave/tide influence at the inlet.
For large values of 𝑟 (tidal prism much larger than the wave-induced littoral drift), the
inlet is tide-dominated. For values of 𝑟 > 300, the bypassing of sand is predominantly
via the inlet. On the other hand, for small values of 𝑟 (wave-dominance) the bypassing
is predominantly via shoals on the ebb-tidal delta.
In Table 9.3 some values for 𝑟 are given. The tidal inlets along the West Frisian Barrier
islands all show a large value fFor 𝑟, which indicates that tidal flow bypassing is dom-
inant. In general, increased values of 𝑟 produce greater seaward displacement of inlet
bars (Oertel, 1988). The first eight inlets in Table 9.3 are Wadden Sea inlets. The given
values for 𝑟 show that 𝑟 decreases in an eastward direction. This seems to correspond
with the occurrence of tidal sand ridges at the more easterly located inlets.

Figure 9.19 shows one of those inlets, viz. the Wichter Ee (𝑟 ∼ 60), one of the meso-tidal
inlets of the eastern (German) Wadden Sea. The wave-induced longshore sediment
transport along most of the North Sea coast is from west to east (left to right in the
photo), as can be seen from the sand abundance at the eastern tip of Norderney and the
lack of sand at the western tip of Baltrum. Apparently, the sand bypasses the inlet via
a number of shoals, which migrate in the outer delta from the updrift to the downdrift
island. This probably forms the principal sediment bypass mechanism across the inlet.
The shoals get welded with the coast of Baltrum somewhere east of the tip, where sand
abundance is found again.
Table 9.3: Values for 𝑟 (Eq. 9.2) of some tidal inlets (from: Bruun, 1978).
Total Littoral
Inlet Tidal Prism 𝑃 Drift 𝑀tot 𝑟
[106 m3 ] [106 m3 /yr]
Inlet of Texel, Holland 1000 ∼1 ∼1000
Wadden
⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃖⃗
Eierlandse Gat, Holland 200 ∼1 ∼ 200

Inlet of the Vlie, Holland 1000 ∼1 ∼1000
west to east
Borndiep, Holland 500 ∼1 ∼ 500

Sea Inlets from
Friesche Zeegat, Holland 300 <1 e ∼ 400

Nordeneyer Seegat, Germany 160 <1 e ∼ 250
Wichter Ee, Germany 42 <1 e ∼ 60
Otzumer Balje, Germany 110 <1 e ∼ 150
John’s Pass, Florida 14 0.1 140
Longboat Pass, Florida 20 0.1 200
Aveiro, Portugal 60 1 60
Big Pass, Florida 10 0.01 100
Masonboro Inlet, Florida 20 0.3 70
Saratosa Pass, Florida 3 0.1 30
Penang Harbour, Malaysia 700 0.6 ∼1000
Krishnapatam, India 10 0.6 17
Thyborøn, Denmarka 100 0.8 120
East Pass, Floridab 40 0.1 400
Oregon Inlet, N. Carolina 60 1 60
Tan My, Vietnam 47 1.6 30
a
jettied inlet
b
weir jetty
e
estimate
The shoals are separated by distinct channels, which all branch off from the main
channel. In this particular case, the latter is strongly deflected eastward, but other
inlets show that the main channel can just as well have a different direction. Fur-
thermore, the main channel tends to migrate and change direction through time. The
main channel and its branches are probably all ebb-dominated (in this case meaning

(a) date: 2.8.1982
(b) date: 10.9.1982
Figure 9.19: Aerial photographs of the morphology of the tidal inlet Wichter Ee, between the
islands Norderney (left) and Baltrum (right). Copyright Ehlers (1988), see ‘Credits’ on page 575.
The date convention as used by Ehlers (1988) is unknown to us, but we expect that the photos
are from 2 August and 10 September 1982 – hence a good month apart – and can be considered
as pre-storm and post-storm, respectively.
that, averaged over the tide, there is a residual flow in the ebb direction). If this is true,
the principle of conservation of water mass requires that other parts of the area are
flood-dominated. Often there are more or less distinct flood channels, e.g. around the
tip of the updrift island. In the case of the Wichter Ee, at least at the time that the
photos of Fig. 9.19 were taken, these flood channels were not very well developed.
From a project at Nerang River, Australia, it becomes clear what can go wrong when
ignoring the nature of this sediment circulation. Here, an inlet to a tidal lagoon was
converted into a shipping channel by building groynes at either side of the inlet and

bypassing the longshore drift via a pumping system. This scheme works well from the
point of view of downdrift erosion, but it does a much worse job for the outer delta
and where the hydrodynamic conditions in the inlet are concerned. By artificially
bypassing all the sediment which arrived on the updrift side of the inlet, the shoal
bypassing system was cut off from its sand supply, the shoals rapidly degraded, and the
outer delta disappeared. As a consequence, the tidal motion into and out of the lagoon
was no longer hampered by the delta and increased considerably. This led to a dramatic
increase of the flow velocities in the inlet, due to which its navigational function (the
prime reason for carrying out the scheme) has become heavily threatened.
9.4.3. Empirical relationships: volume of the ebb-tidal delta

The great importance of the ebb-tidal delta is evident from the large volume of sand
that is accumulated in this delta. This ebb-tidal delta volume (Fig. 9.20) will be lar-
ger for low onshore-directed wave energy and large tidal forces. Under these circum-
stances the ebb-tidal delta can extend far seawards, without a distinct terminal lobe.
Quite often the volume of sand deposited in an ebb-tidal delta is much larger than
the total volume of the adjacent beaches. Although the governing physical processes
(Sects. 9.4.1 and 9.4.2) are complex and not yet fully understood, it is beyond dispute
that sediment exchange takes place between the adjacent barrier beaches and the ebb-
tidal delta.
The volume of sand stored in the ebb-tidal delta has empirically been related to the
tidal prism of the backbarrier system (Fig. 9.21). The relationship in Fig. 9.21 was first
derived for outer deltas in the USA and reads:
𝑉𝑜𝑑 = 𝐶𝑜𝑑 𝑃 1.23 (9.3)
in which:
𝑉𝑜𝑑 sand volume stored in the outer delta m3

𝐶𝑜𝑑 empirical coefficient m−0.69
𝑃 tidal prism m3
Figure 9.21 shows that the coefficient is dependent on the wave climate, in such a
way that for the same tidal prism the volume in the ebb-tidal delta is smaller for more
energetic waves. In Fig. 9.21 the tidal prism is determined based upon the spring or
diurnal tidal range. Using the mean tidal prism 𝑃 rather than the spring tidal prism,
Eysink and Biegel (1992) found 𝑐𝑜𝑑 = 63.3 × 10−4 for the Wadden Sea (𝑃𝑠 /𝑃 ≈ 1.15).
Even small changes in the tidal prism or the wave conditions (by nature or human
interferences) may result in changes in the sand volume of the ebb-tidal delta. For
example, suppose the tidal prism of an inlet is enlarged after flooding of part of the
hinterland. According to the empirical relationships (Fig. 9.21), this will result in an

Figure 9.20: Definition of the ebb-tidal delta volume. A: the no-inlet bathymetry. B: the ebb-
tidal delta bathymetry. The ebb-tidal delta volume is defined as the difference in bottom height
between the actual and the no-inlet bathymetry of the adjacent coast for the entire extent of the
ebb-tidal delta along the shore (see middle panel).
enlargement of the volume of sand in the ebb-tidal delta. The sediment required for
this enlargement may originate from the adjacent barrier coast, the backbarrier system
(i.e. the basin) or from offshore. Most probably it will be a combination of these three
sources, the distribution of which is very hard to determine without thorough know-
ledge of the underlying physics. Probably, erosion of tidal gullies will contribute most
to the ‘sand demand’ of the (ebb)-tidal delta. Because of the relatively small morpho-
logical timescales of such adjustments, they are able to respond fast. Due to increased
flow velocities, the tidal channels will (immediately) start to erode. The eroded ma-
terial will mainly be deposited on the ebb-tidal delta where the flow decelerates. On
the other hand, the contributions of the adjacent barrier coasts may be smaller, but
these coasts are more vulnerable, as small changes in the sediment balance may cause
severe coastline retreats.
Removal of sand from the outer delta

The sand in the outer delta may be used as a sand source for e.g. beach nourishment, or
it may be necessary to dredge navigation channels through the outer delta. In both of
these cases the sand is removed from the inlet system and the system will be disturbed
from its ‘equilibrium state’. Since the tidal prism and the protrusion rate of the delta

1000
500 Texel Inlet

Vlie Inlet
200
Ameland Inlet
100
sediment volume in outer delta [10 6 m 3 ]
Eierland Inlet
50
20
10
1
wave climate class
0.5 mildly exposed H 2T 2 < 2.8 m2s2
moderately exposed H 2T 2 2.8 to 28 m2s2
highly exposed H 2T 2
> 28 m2s2
0.2
Dutch Wadden Sea Inlets
0.1
2 5 10 20 50 100 200 500 1000 2000 5000 10000
tidal prism P based on mean spring or diurnal range [10 6 m 3 ]
Figure 9.21: Empirical relationship between volume of sand in the outer delta and the tidal prism.
The effect of the wave climate on this relationship is also shown; the general trend is that the
ebb-tidal delta sand volume reduces with increasing wave activity. USA data from Walton and
Adams (1976), Dutch data from Eysink and Biegel (1992).
remain constant, the supply of sand into the basin via the flood channels does not
change.
In the first years after the start of dredging operations, the delta will retreat as a result
of sand loss, while the barrier islands are not affected. This trend reverses, however, as
the outer delta will demand sand in order to return to its equilibrium state. Assuming
that the inlet system is a ‘closed system’, the demanded sand will be supplied by the
adjacent coastlines, mainly by the downdrift coastline that will be cut off from its
supply. The erosion of the barrier islands will begin on the foreshore, the beach profile
steepens and gradually the erosion of the beach will become evident. Since the erosion
of the barrier islands occurs tens of years after the dredging, the cause of the erosion
may not be linked to the dredging operation.
In order to minimise the additional erosion it may be wise to consider the option of
dredging sand from the inner (flood-tide) delta instead of from the outer delta. If the

amount of sand removed from the inlet system per year is less than or equal to the
total volume of sand transported into the basin per year via the flood channels, then
the effects of dredging the inner and outer delta on the long term (decades to centuries)
will be the same. However, the outer delta retreat and thus the additional erosion on
the downdrift island will be slower in the case of dredging of the inner delta.
Effects of the construction of a groyne on the updrift island

The construction of a groyne on the updrift island will trap sediment that previously
entered the inlet via the updrift flood channels. Not only can lee-side erosion be ex-
pected, but erosion of the outer delta and the downdrift island as well. The outer delta
acts as a buffer for the other elements of the inlet system. When sand is removed from
one portion of the inlet system, the outer delta will respond by sharing sand with the
other elements. This supply, however, is only a ‘short-term loan’ which will be repaid
by the adjacent coastlines in the longer term. So erosion of the outer delta can be ex-
pected to occur rapidly, while on the downdrift island this will take place after tens
of years. In the case of a longer groyne, more sand will be trapped behind the groyne
and it will also take more time before bypassing begins. Consequently the erosion of
the outer delta and the downdrift island will be more severe.
9.5. Stability of the inlet cross-sectional area

9.5.1. Escoffier’s model
A tidal inlet is not fixed, but a dynamic entity governed by important factors such as
tidal currents, storms, the tidal prism (the storage volume of the estuary between low
tide and high tide level) and littoral sediment transport. Escoffier (1940) was the first to
study the stability of the cross-sectional area of the inlet proper. Because of the littoral
drift leaving and entering the inlet with the tide, there can be considerable variation
in the cross-sectional area.
Escoffier’s predominantly qualitative study led to a relationship for the maximum
cross-sectionally-averaged entrance channel velocity (𝑢𝑒 , with the subscript indicating
the entrance) for a given estuary or inlet (see Intermezzo 9.4 for an approximation). He
related 𝑢𝑒 to the hydraulic radius of the channel (𝑅), its cross-sectional area (𝐴𝑒 ) and
the tidal range in the estuary (Δℎ). Since this calculation is made for a given inlet,
other variables such as the channel bed roughness, its length, the surface area of the
inlet, and the tidal range at sea have then all become more or less constant. Escoffier
combined the variables for a given inlet into a single parameter 𝑥, such that a larger
entrance cross-section results in a larger value of 𝑥. Qualitatively, he found that 𝑢𝑒
varied as a function of 𝑥 more or less as shown in Fig. 9.22.
A curve like in Fig. 9.22 is called a closure curve. In the range from A to C on this
curve, the entrance channel is so small that it chokes off the tidal flow, so that the tidal

9.5. Stability of the inlet cross-sectional area 439
difference within the estuary will be less than at sea. For that reason the channel velo-
city will increase for an increasing cross-section. In terms of Eq. 9.5: with increasing
𝐴𝑒 , 𝑃 increases so much that 𝑢̂𝑒 increases. On section C–E of the curve, the tidal flow
is not choked off any longer (now 𝑃 remains constant for increasing 𝐴), so that the
maximum current velocity decreases as the channel becomes larger. For any estuary
or inlet, a closure curve can be computed using a hydrodynamic model. This can be
either a numerical model or a simpler analytical model. Simplified analytical solutions
can for instance be obtained by assuming a short basin that responds in pumping mode
(a uniformly fluctuating water level, see Sect. 5.7.3).
Intermezzo 9.4 Cross-sectional velocity for a sinusoidal tide
The maximum cross-sectional velocity is the maximum velocity during the tidal
cycle. To understand its behaviour as a function of the cross-sectional area, it can
be approximated as the amplitude 𝑢̂𝑒 of a sinusoidal tidal motion 𝑢. In that case
we can relate the maximum cross-sectionally-averaged entrance velocity 𝑢𝑒 = 𝑢̂𝑒
to the tidal prism 𝑃. The tidal prism 𝑃 is equal to the time integral of the inflow
during flood or to the outflow during ebb (cf. Eq. 9.1):
1/2𝑇 1/2𝑇
2𝜋 𝑇 𝐴𝑒
𝑃 =∫ 𝐴𝑒 𝑢𝑑𝑡 = ∫ 𝐴𝑒 𝑢̂𝑒 sin ( 𝑡) 𝑑𝑡 = 𝑢̂ (9.4)
0 0 𝑇 𝜋 𝑒
and thus:
𝜋𝑃
𝑢̂𝑒 = (9.5)
𝐴𝑒 𝑇
with 𝑇 being the tidal period.
C
max. channel velocity
B D
ble sta
sta ble
un
E
A
0
Figure 9.22: Channel velocity geometry relationship.
Escoffier’s next step was to introduce the concept of an equilibrium maximum velocity
𝑢𝑒𝑞 , below which the velocity in the channel is too low to erode sediment and keep the
entrance channel open. This critical velocity is more or less independent of the channel

geometry, according to Escoffier, and he plotted it as a horizontal line on Fig. 9.22, viz.
independent of the cross-section. In reality 𝑢𝑒𝑞 is generally a weak function of the
cross-sectional area, but at first order this effect can be neglected.
The fate of an estuary or tidal inlet can now be predicted by examining the curve
ACE in relation to 𝑢𝑒𝑞 . Obviously, if 𝑢𝑒 is always less than 𝑢𝑒𝑞 (for all values of 𝑥
the closure curve lies below 𝑢𝑒𝑞 ), then any sediment deposited in the entrance will
remain there and the estuary will be closed off eventually. However, if a curve of 𝑢𝑒
versus 𝑥 intersects the 𝑢𝑒𝑞 line as shown at B and D in Fig. 9.22, then a variety of
situations can exist. If, for example, the channel dimensions place it on section A–
B of the curve in Fig. 9.22, then the channel is too small and the friction too high
to maintain itself; so it will be closed by natural processes. If the channel geometry
places it on section D–E of the curve, it will also become smaller, but as it does so, the
velocity 𝑢𝑒 will increase; sedimentation continues until point D is reached. Lastly, if the
channel configuration places it on section B–D of the curve, then erosion takes place
until point D is again reached; point D represents a stable situation. Since 𝑢𝑒 = 𝑢𝑒𝑞
represents the stable (D) and unstable (B) equilibrium conditions, the relationship for
𝑢𝑒𝑞 is called the equilibrium flow curve or the stability curve. The equilibrium velocity
curve that Escoffier introduced assumes that the equilibrium velocity 𝑢𝑒𝑞 is a constant
that depends only on the sediment diameter, and suggests that a good approximation
of the velocity is 3 ft/s (0.9 m/s).
With this insight, it is now possible to evaluate the influence of changes in an estuary
mouth. Since point D represents a naturally stable situation, most natural estuaries
will tend to lie more or less in that region. Of course, a severe storm can cause severe
sedimentation, largely filling the entrance, which is then suddenly in the state repres-
ented by section A–B of the curve. In such a situation, immediate dredging is called
for to prevent complete closure. It is not necessary to restore the original situation,
however, since, once the entrance geometry places it on section B–C–D of the curve
in the figure, nature will do the rest of the work given enough time. Apparently, Escof-
fier’s model incorporates a feedback between hydrodynamics and morphology and can
therefore be seen as a morphodynamic model for the entrance of tidal basin.
Shipping interests may make it desirable to enlarge the entrance of a given estuary to
accommodate larger ships. If such an expansion scheme places the channel on section
D–E of the curve, continual dredging operations will be necessary. It may be possible
to carry out the expansion and prevent the need for continual dredging by changing
the channel alignment and artificially constricting its width – techniques often used
in rivers – so that the larger channel cross-section remains stable. Translating such
changes into a figure such as Fig. 9.22 means that a new curve of 𝑢𝑒 versus 𝑥 has been
generated, which generally yields a slightly higher value of 𝑢𝑒 for a given 𝑥 value. This
results in point D, the equilibrium situation, being moved to the right in the figure.

9.5.2. Empirical equilibrium cross-sectional area

One of the most important questions to be answered in order to use the approach by
Escoffier outlined above is “what is the stable equilibrium condition of a basin?” or in
other words, “when has point D in Fig. 9.22 been reached?”
Probably, the first relevant reference in this context is LeConte (1905), who, based on
observations of a small number of inlet entrances and harbours on the Pacific coast of
the USA, found an empirical relationship between the inlet cross-sectional area and the
tidal prism. The pioneering work of LeConte (1905) was then followed up by O’Brien
(1931), O’Brien (1969) and Jarrett (1976).
The general form of the empirical relationship for the equilibrium cross-section based
on the tidal prism is as follows:
𝐴𝑒𝑞 = 𝐶𝑃 𝑞 (9.6)
in which:
𝐴𝑒𝑞 the minimum equilibrium cross-section of the entrance channel m2

(throat), measured below MSL
𝑃 the tidal prism, often the spring tidal prism m3
𝑞 coefficient −
𝐶 coefficient m 2−3q
This equation seems to be equally valid for large estuary mouths, bays and tidal la-
goons. The coefficients 𝐶 and 𝑞 are empirical parameters obtained from observational
data. The coefficient 𝐶 is not dimensionless and has dimensions of 𝐿(2−3𝑞) . Several
researchers have reported values for 𝐶 and 𝑞 that vary with the type of inlets con-
sidered (Dutch Wadden Sea inlets give different values than the USA Atlantic coast
inlets, for instance). The coefficient 𝑞 is order of magnitude 1. In metric units, 𝐶 is in
the range 10−4 to 10−5 . For instance, O’Brien (1969) showed that for 28 USA entrances
𝐶 = 4.69 × 10−4 and 𝑞 = 0.85 are best-fit values applicable to all entrances, when 𝑃 is
measured in cubic metres (m3 ) and 𝐴 in square metres (m2 ). But, when limited to 8
non-jettied entrances, he derived 𝐶 = 1.08 × 10−4 and 𝑞 = 1 as best-fit values. The as-
sumption of Escoffier that the equilibrium velocity 𝑢𝑒𝑞 is approximately 0.9 m/s implies
assuming 𝑞 = 1 and 𝐶 = 7.8 × 10−5 m−1 . This is demonstrated in Intermezzo 9.5.
A certain combination of values for 𝐶 and 𝑞 is valid only for a set of inlets that have the
same sediment characteristics (which determine how easily sediment is transported)
and that are subject to the same wave conditions (important for the littoral sediment
transport that reaches the inlet) and tidal conditions (since tidal prism and tidal period
determine the current velocity – see Intermezzo 9.4 – and therefore the tidal transport
capacity). Obviously, the 28 inlets in O’Brien’s dataset do not all have the same littoral

drift and tide conditions, and it is doubtful whether this is the case for the 8 non-jettied
entrances.
The dependencies of 𝐶 and 𝑞 on littoral drift and tidal conditions can be understood
when looking at the equilibrium condition from the perspective of sediment balance.
The sediment that is transported to the inlet entrance by wave-induced longshore cur-
rents is carried into the basin by the flood-tidal currents. When the inlet is in equi-
librium, this sediment is transported back in the seaward direction by the ebb-tidal
currents. The annual mean flux of sand entering the inlet on the flood is 𝑀 (m3 /s). 𝑀
is assumed a constant fraction of the annual mean longshore sediment transport and
is assumed to remain constant over time. The remaining fraction is assumed to bypass
the inlet via the ebb-tidal delta (see Sect. 9.4.2). When in equilibrium, the annual mean
flux 𝑀 equals the annual mean ebb-tidal sediment flux.
Intermezzo 9.5 The 𝐶 and 𝑞 values according to Escoffier’s equilibrium

velocity
In the approximation of a sinusoidal tidal motion, we have according to Eq. 9.5:
𝜋𝑃
𝑢̂𝑒𝑞 = (9.7)
𝐴𝑒𝑞 𝑇
Combining Eq. 9.7 with the equilibrium condition Eq. 9.6 and eliminating P leads
to the following formula for the equilibrium velocity:
1/𝑞−1 −1/𝑞 −1
𝑢̂𝑒𝑞 = 𝜋𝐴𝑒𝑞 𝐶 𝑇 (9.8)
For 𝑞 ≠ 1 the equilibrium velocity must be dependent on the cross-sectional area,

(see also Stive et al., 2009). Escoffier assumed a constant equilibrium velocity 𝑢𝑒𝑞
that depends only on the sediment diameter, and suggested an approximate value
of 0.9 m/s (Sect. 9.5.1). His assumption implies that 𝑞 = 1 and 𝐶 = 𝜋 𝑢̂𝑒𝑞 −1 𝑇 −1 =
3.49 m/s ⋅ 𝑇 −1 . Assuming a semi-diurnal tide with 𝑇 = 44 700 s, we find 𝐶 =

7.8 × 10−5 m−1 . This is the order of magnitude of 𝐶 values found empirically for
𝑞 = 1.
Assuming a power law for the sediment transport and a sinusoidal velocity, the ebb-
tidal sediment transport rate in the entrance 𝑇 𝑅 is taken proportional to the power 𝑛
of the velocity amplitude and the power 𝑚 of a length dimension 𝑙 of the cross-section:
𝑇 𝑅 = 𝑘 𝑢̂𝑒𝑛 𝑙 𝑚 (9.9)
The coefficient 𝑘 is a constant and its value is dependent on the sediment characterist-
ics. The values of 𝑛 and 𝑚 depend on the adopted sediment transport formula, where

𝑛 is in the range 3 to 6 and 𝑚 of order 1, and the length scale 𝑙 is either the annually
averaged width or depth, the final choice depending on the way sediment enters and
leaves the inlet entrance (Van de Kreeke, 1992, 2004). In the following we will adopt
the assumption that for a given offshore tide, 𝑢𝑒 is a unique function of the entrance
cross-sectional area 𝐴𝑒 .
After a change in the entrance cross-sectional area, the shape of the cross-section is
assumed to remain geometrically similar. This allows the length scale to be expressed
as a constant 𝛼 times the square root of the annual mean cross-sectional area 𝐴𝑒 :
𝑙 = 𝛼 √𝐴𝑒 (9.10)
The value of 𝛼 depends on the shape of the cross-section and the definition of 𝑙.
Substituting Eq. 9.10 for 𝑙 in Eq. 9.9 results in a formula for the annual mean ebb-tidal
sediment transport as function of the cross-sectional area:
𝑚/2
𝑇 𝑅 = 𝑘𝛼 𝑚 𝑢̂𝑒𝑛 𝐴𝑒 (9.11)
In this equation 𝑢̂𝑒 is a function of 𝐴𝑒 (see Eq. 9.5). When the cross-section is in equi-
librium:
𝑇𝑅 = 𝑀 (9.12)
The shapes of the functions, which are independent for realistic values of the paramet-
ers 𝑘, 𝛼 and 𝑚, represented by Eqs. 9.11 and 9.12, are plotted in Fig. 9.23. In general,
there will be two values of the cross-sectional area for which the annual mean ebb-tidal
transport equals the mean annual influx of sediment 𝑀 (𝐴𝑒𝑞1 and 𝐴𝑒𝑞2 ).
Figure 9.23: Equilibrium cross-sectional areas (Eqs. 9.11 and 9.12).
For a given inlet situation with a mean annual influx of sediment 𝑀, a difficulty in
determining the value of 𝐴𝑒𝑞1 and 𝐴𝑒𝑞2 from Eqs. 9.11 and 9.12 is ascertaining the
values 𝑘, 𝑛, 𝑚 and 𝛼. An elegant way to circumvent ascertaining the parameters was

put forward by Van de Kreeke (2004). For inlets at equilibrium 𝐴𝑒 = 𝐴𝑒𝑞 and 𝑢𝑒 = 𝑢𝑒𝑞
and we find from Eqs. 9.11 and 9.12:
𝑚/2
𝑀 = 𝑘𝛼 𝑚 𝑢̂𝑒𝑛 𝐴𝑒 (9.13)
Subsequently, Van de Kreeke substitutes the relationship between the velocity amp-
litude of a sinusoidal tide 𝑢̂𝑒𝑞 and the cross-sectional area 𝐴𝑒𝑞 (Eq. 9.7). This yields
after some algebra:
2
( 𝑚−2𝑛 )
𝑀𝑇 𝑛
𝐶 = ( 𝑚 𝑛) (9.14)
𝑘𝛼 𝜋
and
𝑛
𝑞= (9.15)
𝑛 − 𝑚/2
Equations 9.14 and 9.15 imply that for a set of inlets in equilibrium that have the same
values of 𝑀, 𝑇 , 𝑘, 𝛼, 𝑚 and 𝑛, values of 𝐶 and 𝑞 theoretically should be the same.
It also follows that 𝑞 > 1, at least under the present assumptions. This means that
datasets used to determine the constants 𝐶 and 𝑞 in Eq. 9.6 should ideally consist of
inlets that show phenomenological similarity (i.e., have similar values for 𝑘, 𝑛, 𝑚 and
𝛼 for a given inlet situation with mean annual influx of sediment 𝑀), implying that
the datasets should be clustered while ensuring:
• similar wave-driven littoral drift;
• similar tide characteristics (form factor, Eq. 4.1, and amplitude);
• similar grain size and grain density;
• similar shape of the cross-section.
Most of the data sets published do not fulfil these recommendations and should there-
fore be considered with care (see Stive et al. (2009) for a review of the datasets). The
resolution of this issue is important, because slightly different values of 𝑞 result in sig-
nificantly variable values for the equilibrium cross-sectional value of the tidal entrance.
This may have significant implications in determining the true stable equilibrium en-
trance cross-sectional area.
9.6. The inner basin geometry

9.6.1. Complex geometry of tidal basins
The interaction between bottom morphology and tidal motion is the cause of a complex
three-dimensional structure of residual circulations, which are both the cause and the
result of the morphology of a basin. The residual circulations in meandering channels

9.6. The inner basin geometry 445
play an active role in the morphological evolution of channel meanders and of ebb and
flood chutes, as will be described below. Sedimentation, erosion, channels and flats are
connected by these flow structures.
Section 5.7.6 described the transversal secondary flow induced by the curvature of the
tidal current in channel bends and by the Coriolis effect. This secondary flow, though
rather weak compared to the maximum tidal current, can have a significant residual
effect on the current, and also on the sediment transport and the bed topography. The
curvature-induced secondary flow does not change sign as the tide turns: in the upper
part of the water column, it is always directed away from the centre of curvature
of the flow; in the lower part always towards. This means that near the bottom the
secondary flow is towards the inner bend, generating a transport of sediment from
the outer bend, which erodes, towards the inner bend, which accretes. Hence, this
phenomenon has a positive feedback mechanism (see Sect. 1.5.2). Straight channels
are inherently unstable; a small eccentricity in the channel alignment will be inclined
to grow. Similarly, curvature-induced secondary flow components contribute to the
maintenance of shoals, see Fig. 9.13. Note that Coriolis (as well as density gradients)
may substantially alter this simple picture.
As a result of Coriolis, amongst others factors, channels will be inclined to split up
in an ebb-dominated and a flood-dominated channel. In areas where the width of the
basin is not restricted, two mainly independent channel systems can develop; the ebb
current concentrates in one set of channels, while the flood current is often strongest
in a different set of channels. Flood channels can usually be recognised because they
tend to be shallower than ebb channels and they tend to die out; they lead to progress-
ively shallower water and finally spread out on a shoal. Conversely, ebb channels are
continuous and tend to be deeper. Often the maximum ebb current occurs when the
tide level is lower than that corresponding to the maximum flood current (for a ve-
locity leading the surface elevation by less than 90°, see Sect. 5.7.3). Furthermore, in
the case of a river discharge, the total discharge during ebb is larger than during flood.
The combined effect of higher total ebb flow and the lower tidal level during this flow
tends to increase the velocity and intensify erosion in ebb channels.
Due to inertia, the tidal current is inclined to overshoot the bend pathway (or take
a wider bend; in Dutch: uit de bocht schieten), both during ebb and flood. Since the
currents try to take a wider bend, ebb and flood chutes occur at the ending of the
bends and are directed into the flat areas. Due to the fact that the mean water level
during flood (especially near the ending of the flood) is higher than during ebb, the
flood chutes are generally better developed than the ebb chutes. During the flood
tide, the water spreads over the flats, while the ebb current primarily follows the main
channel. Figure 9.24 sketches the characteristic structure of ebb and flood channels
in a meandering channel system, with an indication of the depth-mean residual flow
pattern.

sea side land side
ebb chute
flood chute
Figure 9.24: Meandering tidal channel with ebb and flood chutes (in Dutch: ebscharen and vloed-
scharen).
The flats are mostly fed with sand through the ebb and flood chutes. This feeding
process causes the existence of sills between the ebb and flood chutes. When a sill
breaches, the channel bend is cut off. A new, straighter channel is formed, while the
old main channel accretes. However, in the course of time, the new channel will start to
increase its curvature. Hence, the morphology of tidal basins is not static but dynamic;
cycles of increasing channel curvature and channel cut-off are an important part of
these internal dynamics.
9.6.2. Equilibrium relations for tidal channels and flats

The stability of the various morphological units of the inlet system can be described
with empirical relations. These relationships relate geometric properties to hydraulic
boundary conditions. In Sect. 9.5.2 we discussed the stability of the entrance of the
inlet and described the cross-sectional area of the entrance as a function of the tidal
prism. The sand volume stored in the ebb-tidal delta, which was discussed in Sect. 9.4.3,
is also related to the tidal prism. In this section two other morphological units of an
inlet and basin system are described: the tidal channels and tidal flats in the basin.
Equation 9.6 relates the equilibrium cross-sectional area in the throat, 𝐴𝑒 , to the tidal
prism 𝑃 (with 𝑞 = 1). From data of a tidal channel in the Wadden Sea it appeared that
also along a channel the flow area was related to the tidal volume passing the local
cross-section. Later on it was found that this relation is valid for various tidal channels
in the Wadden Sea and in the estuaries of the Delta in the southern Netherlands. This
relation is given by:
𝐴MSL = 𝐶𝐴 𝑃𝐴𝐵 (9.16)
in which:

9.6. The inner basin geometry 447
𝐴MSL the equilibrium flow area in a certain cross-section AB of the m2

basin, measured below mean sea level
𝑃𝐴𝐵 the tidal prism landward of the cross-section AB under m3
consideration
𝐶𝐴 empirical coefficient m−1
The tidal prism now is the tidal prism behind the cross-section under consideration
(and not the total tidal prism for the entire basin). The flow area below MSL could be
considered as the channel cross-sectional area 𝐴𝑐 . The flats are then defined as lying
above MSL (but note that in other cases, they may be defined as lying above MLW, see
also Fig. 9.25).
In the case that the flood-tidal delta spans the entire basin area (as is the case for e.g.
the Wadden Sea), an empirical relationship for the total basin channel volume is:
𝑉𝑐 = 𝐶𝑉 𝑃 /2
3
(9.17)
in which:
𝑉𝑐 the equilibrium total channel volume below mean sea level m3

𝐶𝑉 empirical coefficient m−3/2
For example, the empirical coefficient 𝐶𝑉 is 65 × 10−6 m−3/2 for the Wadden Sea and
73 × 10−6 m−3/2 to 80 × 10−6 m−3/2 for the Eastern Scheldt and the Grevelingen (Eysink,
1991).
The power 3⁄2 in Eq. 9.17 can be understood as follows. We have found that the cross-
sectional area of the channels is proportional to the tidal prism (Eq. 9.16). Furthermore,
the length of the channels is proportional to the square root of the basin surface area
(channels and flats), which in turn is proportional to the square root of the tidal prism.
We then have:
𝐴𝑐 ∝ 𝑃
⇒ 𝑉𝑐 = 𝐴𝑐 𝐿𝑐 ∝ 𝑃 √𝐴𝑏 ∝ 𝑃 3/2 (9.18)
𝐿𝑐 ∝ √𝐴𝑏 ∝ √𝑃
in which:
𝐴𝑐 the channel cross-section m2

𝐿𝑐 length of channels m
𝑉𝑐 volume of channels m3
𝐴𝑏 the gross basin area (flats and channels) m2

From the proportionalities of Eq. 9.17, we can also deduce a relation between the tidal
flat area and the total basin area. The flats area is equal to the total basin area minus
the channels area (Fig. 9.25):
𝑉𝑐 𝑃 √𝐴𝑏 𝐻 3/2
𝐴𝑓 = 𝐴𝑏 − 𝐴𝑐ℎ = 𝐴𝑏 − ≈ 𝐴𝑏 − 𝛼 ≈ 𝐴 𝑏 − 𝛽 𝑚 𝐴𝑏 (9.19)
𝐷𝑐 𝐷𝑐 𝐷𝑐
in which
𝐴𝑓 flats area, i.e. the area above MSL m2

𝐴𝑐ℎ the horizontal area below MSL covered by all channels m2
𝛼, 𝛽 constants of proportionality m−1
𝐷𝑐 typical channel depth m
𝐻𝑚 mean tidal range m
MHW
MSL Hm
MLW
Ab
Ach,MSL Af,MSL
Ach,MLW Af,MLW
A(z)
AMLW AMSL AMHW
Figure 9.25: Definition of the basin area, channel area and flats area relative to the tidal levels.
Renger and Partenscky (1974) found for the German Bight, with the flats defined as
lying above LW:
3/2
𝐴𝑓 = 𝐴𝑏 − 0.025𝐴𝑏 (9.20)
Figure 9.26 shows similar relationships for the estuaries in the south of the Netherlands
and for the Wadden Sea basins. An explanation for the trend of the relation could be
the increasing activity of local wind waves in larger basins (since waves act as an
eroding agent on the flats).

9.7. Net sediment import or export 449
1
Wadden Sea
SW delta area
0.8 E
P Z1
A
0.6
Af /A b
ES
0.4 V
WS
T2
0.2
0
50 100 500 1000 5000
6 2
Ab [10 m ]
Wadden Sea southwestern (SW) delta area 1
before closure of the Lauwerszee
T Texel Inlet WS Western Scheldt
2
after closure of the Zuiderzee
E Eierland Inlet ES Eastern Scheldt (IJsselmeer); no equilibrium yet
V Vlie Inlet
A Ameland Inlet
Z Zoutkamperlaag Inlet
P Pinke Inlet
Figure 9.26: Relative area of the intertidal zones in the Dutch Wadden Sea (upper two lines)
and the delta area in the southern Netherlands (lower two lines). Some of the inlet systems are
indicated by letters.
9.7. Net sediment import or export

9.7.1. Introduction
Within the basin, the tide is strongly deformed by bottom friction and other non-linear
effects associated with the basin geometry (see Sects. 5.7.3 and 5.7.4). This tidal dis-
tortion can either strengthen or weaken the magnitude of the maximum flood flow
compared to the maximum ebb flow (flood versus ebb dominance) and causes an asym-
metry between durations of the slack water periods as well. These asymmetries greatly
impact the net movement of sediment, both sand and silt, and thus the morphological
development of the basin over time. As mentioned before, due to this feedback mech-
anism the basin in essence determines its own evolution (if we ignore external condi-
tions such as sediment availability, storm surges and the tidal conditions at sea).
So-called hypsometric5 curves give information about the depth of the channels and
the extent of the flats (see Fig. 9.27). In Sect. 5.7.4, these parameters were found to
control ebb and flood dominance in the following way:
• A large tidal amplitude and shallow channels enhance flood dominance;
5
The term hypsometry refers to the topographical characteristics of tidal basins.

• A large intertidal storage volume (as compared to channel volume) enhances

ebb dominance.
The import or export of sand is strongly determined by the velocity signal being flood-
dominant or ebb-dominant, as we will see in Sect. 9.7.2.
z z
MHW MHW
MSL MSL
MLW MLW
channels flats channels flats
A(z) A(z)
(a) (b)
Figure 9.27: Hypsometric curves for two different basin types; the wetted basin surface area
𝑂(𝑧) as a function of the water level. The left figure is representative of a situation with shallow
channels and large intertidal storage areas. The right situation is the opposite: deep channels
and small intertidal areas. Here the flats are defined as lying above MLW.
9.7.2. Tide-induced residual transport of (medium to) coarse sedi-

ment
In this paragraph, we explore the sediment flux averaged over the tide and analyse the
conditions under which a net import or export of (medium to) coarse sediment in the
basin will occur (for fine sediment an additional effect occurs which is discussed in
Sect. 9.7.3). A larger influx than outflux of sediment leads to a net landward directed
sediment flux; as a result, the basin will import sediment. The opposite applies for
basins exporting sediment.
In Ch. 5 we discussed in detail the mechanisms that contribute to a residual sediment
transport, either landwards or seawards:
• The asymmetry of the horizontal tide (Sects. 5.7.4 and 5.7.5);
• Tide-averaged residual currents (Sect. 5.7.6) along the channels.
The residual flow along the channels may be composed of for instance contributions
due to river flow and as compensation for Stokes drift. The secondary flow patterns
as discussed in Sect. 5.7.6 have no direct impact on the import or export of sediment
from the basin.

In the analysis below, we approximate the tide-averaged residual flow by a constant

and depth-averaged residual tidal velocity 𝑢0 in the channel direction. We further
assume that M2 (semi-diurnal, 12.42 h period) is the dominant tidal-current constituent
and that all other constituents are of a lesser order of magnitude. Furthermore, the
residual flow velocity 𝑢0 is assumed to be small compared to the amplitude of the M2
tidal current. These conditions roughly apply in most tidal basins in the Netherlands.
Coarse sediment was defined in Ch. 6 as having a diameter such that 𝑤𝑠 /𝑢∗ > 1, where
𝑤𝑠 is the fall velocity and 𝑢∗ is the shear velocity. Coarse sediment is thought to respond
instantaneously to the flow velocity (or alternatively the bed shear stress). In terms
of the flow velocity we can write for the bed load transport (defined as volumetric
transport excluding pores in m3 /s/m) in the case of small bottom slopes:
𝑆 ≈ 𝑐 |𝑢 𝑛−1 | 𝑢 (9.21)
The coefficient 𝑛 is thought to lie in the range 3 to 5. Here, we propose 𝑛 = 3, consistent

with the Bagnold-type formulation (see Eq. 6.50) for bed load in the case of small bed
slopes. Furthermore 𝑐 = 10−7 m2−n sn−1 to 10−4 m2−n sn−1 . In this transport formula,
initiation of motion is not taken into account. This would further enhance the effect
of asymmetry.
For suspended load transport of medium to coarse, non-cohesive bottom material
(sand), a similar relationship is often used, but with a higher velocity power (𝑛 = 4, in
agreement with Bagnold’s suspended load formulation).
For suspended load transport of fine (cohesive) sediment, other formulations need to
be adopted, which take account of the time lags related to settling and resuspension.
This is treated in Sect. 9.7.3.
The velocity signal u(t) considered by Van de Kreeke and Robaczewska (1993) can be
written as:
𝑢(𝑡) = 𝑢0 + 𝑢̂M2 cos (𝜔M2 𝑡) + ∑ 𝑢̂𝑖 cos (𝜔𝑖 𝑡 − 𝜑𝑖 ) (9.22)

𝑖
in which:
𝑢0 the Eulerian residual flow 1

𝑢̂M2 the amplitude of the M2 tidal current m/s
𝑢̂𝑖 the amplitude of the other tidal-current constituents m/s
𝜔M2 the angular frequency of the M2 constituent rad/s
𝜔𝑖 the angular frequency of the other tidal-current constituents rad/s
𝜑𝑖 the phase lag between M2 and the other tidal constituents rad

By substituting this velocity signal (Eq. 9.22) in Eq. 9.21, Van de Kreeke and Robaczewska
(1993) demonstrated, under the assumption of M2 dominance and 𝑛 = 3 (and thus
𝑆 ∝ 𝑢 3 ), what the most important contributions to net tide-induced bed load transport
of coarse sediment are. Note that their approach is quite similar to the decomposition
of the wave-induced cross-shore transport, as discussed in Sect. 7.5.
The resulting formula for bed load transport averaged over the long term, valid under
the above-mentioned restrictions, is:
𝑆 3 𝑢0 3 𝑢̂ 3 𝑢̂ 𝑢̂
3
= + M4 cos 𝜑M4-2 + M4 M6 cos (𝜑M4-2 − 𝜑M6-2 ) (9.23)
𝑐 𝑢̂M2 2⏟𝑢̂M2 ⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟
4 𝑢̂M2 2 𝑢̂M2 𝑢̂M2
⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟
1 2 3
in which:
𝑢0 the Eulerian residual flow 1

𝜑M4-2 the phase lag 𝜑M4 − 2𝜑M2 between M2 and M4 (cf. Eq. 5.111) rad
𝜑M6-2 the phase lag 𝜑M6 − 3𝜑M2 between M2 and M6 rad
𝑐 coefficient defined through Eq. 9.21
Apparently, the long-term-averaged bed load transport is predominantly determined

by:
• the residual flow velocity 𝑢0 ;
• the amplitude of the M2 tidal current;
• the amplitudes and phases (relative to the M2 tidal current) of the M4 (quarter-
diurnal, 6.21 h period) and M6 tidal-current constituents.
Although higher odd and even overtides also contribute, the first even overtide M4 and
first odd overtide M6 are the most important contributing overtides. The components
K1, S2, N2 and MS4 were also included in the analysis, but were found to only cause
fluctuations of the transport rates that would average out in the longer term. For
example, the effect of inclusion of S2 is only to give a beating of the transport flux
with a period of 14.77 days (cf. Fig. 3.25). The inclusion of the diurnal component
merely gives a daily fluctuation, but does not influence the longer-term net transport
(cf. Fig. 3.26).
The three numbered terms on the right-hand side of Eq. 9.23 represent the net transport
as a result of:
1. The asymmetry introduced by the addition of a small residual flow to the si-
nusoidal M2 tidal current component (cf. Fig. 5.72 for a mean river discharge,
although in that case 𝑢0 is not small);

2. The asymmetric velocity signal of the M2+M4 tidal current combined. The tide-
averaged velocity of the M2 + M4 tidal current is zero. However, due to the non-
linear response of the sediment transport to the velocity, larger (positive and
negative) velocities get relatively more weight in contributing to the transport.
The result is a net transport in the ebb or flood direction respectively, depending
on the phase angle 𝜑M4-2 (cf. Fig. 5.71); For 𝜑M4-2 = 𝜋/2 or 3𝜋/2 the ebb and
flood velocity are of the same size and the signal has a sawtooth shape. The net
transport (averaged over the tidal cycle) is zero. For other values of 𝜑M4-2 the
maximum ebb velocities differ from the maximum flood velocities (either larger
or smaller in the case of ebb or flood dominance respectively), causing a net
sediment transport. The net transport is largest for the maximum skewness of
the velocity signal (𝜑M4-2 = 0 or 𝜋);
3. An interaction term among M2, M4 and M6 (smaller than the first two contribu-
tions). Its importance is governed by the phase angles 𝜑M4-2 and 𝜑M6-2 .
The first two terms in Eq. 9.23 are the most important. Their origin can be further
clarified by considering the effect of the addition to the M2 sinusoidal current signal
of 𝑢0 and the M4 tidal current separately. We will also explore the effect of inclusion
of M6.
Let us first consider 𝑢(𝑡) = 𝑢0 + 𝑢̂M2 cos (𝜔M2 𝑡). If we substitute this in Eq. 9.21 and
use 𝑛 = 3, we find for the time-dependent transport:
3
𝑆(𝑡) ≈ 𝑐𝑢(𝑡)3 = 𝑐 (𝑢0 + 𝑢̂M2 cos (𝜔M2 𝑡)) ⇒
3
𝑆 𝑢
3
=( 0 ) +
𝑢̂M2
𝑐 𝑢̂M2 ⏟⏟⏟⏟⏟⏟⏟⏟⏟
1
2
𝑢 (9.24)
3 ( 0 ) cos (𝜔M2 𝑡) + cos3 (𝜔M2 𝑡) +
𝑢̂M2
⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟
2
𝑢
3 ( 0 ) cos2 (𝜔M2 𝑡)
𝑢̂M2
⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟⏟
3
The term denoted ‘1’ can be neglected relative to the other terms, since we had assumed
that 𝑢0 /𝑢̂M2 is a small quantity (M2 being dominant). The remaining formula could also
be obtained by using a Taylor expansion, as in Intermezzo 7.3, now with the residual
flow velocity being the perturbation. The terms denoted ‘2’ are symmetrical about
the horizontal axis and will not give a contribution when averaged over the M2 tidal

period. The only term of interest for the tide-averaged sediment transport is term ‘3’.
Integration over the tidal period results in:
⟨𝑆⟩ 3 𝑢0
3
= ( ) (9.25)
𝑐 𝑢̂M2 2 𝑢̂M2
which is identical to the first term in Eq. 9.23 and represents the effect of the interaction
of a (small) residual flow and the M2 tidal current.
The next step is to look at the interaction of M2 and M4. The velocity can now be
written as 𝑢(𝑡) = 𝑢̂M2 cos (𝜔M2 𝑡) + 𝑢̂M4 cos (𝜔M4 𝑡 − 𝜑M4-2 ) with 𝜔M4 = 2𝜔M2 . The
phase lag 𝜑M4-2 between M2 and M4 is 𝜑M4 − 2𝜑M2 . The effects of the phase angle on
the velocity signal and 𝑢 3 are illustrated in Figs. 9.28 and 9.29.
uM2 uM2 uM2+M4
φM4-2 = 0 φM4-2 = π/2

u [m/s]
φM4-2 = π φM4-2 = 3π/2

u [m/s]
ωM2t [rad] ωM2t [rad]
Figure 9.28: M2 and M4 tidal-current constituents. Flood dominance is found for −𝜋/2 < 𝜑M4-2 <
𝜋/2 and ebb dominance for 𝜋/2 < 𝜑M4-2 < 3𝜋/2. Flow reversal from flood to ebb (HWS) is of
shorter duration than flow reversal from ebb to flood (LWS) for 0 < 𝜑M4-2 < 𝜋 and of longer
duration for 𝜋 < 𝜑M4-2 < 2𝜋.
We now get:
3
𝑆(𝑡) ≈ 𝑐 (𝑢̂M2 cos (𝜔M2 𝑡) + 𝑢̂M4 cos (𝜔M4 𝑡 − 𝜑M4-2 )) (9.26)

uM2+M4 uM2+M43
u [m/s] or u3 [m3/s3] φM4-2 = 0 φM4-2 = π/2

〈S 〉 > 0 〈S 〉 = 0
φM4-2 = π φM4-2 = 3π/2

u [m/s] or u3 [m3/s3]
〈S 〉 < 0 〈S 〉 = 0
Figure 9.29: 𝑢 and ⟨𝑢 3 ⟩ for 𝑢 consisting of M2 and M4 tidal-current constituents with four different
phase angles between M2 and M4. Since ⟨𝑆⟩ ∝ ⟨𝑢 3 ⟩ we find ⟨𝑆⟩ > 0 for −𝜋/2 < 𝜑M4-2 < 𝜋/2 and
⟨𝑆⟩ < 0 for 𝜋/2 < 𝜑M4-2 < 3𝜋/2. ⟨𝑆⟩ = 0 for 𝜑M4-2 = 𝜋/2 and for 𝜑M4-2 = 3/2𝜋.
The tide-averaged transport can be obtained by integration of Eq. 9.26. Analogous

to the derivation of the M2-residual flow interaction, we neglect the third power of
𝑢̂M4 /𝑢̂M2 . This leads to:
⟨𝑆⟩ 3 𝑢̂M4
3
= ( ) cos 𝜑M4-2 (9.27)
𝑐 𝑢̂M2 4 𝑢̂M2
In this formula we can recognise term ‘2’ from Eq. 9.23. It represents the effect of
the interaction of the M2 tidal current and its M4 overtide. As mentioned already,
for cos 𝜑M4-2 = ±1, this interaction term is at its maximum. This corresponds to the
situation that the ebb and flood velocities differ the most in magnitude. In the case of
cos 𝜑M4-2 = 1 (𝜑M4-2 = 0) the velocity signal is flood-dominant (see Fig. 9.28, top left)
without any sawtooth asymmetry, leading to a maximum net import of (coarse) sedi-
ment (Fig. 9.29, top left). For cos 𝜑M4-2 = −1 (𝜑M4-2 = 𝜋), the velocity is ebb-dominant
and the system exports sediment (see Figs. 9.28 and 9.29, bottom left). There is no
contribution to the net sediment transport for cos 𝜑M4-2 = 0 (𝜑M4-2 = 𝜋/2 or 3𝜋/2),
see the right panels of Figs. 9.28 and 9.29. This corresponds to a velocity signal that
demonstrates sawtooth asymmetry, but has equal flood and ebb current magnitudes
and durations.

Why is a comparable term containing the odd overtide M6 not visible in Eq. 9.23?
Apparently the interaction between the M2 and M6 does not lead to a net sediment
transport, regardless of the phase angle. The combination of M2 and M6 leads to saw-
tooth asymmetry only (see Fig. 9.30), and therefore does not give a residual sediment
transport (Fig. 9.31).
uM2 uM6 uM2+M6
φM6-2 = 0 φM6-2 = π/2

u [m/s]
φM6-2 = π φM6a-2 = 3π/2

u [m/s]
Figure 9.30: M2 and M6 tidal-current constituents for four different phase angles between M2
and M6.
Combining the above with our knowledge about how the basin influences the tidal
asymmetry, we may conclude:
Flood-dominant systems (with shallow channels and limited intertidal storage)

enhance landward near-bed transport and tend to fill in their channels with coarse
material, whereas ebb-dominant systems (with deep channels and large intertidal
storage) enhance seaward near-bed transport and flush coarse sediment seaward.
9.7.3. Fine sediment transport and siltation

For coarse sediment, the bed load as well as the suspended load transport are determ-
ined largely by the hydrodynamic conditions and the sediment properties at the point
of consideration. This can be modelled with a sediment transport formula (e.g. the
Bijker or the Bailard formula, much like in the previous section), or with an intra-
wave model which describes the suspension process during a wave cycle (Ch. 6). In
the case of fine sediment, the suspended sediment transport not only depends on the

uM2+M6 uM2+M63
u [m/s] or u3 [m3/s3] φM6-2 = 0 φM6-2 = π/2

〈S 〉 = 0 〈S 〉 = 0
φM6-2 = π φM6-2 = 3π/2

u [m/s] or u3 [m3/s3]
〈S 〉 = 0 〈S 〉 = 0
Figure 9.31: 𝑢 and ⟨𝑢 3 ⟩ for 𝑢 consisting of M2 and M6 tidal-current constituents with four different
phase angles between M2 and M6. ⟨𝑆⟩ ∝ ⟨𝑢 3 ⟩ = 0 for all phase angles.
local instantaneous flow velocity, but also on the flow conditions upstream and in
the past. We therefore need to include a ‘memory effect’ that makes the sediment
concentration at a certain point respond to the conditions in all points the suspended
sediment has come through in the past (Sect. 6.8.1). The timescales of erosion and
sedimentation become important, a characteristic that causes an essential difference
in the tidal dynamics of fine sediment (silt) compared to coarser sediment (sand). For
sand this timescale is an order of magnitude smaller than the tidal period (so that we
can assume an instantaneous response), whereas for fine sediment the timescales are
of similar order. For fine sediment we denote this timescale as 𝑇𝑠𝑒𝑑 . If we assume that,
under the condition of a constant flow velocity, after some time an equilibrium concen-
tration 𝑐𝑒𝑞 is reached, then 𝑇𝑠𝑒𝑑 is the timescale for which the suspended concentration
𝑐 approaches the equilibrium concentration. In formula form:
𝐷𝑐 1
= (𝑐 − 𝑐) (9.28)
𝐷𝑡 𝑇𝑠𝑒𝑑 𝑒𝑞
Note that this is identical to Eq. 6.53. In principle a distinction must be made between
the relaxation timescales 𝑇𝑠𝑒𝑑 for erosion and for sedimentation, respectively 𝑇𝐸𝑟 and
𝑇𝑆𝑒 . The equilibrium concentration 𝑐𝑒𝑞 can be determined experimentally; it turns out
that often
𝑛−1
𝑐𝑒𝑞 ≈ 𝛽 |𝑢| (9.29)

in which 𝑛 = 3 to 5 (depending on the type of sediment) and 𝛽 a constant. Please note

the resemblance between Eq. 9.21 and Eq. 9.29. A (theoretical) example of the solution
to this equation is shown in Fig. 9.32. The velocity signal 𝑢(𝑡) = cos(𝜔𝑡) + 1/4 cos(2𝜔𝑡 −
3𝜋/2) has been chosen such that the flood velocities are equal in magnitude to the
ebb velocities. Hence, no residual bed load transport of coarse material will occur.
For the equilibrium concentration, Eq. 9.29 has been used with 𝑛 = 5, resulting in
𝑐𝑒𝑞 ∝ 𝑢 4 . In the case of an instantaneous sediment response (coarser material), the
concentration would be equal to the equilibrium concentration and the net suspended
sediment transport would be zero. In the case of finer sediment, the sediment needs
time to respond and will follow the dash-dotted line. This is computed from Eq. 9.28
using 𝑇𝑠𝑒𝑑 /𝑇M2 = 0.2. It can be seen that now the concentration is generally higher
during the flood period than during the ebb period. As a result, the sediment flux 𝑢𝑐
during flood is 31 % larger than during ebb, leading to a net landward-directed flux of
fine sediment.
The asymmetry in the concentration stems from the asymmetry in the slack water
periods. In the schematised situation of Fig. 9.32, the duration of flow reversal from
flood to ebb (HWS) is longer than the duration of flow reversal from ebb to flood (LWS).
Hence, a strong sedimentation occurs around HWS; in the initial ebb phase only a
small amount of material is in suspension. Later in the ebb period, more fine sediment
is suspended. Around LWS part of this sediment settles, but not so much because of the
short slack water duration. When the flood current increases in strength, a relatively
large amount of suspended material is immediately transported. Subsequently, the
suspended concentration increases even further. It is clear that, averaged over the
flood, the concentration of suspended sediment is therefore larger than that during ebb.
Over the whole tidal period, we therefore observe a net landward-directed transport.
In practice, both types of tidal asymmetry can be present at the same time (Fig. 9.33).
A long HWS duration and a net landward-directed transport of fines can be observed
in Fig. 9.33. The slack durations before ebb (𝑇1,slack ) and flood (𝑇2,slack ) are defined by
the flow velocity (𝑢crit,flood , 𝑢crit,ebb ); for smaller velocity magnitudes, sedimentation
takes place. Since 𝑇1,slack > 𝑇2,slack , an import of fines is likely to happen. For coarser
sediment (sand) slack water duration does not play a role of importance because of
the short sedimentation timescale. Generally speaking, we may state that we have
net flood transport for 𝑢max,flood > 𝑢max,ebb , and net ebb transport in the reverse case.
In the present example, the maximum ebb velocity is larger than the maximum flood
velocity. Hence, an export of coarse sediment can be expected.
In Sect. 5.7.4 we discussed ‘sawtooth asymmetry of the horizontal tide’. We concluded
that for a short basin, with shallow channels and little storage-offering flats, the HWS
duration is longer than the LWS duration. On the other hand, in a basin with vast
storage-offering tidal flats and deep channels, the HWS duration is shorter than the
LWS duration. This suggests that import of fines is a characteristic property of a tidal
basin with shallow channels and limited tidal storage flats. Similarly, one may expect

A1 = 1.31 A2
A1
u A2
ceq
c
uc
ωt [rad]
Figure 9.32: Lag effects on the residual sediment transport (𝑢̂M4 /𝑢̂M2 = 0.25 and 𝜑M4-2 = 3𝜋/2).
Positive values for 𝑢 (above the dashed line) indicate flood velocities. Since the duration of HWS
(flow reversal from flood to ebb) is longer than the duration of LWS (flow reversal from ebb to
flood), the sediment concentration is higher during the flood period than during the ebb period
and a residual transport occurs in the flood direction.
umax, flood
T1, slack T2, slack

u [m/s]
ucrit, flood
ucrit, ebb
umax, ebb
ωt [rad]
Figure 9.33: Asymmetry of the horizontal tide. The signal combines a small ebb current with
M2 and M4 tidal-current components with 𝜑M4-2 = 3𝜋/2.
that basins with deep channels and large intertidal storage are likely to export fines.
However, the latter is often not true, due to a counteracting effect that is especially
important when large intertidal flats are present.
This counteracting effect is related to differences of sedimentation at HW and at LW.
The relaxation timescale for sedimentation 𝑇𝑆𝑒 is, amongst other factors, dependent
on the water depth, viz. 𝑇𝑆𝑒 ≈ ℎ/𝑤𝑠 . Hence, the amount of sedimentation at HW can
strongly differ from that at LW. This is particularly the case when the wet surface of
the basin is much larger at HW than at LW (large intertidal storage), implying an on

average smaller water depth at HW than at LW. The result is a larger sedimentation at
HW than at LW, reducing the concentration and sediment flux during the ebb period.
Apparently, a large storage-offering flat area has two opposing effects on the net trans-
port of fine sediment: on the one hand a large storage prism causes a short slack dur-
ation in the flow channel at HWS; on the other hand, in this short period a strong
settling can occur due to the small water depth. If the latter effect dominates this
leads to a net import of fines, even in the case of large storage areas. This may be one
of the reasons that many tidal basins are deposition areas for fine sediments (silt).
Furthermore, wave stirring and storm surges can cause a strong erosion and export of
fine sediment. Waves can be expected to stir up sediment very effectively, and more
so for smaller water depths. This increases fine sediment concentrations in the initial
ebb phase, enhancing an ebb-transport.
9.7.4. Overview of the relation between morphology and sediment

transport
From the above, we may derive qualitative relations between net sediment transport
and the morphological characteristics of tidal basins. Table 9.4 and Table 9.5 give an
overview of, respectively, the conditions for net flood and ebb transport for sediment
in general and for fine sediment only. In the tables, the role of external conditions
(tides, waves) is also indicated.
Table 9.4: Conditions for net import or export of sediment
Net flood transport Net ebb transport Tidal asymmetry

Small storage flat area Large storage flat area 𝑢max flood/ebb
Long shallow channels Long deep channels 𝑢max flood/ebb
Fast tidal rise at sea Fast tidal fall at sea 𝑢max flood/ebb
Table 9.5: Conditions for net transport of fine sediment (silt)
Net flood transport Net ebb transport tidal asymmetry

Small storage flat area Large storage flat area
𝑢max flood/ebb
slack duration HW/LW
Large storage flat area Small storage flat area sedimentation HW/LW
Shallow channels Deep channels 𝑢max flood/ebb
slack duration HW/LW
Long HW period at sea Short HW period at sea slack duration HW/LW
Protected location, few waves Open, many waves sedimentation HW/LW

9.7.5. Large-scale morphodynamics

Basins are said to be flood-dominant or ebb-dominant, which refers to net sand import
or sand export. In the case of neither flood nor ebb dominance, there is no net sediment
gain or loss and the system is said to be in equilibrium (on the scale of the basin).
Morphological equilibrium is dependent on external conditions, but if we assume these
to be neutral, different basin geometries may fulfil the condition for equilibrium:
• Shallow channels and large intertidal (storage) flat area (e.g. Wadden Sea);
• Deep channels and small intertidal (storage) flat area (e.g. Eastern Scheldt);
• Intermediate geometries.
In the equilibrium condition, the geometrical characteristics of the basin are such that
the durations of ebb and flood are approximately equal. (Dronkers, 1998) derived a
formula for this equilibrium condition, based on the assumption that the tidal asym-
metry at sea is not strong. This condition is shown in Fig. 9.34. It indicates that for
Dutch tidal basins the ratio between the average channel depths and flat surface area
is such that net ebb and flood transports are approximately equal, but have a slight
flood-dominant tendency.
The tidal basins in the Wadden Sea (especially in the eastern Wadden Sea) are generally
considered to be in morphological equilibrium. The flood dominance found in Fig. 9.34
could be the result of the neglect of residual currents in the derivation. As a result of
Stokes drift, a compensating Eulerian current will enhance the ebb flow (see Sect. 5.7.6).
It could also be that the flood dominance is the result of a sediment demand of the basin
as a result of sea level rise (Sect. 9.8.3).
ED = Ems Dollard
L = Lauwers Inlet
F = Frisian Inlet
P = Pinke Inlet
P
3 A = Ameland Inlet
V = Vlie Inlet
E = Eierland Inlet
y = h HW /h LW
T = Texel Inlet
WS = Western Scheldt
ES = Eastern Scheldt
L
2 A E
F
V
ED
T WS
ES
1
1 2 3 4
x = Ab, HW /Ab, LW
Figure 9.34: Relation between basin-averaged channel depth and flat area for Dutch tidal basins
(after Dronkers, 2005). The ratio of water depths at HW and LW (𝑦-axis) becomes larger for shal-
lower channels. The wet surface area ratio on the 𝑥-axis increases for larger intertidal storage
areas.

What happens if we disturb this equilibrium slightly? An increase of the channel

depth will enhance the ebb dominance and cause a net sediment export. The result
is a further deepening of the channels; this is a case of an unstable morphological
equilibrium (Sect. 1.5.2). An increase of the channel width is equivalent to a relative
decrease of the intertidal areas. In this case a net sediment import occurs, annihilating
the initial disturbance (stable equilibrium, see Sect. 1.5.2).
Some tidal basins are unstable and silt up (with sand and/or finer fractions) in a re-
latively short period, such as the Zwin that, in the Middle Ages, gave navigational
access to the town of Bruges. In other basins there is a dynamic equilibrium between
import and export of sediment. This is for instance the case for the eastern Wadden
Sea, which has had the same topography for nearly a thousand years. The locations
of flats and channels fluctuate over timescales of tens to hundreds of years.
The fact that tidal basins exist implies that they display a large degree of morphological
stability. This stability requires that channel width, channel depth, flat width and flat
depth are dynamically coupled. Especially the transversal transports (between flats
and channels) in tidal basins are responsible for this dynamic coupling. This internal
dynamic coupling can restore the morphological stability of the basin, for instance
after the channel depth is disturbed by deepening or filling.
9.8. Changes in dynamic equilibrium

The various empirical relationships as described in this chapter can help understand
the large-scale morphological response to changes in the dynamic equilibrium of tidal
basins. A few examples will be given.
9.8.1. Closure of a part of a tidal basin

The closure of a part of a tidal basin will result in a reduction of the channel volume 𝑉𝑐
and the tidal prism 𝑃. With the help of Fig. 9.35 and Table 9.6 we will demonstrate how
the basin will adapt to two different closures. In this example we will neglect the role
of the flats and assume that they are already more or less in equilibrium immediately
after the closure.
The dashed lines in Fig. 9.35 represent the power relationship between the channel
volume 𝑉𝑐 and the tidal prism (Eq. 9.17, with 𝐶𝑉 is 65 × 10−6 m−3/2 ). The solid lines
represent the power relationship between the sand volume of the outer delta 𝑉𝑜𝑑 and
the tidal prism (Eq. 9.3, with 𝐶𝑜𝑑 is 65.7 × 10−4 m−0.69 ). Since both the horizontal and
vertical axes are logarithmic, the power relationships plot as straight lines. The slope
of the lines is determined by the power in the power relationships, which means a
larger slope for 𝑉𝑐 than for 𝑉𝑜𝑑 . The intercept with the vertical axis is given by the
coefficient in the power relationships and is smaller for 𝑉𝑐 than for 𝑉𝑜𝑑 .

9.8. Changes in dynamic equilibrium 463
∆P ∆P
∆Vc ∆Vc
a1 a2
[10 6 m 3 ]
b1 b2
Vc Vc
V
Vod Vod
P [10 6 m 3 ] P [10 6 m 3 ]
Figure 9.35: Effect of closure of part of the basin on the channel volume 𝑉c and the volume of
the outer delta 𝑉od (crosses: initial situation, solid dots: immediately after closure, open dots:
new equilibrium). Closure 1 (left) has a smaller Δ𝑉𝑐 than closure 2 (right). The reduction of the
tidal prism Δ𝑃 is the same for both closures. Corresponding numbers are found in Table 9.6.
Now assume an initial equilibrium situation with tidal prism 𝑃 and corresponding
channel volume 𝑉𝑐 and sand volume of the outer delta 𝑉𝑜𝑑 (crosses in Fig. 9.35). At
a certain moment in time, a part of the tidal basin is closed off, which immediately
results in a reduction of the channel volume Δ𝑉𝑐 and a reduction of the tidal prism Δ𝑃
(solid dots in Fig. 9.35). Closure 1 (Fig. 9.35, left) and closure 2 (Fig. 9.35, right) differ
in the magnitude of Δ𝑉𝑐 . The reduction of the tidal prism Δ𝑃 is the same for both
closures. A new equilibrium for 𝑉𝑐 and 𝑉𝑜𝑑 will arise at the equilibrium lines for tidal
prism 𝑃 − Δ𝑃 (open dots in Fig. 9.35, the same for both closures). Note that the change
in the prism has a larger absolute effect on the equilibrium channel volume than on
the equilibrium volume of the outer delta (𝑉𝑐, before − 𝑉𝑐, after > 𝑉𝑜𝑑, before − 𝑉𝑜𝑑, after ).
Due to the closures the channel volume is 𝑎 m3 too big and the sand volume of the
outer delta is 𝑏 m3 too big. The sand of the outer delta is available for the adaptation
of the channels (to an amount of 𝑏 m3 ). The rest, if 𝑎–𝑏 > 0, has to be supplied from
outside (resulting in erosion of the downdrift coast). The magnitude of 𝑏 is given by
𝑉𝑜𝑑, before − 𝑉𝑜𝑑, after . The magnitude of 𝑎 is determined as 𝑉𝑐, before − 𝑉𝑐, after − Δ𝑉𝑐 and
is therefore different for closure 1 and closure 2. As a consequence 𝑎–𝑏 > 0 for closure
1 and 𝑎–𝑏 < 0 for closure 2 (see Table 9.6).
An example of the above situation is the closure of the Lauwerszee in 1969 (see Fig. 9.2)
as described in Wang et al. (2009), reprinted in App. E. Due to the decrease of the

Table 9.6: Values corresponding to closure 1 (Fig. 9.35, left), closure 2 (Fig. 9.35, right) and
accretion (Fig. 9.36). All variables in 106 m3 .
closure 1 closure 2 accretion

Prism before 600 600 300
Prism after 300 300 225
Δ𝑉𝑐 300 470 0
𝑉𝑐, before 955 955 338
𝑉𝑐, after 338 338 219
𝑎 318 148 118
𝑉𝑜𝑑, before 412 412 176
𝑉𝑜𝑑, after 176 176 123
𝑏 236 236 52
𝑎−𝑏 82 −88 66
tidal basin area, the tidal prism and, as a result, the magnitude of the flow velocity
decreased significantly. The tidal asymmetry changed such that it became more flood-
dominant, favouring sediment input. The effect of the closure therefore was a sediment
deficit that needed to be supplied from outside. Since the closure, the basin has been
accumulating sediment and the ebb-tidal delta has been eroding. The sedimentation
in the basin and the erosion of the ebb-tidal delta are more or less in balance. As a
consequence the closure has not caused erosion of the adjacent coasts.
9.8.2. Accretion of new land

Natural accretion along the borders of the basin, due to the deposition of silt, is a very
slow process. This process causes a gradual reduction of the tidal prism, see Fig. 9.36.
Due to this reduction the channel volume is 𝑎 m3 too large and the sand volume of the
ebb-tidal delta is 𝑏 m3 too big. This results in a sand demand from outside. Also in
this case, not all the sand can be supplied from the outer delta; it will be supplied at
the expense of the downdrift coast.
9.8.3. Relative sea level rise

The morphological response to sea level rise is more difficult to assess. From hypso-
metric curves of the different basins it follows that, without adaptation of the level of
tidal flats, the tidal prism will increase and the area of intertidal zones will decrease.
The latter effect will be most serious in the Wadden Sea, especially in the western part,
where the tidal flats are low. If this scenario, without adaptation of the tidal flats, is
realistic, the channels in the basin will widen and sand will be transported partly to
the outer delta, which will extend, and partly will become available for accretion of
the North Sea coast adjacent to the tidal inlet. However, an increase in sea level will

9.8. Changes in dynamic equilibrium 465
5000
Vc
Vod
1000
a1
[10 6 m 3 ] b1
100
V
10
2
10 100 1000 5000
6 3
P [10 m ]
Figure 9.36: Accretion of new land (crosses: initial situation, open dots: new equilibrium). From
Eqs. 9.3 and 9.17 it can be seen that a change in the prism has a larger effect on the equilibrium
channel volume than on the equilibrium volume of the outer delta, such that 𝑎 − 𝑏 > 0.
also affect the level of the tidal flats. In a relative sense, the disturbance of the charac-
teristic water depth on the flats will be much greater than in the channels. Hence, the
sediment transport in the channels will be far less affected than that on the flats. Con-
sequently, it seems realistic to assume that the response of nature will be strongest
on the tidal flats. If it is assumed that the levels of the tidal flats can follow the sea
level rise, this implies that the tidal prism of the basin remains unchanged, whereas
the volume of the channels increases. Thus, this scenario results in a demand of sand
from outside.
9.8.4. Adaptation time

The above-mentioned relationships only give an indication of the new equilibrium
between morphology and hydrodynamic conditions in the case of changes. They do
not give any information on how the adaptation will take place and what time this
will take. In general, adaptation processes show a logarithmic character, see Fig. 9.37
and Eq. 1.2 in Sect. 1.5.4.

A0 = 335 km2
300
200
A [km²]
100
τ = 350 yr
0
1500 1600 1700 1800 1900 2000 time [yr]
largest extension
of the Dollard 0 100 200 300 400 500 time [yr after disturbance]
Figure 9.37: Accretion of the Dollard (according to Eysink, 1991). In the equation in the figure the
𝑡-origin is the time after the disturbance (extension), 𝐴 is the area.

467
10
Coastal protection
10.1. Introduction
We have seen in the previous chapters how coasts can develop – can erode or accrete.
Unfortunately, these processes may conflict with economic interests. A channel to
a harbour may silt up, making the entrance too shallow for shipping, or a valuable
structure (such as a hotel or jetty) may be washed away as a beach erodes. Beach
accretion is rarely felt as a problem, except when the beach is too wide and the walking
distance to the waterline is too far from a recreational point of view. We will therefore
mainly focus on erosion problems.
Many sandy coasts all over the world suffer from structural erosion and/or dune and
beach erosion during severe storm surges. In coastal engineering practice an import-
ant aim is the proper protection of these threatened coasts. Moreover, newly reclaimed
areas or upgraded beaches have to be protected from the attacks by the sea. Further-
more, construction of harbours and other engineering works inevitably impacts the
coastal system. The mitigation of adverse impacts should be an integral part of the
design of these works (see Sect. 10.5.2).
Strategies and methods for coastal protection are introduced in Sect. 10.2. Next, the
nature of coastal erosion (permanent or structural versus temporary) is extensively de-
scribed in Sect. 10.3. Subsequently, Sect. 10.4 discusses the principle of interference in
longshore transport rates to defend a structurally eroding stretch of coast. Section 10.5
deals with structures that are designed to interfere in longshore transport rates. Sec-
tion 10.6 is devoted to structures that literally act as a barrier between the sea and
the land. They prevent storm-induced, temporary loss of material from the dunes or
land as well as flooding of the hinterland. Nourishment, a so-called ‘soft’ protection
method, is treated in Sect. 10.7.

468 10. Coastal protection
In the sections on structures, we focus on the functional design aspects of these struc-
tures. We will therefore discuss their effectiveness in coping with certain coastal prob-
lems and (briefly) the appropriate choices regarding position, crest height, length, etc.
No attention will be paid to constructive and technical design aspects, like for example
the required mass of stones of armour layers, thickness of various layers in the struc-
tures, etc. Other TU Delft courses like Bed, Bank and Shoreline Protection (CIE4310)
and Breakwaters and closure dams (CIE5308) deal with this. We have, however, high-
lighted some interesting similarities between sediment transport and breakwater dam-
age in App. D. Also, an overview of these topics can be found in Thorne et al. (1995)
and in the CEM (see Sect. 1.7.3).
Although the term ‘coastal protection’ is usually not only used to mean protection
against erosion, but against flooding as well, the determination of safety levels against
flooding is not discussed in this book. The same holds for measures against sediment-
ation in approach channels to harbours. Dredging, the ‘soft’ solution to the latter
problem, is discussed in detail in the course CIE5300 on Dredging Technology.
10.2. Coastal protection strategies and methods

10.2.1. Management strategies
In the last decade of the last century, especially in the context of raising awareness
of the increased pressures on the coastal zone due to possibly accelerating climate
change, management strategies to deal with coastal erosion were broadly divided into
three categories (see Fig. 10.1):
Retreat do nothing and accept retreat and possibly flooding (as happened during the
successive deglaciation periods since the earth came into existence).
Accommodate adapt coastal infrastructure to resist the increased risk of erosion and
flooding.
Protect take protective measures to counteract erosion and flooding.
The management strategy ‘retreat’ may be applied in situations with strong coastline
fluctuations without a clear long-term trend, or in situations where postponement of
measures leads to more simple solutions in the future, because of the fact that erosion
rates decrease as a function of time.
The chosen strategy as well as the chosen safety levels against flooding determine the
methods to be used. Since we focus, in the present chapter, on protective measures
to counteract erosion, we have implicitly assumed that the choice for protection as a
management strategy has been made. However, protecting the coast is not necessarily
the most appropriate strategy in all situations. The appropriate management strategy
is closely linked to both the level of vulnerability and the land use (infrastructure,
living, recreation, agriculture etc.) and thus to the social, economic and cultural value
of the coast and the amount of available budget.

10.2. Coastal protection strategies and methods 469
(a) retreat (b) accommodate
(c) protect (d) 1990
Figure 10.1: Management strategies for coastal erosion according to IPCC (1992) and World
Coast Conference (1993): (a) retreat, (b) accomodate and (c) protect. Panel (d) refers to a
specific situation in which a certain coastline (in this case the 1990 coastline in the Netherlands)
is preserved using mainly sand nourishments.
10.2.2. Selection of protection method

In general, there are two possible solutions to coastal morphological problems, viz.
‘hard’ measures (coastal structures) and ‘soft’ (natural) measures.
‘Soft’ methods
The principle of ‘soft’ measures like beach or foreshore nourishment is to compensate
for the eroded sand by nourishing sand, without great interference in the sediment
transport patterns. This implies that the ‘natural’ erosion processes are allowed to
continue. The eroded material is replaced on a regular basis with sand from some-
where else. This sand could be supplied, for instance, from deep-water borrow areas,
or from the sediment deposited in an approach channel or upstream of a port. So the
‘soft’ solution to the erosion problem is in fact pumping sand towards the erosion area.
Another example of a ‘soft’ solution is managed retreat in combination with protecting
the retreated shoreline with a restored marsh.
Bypass systems form a special category of ‘soft’ solutions that artificially restore a
(human-induced) blockage of the sediment transport. An example is a sand bypass
system that brings sand from the accreting updrift shoreline to the eroding downdrift
shoreline (see Fig. 10.2). In this way, the interruption of the sediment transport by the
harbour is taken away.
Another example is a system in a river, which bypasses sediment from the upstream
side of a hydropower dam to downstream in order not to interrupt the sediment supply
to a river delta. The bypass system at the dam may take the form of a small cutter
dredge that is continuously cleaning the basins upstream of the dam. The spoil is
transported by a pipeline downstream of the dam into the river. Without such a system,
river deltas may retreat after the construction of the dam.

S=0
S [m3/yr] accretion
erosion
Figure 10.2: Typical structural downdrift erosion and updrift accretion.
‘Hard’ methods
Examples of ‘hard’ measures are series of groynes, series of offshore breakwaters, sub-
merged breakwaters and revetments or seawalls (see Sect. 10.5.1 and Sect. 10.6). They
ensure that sediment is not eroded, or to a lesser extent, by interfering in the sediment
transports in alongshore and cross-shore directions of the coasts. This can be achieved
in many ways, depending on the problem. For instance, undesired sedimentation can
be prevented by structures that guide the currents in such a way that velocities and
thus sediment transport rates increase. The ‘hard’ measures that counteract erosion
can broadly be subdivided in:
1. Structures that primarily influence the rate of longshore transport under both
normal and extreme conditions (groynes, a (long) dam to divert the wave- and
tide-driven currents, detached breakwaters1 );
2. Structures that prevent erosion during extreme storm events (seawall, revet-
ment, sea dike).
The choice between a ‘hard’ or a ‘soft’ measure depends both on the characteristics of
the problem concerned and on economic considerations. In Coastal Zone Management
practice, the use of beach nourishment is becoming increasingly popular. Many of the
frequently occurring adverse side effects of ‘hard’ measures can be avoided by using
beach nourishments. It is noted, though, that beach nourishment in coastal zones that
have a rich nearshore ecological value (e.g. sea grass meadows) is often problematic
because of environmental reasons.
Nonetheless, the possible use of ‘hard’ measures for coastal protection still cannot be
disregarded. A coastal engineer should at least have proper insight into the physical
processes related to ‘hard’ measures. The desired effects (often: reduction or mitig-
ation of the erosion potential in a given stretch of coast), and the often detrimental
effects on adjacent coasts (lee-side erosion), have to be considered with care. Only
then an appropriate choice can be made between the many coastal protection meth-
ods.
1
Emerged detached breakwaters may also reduce storm-induced beach and dune erosion to a certain
extent, see Sect. 10.5.4.

10.3. Coastal erosion 471
10.2.3. From problem definition to realisation

Coastal protection is an important task that requires skilled and experienced profes-
sionals. In some countries special institutes or Coastal Zone Authorities have been
appointed to carry out the tasks involved. It is obvious that such authorities can only
adequately operate if provided with governmental support and legal backing.
A few remarks can be made about the process from problem definition to implement-
ation:
• The solution of a coastal erosion problem always starts with a clear understand-
ing of the coastal processes that cause the coastal erosion problem (see for in-
stance the erosion example of Fig. 10.7);
• Subsequently, the requirements for a possible solution have to be clearly defined.
Should the erosion be prevented along the entire coast or only in a limited area?
Is halting the erosion in a limited area enough or is recovery (accretion) in that
area desirable? And are any changes in the future expected, which require ad-
aptive intervention?
• In the next phase, different alternatives have to be analysed. Which alternatives
meet the requirements? What are possible detrimental side effects? What are
the costs involved?
• In the final selection phase, the best alternative has to be chosen, which is then
designed in more detail, with respect to engineering, cost and implementation;
• After the implementation of the selected protection measure, it is strongly re-
commended to monitor the actual behaviour of the coast. Since ‘art’ and ‘sci-
ence’ are still closely related in coastal engineering practice, the chosen solution
may sometimes turn out to be far from ideal. The experience gained in this way
can be very helpful in designing new projects.
10.3. Coastal erosion

10.3.1. Structural erosion of coasts
We speak about structural erosion in the case of a clear long-term erosional trend.
Hence, structural erosion is a permanent erosion phenomenon. A typical example
of a (human-induced) structural erosion problem is the erosion of a sandy coast on the
lee side of port entrances sheltered by two breakwaters (see Fig. 10.2). As mentioned
before, structural erosion also occurs at river deltas as a result of upstream damming
of the river. This reduces the sediment supply from the river to the coastal zone and
causes long-term retreat of the delta. Sometimes structural erosion problems have
(partly) natural reasons, for instance if the wave climate varies along the coast due to
sheltering effects or in the case of interruptions in the coastline (see Sect. 10.3.3) or if
delta lobe switching occurs (see Intermezzo 2.2).

Structural coastal erosion means that the volume of sand in an arbitrary cross-section
between well-chosen boundaries in that cross-section (control volume in m3 /m, see as
shown in Fig. 10.4) gradually reduces as a function of time. The volume reduction is
typically in the order of magnitude of 10 m3 /m to 50 m3 /m per year and the structural
coastal retreat in the order of a few metres per year (Fig. 10.3). Intermezzo 10.1 provides
some more information on the rate of recession versus the volume rate of erosion.
[m 3 /m]
m
V
time [yr]
Figure 10.3: Volume loss out of control volume due to structural erosion.The slope 𝑚 = 𝑑𝑉 /𝑑𝑡
[m3 /m/yr].
Intermezzo 10.1 Rate of recession versus volume rate of erosion
The two cross-shore profiles in Fig. 10.4 represent the position of the profiles at
two different moments. As a first approximation it is assumed that the shape of
the profiles remains identical (since the boundary conditions, e.g. wave climate
and tides, are unchanged, see Sect. 7.2). The new profile can be found by shifting
the old profile horizontally. It is assumed that not only the waterline shifts in the
landward direction with a certain speed but that this holds in fact for all depth
contours (cf. Sect. 7.4).
If the waterline shows a recession of 𝑅 m/yr on average, the annual loss of volume
out of the control volume area Δ𝑉 (m3 /m/yr) is 𝑅 × (ℎ + 𝑑); see Fig. 10.4 for the
definition of ℎ and 𝑑. The dune height 𝑑 with respect to MSL is easy to measure.
The underwater part of the cross-shore profile ℎ which has to be taken into account
is more difficult to determine. Often an empirical relation for the depth of the so-
called active part of the profile is taken as representative (depth of closure, see
Sects. 1.5.1 and 7.2.3). This relates the depth at the end of the active part of the
profile to the annual wave climate. A first approximation, the depth ℎ can be
found by multiplying the significant wave height which is exceeded for one day
a year, with a factor of 2 to 3 (see the wave-breaking criterion in Sect. 5.2.5 and
Hallermeier’s closure depth in Sect. 7.2.3). With actual profile measurements often
a more accurate estimate of ℎ can be determined.

control volume
Figure 10.4: Retreat of the coastal profile due to structural erosion.
In Fig. 10.3, the development with time of the volume 𝑉 in the cross-shore profile
has been given. The slope 𝑚 = 𝑑𝑉 /𝑑𝑡 (m3 /m/yr) is in this case a measure of
the severity of the erosion problem. Of course a figure like Fig. 10.3 can also be
made for an eroding profile in a structurally eroding part of the coast, by showing
the recession of the position of the waterline (in metres with respect to a reference
point) over time. The slope 𝑅 (m/yr) is then a measure of the erosion rate. However,
from a morphological point of view, it is preferable to represent erosion rates in
terms of volumes instead of distances.
Although beach and dune erosion during a severe storm surge (see Sect. 10.3.2) is
also considered as a typical erosion problem, it is not necessarily a structural erosion
problem. Indeed, after the storm event the dunes and/or upper parts of the beaches
may have lost sediment, which has disappeared from its pre-storm position. Often,
however, the lost volume is found on the foreshore in the nearshore area (Fig. 10.5).
Essentially, the total volume of sand in a cross-shore profile has not changed during
the storm surge. In fact, only a redistribution of the sediment masses over the cross-
sectional area has taken place. Depending of course on the severity of the storm surge,
the volume of sand lost from the upper parts of the cross-section associated with this
type of erosion is in the order of magnitude of 10 m3 /m to 100 m3 /m per event (per
storm, that is to say, per day to a few days). In the period after the storm surge event,
the normal natural conditions often force a recovery of the beaches and dunes. Thus,
dune erosion during a severe storm surge is a temporary instead of a permanent erosion
phenomenon.
In case of typical structural erosion problems, both normal conditions and storm condi-
tions contribute to the eventual loss of sediment from a cross-section. Often a gradient
in the longshore sediment transport is the main reason (gradient 𝑑𝑆/𝑑𝑥 ≠ 0 with 𝑆 the
net yearly longshore sediment transport and 𝑥 the coordinate directed along the coast-
line).

Structural erosion under normal conditions entails that the upper part of the profile
(dry beach and slope of dune or mainland front) does not participate in the transport
processes; the water and the waves do not reach this part of the cross-section. How-
ever, the erosion of the foreshore will continue under these normal conditions.
By contrast, under storm conditions, when the waves reach the dunes due to higher
water levels and higher waves, the upper part of the cross-section forms an integral
part of the entire active profile. In that case, the erosion of dunes and possibly hinter-
land will occur.
In an essentially stable situation, the storm-induced erosion of dunes and hinterland
is only temporary. In a structurally eroding case, however, this erosion is partly per-
manent. This is because the eroded sediments from the upper part of the cross-section
will not fully return under normal conditions, but will be removed in an alongshore
direction. Hence, at the end of the day, by means of cross-shore transport processes,
the structural erosion of the beach and the foreshore will also cause erosion of the
dunes or the hinterland. This distinction is often not clear to the general public. Often
permanent losses of dunes and firm land are incorrectly associated with storm surge
events, while the basic problem is in fact structural erosion.
Structurally eroding coasts are often the source of serious problems for the various
users of the coastal zone. Properties built close to the sea are lost; roads in the area
disappear into the sea. Often society calls for action to be taken by the Coastal Zone
Authorities in order to prevent the detrimental effects of structural erosion.
In summary, the general public often thinks that storms are the reason for (per-
manent) beach and dune erosion problems, while the real reason is a gradient in
longshore sediment transport rates. The storm just provides the necessary link in
a chain of processes, by supplying sediments from the upper part of the profile to
the active littoral zone.
10.3.2. Beach and dune erosion during severe storm surges

Coasts which over a number of years seem to be stable may suffer from the effects
of severe storm surge events. During a severe storm, sediments from the dune and
upper parts of the beach are eroded and deposited in deeper water within a short time
period. This is a typical cross-shore sediment transport process. Severe storms are
usually accompanied by higher water levels and higher waves. Under normal condi-
tions, the shape of a cross-shore profile might be considered to be in equilibrium (cf.
Sect. 7.2). During a storm, the initial profile shape (viz. the pre-storm profile) is far
from the equilibrium shape that corresponds to the severe storm conditions. As a res-
ult, processes occur to reshape the profile, causing erosion of the upper part of the
cross-shore profile, while the foreshore is accreting, resulting in flatter slopes in the
post-storm profile (Fig. 10.5). Note that the control volume does not change during

the storm surge. This is because no sand is lost from the profile; sand is merely re-
distributed in the cross-shore direction. During the storm, while the profile flattens,
the erosion process slows down. After the storm, generally a recovery towards the
original situation occurs under the combined action of moderate waves and wind (the
latter resulting in aeolian transport rebuilding the dunes).
post-surge dune foot SSL

pre-surge dune foot
MSL
before surge
after surge
Figure 10.5: Dune erosion during a severe storm surge redistributes sediments in the cross-shore
direction. There is no loss of sediment in the alongshore direction and hence the sediment will
return under moderate conditions, through the combined action of small waves (under water)
and onshore wind (on the beach).
These episodic events do occur along all types of coasts (along structurally eroding,
stable and even along accreting coasts). As mentioned in the previous section, the rate
of erosion can be very high (of course depending on the actual conditions, erosion
rates of several metres per event, say per day to a few days). Associated with the loss
of volumes of sediment from dune or firm land, a retreat 𝑅 of the coast (see Fig. 10.5)
of many metres may occur during a single event. Nowadays methods are available to
reliably quantify the rate of dune erosion during arbitrary boundary conditions (see
Sect. 7.3.5).
The effects of structural erosion or temporary storm-induced erosion on properties
built too close to the shoreline are eventually the same. In both cases, the properties
may be lost (see Fig. 10.6). However, it is beyond doubt that the countermeasures
used to resolve or mitigate these different types of erosion must be quite different (see
Sects. 10.5.1 and 10.6.1).
10.3.3. Dynamic behaviour of tidal inlets

Coasts near tidal inlets are often very dynamic. This is the result of a complicated
sediment transport pattern, induced by tidal and wave-driven currents. Sediments
from one side are imported into the tidal basin or are bypassed along the tidal bars,
whereas the other side may receive sediments exported from the tidal basin or from
direct bypass (see Sect. 9.4.2). The position of the main channel system may oscillate
or gradually shift in one specific direction due to the accumulation of sediments at one

Figure 10.6: Historical photo of damage to a hotel at Schiermonnikoog due to dune erosion, circa
1923. From Rijkswaterstaat (‘Credits’ on page 575).
side of the inlet and erosion on the other side of the inlet. This may eventually lead to
spit formation (cf. Fig. 8.28). This dynamic behaviour often hampers safe navigation
and may cause damage to the properties adjacent of the inlet. The inlet may act as a
sediment sink for longshore sediment transport, thus depriving the downdrift coastline
of sediment.
10.4. Modification of longshore transport processes

In most cases, the use of structures for coastal protection relies on the ability of such
structures to interfere with the existing sediment transport processes. The main reason
for structural erosion is often a gradient in the net longshore sediment transport along
a stretch of coast. Structures counteracting this erosion are designed to change the
curve of the longshore transport 𝑆 versus distance along the shore 𝑥, such that the
transport gradients become zero and hence erosion is stopped. This situation is dis-
cussed in more detail below.
In Fig. 10.7, line 𝑎 indicates the net yearly longshore sediment transport 𝑆 along a
coast (𝑆: expressed in m3 /yr). The 𝑥-axis is the alongshore coordinate, with the posit-
ive transport direction coinciding with the positive 𝑥-direction. The increasing trans-
port from A to B (difference 𝑉 ) causes an erosion problem in stretch A–B (along A–B
𝑑𝑆/𝑑𝑥 ≠ 0). Such a situation may occur for a straight coastline that is subject to waves
that increase in height in the positive 𝑥-direction along the coast.

10.4. Modification of longshore transport processes 477
S [m3/year]
b S
c
x
no further erosion allowed
A B
Figure 10.7: Longshore sediment transport distribution along eroding coast.
Assume that one wishes to protect only stretch A–B of the coast, for instance because
important investments have been made in section A–B, which are at stake due to struc-
tural erosion. The erosion can be stopped by ensuring that the existing sediment trans-
port distribution line 𝑎 in Fig. 10.7 is changed to line 𝑏 (along section A–B 𝑑𝑆/𝑑𝑥 = 0
in that case). At least in section A–B the erosion would stop if distribution 𝑏 could be
achieved. In the section to the left of A, the erosion will just continue, since the sedi-
ment transports have not changed. In the section to the right of B the existing erosion
also continues, but at an even higher ratethan before, since the throughput of sediment
through the cross-section in point B has been reduced, yielding steep gradients in the
longshore sediment transport distribution. At the right-hand side of B, consequently,
lee-side erosion occurs.
The formulation of the requirements of the sediment transport distribution in section
A–B is rather simple; however, it is rather difficult to acquire curve 𝑏. Structures that,
in principle, interfere with the process of longshore sediment transport rates can be
used. Series of groynes or series of breakwaters with a crest either above (emerged) or
below sea level (submerged) will undoubtedly affect the existing longshore sediment
transport, but it is very difficult to ensure that these countermeasures are appropriate
for specific cases. If the effectiveness of the countermeasure does not meet expecta-
tions, the erosion will be reduced but not entirely stopped. If the countermeasure is
too effective in reducing the sediment transport, or in other words if the reduction of
the sediment transport in stretch A–B is too great (see line 𝑐 in Fig. 10.7), accretion
in stretch A–B will be unavoidable. This might be beneficial for section A–B (accre-
tion instead of the desired stabilization), but the lee-side erosion in the section at the
right-hand side of B will grow worse. In fact, achieving line 𝑐 in Fig. 10.7 represents
an ‘overkill’ operation.
Even if the countermeasures in section A–B are well tuned, lee-side erosion is usually
unavoidable. In fact, a solution for the erosion problem in section A–B, with the help
of structures that decrease the sediment transport along A–B, always comes at the
expense of the stretch of coast beyond B; the problem has simply been shifted. If the

extra lee-side erosion beyond B becomes unacceptable, countermeasures in this section

are necessary as well. The lee-side erosion is then shifted further down the coast. Only
if there is an accreting stretch of coast beyond B, or if position B represents the very
end of a stretch of coast (e.g. if there is a tidal inlet beyond B), lee-side erosion is less
obvious.
Alternatively, artificial nourishment could be selected as countermeasure in section
A–B. The erosion problem is not permanently solved by artificial nourishment; the
nourishment does not reduce the sediment transport involved, but eliminates the ef-
fects of the erosion instead. Since the erosion continues, the nourishment has to be
repeated on a regular basis. In the example of Fig. 10.7, a volume 𝑉 (in m3 /yr) must
be nourished. It is not practical or economical to nourish every year, so usually a
nourishment lifetime of 5 to 10 years is chosen.
At first sight, the limited lifetime of a nourishment may seem to be a drawback. In
many cases, however, artificial nourishment is more cost-effective than solutions with
permanent structures. This is certainly the case when the present-day value of the
measures is calculated (the cost of future nourishment contributes little to the present-
day cost). An additional advantage of artificial nourishment is that extra lee-side
erosion does not take place.
10.5. Structures influencing longshore transport rates

10.5.1. Introduction
Structures of which the primary aim is to change the longshore transport rates under
both normal and extreme conditions are:
• Jetties/shore-normal breakwaters;
• Series of groynes;
• Detached shore parallel offshore breakwaters (emerged and/or submerged).
The above-mentioned structures can all be observed along a Mediterranean coastal
section of about 5 km in Sitges, Spain (Fig. 10.8). Open groyne cells with a spacing
of about 500 m can be observed at the northern part of the beach and partly closed
cells are present along the southern side of the beach. The harnessed solution with
T-head groynes and detached breakwaters on the southern side is necessary to retain
the beach sand within the cells (see further Sects. 10.5.3 and 10.5.4).
10.5.2. Jetties or shore-normal breakwaters

A jetty is a structure perpendicular to the coast that extends out into the sea and pro-
tects a harbour or stabilises a coastline. The crest heights of jetties are often well above
MSL (see Figs. 10.8 and 10.9). They often function as breakwaters, in that they reduce

10.5. Structures influencing longshore transport rates 479
(a) south
(b) north
Figure 10.8: Coastal structures in Sitges, Spain. Orthophotos from ICGC obtained in May 2019
the wave action behind it. In the following, the morphological functions of these jetties
or breakwaters are explored.
Figure 10.9: Port of Scheveningen in August 2010. From Rijkswaterstaat (‘Credits’ on page 575).

Function 1: Blocking the longshore transport of sand, which would otherwise

settle into a dredged approach channel
When a longshore transport threatens to cause sedimentation in an entrance channel
to a harbour or a river, this process can be interrupted by constructing a jetty slightly
‘updrift’ from the harbour or river (that is to say, where the sand is coming from).
This jetty or breakwater should extend at least through the breaker zone, even during
storms, and even after the coast has moved forward as a result of accretion. Material
transported along the coast will accumulate against the jetty on the updrift side.
Generally, dredging is required to maintain an approach channel at the required depth
for shipping. For example, the breakwaters of the Dutch port of IJmuiden extend about
2.5 km into the sea, thus (partly) preventing sedimentation in the approach channel.
In many cases, the approach channel extends seaward of the end of the breakwaters.
Here, the maintenance of an artificial channel is more expensive than in between the
breakwaters. This is because large sedimentation may occur in the unprotected chan-
nel and dredging operations are less efficient here due to wave action. The optimum
length of the breakwaters follows from a balance between the construction costs of
the breakwater and the maintenance costs of the entrance channel.
The impact of long structures on the morphodynamics of the adjacent coasts can be
very large. Since wave-induced longshore sediment transports are interrupted, sedi-
mentation and coastline accretion on the updrift side is expected. On the downdrift
side, erosion and coastline retreat will occur (see also Sect. 8.4.2). This sedimenta-
tion/erosion pattern is avoided with the implementation of an adequate sand bypass
system.
Also, the further the breakwaters extend seaward, the more the (tidal) currents are
affected (see Sect. 5.7.2). Contraction of the currents (high velocities) just near the en-
trance to the port hampers vessels to safely enter and depart the port. The contraction
may further lead to local scour (possibly undermining the stability of the breakwaters).
With a well-chosen layout of the breakwater, the currents can be guided such that the
nuisance for shipping operations will be acceptable and scour is reduced as much as
possible.
Figure 10.10 shows a nature-based port solution on the West African coast, which
has been named the Sandbar Breakwater. Along the Gulf of Guinea coast, longshore
sediment transport is unidirectional, and obstructions such as a breakwater result in
strong sedimentation updrift and continuous coastal erosion on the downdrift side.
The Sandbar Breakwater concept makes the inevitably growing sandbar updrift the
basis for the port protection, while restoring the natural sand balance by repetitive
nourishments downdrift, at the Sand Engine (Van der Spek et al., 2020).
Long structures extending into open sea have an impact not only on the coastal mor-
phology, but may also affect the transport patterns of silt and e.g. fish larvae. These
impacts may be felt at rather large distances from the structures due to the low settling

Figure 10.10: Westward-looking aerial photo of the Sandbar Breakwater, including the sand en-
gine, Lekki, Nigeria. Image courtesy of CDR International B.V.
rates. These aspects must be taken into account in the decision making process. Com-
pare the discussion in The Netherlands of the impact of extensions of the Maasvlakte
(Port of Rotterdam) on morphological processes in the Wadden Sea.
Function 2: Stabilizing a natural river mouth or coastal inlet

Natural river mouths and coastal inlets fronted by barrier spits typically have en-
trances that migrate over time (see Sect. 8.4.5). In order to protect existing infra-
structure and property at the eroding side of the mouth and to accommodate shipping
(yachts; fishery vessels), these entrances may be fixed at a certain position. This can
be accomplished by constructing two jetties at either side of the mouth (see Fig. 10.11).
jetties
S
Figure 10.11: Stabilised river mouth.
Since in most cases these jetties will interrupt the natural longshore sediment trans-
port, a proper design calls for additional measures in order to restore the sediment
transport across the river mouth or tidal inlet. This can be done with a sand bypass
system, which in fact should form an integral part of the design of the stabilizing struc-
ture.
Estimation of the required capacity of a sand bypass system is a difficult task. This
is because the capacity should be based not only on the annual sediment transport

capacity but also on the fluctuation in time and the spatial distribution. Well-calibrated
mathematical models may provide a good first estimate, based on the calculated actual
sediment transport rates and their distribution over space and time.
It is tempting to consider the (yearly) growth of the spit (the migration speed of the
mouth) as a measure for the annual net sediment transport. Be aware, however, that in
natural systems often only a part of the annual net sediment transport contributes to
the growth of the spit; the other part bypasses in a natural manner. If only the growth
of the spit is taken into account, the rates of accretion and erosion after stabilization
and the necessary capacity of a bypass system may be greatly underestimated. There-
fore one may want to postpone the design of a bypass system until nature shows the
real quantities involved, by accumulation on one side and by erosion on the other side.
However, postponing the design of the bypass system (“let’s wait and see”) means in
many cases that at the end of the day there is no money left to build a proper system.
Without an adequate sand bypass system, the negative impact of the stabilization pro-
ject, that is, erosion on the downdrift side of the jetties, is not mitigated.
As in the previous cases, stabilization of a river mouth or tidal inlet requires a compre-
hensive decision-making process, taking both the planned benefits as well as possible
adverse consequences into account.
Function 3: Flushing the entrance channel to a harbour or river by constriction

of the entrance
When a river with a sufficiently large sediment supply debouches into the sea, shoal
formation may be expected slightly offshore from the river mouth. This shoal can be a
major obstacle for shipping, especially during severe storms when waves are breaking
on the shoal. In order to push the shoal formation to deeper water, two jetties can be
built at the river mouth, in such a way that the entrance is kept narrow until deeper
water is reached. In this way, the flow velocities in the entrance are increased, resulting
in an increased sediment transport capacity, which tends to keep the entrance open.
Function 4: Preventing structural erosion of a sandy coast near a tidal inlet

The stretches of coasts near tidal inlets are generally very dynamic (see Sects. 9.4.2
and 10.3.3). Often one side of the inlet shows an accretive and the other side (the
downdrift side) an erosional tendency. By building a long pier (dam) near the inlet
at the end of the eroding coast (the crest level of the dam should be high enough to
catch the sediments), one prevents that sediments disappear into the tidal basin. The
construction of an 800 m long dam at the northern part of the Dutch island of Texel in
1995 is such an example (Fig. 10.12). Although this dam has stopped coastal erosion
(see also Fig. 9.14), it is still unclear how the tidal inlet and tidal basin will respond on
the long term to the reduction of the annual sediment supply. Due to the large surface
area of the tidal basin, the short-term effects have been relatively small.

Figure 10.12: Long dam at the eastern tip of Wadden Island of Texel in August 2011. Photo from
Rijkswaterstaat (see ‘Credits’ on page 575).
10.5.3. Groynes
A jetty only prevents accumulation of material in a small area or stimulates accumula-
tion in another rather restricted area. Its influence is mostly local. A field of groynes,
on the other hand, is a series of smaller jetties extending into the surf zone and spaced
at relatively short intervals along a beach (see Fig. 10.13). They can be very effective
in reducing the existing longshore sediment transport rate2 along a coast. By keeping
the coastal sand trapped between two adjacent groynes, they tend to stabilise the en-
tire coast along which they are built. As such, they can be used to defend an eroding
coast (for instance over a length of 5 km), to widen a beach or to extend the lifetime of
beach fills.
Two main types of groynes can be distinguished:
Impermeable, high-crested structures with crest levels above MSL +1 m; these types
of groynes are used to keep the sand within the compartment between adjacent
groynes. The shoreline will be oriented perpendicular to the dominant wave
direction within each compartment (sawtooth appearance of overall shoreline).
Permeable, low-crested structures with crest levels between the MLW and MHW
lines, such that structure-induced eddy generation is reduced, at least at high
tide (cf. Fig. 5.45). These types of groynes are generally used on beaches with
a small sediment deficit; the function of the groynes is to slightly reduce the lit-
toral drift in the inner surf zone and to create a more regular shoreline (without
sawtooth effect). They are for instance made from sheet piles: a row of vertical
2
Not only wave-induced longshore transport rates will be reduced, but tide-induced longshore transport
rates as well, since groynes push the tidal current away from the coast.

(wooden or steel) piles purposely spaced so as to make a porous barrier, thus

reducing but not totally blocking the longshore transport (Fig. 10.14).
Figure 10.13: Groynes on the Dutch southwestern delta coast, near Domburg, in November 2005.
Photo from Rijkswaterstaat (‘Credits’ on page 575).
Figure 10.14: A row of piles serving as a groyne at the Zeeland coast in March 2004. This is only
one of many types of groynes. Photo by Ad Reniers (see ‘Credits’ on page 575).
If properly designed, groynes can be used to achieve sediment transport curve 𝑏 in

Fig. 10.7, which is the ideal situation in that case. Proper design implies that the right
choices for groyne length, height and permeability to sand are made. The finetuning
problem (length and length/mutual spacing ratio) is, however, a difficult problem to

resolve. The physics of groyne systems are not completely understood, making the
successful design of a groyne system more an art than a science. No generally applic-
able design rules are available, although there are some design guidelines (see below).
A spacing equal to a few times the length of the groynes is common.
Since for solution 𝑏 only a small portion of the sand transport must be stopped, the
groynes should be short in this case, viz. shorter than the width of the breaker zone.
Long groynes extending through the breaker zone tend to reduce the sediment trans-
port curve to (nearly) zero in stretch A–B, which will unnecessarily maximise the lee-
side erosion. A partial reduction can also be achieved with impermeable groynes.
In the example case of Fig. 10.7, in the cross-section through point B, the existing
sediment transport had to be reduced by a factor of approximately 0.5 in order to fulfil
the requirements (i.e. achieving line b in Fig. 10.7; 𝑆B,new ≈ 0.5𝑆B,old ). Note that for
arbitrary cross-sections between A and B different, larger ratios are required. Let point
C be half way between A and B. Then, according to Fig. 10.7, 𝑆C,new ≈ 0.7𝑆C,old .
Another point of practical concern is the absolute magnitude of the net yearly long-
shore sediment transport involved. Although the transport gradient over section A–B
(viz. the difference 𝑉 between 𝑆B and 𝑆A over the considered length of coast) can be
measured quite accurately (via measuring the volume loss in section A–B), the absolute
values of the net yearly sediment transport of either 𝑆B or 𝑆A are in fact not automat-
ically known. Since it is difficult to calculate these sediment transport rates, errors
are easily made in the proper quantification of 𝑆B and 𝑆A . If in Fig. 10.7 both 𝑆A and
𝑆B are increased by Δ𝑆, the difference 𝑉 remains the same. But to achieve a constant
sediment transport in section A–B, quite different reduction factors than mentioned
previously are necessary.
Consider for example the following two situations:
1. Stretch of coast, 5 km: 𝑆in = 100 000 m3 /yr; 𝑆out = 200 000 m3 /yr;
2. Stretch of coast, 5 km: 𝑆in = 200 000 m3 /yr; 𝑆out = 300 000 m3 /yr.
In both cases, the annual loss is 100 000 m3 /yr, which amounts to a rate of structural
erosion of 20 m3 /m/yr, a quite usual annual structural loss. However, an adequate
groyne system for case 1 is quite different from an adequate groyne system for case
2. In case 1, the net sediment transport near the downdrift side of the stretch of coast
must be reduced by 50 % in order to ensure that 𝑆out equals 𝑆in . In case 2 the reduction
should be only 30 %.
The rate of reduction of the sediment transport depends among others of the length
of the groynes. For an impermeable groyne, a rough estimate of the required length
can be made by taking the cross-shore distribution of the undisturbed wave-induced
longshore sediment transport as a starting point (see Fig. 10.15).
The transport reduction can now be estimated by simply assuming that the groyne
completely blocks the transport and that seaward of the groyne the sediment transport

S groyne surf zone
s [m3/m/s]
part of sediment transport
interrupted by groyne
Figure 10.15: Rate of interruption of sediment transport by a groyne.
is unaffected by the groyne. Less crude estimates can be made by applying coastline
models, such as Unibest-CL+ or Genesis (see Sect. 8.3.2). Such models take the impact
of the shoreline development inside the groyne bays on the longshore sediment trans-
port into account. Structure-induced eddies can also significantly impact the sediment
transport patterns. These effects can only be taken into account in area models (2DH
or 3D).
In the framework of Conscience, a European Union-funded research program, the ef-
ficiency of groyne systems was investigated and a number of design guidelines were
presented (Van Rijn, 2010). It was concluded, amongst others, that:
• Nowadays, the design of groyne fields is generally combined with beach nour-
ishment inside the groyne compartments in order to widen the beaches for re-
creation and to reduce the downdrift impacts;
• Long, curved groynes can be used to protect a major beachfill at both ends, cre-
ating a wide beach for recreational purposes (pocket beach) at locations where
lee-side erosion is acceptable or manageable;
• The trapping efficiency can be enhanced by using, for instance, L-head or T-head
groynes (see also Fig. 10.8) instead of straight groynes;
• Groynes preferably only extend over the inner surf zone (up to the landward
flank of the inner bar trough), crest levels should be relatively low and spacings
in the range of 1.5 to 3 times the length. This ensures sufficient sediment by-
passing, so that lee-side erosion is prevented as much as possible.
• By reducing lengths at the downdrift end of the groyne field, downdrift erosion
can be controlled and reduced (groyne tapering);
• By constructing groynes from downcoast to upcoast, initial erosion of the area
to be protected is avoided.

10.5.4. Detached shore-parallel offshore breakwaters

Detached breakwaters are breakwaters parallel to the coast at a certain distance from
the coastline (thus, ‘detached’). They are often built from stone, just like ordinary
harbour breakwaters or jetties, and may be segmented (Fig. 10.17). Disadvantages are
the relatively high construction and maintenance costs and the inconvenience or even
danger to swimmers and small boats. The breakwaters are either emerged (crest above
MSL) or submerged (crest below MSL), see Fig. 10.16.
Emerged detached breakwaters with crest levels well above MSL are known to enhance
accretion in their lee and have successfully been adopted for coastal protection for
many decades. From an aesthetic point of view, submerged breakwaters are preferred
to surface-piercing structures (see Fig. 10.18). However, submerged breakwaters have
resulted in worse shoreline erosion in a number of cases (Ranasinghe & Turner, 2006;
Van Rijn, 2010). Below, we separately discuss the morphological impact of emerged
and submerged breakwaters.
Emerged detached breakwaters

Most emerged breakwaters have been built along micro-tidal beaches in Japan, in the
USA and along the Mediterranean. Few have been built along open, exposed meso-
tidal and macro-tidal beaches. In the shadow zone behind the structures, tombolo
development is stimulated. The governing mechanisms for this shoreline accretion
were treated in Sect. 8.4.3.
emerged
submerged
seabed
(a) emerged (b) submerged
Figure 10.16: Detached shore parallel breakwaters (not to scale: slopes in figure are too steep).
Figure 10.17: Series of emerged detached breakwaters in Sea Palling, England. Note the sali-
ents and tombolos. Vertical aerial photography obtained in May 2019. From UK Environmental
Agency (‘Credits’ on page 575).

Figure 10.18: A coupled breakwater system containing two detached shore-normal submerged
breakwaters, Sunny Isles, USA. Note the salients at the coast, behind the breakwaters. Ortho-
photos obtained in January 2009. From U.S. Geological Survey (‘Credits’ on page 575).
Since emerged breakwaters are effective in reducing the longshore sediment trans-
port capacity in their shadow, structural coastal erosion can be solved with a series
of emerged shore-parallel offshore breakwaters, as in Fig. 10.17. In the example of
Fig. 10.7, the sediment transport in the cross-section through point B must be reduced
from 𝑆𝐵 to 𝑆𝐴 (but not to zero, to avoid unnecessarily large lee-side erosion). Therefore,
either in the area in the lee of an offshore breakwater near B or in the area seaward
of that breakwater, a non-zero sediment transport is required. A salient occurs for a
not too large negative longshore transport gradient behind the breakwater, but still
allows sediment to be transported via the shadow zone behind of the breakwater to
downdrift beaches. A tombolo prevents this.
What type of beach planform (the beach as viewed from above) evolves, strongly de-
pends on dimensions and geometry (breakwater length 𝐿, offshore distance to original
shoreline 𝐷 and length of gap between segments 𝐿gap , see Fig. 10.19). Bricio et al. (2008)
analysed 27 detached breakwater projects along the northeast Catalonian coastline of
Spain (which is almost tideless), based on pre- and post-project aerial photographs.
The offshore distances 𝐷 were in the range of 80 m to 230 m and the (emerged) break-
water lengths 𝐿 in the range of 60 m to 240 m. Tombolos were found for 𝐿/𝐷 > 1.3 and
salients for 0.5 < 𝐿/𝐷 < 1.3. These findings are broadly in line with empirical rela-
tions found in other studies. In Sect. 8.4.3, we assumed tombolo formation for 𝐿/𝐷 > 2,
based on non-interfering diffraction patterns.
Opposite the gaps, shoreline erosion may occur if the gaps are sufficiently large (say
𝐿gap /𝐿 > 1 to 1.5).

L gap length L
diffraction offshore
distance D
original shoreline
tombolo salient
leeside erosion
Figure 10.19: Detached shore-parallel breakwaters.
From the above, it follows that the optimal emerged breakwaters in terms of coastal
protection are built close to the shore, with a high crest level and small gap lengths,
but such a structure will block the horizon and is not attractive in terms of beach
recreation.
Emerged breakwaters also reduce storm-induced beach and dune erosion to a certain
extent. However, the still water level during a storm often increases (storm surge, see
Sect. 5.6). Therefore, depending on the the height of the breakwaters above MSL, tidal
range and storm surge levels, large storm waves may still pass over the structure, such
that storm erosion cannot be completely stopped.
Submerged detached breakwaters

Examples are known from field and laboratory tests in which severe erosion is gener-
ated landward of submerged breakwaters. For example Dean et al. (1997) describe a
striking example where the application of offshore breakwaters created more erosion
than without this ‘protection’ measure: a single submerged breakwater built on the
lower east coast of Florida (USA), approximately 7 km south of the entrance of the Port
of Palm Beach, was later (in 1995) removed because of excessive erosion problems in
the lee of the breakwater.
A submerged breakwater parallel to and at some distance from the shoreline will un-
doubtedly reduce the wave heights landward of the submerged breakwater (depend-
ing on the wave climate, the positioning and the crest height relative to the still water
level). As in the case of emerged structures, this may result in accretion behind sub-
merged structures. However, wave-breaking on the structure drives a current pattern,
with diverging shore-parallel currents in the shadow zone of the submerged breakwa-
ter (see Sects. 5.5.7 and 8.4.3). Along the end sections of (segmented) breakwaters, the
currents escape in the seaward direction (see Fig. 5.49). This current pattern can trans-
port large amounts of sediment outside the area and counteract the accretional tend-
ency. The erosional current pattern becomes stronger for decreasing distance between
the breakwater and the shore. Hence, shoreline erosion may occur, for instance, if a
submerged breakwater is built too close to the shore.

Unfortunately, very few guidelines are available for submerged breakwater design, as
yet. Clearly, the successful application of submerged breakwaters requires a better
understanding of the shoreline response. Any attempt to understand and model the
morphodynamic impact of (submerged) structures requires the explicit consideration
of the resulting complicated circulation currents.
Submerged structures have only a limited effect on storm-induced erosion, as most of
the storm waves can be expected to pass over the structure and attack the dune or cliff
front. Supplementary beach nourishments may be required to deal with local storm-
induced shoreline erosion, especially opposite to gaps, where under storm conditions
large amounts of sediment may irrevocably be carried seaward.
Sometimes, submerged breakwaters are constructed as sills between the tip of groynes
to support the seaward toe of beach fills (perched beaches, see Fig. 10.39).
In conclusion, in this paragraph the morphological consequences of detached

shore parallel offshore breakwater were discussed. The finetuning of series of
offshore breakwaters is a very difficult task, especially for submerged breakwa-
ters. Waves from different directions and tidal effects (water level and currents)
also seriously complicate the situation. A successful project may involve an ini-
tial design phase based on mathematical and physical modelling, the testing of
the design by means of a field pilot project, including a detailed monitoring pro-
gramme, and the finetuning of the design by modification of breakwater lengths
based on practical experience.
If a submerged or low-crested emerged breakwater is not designed properly, additional

negative morphological effects, such as local scour and shoreline erosion, may easily
occur. Attention should further be paid to mitigation of downdrift (lee-side) erosion.
This can be established by creating a transitional zone with gradually increasing gap
lengths and/or decreasing crest levels. In addition, protection measures may be re-
quired against local toe scour in front of the breakwater, just as for seawalls (see
Sect. 10.6.2).
10.5.5. Piers and trestles

Piers and trestles are rather long structures with a horizontal deck on a series of piles
extending perpendicular to the coast into the sea (see Fig. 10.20). These structures
serve as a landing place for vessels, as recreation facilities, as measuring facilities for
coastal processes or as part of a sand bypass facility.
The supporting piles might impact the adjacent coast. Especially if a large number of
large diameter piles have been applied, obliquely arriving waves will cause, in the lee of
the rows of piles, an area with reduced wave heights. The sediment transports may be
reduced as well; spots with some accretion might occur (in general at both sides, since

waves will approach from both sides). In this way a measuring pier creates atypical
measuring conditions.
Figure 10.20: Recreation pier in Scheveningen in August 2010. Photo from Rijkswaterstaat (see
‘Credits’ on page 575).
10.5.6. Concluding remarks

Long breakwaters, series of groynes or detached offshore breakwaters can be used to
solve structural erosion problems. The principle of these measures is that they locally
reduceexisting longshore sediment transports and hence change the longshore trans-
port gradients. Note that even for a well-designed protection scheme, lee-side erosion
is unavoidable; downdrift of the protection scheme, longshore transport gradients will
increase, leading to (additional) erosion. Coastal zone managers have to take these ad-
verse consequences fully into account in the decision-making process.
The use of seawalls or revetments parallel to the shore, built along the front slope
of the dunes or the firm land, does not provide an adequate solution to a structural
erosion problem. This is because the cause of the erosion problem, namely longshore
transport gradients, is not taken away. Although initially the loss of dunes or land is
indeed prevented, the erosion of the beach and underwater profile will continue. The
end result may be a complete loss of beaches and damage or failure of the seawall or
revetments, exemplified in Fig. 10.21 and further discussed in Sects. 10.6.2 and 10.6.3.
Figure 10.21 shows the cross-shore and longshore damage of an incorrectly applied
revetment. The picture is looking towards the north; to the south of this location,
a non-periodic, episodic high river discharge event 25 years earlier had served as a
source of sediment for two decades, counteracting sediment losses from Cua Dai beach
in the northward direction. Around the beginning of the 21st century, the provincial
authorities allowed resorts to be developed to the south of this location. However,

the absence of a new episodic source event eroded the beaches in front of the resorts
and the resorts started to act as revetments and/or groynes. Hence, the coast north
of the resorts started to erode, so that the beaches – important for tourism – were
disappearing. Although this was a problem of alongshore transport gradients causing
erosion, it was attempted to halt this erosion by geotextile bags, preventing the cross-
shore supply of sediment to resolve the longshore transport erosive gradients. As a
result two types of damage occurred: 1) although the erosion was diminished by the
geotextile bags, wave attack on the bags caused toe erosion and damage to the bags’
stability and the beaches in front disappeared; 2) the geotextile bags were initially
successful in preventing further feeding of sediment to the alongshore transports, but
shifted the alongshore transport gradients further to the north, causing erosion.
Figure 10.21: Cross-shore and longshore damage to an incorrectly applied revetment, Cua Dai
beach, Hoi An, Vietnam, around 2016. Photo by Marcel J.F. Stive (‘Credits’ on page 575).
10.6. Structures protecting against storm-induced erosion

If the rate of erosion due to a severe storm surge is unacceptably large during design
conditions, the use of structures may be helpful in reducing the rate of erosion.
Series of groynes do not help to reduce the associated offshore-directed sediment trans-
port. In principle, series of emerged breakwaters or submerged breakwaters reduce the
wave heights landward of these structures and consequently may have the effect of re-
ducing the rate of dune erosion. However, due to the increase of the still water level

10.6. Structures protecting against storm-induced erosion 493
associated with a storm (the surge), the effectiveness of these types of structures in re-
ducing the wave heights under storm conditions is limited (see also Sect. 10.5.4). Also,
wave overtopping of the breakwater segments may lead to concentrated rip currents
in between the breakwater segments and therefore quite adverse effects (as is also
the case for permanently submerged breakwaters). Furthermore, the use of detached
breakwaters for the reduction of storm erosion may have undesirable consequences
because of the inevitable impact on longshore transport gradients.
By contrast, the structures that can be effective in protection of the mainland against
storm-induced erosion and flooding are:
• Seawalls;
• Revetments;
• Sea dikes.
10.6.2. Seawalls
A seawall is a shore-parallel and (nearly) vertical structure at the transition between
the low-lying (sandy) beach and the (higher) mainland or dune (see Fig. 10.22). The
seawall bridges the total height difference between beach and surface level of the main-
land (often a boulevard, road or parking area). The seaward-sloping surface of a sea-
wall is generally smooth and impermeable.
Seawalls are considered a relatively easy to build coastal protection measure. The
philosophy behind the concept is that storm erosion will be prevented by simply cut-
ting off the local supply of material. While in a situation without a seawall even a mod-
erate storm (surge) will attack and erode the mainland, in the situation with a seawall
this is prevented. Unfortunately, a rigid, massive seawall tends to reflect the incoming
waves. The increased turbulence from this reflection may erode a deep trough along
the toe of the seawall (as in Fig. 10.23a). The presence of this trough endangers the
foundation, with the risk that the wall will fail by collapsing into the scour trough.
This can be prevented by maintaining a beach in front of the seawall using some other
means (e.g. regular artificial beach nourishments). If this is done, however, the logical
question is: “why build a seawall, then?” The only purpose of the seawall is then to
provide a clear and fixed distinction between the beach and the mainland (a boulevard
or road), see Fig. 10.23b. Staircases are needed to facilitate access to the beach.
The design conditions for the seawall have to be properly chosen to provide the de-
sired protection of infrastructure and buildings situated close to the edge of mainland
or dunes. The larger the design wave conditions, the heavier the seawall must be; es-
pecially the foundation depth will increase accordingly. The crest height of a seawall
determines (together with the boundary conditions at sea) to a great extent the rate
of overtopping (water reaching the mainland by wave run-up, breaking waves and
splashwater transported by landward-directed wind). With an additional wall and/or
a slightly curved front, rates of overtopping might be reduced (see Fig. 10.24).

(a) normal conditions (photo taken in September 2015 by Wayland Smith ‘Credits’ on
page 575)
(b) storm conditions (photo taken in 2014 by Carol Blaker (‘Credits’ on page 575)
Figure 10.22: A seawall in Dawlish, UK.
Let us consider a stretch of sandy coast. If unprotected, an extreme storm surge may
cause a rate of mainland erosion of e.g. 40 m. With a seawall which is able to with-
stand these conditions, the erosion of the mainland will be zero. Instead, in front of the
seawall a deep scour hole will be formed. The scour does not present any problems as
long as the seawall does not fail; the scour hole will slowly be re-filled during moder-
ate wave and wind conditions. If, however, the seawall partly collapses and locally a
gap in the seawall is formed, a rather dangerous situation will occur. A large volume
of sediment from the mainland may disappear through the gap and will flow in the

boulevard
storm surge
ll
a
seaw
scour
hole
depth of
beach foundation
(a) scour in front of a seawall or revetment

boulevard
waallll
sseeaaw
depth of
beach foundation
(b) maintaining a beach requires e.g. regular sand nourishments
Figure 10.23: Seawall or revetment separating the beach from the mainland (a boulevard in this
case).
ll
a
seaw
storm surge
MSL
storm surge
MSL
Figure 10.24: Seawall with measures to reduce overtopping.
alongshore direction along the sections of the seawall which are still in good condi-
tion, filling the scour trough. In this case, the final rate of erosion of the mainland
behind the gap may be more than the 40 m mentioned for the unprotected case.

Similarly, the unprotected coastline adjacent to a seawall may suffer from extra storm
erosion, since large amounts of sediment are lost to the scour trough.
Other than to prevent storm erosion, a seawall can be used to prevent flooding of low-
lying hinterland, if the existing row of dunes does not meet safety requirements. The
dunes may for instance be too slender to withstand the design wave and water level
conditions.
Neither seawalls nor revetments can protect a coast from structural erosion due to
longshore transport gradients. Nevertheless, a large number of cases are known world-
wide where seawalls have been applied – unsuccessfully, however – to protect structurally-
eroding coastlines.
Structural erosion as a result of a gradient in longshore sediment transport means that
volumes of sediment are irrevocably lost from the area under consideration. Due to the
low frequency of occurrence of extreme conditions, the majority of the longshore sed-
iment transport is due to moderate wave conditions and occurs in the area below MSL.
The presence of the seawall does not prevent this (see Fig. 10.25). Due to cross-shore
redistribution of sediments during high tides and/or modest storms, the alongshore
loss of sediment causes a recession of the entire cross-shore profile in front of the sea-
wall. This causes loss of the beach and deepening in front of the seawall, which in turn
results in an increased wave attack on the seawall and possibly in damage and failure
of the seawall.
MSL
Sx along coast:
beach
initial erosion due to gradient
in longshore transport
Figure 10.25: Seawalls and revetments cannot prevent structural erosion due to longshore trans-
port gradients, since the longshore transports and transport gradients occur below the waterline
in the surf zone.
In the short run, seawalls as a protective measure seem to solve erosion problems.
Indeed, local citizens – whose property is endangered – have noticed in the past that
every storm surge resulted in a loss of land. Just after the construction of a seawall, the
‘protected’ parts of the coast no longer show any signs of further erosion, whereas in
the unprotected parts the erosion of the mainland continues. The seawall is therefore
believed to function well. Unfortunately, after some time it will become clear that
structural erosion has not been countered at all. If coastal zone managers decide to
apply seawalls as a countermeasure against structural erosion, this must be due to
either lack of knowledge or pressure by local citizens.

10.6.3. Revetments
A revetment (Fig. 10.26) is similar to a seawall in the sense that it is a shore-parallel
structure that can prevent storm erosion, but is ineffective against structural erosion.
Compared to a seawall, a revetment is more gently sloping (e.g. 1:2 or 1:4), can have
either a rough or smooth surface and can also be applied over a limited vertical dis-
tance.
Two aspects call for some remarks, viz.:
1. Slope of revetment in relation to the depth of the scour hole;
2. Level of upper end of revetment.
(a) (b)
Figure 10.26: Wooden revetment and eroding cliffs in Sheringham, UK, in May 2008. This is a
good example of a bad revetment; the revetment extends from the cliff to the waterline, whereas
erosion can be expected below the waterline. Moreover, the revetment prevents supply of ma-
terial from the cliff. Photos by Evelyn Simak (‘Credits’ on page 575).
For rather smooth revetments it has been demonstrated that the depth of the scour
trench depends on the slope characteristics. During tests in the Delta Flume of Delft
Hydraulics (now Deltares) it turned out that with a slope of 1:3.6, a deeper scour hole
was created than with a slope of 1:1.8. The deepest point of a scour hole is not al-
ways found at the intersection point between revetment and cross-shore profile (see
Fig. 10.27 for a sketch). For a rough slope, a smaller scour depth may be expected than
for a smooth slope.

SSL
deepest point
scour hole
st-storm profile
po
Figure 10.27: Scour in front of a revetment. Not to scale.
The level of the upper end of a revetment, above which the natural dune or cliff is
unprotected, determines the amount of dune erosion above that level, but also to some
extent the depth of the scour hole. The higher the level to which the revetment is
applied, the smaller the dune erosion, but by implication also the deeper the scour
trough (since no sediment is supplied from the dunes). If the level of the upper end of
a revetment is equal to the storm surge level (or lower than that level), no reduction
in dune erosion is found compared to a situation without any protection.
Summarizing, in a situation without longshore transport gradients, revetments as well
as seawalls can be used to limit the rate of storm-induced erosion. These structures
physically prevent the loss of material from dunes or land and thus reduce dune or
mainland retreat. However, potential scouring in front of the structures has to be
taken into account in the design.
10.6.4. Sea dikes

Sea dikes differ from revetments in the sense that a beach in front of the structure is
absent (see Fig. 10.28). Just like dikes along e.g. a river, sea dikes are often meant to
prevent flooding.
The design crest height of a sea dike has to be carefully chosen in order to prevent
too much overtopping. In the Netherlands the former3 Hondsbossche and Pettemer
sea defence is an example of a sea dike (see Fig. 10.29). Because of the flat slope, the
wave run-up under design conditions for the Dutch coast reaches rather high levels
(about 8 m above design storm surge level). New (preliminary) insights in likely wave
characteristics during design conditions showed that the crest level of the existing sea
defence was too low. As a ‘no-regret’ measure, that is, a matter worth implementing
in any case, the sea defence was partly supplemented with additional sheet piling (see
Fig. 10.30). Note that the additional measures only prevent wave overtopping; the
sheet piling cannot withstand hydraulic pressures in the case of high surge levels.
3
Before the sandy reinforcement of the sea dike was completed, see also Fig. 7.14 and Sect. 10.7.1.

Figure 10.28: Sea dike, Westkapelle, the Netherlands, in June 1993. From Rijkswaterstaat (‘Cred-
its’ on page 575).
Figure 10.29: Hondsbossche and Pettemer sea defence, The Netherlands, in June 1993. From
Rijkswaterstaat (‘Credits’ on page 575).
An additional point of concern is the position of the (sandy) bottom just in front of the
sea defence. The wave run-up depends on the local wave height at the toe of the sea
defence, which is limited by the local depth. Therefore, when during design conditions
erosion of the sandy bottom just in front of the sea defence is expected, the water depth
and hence the typical wave height might be higher than expected.

Figure 10.30: Additional sheet piling at crest of dike, Hondsbossche and Pettemer sea defence.
From Wikipedia, uploaded in 2005 (‘Credits’ on page 575).
10.7. Nourishments
In this section, we discuss how coastal erosion problems can be solved by nourish-
ments with sand instead of by building structures of quarry stone or concrete. These
types of solutions are called ‘soft’ solutions. The basic idea is to supplement sand
by artificial means (dredge, truck) in places where the loss or lack of sand is causing
problems. Sand nourishment leaves the coast in a more natural state than permanent
structures and preserves its recreational value. The method works if ample quantities
of sand are available from a borrow area at a relatively short distance from the problem
area.
An example is the sandy reinforcement of the Hondsbossche and Pettemer sea defence,
completed in 2018 (Fig. 7.14). In order to meet current safety standards, the sea dike
was reinforced in 2015 with a soft, natural barrier of 30 × 106 m3 of sand on the sea side
of the dike. It was renamed ‘Hondsbossche Dunes’ (‘Hondsbossche Duinen’ in Dutch).
The design consists of a soft shallow foreshore (the beach) and a varied artificial dune
landscape. In this way, the spatial quality of the primary flood defence was greatly
improved.
Artificial nourishments can be applied for various reasons:
1. to compensate for losses as a result of structural erosion;

10.7. Nourishments 501
2. to enhance the safety of the hinterland against flooding and to protect the beach
and dune area and properties built close to the edge of the dunes against storm
erosion;
3. to broaden a beach, create new beaches (for recreation) or reclaim large areas of
new land, such as artificial islands.
The second and third applications refer to situations in which the nourishment is, in
principle, a once-only measure. However, in the case of structural erosion, which
is an ongoing process, the nourishment will have to be repeated from time to time.
The interval between successive supply operations depends on the rate of erosion and
on the cost of mobilising the dredging equipment. Generally, an interval of 5 years
between successive operations is considered acceptable.
With respect to structural erosion (application 1), a distinction between the ‘hard’ and
the ‘soft’ methods is that the latter must be repeated from time to time, that it is flex-
ible (in the sense that it is easy to modify the scheme if the results are not as expected)
and that the cost of the operation is deferred, in the sense that it is spread out over
a longer time. For conditions along the Dutch coast this makes ‘soft’ solutions more
economical than ‘hard’ solutions. One could also argue that, since sediment is always
added to the system, a nourishment can never go really wrong, although a badly de-
signed nourishment may not be as effective as expected in counteracting the erosion
problem under consideration, which may lead to damage to properties. Last but not
least, lee-side erosion does not occur with ‘soft’ solutions, as would be the case with
solid structures, like a series of groynes.
In Sect. 10.7.2 some design considerations are given. Next, the three different applica-
tions listed above are discussed in Sects. 10.7.3 to 10.7.5.
For further reference: many details related to the use of artificial nourishment are
found in e.g. Dean (2002) and Van der Graaff et al. (1991) and in a reportVan Rijn (2010),
prepared in the context of the European Union-funded project Conscience (Concepts
and Science for Coastal Erosion Management). These publications also contain many
references to papers specifically devoted to artificial nourishment.
10.7.2. Design aspects

Origin of the sand
The required volumes of sand are in the order of magnitude ranging from several
hundreds of millions m3 per project for large-scale reclamation projects to 10 million
m3 per project for smaller-scale nourishments.
The sand used for nourishments (borrow sand) can be obtained from land-based sources
or from marine sources. Land-based sources may be river beds or dry sand deposits.
Marine sources can be estuaries or the seabed. It is often attractive to try to combine
excavation works with nourishment activities, to reduce the overall cost.

When marine material is used, it must be dredged at a sufficiently large distance from
the shore to prevent extra erosion due to the presence of the borrow pits. Specifically,
when material is dredged from the seabed, the question arises of whether it is better to
dredge the material from small, deep borrow pits, or to dredge thin layers of material
from an extended borrow area. Dredging thin layers means disturbing the biologically
active surface layer over a large area, whereas creating deep borrow pits enhances the
risk of stagnant water of poor quality remaining in the deeper parts. Although research
into the effects of sand borrowing is being carried out, there is no clear conclusion
yet. In the Netherlands, some sand is obtained by maintenance dredging in the access
channels to the ports of Rotterdam and IJmuiden. The remaining amount of sand is
extracted by dredging thin layers at a distance of at least 20 km from the shore.
Although sand nourishment may offer significant benefits, it may also be a costly
method if life spans are fairly short or if the long-term availability of adequate volumes
of compatible borrow sand at nearby (economic) locations is problematic. For example,
suitable material cannot easily be found at most Italian and Spanish sites along the
Mediterranean.
A special application of an artificial nourishment project is a sand-bypass system. A
port built along a sandy coast or jetties to stabilise the mouth of a river flowing out
into the sea often cause updrift sedimentation and downdrift erosion of the coast. To
avoid erosion at the lee side, a sand-bypass system is a perfect tool. It artificially
restores the blockage of the littoral drift by transporting sand from one side to the
other. In Australia various examples are known of adequate and cost-effective systems
(e.g. Nerang River project and Tweed River project).
Quality requirements for the sand

When sand is supplied to a(n) (eroding) stretch of coast, the newly supplied sand will
tend to form a crust or blanket over the existing coastal formations. The supplied
sand will follow the same laws of physics as the existing sand. For this reason, major
changes in the slopes and other coastal features are to be expected when the grain size
of the supplied sand differs from the original material. Usually, this is not acceptable,
and one of the basic rules for beach nourishment is to use material that is relatively sim-
ilar in grain size (and preferably grading) to the existing material (see also Sect. 10.7.3).
However, in some cases the use of coarser material is preferable in order to diminish
losses.
Another point of concern is the silt content of the borrow material. The borrow mater-
ial sometimes contains some (immobile) silt (2 % silt is a quite normal percentage). Dur-
ing the dredging operation, silt is brought into the sea water via the overflow system of
the dredge. This may have a negative impact on the marine environment. Moreover,
any fines still present in the material when fed to the coast may disrupt the ecolo-
gical equilibrium of the coastal zone. Since the bottom material of a sandy shore is
extremely mobile, specifically in the breaker zone, fines will have been washed out

long ago, so that the water is relatively clear. It must therefore be ascertained that the
beach nourishment material contains little or no fines. In the worst case, measures
have to be taken to wash out the fines before placing the material on the beach.
Note that in Europe large areas have been declared Natura 2000 sites including dune
and coastal areas. For these areas, environmental protection and prosperous develop-
ment are policy aims. If a nourishment project is foreseen in the vicinity of Natura
2000 areas, an assessment has to be made whether negative effects are to be expected.
Location of the nourishment

Sand nourishments can be placed at different locations in the coastal profile, namely
(see Fig. 10.31):
1. On the inner slope of the dunes;
2. On the outer slope of the dunes;
3. On the dry beach;
4. On the shoreface.
In theory, as long as the nourishment location is within the active zone, waves will
redistribute material over time from the nourishment location over the entire coastal
profile in order to achieve an equilibrium profile again. Hence, nourishments of the
shoreface and beach can be regarded as local perturbations of the profile, which will
be flattened out by cross-shore sediment transport processes.
1. dune
2. dune
3. beach
4. shoreface
Figure 10.31: Locations of nourishments.
If a sand nourishment is carried out at locations 1 or 2, it is called a dune nourishment.

The use of sand from land-based sources is a reasonably cost-effective option in these
cases. This is even more important if the use of marine sand would cause a salt in-
trusion problem in a vulnerable region. A nourishment at location 3 is called a beach
nourishment and at location 4 a shoreface nourishment. Placing sand on the dry beach
is sometimes complicated, because it is necessary to cross the breaker zone with the
dredging equipment (see Fig. 10.32). Since the breaker zone is a difficult place to work
in, the costs of this option can be high. Furthermore, beach recreation is hampered as

a result of the pipelines that are needed to bring the sand to the beach and the noisy
shovels needed to properly distribute the sand over the beach.
Figure 10.32: Beach suppletion during construction of the Sand Engine using pipelines (indicated
by the black arrows) that cross the breaker zone in March 2011. From Rijkswaterstaat (‘Credits’
on page 575).
If compatible sand is available in nearby borrow areas, shoreface nourishments are

more economical and recreation-friendly than beach nourishments, because the sedi-
ment can be placed at the seaward edge of the surf zone, where the navigational depth
is sufficient for hopper dredgers. During calm wave conditions, material is dumped by
shallow draft dredges (through doors in the bottom of the dredge or through rainbow-
ing) at the shortest possible distance from the beach (see Fig. 10.33). Often shoreface
nourishments are about half as expensive as beach nourishments.
The choice of the nourishment location further depends on the purpose of the nour-
ishment. This is discussed in the next sections.
10.7.3. Counteracting structural erosion of coasts

When a beach suffers from structural erosion, artificial nourishments can be applied
as a ‘soft’ remedy. Losses that occur are replenished from time to time. Regularly
applying nourishments with borrow material that has the same size as the native ma-
terial will, to a first approximation, not interfere in the longshore sediment transports
and thus not influence the losses. Hence, the erosion does not stop and after a certain
period the nourishment has to be repeated. The example of the Dutch coast shows
that long-term and large-scale erosion can be stopped by massive beach and shoreface
nourishment over long periods of time.

Figure 10.33: Shoreface nourishment at the Dutch coast through rainbowing in September 2005.
From Rijkswaterstaat (‘Credits’ on page 575).
If a well-defined part of the coast has been assigned a minimum sand volume, then a
re-nourishment project has to be carried out as soon as this minimum is reached (at
𝑡/𝑡𝐿 = 1 in Fig. 10.34).
actual volumes
minimum
[m 3 /m]
volume
lifetime
V
Db > Dn
Db = Dn
Db < Dn
0 0.5 1 1.5 2
t/t L
Figure 10.34: Schematised behaviour of an artificial nourishment with grain size of the borrow
material 𝐷𝑏 and grain size of the native material 𝐷𝑛 . 𝑉 is a sand volume per m along the shore.
If the borrow sand is the same as the native sand and if it is assumed that the morpho-
logical processes do not change, the erosional tendency after the nourishment is the
same as before the nourishment (see the solid line in Fig. 10.34, which has the same
slope before and after the nourishment). The expected lifetime of the nourishment pro-
ject then is 𝑡2 − 𝑡1 . Generally lifetimes of 5 to 10 years are strived for, since the initial
costs of a nourishment operation are often rather high due to mobilisation costs.

By contrast, if the grain size of the borrow sand is larger than that of the native sand,
the longshore transport rate will be smaller than before. The time over which the
beach volume will decrease to the minimum volume will be larger. The lifetime of
the nourishment increases (see the dash-dotted line in Fig. 10.34). However, when the
grain size of the borrow sand is smaller than the native sand, the longshore transport
in the nourished area increases. Therefore, the lifetime of a nourishment with finer
sand is smaller than the lifetime of a nourishment with native sand (see the dashed
line in Fig. 10.34).
With an average sand loss Δ𝑉 of 20 m3 /m per year (annual retreat of coastline of about
1 m per year; see Intermezzo 10.1) a time period between nourishments of 5 years
means that about 100 m3 /m has to be replenished every 5 years. This is a quite normal
volume. If the relevant stretch of coast is 5 km long, a total volume of 500 000 m3 has
to be nourished every 5 years. This volume must be increased with 10 % to 20 % to
account for additional losses of the fine-grained fraction, which may have been washed
out during the execution of the nourishment works.
Artificial nourishments to combat structural erosion of a stretch of coast can be placed
at various positions in a cross-shore profile (Fig. 10.35), viz. at the dry beach (Fig. 10.35b)
or at the shoreface (Fig. 10.35a). In both cases cross-shore redistribution of sand will
disperse the sediments. The dune reinforcement (Fig. 10.35d) and mega-nourishment
(Fig. 10.35c) are described in the following sections.
In the case of a beach nourishment, sand is placed between the LW line and the dune
foot. Eventually, the quantity supplied will be evenly distributed over the full height
(and length) of the shoreface slope, following the equilibrium rules dictated by wave
climate and grain size. This means that after a beach nourishment operation, large
quantities of newly supplied material will soon disappear under water. The general
public tends to call this phenomenon “erosion”, though we must understand that it is
initially no more than a re-distribution of material within the natural cross-section of
the coast.
For shoreface nourishments, the nourishment volume often is of the order of the
volume of the outer breaker bar (ca. 300 m3 /m to 500 m3 /m). The length scale (along-
shore 2 km to 5 km) of a shoreface nourishment is of the order of several times the
width of the surf zone. Relatively large nourishment volumes are required, as only
part of the nourishment volume (approximately 20 % to 30 %) will reach the beach zone
after 5 years. Besides, because the costs per m3 of a shoreface nourishment project are
far less than the costs of a beach nourishment project, much larger volumes are often
added to the system in the case of shoreface nourishments than in the case of beach
nourishments (also called stockpile nourishments). Such large shoreface nourishments
can be expected to significantly impact sediment transport processes. A large shore-
face nourishment may behave in the same way as a submerged breakwater, although
the effect diminishes over time. Examples are also known where the bar motion in

(a) shoreface nourishment
(b) beach nourishment
(c) mega-nourishment
(d) dune reinforcement
Figure 10.35: Different types of nourishments.

the seaward direction is disturbed by a large shoreface nourishment project; see e.g.
Spanhoff and Van de Graaff (2007).
Artificial nourishments can also be applied to counteract structural coastal retreat due
to sea level rise, see also Sect. 7.4. If one chooses to maintain the current coastline (in
the Netherlands the 1990 coastline is maintained, Fig. 10.1), nourishments should fill
in the space created by sea level rise (between profile 1 and 2 in Fig. 10.36). A volume
equal to the magnitude of sea level rise (SLR) times the fill distance 𝐿 is necessary.
With e.g. SLR = 1.0 m per century and 𝐿 = 1000 m, a volume of 1000 m3 /m is required
in 100 years. This volume of 10 m3 /m per year is a rather normal nourishment volume.
a MSL after sea level rise

h
SLR original MSL
d
d 2
1 3
original profile
after sea level rise
fill distance L
Figure 10.36: Profile adaptations and sea level rise (SLR).
Profile 3 in Fig. 10.36 represents the resulting equilibrium coastal profile without hu-
man interference; the new equilibrium is achieved by a horizontal shift a of profile 2
towards profile 3 in Fig. 10.36. The horizontal shift a can be determined from Eq. 7.13:
𝑎 = (SLR × 𝐿)/(𝑑 + ℎ). With SLR = 1.0 m (e.g. per century), 𝐿 = 1000 m, 𝑑 = 10 m and
ℎ = 10 m, 𝑎 becomes 50 m. Whether such a gradual retreat of the coast (in the example
0.5 m per year) is acceptable depends on the situation. Note that in the situation of
a very flat coastal profile (see Fig. 2.25, Bangladesh), the retreat will be significantly
larger.
10.7.4. Dune reinforcement

Nourishment of the dune area is usually done if the volume of material in the dune
ridge is insufficient to cope with dune erosion during the design storm, the hypothet-
ical extreme storm (consisting of a design wave condition, a design water level and a
duration) a coastal protection structure is designed to withstand. In such cases, the
hinterland may be exposed to flooding, not because of ongoing erosion, but simply be-
cause the existing dunes are not strong enough to withstand extreme conditions. The
protection level of the land behind the dunes is increased more effectively by making
a row of dunes wider than by making them higher.

Figure 10.35d repeats the two options to reinforce a row of dunes with sand. Rein-
forcement at the back of the dunes increases the safety level, but has no morphological
impact. Often it may not be feasible to place sand at the landward side of the dune due
to existing infrastructure and properties.
In the case of seaside reinforcement, it seems that only a relatively small volume of
sand is required. However, the widening of the dunes in the seaward direction disturbs
the dynamic equilibrium of the cross-shore profile (the dune foot being suddenly too
close to the MSL). The same holds if the dry beach is nourished to compensate for
storm losses. As a result, sand from the new dune or beach area is transported in the
seaward direction and the new sand will be redistributed over a large part of the cross-
shore profile (e.g. up to the active part of the cross-shore profile). Therefore, relatively
large volumes of sand are required, since not only the face of the dune has to be shifted
in the seaward direction, but also the active part of the cross-shore profile.
A similar redistribution of sand needs to be taken into account in the alongshore dir-
ection, as shown in Fig. 10.37. A purely local dune reinforcement will therefore have
a short lifetime.
shoreline after nourishment
~40m shoreline after some time
original shoreline
~2km
Figure 10.37: Dune reinforcement over a limited alongshore distance of 2 km, with a 40 m cross-
shore extent, showing the expected alongshore redistribution. Cross-shore redistribution is not
shown.
Widening and heightening of the dunes may not always be the preferred option of
property owners close to the dune, because houses may end up further from the sea
and sea views may be lost.
10.7.5. Beach widening and creation

Increasing urbanization of coastal areas often conflicts with natural shoreline fluctu-
ations. An example is the Mediterranean, where overdevelopment often leads to a
need to widen the beaches by nourishments, which in principle is a once-only meas-
ure. Shoreface nourishments mainly contribute to the sediment balance of the active
surf zone, but are not very efficient for immediate beach widening for recreational pur-
poses. Beach nourishments are about twice as expensive as shoreface nourishments,
but directly benefit the beach. The lifetime of a beach fill can be extended by using
coarser sediments.
More far-reaching is the situation in which a coast has to be extended over a con-
siderable distance (for example in the order of magnitude of 2 km in the cross-shore
direction and over 20 km in the alongshore direction). It is assumed that no serious

structural erosion occurs in the existing situation. The area is intensively used as a
beach recreation area. One of the requirements is that, after the extension of the coast,
recreation beaches are again available. This means that protection of the newly re-
claimed area by a dike or revetment is not an acceptable option.
The most straightforward option for land reclamation would be an entire shift of the
cross-shore profile of 2 km in the seaward direction. This applies not only to the wa-
terline, but in principle also for all other depth contours to a water depth where nat-
ural adaptations of the profile are hardly to be expected (say 15 m below MSL). In
order to achieve a 2 km shift of the coastline, a volume, per metre along the shore,
of 2000 m × 20 m = 40 000 m3 /m is required. The factor 20 m in the calculation is
found by assuming a lower limit of MSL −15 m and an upper limit above MSL of 5 m.
With an alongshore extension over 20 km, the total required volume of sediment is
800 million m3 , which means a very large project.
The large volume of sediment needed is the result of the new coast having a founda-
tion that is identical to the old coast; a large part of the calculated volume is needed to
make the new foundation. In order to restrict the volume of sediment needed for re-
clamation, one could consider an alternative. For example, the upper part of the cross-
shore profile after reclamation is ‘supported’ with the help of a submerged breakwater
(Fig. 10.38). The lower end of the profile can then be left as it was; a large reduction of
the volume of sand is achieved in this way. This solution is called a ‘perched beach’.
a
volume saved
Figure 10.38: Cross-shore profiles in the case of large-scale land reclamation. The shaded
volume is saved if a ‘perched’ beach supported by a submerged breakwater is chosen.
To prevent large alongshore losses, groynes can be applied. The seaward limit of the
reclaimed area can also be moulded with series of detached breakwaters to confine
sediment (Figs. 10.39 to 10.41). The design of this type of reclamation should include
a thorough analysis of all relevant coastal processes and impacts.

artificial nourishment
original coastline
Figure 10.39: Application of structures to confine sand.
Figure 10.40: Submerged breakwater serving to perch the beach nourishment at the Pellestrina
barrier island, in Venice, Italy. On the left side the Venetian Lagoon and on the right side the
Adriatic Sea. Photo by Ni_Giri (‘Credits’ on page 575).
Figure 10.41: Real-life example: tombolos behind emerged (i.e. having their crests above MSL)
offshore breakwaters in Malaga, Spain. Photo from SCNE (‘Credits’ on page 575).

10.7.6. A new nourishment strategy: the Sand Engine

Nourishment strategies applied in the Netherlands have historically developed as fol-
lows. From the 1970s onward, beach and dune nourishments, in which sand was placed
directly on the beach and dunes, were used all over the world (Figs. 10.35b and 10.35d).
Shoreface nourishments, initiated in the 1990s, make use of natural marine processes
to redistribute the sand that is placed under water in the cross-shore direction and
gradually create a wider coastal defense over time (Fig. 10.35a). This strategy is not
very common globally. In the Netherlands nourishments over the last few decades are
regularly executed to maintain the coastline position of 1990. Of these nourisments,
roughly 50 % are traditional beach and dune nourishments and roughly 50 % shore-
face nourishments. The average volume was 30 m3 /m/yr for the whole coast, while
volumes per nourishment intervention varied roughly from 200 000 m3 to 2 000 000 m3
or 200 m3 /m to 500 m3 /m with typical time intervals of 3 to 5 years.
A new nourishment intervention, comprising an unprecedented 21.5 × 106 m3 concen-
trated mega-nourishment known as the Sand Engine, was implemented in the Nether-
lands in the summer of 2011 (Stive et al., 2013). The nourishment advanced the coast
locally in a bell shape with a maximum advancement of more than a km into the sea
over a longshore distance of 1.5 km (Fig. 10.42).
This concentrated mega-nourishment exploits both marine and aeolian processes to
redistribute the sand both in cross-shore and alongshore directions. The Sand En-
gine nourishment is a pilot project to test the efficacy of a local mega-nourishment
as a countermeasure against the anticipated enhanced coastal recession in the 21st
century. The proposed concept, a single mega-nourishment, is expected to be more ef-
ficient, economical, and environmentally-friendly in the long term than the traditional
beach and shoreface nourishments presently in use to negate coastal recession. This
local mega-nourishment is designed to widen the beach along a 10- to 20-km stretch
of coastline and a beach area gain of 200 ha over a 20-year period. Observations since
its construction show indeed a redistribution of the sand feeding the adjacent coasts,
roughly 40 % toward the south and 60 % toward the north. While the jury is still out on
this globally unique intervention, if proven successful it may well become an alternat-
ive worldwide solution for combatting coastal recession on open coasts driven by sea
level rise.

(a) just after completion in 2011
(b) at low water on 27 March 2013
(c) at high water on 27 March 2013
Figure 10.42: Sand Engine mega-nourishment, the Netherlands, realised in 2011. Photos from
Rijkswaterstaat (see ‘Credits’ on page 575).

515
A
Linear wave theory
In the theory for linear waves, equations are given for properties like:
the particle velocity (𝑢, 𝑤) at any height in the water column m/s
the particle acceleration (𝑎𝑥 , 𝑎𝑧 ) at any height in the water column m/s2
the particle displacement (𝜉 , 𝜁 ) at any height in the water column m
the pressure (𝑝) at any height in the water column N/m2
the wave speed (𝑐) m/s
the wave group speed (𝑐𝑔 ) m/s
the wavelength (𝐿) m
the wave profile (𝜂) m
the wave energy per wavelength per unit crest length (𝐸𝑡 ) J/m
the energy per unit water surface area (𝐸) J/m2
the wave power (𝑈 ) J/m×s
The general equations for these properties can be found in Table A.2 in the ‘Trans-
itional Water Depth’ column. The equations for transitional water depth contain hy-
perbolic functions. It may be helpful to use standard tables containing the values of
these hyperbolic functions, which can be found in the CEM. An extract of these tables
is given in Table A.3.
For the specific cases of deep and shallow water (see Table A.1), simplifications can be
made in the hyperbolic functions (see also Fig. A.1).

516 A. Linear wave theory
Table A.1: Criteria for deep- and shallow-water waves with an error of the order of 1 %
Shallow water Deep water

ℎ/𝐿0 ℎ/𝐿 𝑘ℎ ℎ/𝐿0 ℎ/𝐿 𝑘ℎ
< 0.015 < 1/20 < 𝜋/10 > 0.5 > 0.5 >𝜋
10
6
y
x)
h(
s
in
y= y=
x
2 h(x)
y = c os
y = tanh(x)
0
0 1 2 3
x
Figure A.1: Simplified hyperbolic functions.
For relatively deep water (ℎ > 𝐿/2 so 𝑘ℎ > 𝜋 where 𝑘 = 2𝜋/𝐿 is the wavenumber):
1 𝑘ℎ 1
sinh 𝑘ℎ = (𝑒 − 𝑒 −𝑘ℎ ) = 𝑒 𝑘ℎ for 𝑘ℎ → ∞ (A.1a)
2 2
1 𝑘ℎ 1
cosh 𝑘ℎ = (𝑒 + 𝑒 −𝑘ℎ ) = 𝑒 𝑘ℎ for 𝑘ℎ → ∞ (A.1b)
2 2
tanh 𝑘ℎ = 1 for 𝑘ℎ → ∞ (A.1c)
For relatively shallow water (ℎ < 𝐿/20 so 𝑘ℎ < 𝜋/10):

sinh 𝑘ℎ = 𝑘ℎ for 𝑘ℎ → 0 (A.2a)
cosh 𝑘ℎ = 1 for 𝑘ℎ → 0 (A.2b)
tanh 𝑘ℎ = 𝑘ℎ for 𝑘ℎ → 0 (A.2c)
Of the equations in Table A.2, the following are frequently used in this book:
2
2𝜋
𝜔2 = ( ) = 𝑔𝑘 tanh 𝑘ℎ dispersion relation [1/s2 ] (A.3a)
𝑇
2𝜋
𝑘= wavenumber [1/m] (A.3b)
𝐿
𝑐𝑔 2𝑘ℎ
𝑛= = 0.5 (1 + ) [−] (A.3c)
𝑐 sinh 2𝑘ℎ

517
The dispersion relation is implicit in terms of the wavenumber. This means that in or-
der to solve the equation for transitional water depth, an iterative process is required
(to calculate 𝐿 from given ℎ and 𝑇 ). There are also explicit formulas available that
directly calculate 𝐿 without iteration and approximate the solution closely. Alternat-
ively, a look-up table may be used, like Table A.3 that gives ℎ/𝐿 for given ℎ/𝐿0 (with
the deep-water wavelength 𝐿0 = 𝑔𝑇 2 /2𝜋).
The water particle displacement is shown in Fig. A.2 for a wave in shallow water and
for a wave in deep water. In deep water the wave motion does not extend down to the
bed; in shallow water the water makes an oscillating movement over the entire depth.
Near the surface the water particles describe an elliptical path; near the bottom the
water particles make a horizontal oscillating movement.
Figure A.3 shows the relation between the direction of the velocity and the acceleration
of water particles at certain phases in the wave period.
2A 2A
mean ς
H SWL H SWL
position ξ z=h z=h
elliptical orbits circular orbits

2A > H 2A = H
u
w
z z
u bottom z = 0 bottom z = 0
x x
w=0 w=0
u≠0 u=0
shallow-water wave deep-water wave
Figure A.2: Orbital motion under a shallow-water wave and a deep-water wave. Note that, as
opposed to the definitions in Chapters 3 and 5, 𝑧 = 0 at the bottom (instead of 𝑧 = −ℎ) and 𝑧 = ℎ
at the mean water level (instead of 𝑧 = 0).

518
Table A.2: Formulas for shallow-, transitional- and deep-water wave computations according to linear wave theory. Note that as opposed to the definitions in
Chapters 3 and 5, 𝑧 = 0 at the bottom (instead of 𝑧 = −ℎ) and 𝑧 = ℎ at the mean water level (instead of 𝑧 = 0).
Parameter Units Shallow water Transitional water depth Deep water

𝐻
wave profile m → 𝜂= 2
cos(𝑘𝑥 − 𝜔𝑡) ←
𝐿 𝜔 𝐿 𝜔 𝑔 𝑔𝑇 𝐿 𝑔𝑇 𝑔
wave celerity m/s 𝑐= 𝑇
= 𝑘
= √𝑔ℎ 𝑐= 𝑇
= 𝑘
= tanh 𝑘ℎ = 2𝜋
tanh 𝑘ℎ 𝑐 = 𝑐0 = 𝑇
= 2𝜋
= 𝜔 (≈ 1.56𝑇 )
√𝑘
1 2𝑘ℎ 𝑔𝑇
wave group celerity m/s 𝑐𝑔 = 𝑐 = √𝑔ℎ 𝑐𝑔 = 𝑛𝑐 = 2 [1 + sinh 2𝑘ℎ
]𝑐 𝑐𝑔0 = 12 𝑐0 = 4𝜋
𝑔𝑇 2 𝑔𝑇 2
wavelength m 𝐿 = 𝑐𝑇 = √𝑔ℎ𝑇 𝐿 = 𝑐𝑇 = 2𝜋
tanh 𝑘ℎ 𝐿 = 𝐿0 = 𝑐0 𝑇 = 2𝜋
(≈ 1.56𝑇 2 )
𝜔𝐻 𝜔𝐻 cosh 𝑘𝑧 𝜔𝐻0 𝑘0 (𝑧0 −ℎ)
wave particle velocity m/s 𝑢= 2𝑘ℎ
cos(𝑘𝑥 − 𝜔𝑡) 𝑢= 2 sinh 𝑘ℎ
cos(𝑘𝑥 − 𝜔𝑡) 𝑢= 2
𝑒 cos(𝑘0 𝑥 − 𝜔𝑡)
horizontal
𝜔𝐻 𝑧 𝜔𝐻 sinh 𝑘𝑧 𝜔𝐻0 𝑘0 (𝑧−ℎ)
… vertical m/s 𝑤= 2ℎ
sin(𝑘𝑥 − 𝜔𝑡) 𝑤= 2 sinh 𝑘ℎ
sin(𝑘𝑥 − 𝜔𝑡) 𝑤= 2
𝑒 sin(𝑘0 𝑥 − 𝜔𝑡)
2 2
𝜔2 𝐻 cosh 𝑘𝑧
wave particle acceleration m/s2 𝑎𝑥 = 2𝑘ℎ sin(𝑘𝑥 − 𝜔𝑡) 𝑎𝑥 = 𝜔 2𝐻sinh 𝑘ℎ
sin(𝑘𝑥 − 𝜔𝑡) 𝑎𝑥 = 𝜔 2𝐻0 𝑒 𝑘0 (𝑧−ℎ) sin(𝑘0 𝑥 − 𝜔𝑡)
horizontal
−𝜔 2 𝐻 𝑧 −𝜔 2 𝐻 sinh 𝑘𝑧 −𝜔 2 𝐻0 𝑘0 (𝑧−ℎ)
… vertical m/s2 𝑎𝑧 = 2ℎ
cos(𝑘𝑥 − 𝜔𝑡) 𝑎𝑧 = 2 sinh 𝑘ℎ
cos(𝑘𝑥 − 𝜔𝑡) 𝑎𝑧 = 2
−𝐻 −𝐻 cosh 𝑘𝑧 −𝐻0 𝑘0 (𝑧−ℎ)
wave particle displacement m 𝜉 = 2𝑘ℎ
sin(𝑘𝑥 − 𝜔𝑡) 𝜉 = 2 sinh 𝑘ℎ
sin(𝑘𝑥 − 𝜔𝑡) 𝜉 = 2
𝑒 sin(𝑘0 𝑥 − 𝜔𝑡)
horizontal
𝐻𝑧 𝐻 sinh 𝑘𝑧 𝐻0 𝑘0 (𝑧−ℎ)
… vertical m 𝜉 = 2ℎ
cos(𝑘𝑥 − 𝜔𝑡) 𝜉 = 2 sinh 𝑘ℎ
cos(𝑘𝑥 − 𝜔𝑡) 𝜉 = 2
𝜌𝑔𝐻 𝜌𝑔𝐻 cosh 𝑘𝑧 𝜌𝑔𝐻
subsurface pressure N/m2 𝑝 = 𝜌𝑔(ℎ − 𝑧) + 2
cos(𝑘𝑥 − 𝜔𝑡) 𝑝 = 𝜌𝑔(ℎ − 𝑧) + 2 cosh 𝑘ℎ
cos(𝑘𝑥 − 𝜔𝑡) 𝑝 = 𝜌𝑔(ℎ − 𝑧) + 2 0 𝑒 𝑘0 (𝑧−ℎ) cos(𝑘𝑥 − 𝜔𝑡)
2 2
wave energy per J/m 𝐸𝑡 = 1/8𝜌𝑔𝐻 𝐿 𝐸𝑡 = 1/8𝜌𝑔𝐻 𝐿 𝐸𝑡0 = 1/8𝜌𝑔𝐻02 𝐿0
wavelength per unit
crest length
specific wave energy J/m2 𝐸𝑡 = 1/8𝜌𝑔𝐻 2 𝐸𝑡 = 1/8𝜌𝑔𝐻 2 𝐸0 = 1/8𝜌𝑔𝐻02
wave power J/ms 𝑈 = 𝐸𝑐𝑔 = 𝐸𝑛𝑐 = 𝐸𝑐 𝑈 = 𝐸𝑐𝑔 = 𝐸𝑛𝑐 𝑈0 = 𝐸0 𝑛0 𝑐0 = 1/2𝐸0 𝑐0

A. Linear wave theory
Table A.3: Linear wave functions
ℎ/𝐿0 tanh(𝑘ℎ) ℎ/𝐿 𝑘ℎ sinh(𝑘ℎ) cosh(𝑘ℎ) 𝐾𝑠ℎ 𝑛 ℎ/𝐿0 tanh(𝑘ℎ) ℎ/𝐿 𝑘ℎ sinh(𝑘ℎ) cosh(𝑘ℎ) 𝐾𝑠ℎ 𝑛
0.000 0.000 0.0000 0.000 0.000 1.000 ∞ 1 0.200 0.888 0.225 1.41 1.926 2.170 0.918 0.6687
0.002 0.112 0.0179 0.112 0.112 1.006 2.12 0.9958 0.210 0.899 0.234 1.47 2.060 2.290 0.920 0.6559
0.004 0.158 0.0253 0.159 0.160 1.013 1.79 0.9917 0.220 0.909 0.242 1.52 2.177 2.395 0.923 0.6458
0.006 0.193 0.0311 0.195 0.196 1.019 1.62 0.9875 0.230 0.918 0.251 1.57 2.299 2.507 0.926 0.6362
0.008 0.222 0.0360 0.226 0.228 1.026 1.51 0.9834 0.240 0.926 0.259 1.63 2.454 2.650 0.929 0.6253
0.010 0.248 0.0403 0.253 0.256 1.032 1.43 0.9793 0.250 0.933 0.268 1.68 2.590 2.776 0.932 0.6169
0.015 0.302 0.0496 0.312 0.317 1.049 1.31 0.9690 0.260 0.940 0.277 1.74 2.761 2.936 0.936 0.6073
0.020 0.347 0.0576 0.362 0.370 1.066 1.23 0.9588 0.270 0.946 0.285 1.79 2.911 3.078 0.939 0.5999
0.025 0.386 0.0648 0.407 0.418 1.084 1.17 0.9488 0.280 0.952 0.294 1.85 3.101 3.259 0.942 0.5915
0.030 0.420 0.0713 0.448 0.463 1.102 1.13 0.9389 0.290 0.957 0.303 1.90 3.268 3.418 0.946 0.5851
0.035 0.452 0.0775 0.487 0.506 1.121 1.09 0.9289 0.300 0.961 0.312 1.96 3.479 3.620 0.949 0.5778
0.040 0.480 0.0833 0.523 0.547 1.140 1.06 0.9193 0.310 0.965 0.321 2.02 3.703 3.835 0.952 0.5711
0.045 0.507 0.0888 0.558 0.587 1.160 1.04 0.9095 0.320 0.969 0.330 2.08 3.940 4.065 0.955 0.5649
0.050 0.531 0.0942 0.592 0.627 1.180 1.02 0.8998 0.330 0.972 0.339 2.13 4.148 4.267 0.958 0.5602
0.055 0.554 0.0993 0.624 0.665 1.201 1.01 0.8905 0.340 0.975 0.349 2.19 4.412 4.524 0.961 0.5549
0.060 0.575 0.104 0.655 0.703 1.222 0.993 0.8812 0.350 0.978 0.358 2.25 4.691 4.797 0.964 0.5500
0.065 0.595 0.109 0.686 0.741 1.245 0.981 0.8719 0.360 0.980 0.367 2.31 4.988 5.087 0.967 0.5455
0.070 0.614 0.114 0.716 0.779 1.267 0.971 0.8627 0.370 0.983 0.377 2.37 5.302 5.395 0.969 0.5414
0.075 0.632 0.119 0.745 0.816 1.291 0.962 0.8538 0.380 0.984 0.386 2.43 5.635 5.723 0.972 0.5377
0.080 0.649 0.123 0.774 0.854 1.315 0.955 0.8448 0.390 0.986 0.395 2.48 5.929 6.013 0.974 0.5348
0.085 0.665 0.128 0.803 0.892 1.340 0.948 0.8358 0.400 0.988 0.405 2.54 6.300 6.379 0.976 0.5316
0.090 0.681 0.132 0.831 0.930 1.366 0.942 0.8272 0.410 0.989 0.415 2.60 6.695 6.769 0.978 0.5287
0.095 0.695 0.137 0.858 0.967 1.391 0.937 0.8188 0.420 0.990 0.424 2.66 7.113 7.183 0.980 0.5260
0.100 0.709 0.141 0.886 1.007 1.419 0.933 0.8102 0.430 0.991 0.434 2.73 7.634 7.699 0.982 0.5232
0.110 0.735 0.150 0.940 1.085 1.475 0.926 0.7937 0.440 0.992 0.443 2.79 8.110 8.171 0.983 0.5211
0.120 0.759 0.158 0.994 1.166 1.536 0.920 0.7775 0.450 0.993 0.453 2.85 8.615 8.673 0.985 0.5191
0.130 0.780 0.167 1.05 1.254 1.604 0.917 0.7611 0.460 0.994 0.463 2.91 9.151 9.206 0.968 0.5173
0.140 0.800 0.175 1.10 1.336 1.669 0.915 0.7468 0.470 0.995 0.472 2.97 9.720 9.772 0.987 0.5156
0.150 0.818 0.183 1.15 1.421 1.737 0.913 0.7329 0.480 0.995 0.482 3.03 10.324 10.373 0.988 0.5141
0.160 0.835 0.192 1.20 1.509 1.811 0.913 0.7195 0.490 0.996 0.492 3.09 10.966 11.011 0.990 0.5128
0.170 0.850 0.200 1.26 1.621 1.905 0.913 0.7041 0.500 0.996 0.502 3.15 11.647 11.689 0.990 0.5116
0.180 0.864 0.208 1.31 1.718 1.988 0.914 0.6918 1.000 1.000 1.000 6.28 266.893 266.895 1.000 0.5000
0.190 0.877 0.217 1.36 1.820 2.076 0.916 0.6800 ∞ 1.000 ∞ ∞ ∞ ∞ 1.000 0.5000
0.200 0.888 0.225 1.41 1.926 2.170 0.918 0.6687

519
520 A. Linear wave theory
wave propagation
velocity
u=+ u=0 u=− u=0 u=+

w= 0 w=+ w= 0 w=− w= 0
acceleration
ax = 0 ax = + ax = 0 ax = − ax = 0
az = − az = 0 az = + az = 0 az = −
Figure A.3: Local fluid velocities and accelerations at certain phases in the wave period.

521
B
Waves breaking on a beach
B.1. Scale comparison

Stive (1985) performed flume experiments for periodic and random waves breaking
on a gently sloping beach. Measurements were conducted in both a small-scale wave
flume and a large-scale wave flume, such that a scale comparison could be performed.
To that end, the small-scale measurements were scaled up geometrically by the length
scale relation 𝑛length and dynamically by the velocity scale relation according to Froude
0.5 .
𝑛velocity = 𝑛length
B.2. Periodic wave results

The small-scale and large-scale results for periodic waves are compared in Fig. B.1 and
Fig. B.2. Figure B.1 shows the wave height variation and the set-up and Fig. B.2 shows
the characteristics of the velocity field. The small flume has a rigid bed and the large
flume a sandbed, on which a breaker bar was formed. The variability in the large-scale
results, amongst others the shift of the breakpoint during the first and second day, is
related to the morphological development in these large-scale tests.

522 B. Waves breaking on a beach
breakpoint 1st day

wave height breakpoint 2nd day
2
large-scale results
small-scale results
H [m]
0
set-up
0.2
}
0.1
large-s variation
lume
η̅ [m]
cale f
0
line
–0.1 water
bottom profile
large-scale profile after 1st day
6 after 2nd day
small-scale profile
+4.19 m
z [m]
0
0 20 40 60 80 100 120 140 160 180 200 220
x [m] large-scale
Figure B.1: Comparison between small-scale and large-scale flume measurements of wave
height and set-up in periodic waves.

B.2. Periodic wave results 523
zone of initial breaking

time-averaged velocities
z [m]
4.0
2.0 0.1 m/s
max. shoreward velocities (under crest)

z [m]
4.0
2.0 0.5 m/s
max. seaward velocities (under trough)

z [m]
4.0
2.0 0.5 m/s
large-scale measurements
cross-section of velocity measurements small-scale measurements
70 80 90 100 110 120 130 140 150 160 170
Figure B.2: Comparison between small-scale and large-scale flume measurements of time-
averaged velocities, maximum shoreward velocities (under the wave crest) and maximum sea-
ward velocities (under the wave trough).

524 B. Waves breaking on a beach
B.3. Random wave results

For random waves, Fig. B.3 shows the wave height and set-up.
RMS wave height

1.5
1
Hrms [m]
0.5
0
set-up
0.2
large-scale (Hrms/Lp)0 = 0.023
small-scale (Hrms/Lp)0 = 0.010
0.1 small-scale (Hrms/Lp)0 = 0.038
η̅ [m]
–0.1
bottom profile
6
+4.19 m
4
z [m]
2 large-scale profile
small-scale profile
0
0 20 40 60 80 100 120 140 160 180 200 220
x [m] large-scale
Figure B.3: Comparison between small-scale and large-scale flume measurements of root-
mean-square wave height and set-up in random waves.

525
C
Hydrographic charts
C.1. Introduction
Maps are a very important source of information for coastal engineers. This is true for
both maps of the land adjacent to the coast and charts of the seas and oceans. We ex-
pect that these maps will give accurate information about the topography of the area,
but often additional information relating to land use, infrastructure, elevation, etc. is
provided. For the coastal engineer, charts of the seas and oceans are of particular in-
terest. Such charts have been produced for many centuries to provide information to
seafarers. The production of these charts was originally in the hands of private enter-
prises that had an interest in the trade between Europe and the East and West Indies.
In the early days of this trade, maps and charts represented a great commercial value
and they were kept secret by institutes like VOC and the British East India Company.
Later, from the early 19th century, with the formal establishment of the colonies, the
role of governments in various countries became more important. The task of making
proper maps of sailing routes and ports was then transferred to the various navies. Up
to today, in most countries the national navy has a hydrographic department that is
responsible for providing up to date information for ocean navigation. An important
part of that information is contained in hydrographic charts that give an impression of
the local situation, including topography, bottom material, depths, sea levels, currents
etc.
Such hydrographic charts are indispensable to sailors, and the presence of up-to-date
charts is mandatory on board of seagoing vessels. When people cannot easily see what
is below the surface of the water, maps and charts provide the only way for navigators
to find out where it is safe for the ship to go and where it would be unwise to venture.
Hydrographic charts are also an important tool for the coastal engineer, because these
charts give reliable information on the conditions of the coastal zone. For engineers,
however, not only the latest charts are of interest, but certainly also older maps that can

526 C. Hydrographic charts
still be obtained from the archives of the various hydrographic institutes. A sequence
of maps gives a good impression of long-term morphological developments.
This section shows the general principles governing the handling of maps and, more
specifically, hydrographic charts, and indicates roughly what information can be ob-
tained from them.
C.2. Units and their background

Hydrographic charts were meant to provide assistance to the navigators on board sail-
ing vessels, who had little more in the way of instruments than a clock and a sextant.
Positions were determined with respect to the position of the sun and the stars. The
grid of the hydrographic chart is therefore the grid of the degrees latitude and lon-
gitude as drawn on the globe. Transformation of this spherical grid to a plane map
causes distortions, either in the centre or in the corners of the map. This means that
the coordinates of the grid as indicated along the borders of the map are not linear.
Since the mutual distance between the longitudinal coordinates (meridians) varies
(they are long at the equator and zero at the poles), only the degrees of latitude (paral-
lels) give a proper indication of the scale of the map. The circumference of the earth is
40 000 km, which is divided into 360° (degrees), each consisting of 60′ (minutes). This
means that the 40 000 km are equal to 360 × 60 = 21 600′ . The sailors used the minute
as the basis for their unit of distance, the nautical mile, which thus equals slightly
more than 1850 m. Early hydrographic charts were not based on the metric system.
Their scales therefore appear unusual to people who are familiar only with the metric
system of measurement.
The speeds of vessels and the velocities of currents are often expressed in nautical
miles per hour (also called knots), which is slightly more than 0.5 m/s.
Water depths (soundings) are expressed either in the traditional nautical system, or in
the metric system. This is always indicated on the map. The nautical system uses feet,
fathoms (Dutch: vadem), or fathoms plus feet. A foot is equal to 0.3048 m; a fathom is
equal to 6 feet or 1.83 m.
C.3. Explanatory notes

The first thing to do when looking at a map is to study the key. The following inform-
ation may be found:
Horizontal scale The scale indicating the dimensions of details at the map. Most sail-
ors look at the latitude border scales at the side of the map. 1 Minute equals 1
nautical mile. Most coastal engineers look at the linear scale. As a result of the
projection of the globe on a plane, the scale changes over the map. The larger
the area covered by the chart, the larger the deviations.

C.3. Explanatory notes 527
Vertical scale Depth on a hydrographic chart may be shown in metres, feet or fathoms.
The reference level (Chart Datum) of the depths is also important. Chart Datum
(CD) is often related to specific tidal data. This may be LAT, MSL or any other
reference level. Different countries use different definitions of CD! A Belgian
map of the Western Scheldt may thus give different depth values than a Dutch
map of the same area. Even on one map, the CD may differ for different locations,
because the tidal data differ from place to place. This poses a risk for coastal
engineers, since we usually assume that the datum level of a map is horizontal.
Specifically when we make hydraulic calculations, we must make sure that we
use levels that find a reference to a horizontal plane, in order to eliminate errors
in the gravitational forces. A striking example is given in Fig. C.1. The result of
misinterpretation is given in Fig. C.2.
Sharpness Dock
elevation [m]
Aust MLWN
n MLWS
ver
Se LAT
Chart Datum
upstream
Figure C.1: River Severn (England). When CD is related to LAT, or another tidal level, it increases
in upstream direction.
• What happens to levels above the reference level, like mud flats or sandbanks?
These levels are underlined and refer to CD. Levels at the shore may also be
shown. For these levels, different reference levels are usually used, rather than
CD;
• Tidal streams/currents are sometimes shown;
• Dates of publication and dates of smaller or larger corrections;
• Tidal levels are usually shown at some specific locations, as shown in Table C.1.
Terms related to tidal levels are summarised in Table C.2.

Figure C.2: If a horizontal free-surface level is determined relative to a chart datum which is not
horizontal, the free-surface level may be misinterpreted as non-horizontal.
Table C.1: Tabular statement of semi-diurnal or diurnal tides.
place Lat. Lon. heights in metres/feet above datum

(N/S) (E/W)
(position for MHWS MHWN MLWN MLWS
which tidal
levels are
tabulated) MHHW MLHW MHLW MLLW
Table C.2: Terms related to tidal levels.
CD Chart Datum
LAT Lowest Astronomical Tide
HAT Highest Astronomical Tide
MLW Mean Low Water
MHW Mean High Water
MSL Mean Sea Level
MLWS Mean Low Water Spring
MHWS Mean High Water Spring
MLWN Mean Low Water Neap
MHWN Mean High Water Neap
MLLW Mean Lower Low Water
MHHW Mean Higher High Water
MHLW Mean Higher Low Water
MLHW Mean Lower High Water

C.3. Explanatory notes 529
MSL – Mean Sea Level the average height of the sea measured over a long period of
time. This is the average level which would exist in the absence of tides.
LAT – Lowest Astronomical Tide the lowest tide level which can be predicted to oc-
cur under average meteorological conditions and under any combination of as-
tronomical conditions.
HAT – Highest Astronomical Tide the highest tide level which can be predicted to
occur under average meteorological conditions and under any combination of
astronomical conditions.
MHW – Mean High Water the average height of all high tides.
MHHW – Mean Higher High water the average height of the higher of the two daily
high tides over a long period of time. When only one high water occurs per day,
this is taken as the higher high water.
MLW – Mean Low Water the average height of all low tides.
MLLW – Mean Lower Low Water the average height of the lower of the two daily
low waters over a long period of time. When only one low water occurs per day,
this is taken as the lower low water.
MHWS – Mean High Water Spring the long-term average of all high-water observa-
tions of the heights of two successive high waters during those periods of 24
hours (approximately once a fortnight) when the range of tide is greatest, at full
and new moon.
MLWS – Mean Low Water Spring the long-term average of all low-water observa-
tions of the heights of two successive low waters during those periods of 24
hours (approximately once a fortnight) when the range of tide is greatest, at full
and new moon.
MHWN – Mean High Water Neap the long-term average of all high-water observa-
tions of the heights of two successive high waters when the range of tide is the
least, at the time of first and last quarter of the moon.
MLWN – Mean Low Water Neap the long-term average of all low-water observations
of the heights of two successive low waters when the range of tide is the smallest,
at the time of first and last quarter of the moon.
MLHW – Mean Lower High Water the mean of the lower of the two daily high wa-
ters over a long period of time. When only one high water occurs on most days,
no value is printed in the MLHW column, indicating that the tide is usually
diurnal.
MHLW – Mean Higher Low Water the mean of the higher of the two daily low wa-
ters over a long period of time. When only one low water occurs on most days,
no value is printed in the MHLW column, indicating that the tide is usually di-
urnal.

C.4. The map itself

Once familiar with this information, the map itself may be studied. A complete list of
symbols used on Admiralty Charts1 may be found in Symbols and Abbreviations Used
on Admiralty Charts (NP5011) (2020). A summary of frequently used symbols is shown
in Table C.3.
Table C.3: A summary of frequently used symbols and abbreviations on Admiralty charts
Natural features
steep coast, cliffs
flat coast
sandy shore
stones stony shore,

shingle shore
sand
dunes dunes
mangrove
marsh
swamp,
marsh
salt marsh
Cultural features
buildings
bridges
7-8
cables
20
pipelines
1
Admiralty charts are hydrographic charts issued by the United Kingdom Hydrographic Office.

C.4. The map itself 531
Landmarks
factory hotel examples of

landmarks
Artificial features
dykes
seawall
causeway
causeway
breakwater
groyne
mole
wharf
pier
pier, jetty
pontoon
Dn Dns dolphins
ramp
Nature of the seabed

S Sand
M Mud
Cy Clay
Si Silt
St Stones
G Gravel
P Pebbles
Cb Cobbles
R Rock
Co Coral
Sh Shells

C.5. Interpretation
Maps may also provide information about coastal processes like wind and wave dir-
ections and heavy breaking of waves. Many other phenomena can be derived by in-
terpreting the coastal forms on the map: a spit indicates the direction of the long-
shore transport, and thus the dominant wave direction, and river sediment transport
is shown by the presence of shoals and bars. If the map shows only a long, straight,
sandy shore, little can be concluded about wave and wind direction and intensity. Only
when there is an interruption to this shore, it is possible to determine the prevailing
wave and wind direction and resulting sediment transport. For example, at a river
mouth one may find out whether the river or the sea is dominant and the magnitude
and direction of longshore and river sediment transport may be deduced. Another ex-
ample is formed by protruding rocks or artificial features like groynes or breakwaters
on a sandy coast. Here one may also find indications of the presence of longshore
sediment transport (magnitude and direction) and, hence, the wind and wave direc-
tion. Detached obstacles (rocks or breakwaters) may give even more exact informa-
tion about wave direction. In the lee of these obstacles sediment tends to settle, so the
position of shoals indicates the lee side and thus the wave direction.
If dredged channels are present, it is clear that sediments have to be removed on a
more or less regular basis. The location of the dumping grounds of dredged material
sometimes gives an indication of the dredging method that is commonly used.
In large sandy areas, the bottom contours also indicate possible current patterns. Places
with great depths probably indicate areas where the current will be concentrated. Shal-
low areas indicate low current velocities.
C.6. Limitations
Although hydrographic charts provide a very valuable contribution to our knowledge,
the purpose of these charts is to assist navigators, rather than engineers. Soil data
mentioned on the chart are generally only an indication of the surface of the seabed;
they cannot be used in designing a foundation. The charts, and certainly the portions
close to the shore, are meant to warn sailors against running aground. Relatively
more attention is therefore paid to shoals and low water conditions, rather than to
gullies and extremely high water levels. Moreover, the scale of the charts is generally
unsuitable for construction work. For the specification needed in tenders and project
drawings, more detailed maps are required.
When comparing old and recent charts, one must be aware that the locations of buoys
and lighthouses may have changed since the survey or the drawing of the original
chart. Therefore, one should remain vigilant when using older maps for comparison.
The most recent maps will probably be based on positioning with DGPS, an electronic
positioning system using satellites as beacons. This eliminates most errors.

533
D
Stability of structures
D.1. Introduction
In this book only the functional design of coastal engineering structures is treated,
such as jetties/breakwaters, groynes, detached shore-parallel offshore breakwaters,
seawalls and revetments (Ch. 10). We do not discuss the technical and constructive
design of these structures, but refer to other TU Delft courses like Bed, Bank and
Shoreline Protection (CIE4310) and Breakwaters and Closure Dams (CIE5308). Yet,
we would like to point out some similarities between bottom sediment transport on
the one hand and stone stability and structural damage approaches for slopes made of
loose rock and of breakwater elements on the other hand in a more general way.
D.2. Initiation of transport and damage

In Sect. 6.3 the ‘initiation of transport’ concept of Shields (1936) was introduced for
sediment particles under the impact of uniform flow. The parameter (𝜃) that he in-
troduced is the ratio of the drag force due to the flow over the resisting force due to
the underwater weight of the sediment. Initiation of motion occurs for a value of this
parameter (𝜃𝑐𝑟 , Eq. 6.12) between 0.03 and 0.05 in practice. When, at a later stage, sed-
iment transport formulas for uniform flow started to be developed, it turned out that
Shields’ parameter was not only a good parameter to describe threshold of motion, but
also a fundamental parameter in empirical transport formulas, which nearly all can be
expressed as a power function of 𝜃 or (𝜃 − 𝜃𝑐𝑟 ), with the power ranging from 1.5 to 3
(see Sect. 6.5). That is why 𝜃 or (𝜃 − 𝜃𝑐𝑟 ) is often referred to as the mobility number.
Sleath (1978) evaluated initiation of sediment motion under oscillatory flow from lit-
erature and his own experiments, and showed that by using the velocity amplitude

534 D. Stability of structures
(specifically the orbital velocity amplitude) a Shields parameter for waves could be de-
rived, where the empirical values for the initiation of motion only differ marginally
from 𝜃𝑐𝑟 , viz. in the range 0.02 to 0.05.
Interestingly, Iribarren (1938), two years after Shields’ publication and most likely in-
dependently from him, proposed a stability criterion for a sloping rubble mound break-
water. The breakwater is stable if:
3
𝑁 𝜌𝑠 𝑔𝐻 3
𝜌⏟
𝑠 𝑔𝐷 ≥ (D.1)
Δ3 (𝜇 cos 𝛼 ± sin 𝛼)3
stone weight
in which the friction coefficient 𝜇 = tan 𝜑𝑟 with 𝜑𝑟 is the natural angle of repose,
Δ = (𝜌𝑠 − 𝜌)/𝜌, 𝛼 is the angle of the bed slope, 𝐻 is the wave height and 𝑁 is a
coefficient. The density 𝜌𝑠 should now be interpreted as the density of the blocks or
stones and 𝐷 as a typical size.
We can rewrite this equation as:
𝐻 𝜌𝑔𝐷 2 𝐻 1
= 3
≤ 3 (D.2)
Δ𝐷(𝜇 cos 𝛼 ± sin 𝛼) (𝜌𝑠 − 𝜌) 𝑔𝐷 (tan 𝜑𝑟 cos 𝛼 ± sin 𝛼) √𝑁
In the second term of this equation, we can recognize the ratio of drag force over the
resisting force, just as Shields used (see Sect. 6.3.1). The numerator represents the
drag force 𝜌𝑢 2 𝐷 2 taking √𝑔𝐻 as a typical value for 𝑢, assuming that the wave height
of the order of magnitude of the water depth. The denominator reflects the resisting
force proportional to (𝜌𝑠 − 𝜌)𝑔𝐷 3 . We furthermore find in the denominator the effect
of the bed slope, where the plus sign is for uprush and the minus sign for downrush.
Once 𝛼 approaches the natural angle of repose 𝜑𝑟 , the resisting force approaches zero
for downrush. Furthermore, note the correspondence between the slope correction
tan 𝜑𝑟 cos 𝛼 ± sin 𝛼 and the denominator of Eq. 6.48a.
We can write Eq. D.2 as a slope-dependent criterion for the initiation of movement of
a block of a certain size:
𝐻𝑐𝑟 (𝜇 cos 𝛼 ± sin 𝛼)

= (D.3)
Δ𝐷 3
√ 𝑁
Later, many experiments were performed, mostly by Hudson (1952), to find the con-
stants of proportionality in Eq. D.3. For practical reasons, Hudson finally proposed a
criterion in which he indicated the proportionality constant as 𝐾𝐷 and in which he
changed the slope correction term, not based on physics, to cot 𝛼.
𝐻𝑐𝑟 1/3
= (𝐾𝐷 cot 𝛼) (D.4)
Δ𝐷

D.2. Initiation of transport and damage 535
Note that what 𝜃𝑐𝑟 is for the initiation of motion of sediment, (𝐾𝐷 cot 𝛼)1/3 is for ini-
tiation of damage on a rocky slope. Both parameters are based on the same stability
principle. Hudson’s slope correction term limits the validity of this generally-applied
formula to the range of 1.5 < cot 𝛼 < 4.
Just like the values of the Shields parameter, the values of 𝐾𝐷 are based on experiments.
Because structures represent a physically much more complex system than a flat sandy
bed, the range of 𝐾𝐷 values is larger. For instance, for natural rock 𝐾𝐷 is 3–4 and for
artificial elements, like Tetrapods or Xblocs®, it is 8–10. The higher values for artificial
elements are due to the interlocking effect.
This would lead to values for (𝐾𝐷 cot 𝛼)1/3 within the range of 1.5 to 3.5 and thus two
orders of magnitude larger than the critical Shields parameter. What is the reason
behind this difference? To answer this question, we rewrite the Shields parameter
using 𝑢 = √𝑔𝐻 as follows:
𝜏𝑏 𝜌𝑐𝑓 𝑢 2 𝐻
𝜃= = = 𝑐𝑓 (D.5)
(𝜌𝑠 − 𝜌) 𝑔𝐷 (𝜌𝑠 − 𝜌) 𝑔𝐷 Δ𝐷
Hence, the Shields parameter is related to 𝐻 /Δ𝐷 through a friction factor 𝑐𝑓 (order
of magnitude 10−2 ). This would lead to values for the critical Shields parameter that
are two orders of magnitude smaller than critical values of (𝐾𝐷 cot 𝛼)1/3 , which agrees
with the above-mentioned typical ranges.
The simplicity of Hudson’s formula is very attractive, but the complex physics of
rubble mound structures makes that these physics can only be included empirically
in the 𝐾𝐷 coefficient. These physics are in the loading, the structure’s geometry and
its permeability. Van der Meer (1988) conducted a lot of empirical research under vari-
ous conditions and arrived at expressions for (𝐾𝐷 cot 𝛼)1/3 , where this parameter is
a function of the structure’s slope, the type of wave-breaking (Iribarren number), the
number of waves, the permeability and the progressive damage.
As an example we present the equation of Van der Meer (1988) for plunging breakers
on a loose rock slope, in which the right-hand term is a fit to (𝐾𝐷 cot 𝛼)1/3 that includes
the permeability (𝑃), the damage (𝑆), the number of waves (𝑁 ) and the type of wave-
breaking (𝜉 , the Iribarren parameter Eq. 5.20, which includes the slope as a property
of the structure’s geometry):
𝐻𝑠,𝑐𝑟 /(Δ𝐷) = 6.2𝑃 0.18 (𝑆/√𝑁 )0.2 𝜉 −0.5 (D.6)
Typically this leads to a progressive damage curve as shown in Fig. D.1. Note 𝑆 = 2
means very little damage and 𝑆 = 20 means serious damage.

536 D. Stability of structures
serious damage
S [-]
very little damage
Hs/ΔD [-]
Figure D.1: Damage curve according to Eq. D.6 with 𝑃 = 0.5 for a permeable core, 𝑁 = 3000
(which for an average period of 6 s represents a storm of 5 hours duration) and 𝜉 = 2.
In sediment transport terms, i.e. we solve for the damage that may be interpreted as
transport, we may rewrite Eq. D.6 to:
−5 5
𝑆 = (6.2𝑃 0.18 ) √𝑁 𝜉 2.5 (𝐻𝑠 /Δ𝐷) (D.7)
Note that damage now takes on a wider meaning: damage can now vary from initiation
of motion to larger amounts of motion i.e. ‘transport’. It appears that the damage is
proportional to the power 5 of the Hudson parameter. Again we see a similarity with
sediment, in that 𝐻𝑠 /Δ𝐷 is used as a mobility number, so that we are talking about
transport.
D.3. Other protections

In the above discussion, we assumed that the main loading of the structures was ex-
ternal, because of the large permeability of these rocky slope structures. When we
consider coherent, semi-permeable, placed block revetments or impervious asphalt or
concrete structures, the external loading also leads to internal loadings. The failure
mechanisms thus become much more complex. We will not consider this here, but it
is interesting to note that the empirical formulas for semi-impermeable structures also
fall back on Iribarren’s mobility parameter.

537
E
Responses to the closures of Dutch
tidal basins
Reprinted with permission from: Wang, Z. B., De Ronde, J. G., Van der Spek, A. J. F.
& Elias, E. P. L. Responses of the Dutch coastal system to the (semi-) closures of tidal
basins. In: International Conference on Estuaries and Coasts, Sendai, Japan. 2009.
E.1. Introduction
The Dutch coastal system consists of the North Sea coast, a series of Wadden Sea
tidal inlets in the north and various estuaries in the southwestern delta area. During
the last century, various engineering works in the Dutch coastal system were carried
out for flood defence and/or land reclamation purposes. Among these engineering
works, two of the tidal inlets in the Dutch Wadden Sea and three of the estuaries in
the southwestern delta were entirely or partly closed or semi-closed (Fig. E.1). The first
closure is the Afsluitdijk, a dam of 30 km finished in 1932 and separating the former
Zuiderzee from the Wadden Sea. The last of these closures is the Eastern Scheldt Storm
Surge Barrier, finished in 1986, which semi-closed the Eastern Scheldt estuary.
These closures differ in type and location within the corresponding basins. Three
of the five are fully closed dams: the Afsluitdijk, the closure of Lauwerszee and the
closure of the Grevelingen. Haringvliet is closed by a dam in combination with sluices
which only allow discharge of freshwater from the river upstream. The Eastern Scheldt
Storm Surge Barrier only closes the basin under extreme conditions and allows tidal
flow, although with reduced cross-section. The two closures in the Wadden Sea only
close the landwards part of the tidal basin. The closures in the southwestern delta area
almost completely close the entire corresponding basins, but their locations relative to
the mouths of the basins differ from each other.

538 E. Responses to the closures of Dutch tidal basins
Figure E.1: Closures of tidal basins in the Dutch coastal system.
These closures of the tidal basins have impacted the development of not only the (semi-
)enclosed basins themselves, but also the adjacent coast and tidal basins. They influ-
ence large-scale morphological developments and the sediment budget, and thereby
impact the maintenance of the coast by sand nourishment. On a smaller scale they
influence the developments of channels, intertidal flats and other morphological ele-
ments. Specifically the morphological developments of the intertidal flats make an
impact on the ecological system.
In this paper we evaluate the effects of all these closures on the morphological devel-
opment of the various parts of the Dutch coastal system. The evaluation is mainly
based on analysis of field data. Bathymetric data have been collected since 1926. Fur-
thermore, the results of earlier modelling studies are used. Special attention will be
paid to the influence of the location of the closure and the type of the closure. It will
e.g. be shown that the total sediment deficit for establishing the new morphological
equilibrium caused by a closure is very much dependent on the location of the closure
relative to the mouth of the tidal basin. We will also show the different environmental
problems caused by the different types of closures. We believe that the lessons learned
from the evaluation can also be relevant elsewhere in the world.

E.2. Closures in the Wadden Sea 539
E.2. Closures in the Wadden Sea

E.2.1. Closure of Zuiderzee
The Afsluitdijk, a 30-km dam separating the former Zuiderzee from the Wadden Sea,
was finished in 1932. After the closure, part of the Zuiderzee was reclaimed and the
remaining part became a freshwater lake, the IJsselmeer.
Figure E.2 shows schematically how the closure dam influenced the vertical and ho-
rizontal tide. Before the closure, the basin was relatively long compared to the tidal
wavelength. There was a place in the system where the tidal flow before the closure
was minimal. The closure dam is located just seawards of this place. As the tidal range
in the remaining basin became higher, the tidal prisms at the inlets (Texel and Vlie)
even became slightly larger than before the closure.
Wadden Sea Zuiderzee

(II)
q=0
envelope water level
(I)
B A
Texel Inlet closure dam southern boundary
pre-closure post-closure
Figure E.2: Influence of the closure of the Zuiderzee on the tide (from Elias et al., 2003).

Although the closure has not caused significant changes in the total tidal prisms of the
basins, it has induced morphological changes inside and outside the basins, especially
the two large basins: Texel and Vlie. Figure E.3 shows the sediment budget of the
three inlet systems. Note that the ebb-tidal delta is also included in the area indicated
by ‘coast’. Inside the two large (remaining) basins Texel and Vlie sedimentation has
taken place since the closure. The sedimentation rates are much higher than necessary
for keeping pace with the relative sea level rise. See also the total changes in the period
1927-2000 given in Table E.1. Outside the basins, i.e. in the coastal areas, erosion occurs
and the total amount is more or less the same as the sedimentation inside the basins
(Elias, 2006). Apparently, the closure has caused sediment deficits in the basins, which
has driven sediment import from the coastal area into the basins.
Texel Inlet (T) Vlie Inlet (V)
200 200 coast
ebb-tidal delta
100 100 basin
0 0
-100 -100
-200 -200
1925 1945 1965 1985 2005 1925 1945 1965 1985 2005
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
Eierland Inlet (E)
200 KB0908
KB1110
100 E-D
KB1312 F
A
0 KB1514 V
KB1716 E
-100
KB1918
T
-200 KB2120
KB2322
1925 1945 1965 1985 2005
Figure E.3: Sediment budget of the three inlet systems influenced by the closure of Zuiderzee
(Elias et al., 2012). On the vertical axes, cumulative sedimentation since 1927 in million m3 is
given.
Table E.1: Total changes in sediment budget of the Texel, Eierland and Vlie inlet systems in the
period 1927-2000 (ebb-tidal delta forms part of coast).
Coast [106 m3 ] Ebb-delta [106 m3 ] Basin [106 m3 ]

Texel Inlet (T) −240 −189 198
Eierland Inlet (E) −26 5 −28
Vlie Inlet (V) −175 −161 198
The erosion of the coastal area has caused retreat of the coastline, especially around
the Texel Inlet. Since 1990 the Dutch coastline is maintained by sand nourishment.
The coast around the Texel Inlet requires the largest volume of sand nourishment.

E.2. Closures in the Wadden Sea 541
The morphological development within the basins has also caused movement of the
tidal divide, i.e. the boundary between basins, between the Texel and Vlie basins. The
movement is eastwards, enlarging the Texel basin at the cost of the Vlie basin. In the
closed part of the basin, i.e. IJsselmeer, mud supplied by the IJssel River flowing into
the basin accumulates. Locally this has resulted in a polluted lake bottom.
E.2.2. Closure of Lauwerszee

The Lauwerszee was closed in 1969 by a dam and it is nowadays a freshwater lake,
Lauwersmeer. Before the closure it was a part of the Frisian Inlet basin, which is di-
vided by the Engelsmanplaat into Zoutkamperlaag and Pinkegat (Fig. E.4). The closure
caused a decrease of the basin area of the Frisian Inlet by about one third.
-15m
-5m
-10m
Zo
-5m
u
tka
Pinkegat
mp
erl
aga
Engelsmanplaat
Lauwerszee
North Sea
0 5 10 km
NAP -5 m
study area
Figure E.4: Frisian Inlet and closure of Lauwerszee.
Figure E.5 shows the effects of the closure on the water level, flow velocity and sedi-
ment concentration at the centre of the Zoutkamperlaag Inlet, as calculated by Wang
et al. (1995) using a 2DH model. The closure caused a small increase of the tidal range.
However, due to the decrease of the tidal basin area, the tidal prism and, as a result,
the magnitude of flow velocity decreased significantly. The tidal asymmetry changed
such that it became more flood-dominant favouring sediment import.
Since the closure the tidal basin of Zoutkamperlaag has been accumulating sediment
and the ebb-tidal delta has been eroding (Fig. E.6). The sedimentation in the basin and
the erosion of the ebb-tidal delta are more or less in balance. As a consequence the
closure has not caused erosion problems on the adjacent coasts, in contradiction to
the closure of the Zuiderzee. However, to a lesser extent the closure has also caused
movements of tidal divides, enlarging the tidal basin of Zoutkamperlaag.

after closure
before closure
1.5
1
water level [m] 0.5
0
-0.5
-1
-1.5
360
270
direction [o ]
180
90
0
1.5
magnitude [m/s]
1.25
1
0.75
0.5
0.25
0
150
concentration [mg/l]
125
100
75
50
25
0
800 1000 1200 1400 1600 1800 2000 2200
time [min]
Figure E.5: Influence on tide and sediment transport at the centre of Zoutkamperlaag Inlet.
50
cum. sedimentation [106 m 3 ]
40
30
20
10
0
-10
-20
-30 ebb-tidal delta
-40 basin
-50
1927 1937 1947 1957 1967 1977 1987
year
Figure E.6: Sediment budget of Zoutkamperlaag.

E.3. Closures in the delta area 543
An environmental problem caused by the closure is the erosion of the Engelsmanplaat,

the large intertidal flat between the Pinkegat and the Zoutkamperlaag. This is due to
the fact that the tidal flow in the Zoutkamperlaag channel, which is a building force
for the flat, became weaker due to the closure. The eroding force for the flat, the wave
action, did not change due to the closure.
E.2.3. Discussion
It is interesting to compare the two closures in the Wadden Sea. The closure of the
Zuiderzee took away a major part of the original tidal basin, whereas the closure of
the Lauwerszee removed a relatively smaller part of the basin. However, the closure
of the Lauwerszee caused a significant decrease of the tidal prism, while the closure
of the Zuiderzee did not. The reason is that the Frisian Inlet is a short basin, while the
Zuiderzee was originally a long basin, as well as the special location of the Afsluitdijk.
As for the morphological developments, both closures have in common that they caused
sedimentation in the remaining tidal basins and erosion outside the inlet. The differ-
ence is that the closure of the Zuiderzee caused serious erosion of the coasts adjacent
to the inlet, while the closure of the Lauwerszee did not. This can be explained by the
fact that the equilibrium size of the ebb-tidal delta is related to the tidal prism. As the
tidal prism did not decrease after the closure of the Zuiderzee, the equilibrium size of
the ebb-tidal delta remained the same. This means that there is no sediment surplus
in the ebb-tidal delta. The sediment deficit in the tidal basin can be the most easily sat-
isfied by eroding the coast, as the size of the ebb-tidal delta would effectively increase
when the coastline retreats. In other words, by eroding the coast and the ebb-tidal
delta at the same time, the effective size of the ebb-tidal delta can remain the same.
In the case of the closure of the Lauwerszee, the tidal prism decreased. This caused a
sediment deficit in the basin, and at the same time a sediment surplus in the ebb-tidal
delta area. The sediment deficit in the basin can then simply be satisfied by eroding
the ebb-tidal delta. It is thus important to note that the tidal basin and the ebb-tidal
delta form a sediment-sharing system. The closure of the Lauwerszee did not cause
a sediment deficit in this sediment-sharing system as a whole, whereas the closure of
the Zuiderzee did cause a sediment deficit of this sediment-sharing system.
E.3. Closures in the delta area

E.3.1. Overview of the closures
The Delta Works consist of a series of engineering works as shown in Fig. E.7. We
consider the three closures at the mouths of the three estuaries: Haringvliet, Grevelin-
gen and Eastern Scheldt Oosterschelde in Dutch. An overview of the characteristics of
these three closures is listed in Table E.2. In the following two sections, the effects of

these closures on the development of the area seawards of the closures and the effects
in the enclosed basins are discussed.
Maeslantkering
Hollandse IJssel
Kering
Haringvlietdam Hartelkering
Brouwersdam
Volkerakdam
Grevelingendam
Oosterscheldedam Philipsdam
Veersegatdam Zeelandbrug
Zandkreekdam
Oesterdam
Bathse Spuisluis
Figure E.7: Delta Works. The Eastern Scheldt, Grevelingen and Haringvliet are (semi-)closed off
by the Oosterscheldedam, Brouwersdam and Haringvlietdam, respectively.
Table E.2: Characteristics of the three closures under consideration.
Closure Type Closed basin Position Year

Haringvliet sluices, freshwater more 1969
freshwater discharge reservoir landwards
Grevelingen dam, fully closed salt water lake near mouth 1971
Eastern Scheldt storm surge barrier, estuary/bay near mouth 1986
tidal flow
E.3.2. Developments outside area

The influence of the Delta Works is reflected by changes in delta topography. The most
important factors influencing the morphological geometry in ebb-tidal deltas are: 1)
the relative influence of waves versus tidal flow, 2) the average wave direction and 3)
the interaction between tidal flow offshore and in the nearshore channels (Sha & Van
den Berg, 1993). The sedimentation-erosion pattern is shown in Fig. E.8.
The Haringvliet and the Grevelingen have in common that they are closed for tidal flow
by the Delta Works. The Haringvliet Sluices only allow discharge of freshwater from
the former estuary during periods with high river discharges. The Brouwersdam even
forms a complete closure for the estuary. In both ebb-tidal delta areas, the influence
of the tidal flow has significantly decreased after the closures, whereas the wave influ-
ence has not changed much. The morphological developments in the two areas also

E.3. Closures in the delta area 545
4.4
4.3
y -distance [104 m]
4.2
4.1
4.0
3.9
3.8
1 2 3 4 5 6
x -distance [104 m]
20 18 16 14 12 10 8 6 4 2 0 -2 -4 -6 -8 -10 -12 -14 -16 -18 -20 [m]
Figure E.8: Sedimentation and erosion in the period 1965-2006.
show similar patterns: erosion at the outer edge of the deltas (shoreface), formation of
sand bars that retreat landwards at the north side of the deltas, and sedimentation in
the former tidal channels. These developments can mainly be explained by the disap-
pearance of the tidal flow in the cross-shore direction, and thus a significant decrease
of the tidal volume. A smaller tidal volume means a smaller ebb-tidal delta and smaller
tidal channels for the new equilibrium. This explains the erosion at the shoreface and
the sedimentation in the tidal channels.
A difference in overall sediment balance exists due to the difference in relative position
of the closures. The Haringvliet Sluices are located further landwards, partly due to
the construction of the Maasvlakte (extension of the Port of Rotterdam). This means
that the area outside the sluices with sediment-deficiting channels is relatively large.
As a result, the sedimentation in the channels is dominating with respect to shoreface
erosion, resulting in a positive overall sediment balance. The Brouwersdam is located
relatively close to the sea, making the shoreface erosion dominant with respect to the
sedimentation in the channels, resulting in a negative overall sediment balance. The
morphological development in these two areas in fact tends to establish a continuous
smooth coastline, because of the disconnection with the estuaries.
The influence of the tidal volume on the morphological development can also be ob-
served at the Eastern Scheldt delta. The delta front propagated seaward from 1969 to
1980 due to an 8 % increase in tidal volume by the closure of Volkerak in 1969. After

the construction of the storm surge barrier in 1986, the tidal volume decreased, caus-
ing erosion of the shoreface (Aarninkhof & Van Kessel, 1999). The three remaining
channels (Hammen, de Schaar van Roggenplaat and the Roompot), which form inlets
in the barrier, deepened, and the lee side of the dam accreted. The other channels
decreased in size and found a new orientation due to reduction and rotation (more in
the alongshore direction instead of cross-shore) of the tidal flow.
E.3.3. Impact on the (semi-)closed basins

Regarding the effect within the (semi-)closed basins, only the most important environ-
mental problems are discussed here. The problems appear to be mainly dependent on
the type of closure.
The Haringvliet estuary is closed by sluices which only allow fresh river water to flow
out, but no flow from the sea to the former estuary can occur through the sluices. The
basin thus became a freshwater reservoir. As the cross-sectional area of the former
estuary is too large to only discharge river water, accumulation of fluvial sediment
occurred in the basin after the closure. In the first years after the closure, the quality
of the fluvial sediment was poor. Therefore, polluted sediments are now present in the
bottom of the basin, which creates an environmental problem.
In case of Grevelingen, the former estuary was fully closed by a dam, and became a
saltwater lake. The only exchange of water between the lake and the sea takes place
via a siphon structure. Due to the lack of sufficient refreshment of the water in the
basin, water quality problems have started to occur in the basin in recent years.
The Eastern Scheldt has kept its estuarine characteristics because it is semi-closed by
a storm surge barrier. Under normal conditions the barrier is open, allowing tidal
flow through it. Only during severe storms the barrier is closed protecting the area
behind it from the high sea. However, the construction of the barrier has reduced the
cross-sectional area. In order to limit the decrease of the tidal range in the estuary,
the basin area has been reduced by additional engineering works. The end result is
that only limited change of the tidal range have ocurred after the closure. However,
the tidal prism and the strength of the tidal flow have significantly decreased. As a
consequence, the intertidal flats in the estuary suffer from serious erosion (Fig. E.9).
This is an environmental problem, as the intertidal flats are used by birds for feeding
and resting, when they are dry.
E.4. Conclusions
Five closures or semi-closures of tidal basins in the Dutch coastal system were con-
sidered in this paper. The two closures in the Wadden Sea separate part of the cor-
responding basin from the Wadden Sea. The three closures in the delta area close or
semi-close practically the entire corresponding tidal basin, but they differ in the exact

E.4. Conclusions 547
MSL
cross-section 1986
LWL cross-section 2001
predictions 2015
flat
channel
channel
erosion
sedimentation
predicted change 2001-2015
Figure E.9: Erosion of the intertidal flat in the Eastern Scheldt basin.
location with respect to the coastline and in the type of closures, varying from fully
closed to allowing tidal flow under normal conditions. The effects of these closures
have been analyzed from two management points of view, viz. the maintenance of the
coast and environmental problems in the remaining and/or the closed basin.
From the coastal maintenance point of view, the ebb-tidal delta and the remaining
tidal basin together form a sediment-sharing system. The sediment deficits and/or
surpluses in the two elements together determine if the tidal inlet or estuary will be a
source or sink for the coastal system after the closure. It is concluded that the position
of the closure structure with respect to the coastline is important in this regard. When
only part of the tidal basin is closed, as in the case of the two closures in the Wadden
Sea, the tidal wavelength is also a relevant length scale to consider. The Texel Inlet
became a sink of sediment after the closure of the Zuiderzee, causing serious erosion
in the adjacent coasts. The sediment deficit in the basin of the Zoutkamperlaag and
the sediment surplus in its ebb-tidal delta are almost in balance. When almost the
entire basin is closed, as is the case for the closures in the delta area, the area outside
the closure becomes a sink of sediment when the closure is located further landwards
and vice versa.
The type of environmental problems in the remaining and/or closed basin caused by
the closure mainly depends on the type of closure. When the closed basin becomes a
freshwater lake/reservoir, accumulation of fluvial sediment can cause pollution of the
bottom, as in the case of the closures of Zuiderzee and Haringvliet. If the refreshment
of the water in the closed basin is too limited, as in the case of Grevelingen, water
quality problems can occur in the long term. When the tidal flow is weakened, as in
the case of Zoutkamperlaag and the Eastern Scheldt estuary, serious erosion of the
intertidal flats can take place.

549
Acronyms and abbreviations
2DV 2D-vertical LAT Lowest Astronomical Tide

LOWT Large Oscillating Water Tunnel
AGU American Geophysical Union LW Low Water
ASBPA American Shore and Beach Preser- LWL Low Water Level
vation Association LWS Low Water Slack
ASCE American Society of Civil Engin-
eers MDPI Multidisciplinary Digital Publish-
ing Institute
CD Chart Datum MHW Mean High Water
CEM Coastal Engineering Manual MHWN Mean High Water Neap
CERC Coastal Engineering Research MHWS Mean High Water Spring
Council MLLW Mean Lower Low Water
COPEDEC Conference on Coastal and MLW Mean Low Water
Port Engineering in Developing MLWN Mean Low Water Neap
Countries MLWS Mean Low Water Spring
MSL Mean Sea Level
DGPS Differential Global Positioning Sys-
tem NAP Normaal Amsterdams Peil
NH Northern Hemisphere
GPS Global Positioning System
RSL Relative Sea Level
HW High Water RSLR Relative Sea Level Rise
HWL High Water Level RWS Rijkswaterstaat
HWS High Water Slack
SH Southern Hemisphere
ICCE International Conference on SSL Storm Surge Level
Coastal Engineering SWL Still Water Level
IPCC International Panel on Climate
TU Delft Delft University of Technology
Change
ITCZ Intertropical Convergence Zone USACE US Army Corps of Engineers
JONSWAP Joint North Sea Wave Observa- WMO World Meteorological Organisa-
tion Project tion

551
Bibliography
Aarninkhof, S. G. J. & Van Kessel, T. (1999). Data analyse Voordelta. Grootschalige mor-
fologische veranderingen 1960–1996. (tech. rep. Z2694). Delft Hydraulics.
Ahrens, J. (2000). A fall velocity equation. Journal of Waterway, Port, Coastal and Ocean
Engineering, 126(2), 99–102. https://doi.org/https://doi.org/10.1061/(ASCE)
0733-950X(2000)126:2(99)
Apel, J. R. (1987). Principles of Ocean Physics. Academic Press, London.
Apotsos, A., Raubenheimer, B., Elgar, S., Guza, R. & Smith, J. (2007). Effects of wave
rollers and bottom stress on wave setup. Journal of Geophysical Research, 112(C02003).
Ashton, A. D. & Murray, A. B. (2006). High-angle-wave instability and emergent shoreline
shapes: 1. Modeling of sandwaves, flying spits, and capes. Journal of Geophys-
ical Research, 111(F04011). https://doi.org/10.1029/2005JF000422
Bagnold, R. A. (1956). The flow of cohesionless grains in Fluids. Proceedings of the Royal
Society of London, Series A, 249(964), 235–297.
Bagnold, R. A. (1962). Auto-suspension of transported sediment; turbidity currents.
Proceedings of the Royal Society of London, Series A, 265(1322), 315–319.
Bagnold, R. A. (1963). Mechanics of marine sedimentation. In M. N. Hill (Ed.), The Sea
(pp. 507–528). Interscience.
Bagnold, R. A. (1966). An approach to the sediment transport problem from general phys-
ics. (tech. rep. No. 422-1). US Geological Survey Professional Paper. Washing-
ton, DC, US Government printing office.
Bailard, J. A. (1981). An energetics total load sediment transport model for a plane
sloping beach. Journal of Geophysical Research: Oceans, 96(C11), 10938–10954.
Bailard, J. A. (1982). Modelling on-offshore sediment transport in the surf zone. Pro-
ceedings of the International Conference on Coastal Engineering, 11.
Bailard, J. A. & Inman, D. L. (1981). An energetics bedload model for a plane sloping
beach: local transport. Journal of Geophysical Research: Oceans, 86(C11), 2035–
2043.
Bakker, W. T. (1968). A mathematical theory about sandwaves and its application on
the Dutch Wadden Isle of Vlieland. Shore and Beach, 36(2), 4–14.
Battjes, J. A. (1974). Surf similarity. Proceedings of the 14th International Conference on
Coastal Engineering, 1, 467–479. https://doi.org/10.1061/9780872621138.029
Battjes, J. A. (2006). Short Waves. Lecture Notes for CT4320, Technische Universiteit
Delft.
Battjes, J. A. & Janssen, J. P. F. M. (1978). Energy loss and set-up due to breaking of
random waves. Proceedings of the 16th International Conference on Coastal En-
gineering., 569–587.

552 Bibliography
Bayram, A., Larson, M. & Hanson, H. (2007). A new formula for the total longshore
sediment transport rate. Coastal Engineering, 54(9), 700–710.
Bayram, A., Larson, M., Miller, H. C. & Kraus, N. C. (2001). Cross-shore distribution
of longshore sediment transport: comparison between predictive formulas and
filed measurements. Coastal Engineering, 44(2), 79–99.
Bijker, E. W. (1967). Some considerations about scales for coastal models with movable
bed (Doctoral dissertation) [Also: Publication No.50, Delft Hydraulics]. Delft
University of Technology. Delft, The Netherlands.
Bijlsma, A. C., Bruinsma, R. & Vatvani, D. K. (1989). Investigation of surge-tide interac-
tion in the storm surge model CSM-16 (tech. rep. Z0311). Deltares.
Bird, E. C. F. (1985). Coastline Changes: A Global Review. Wiley Interscience.
Bosman, J. J. (1982). Concentration measurements under oscillatory water motion (Rep.
M-1695 Part II). Delft Hydraulics Laboratory.
Bowen, A. J. (1980). Simple models of nearshore sedimentation; beach profiles and
longshore bars. In S. B. McCann (Ed.), The coastline of Canada. (pp. 1–11).
Bowen, A. J. & Inman, D. L. (1971). Edge waves and crescentic bars. Journal of Geophys-
ical Research, 76(36), 8663–8671.
Boyd, R., Dalrymple, R. W. & Zaitlin, B. A. (1992). Classification of clastic coastal de-
positional environments. Sedimentary Geology, 80(3–4), 139–150.
Breusers, H. N. C. (1983). Lecture notes on sediment transport. International course in
hydraulic engineering.
Bricio, L., Negro, V. & Diez, J. J. (2008). Geometric detached breakwater indicators on
the Spanish northeast coastline. Journal of Coastal Research, 24(5), 1289–1303.
https://doi.org/10.2112/07-0838.1
Bruun, P. (1954). Coast erosion and the development of beach profiles (Technical Memor-
andum No. 44). Beach Erosion Board, US Army Corps of Engineers. Vicksburg,
MS.
Bruun, P. (1962). Sea level rise as a cause of shore erosion. Journal of the Waterways
and Harbors Division, 88(1), 117–130.
Bruun, P. (1978). Stability of Tidal Inlets: Theory and Engineering. Developments in
Geotechnical Engineering, 23, 13–38.
Bruun, P. & Gerritsen, F. (1959). Natural By-Passing of Sand at Coastal Inlets. Journal
of the Waterways and Harbors Division, 85(4), 75–108.
Bruun, P. & Gerritsen, F. (1960). Stability of coastal inlets. Coastal Engineering Proceed-
ings, (7), 23–23.
Caballeria, M., Coco, G., Falqués, A. & Huntley, D. A. (2002). Self-organization mech-
anisms for the formation on nearshore crescentic and transverse sand bars.
Journal of Fluid Mechanics, 465, 379–410. https://doi.org/10.1017/S002211200200112X
Carter, R. W. G. (1988). Coastal environments: An introduction to the physical, ecological,
and cultural systems of coastlines. Academic Press.
Clark, J. A., Farrell, W. E. & Peltier, W. R. (1978). Global changes in postglacial sea level:
A numerical calculation. Quaternary Research, 9(3), 265–287.

Bibliography 553
Coco, G. & Murray, A. B. (2007). Patterns in the sand: From forcing templates to self-
organization. Geomorphology, 91(3-4), 271–290.
Curray, J. R. (1964). Transgressions and regressions. In R. L. Miller (Ed.), Papers in
marine geology: Shepard commemorative volume (pp. 175–203). Macmillan.
Dalrymple, R. W., Zaitlin, B. A. & Boyd, R. (1992). Estuarine facies models; concep-
tual basis and stratigraphic implications. Journal of Sedimentary Research, 62(6),
1130–1146.
Davies, A. G., Van Rijn, L. C., Damgaard, J. S., Van de Graaff, J. & Ribberink, J. S. (2002).
Intercomparison of research and practical sand transport models. Coastal En-
gineering, 46(1), 1–23.
Davies, J. L. & Clayton, K. M. (1980). Geographical variation in coastal development
(2nd ed.). Longman.
Davis Jr., R. A. (1994). The Evolving Coast (Vol. 48). Scientific American Library.
Davis Jr., R. A. & Hayes, M. O. (1984). What is a wave dominated coast? Marine Geology,
60(1), 313–329.
De Blij, H. J. & Muller, P. O. (1993). Physical geography: The global environment. John
Wiley; Sons, New York, USA.
De Vriend, H. J., Bakker, W. T. & Bilse, D. P. (1994). A morphological behaviour model
for the outer delta of mixed-energy tidal inlets. Coastal Engineering, 23, 305–
327.
Dean, R. G. (1973). Heuristic models of sand transport in the surf zone. Proceedings
of the First Australian Conference of Coastal Engineering, 1973: Engineering Dy-
namics of the Coastal Zone, 215–221.
Dean, R. G. (1977). Equilibrium beach profiles: US Atlantic and Gulf coasts (Ocean Engin-
eering Report No. 12). Department of Civil Engineering and College of Marine
Studies, University of Delaware.
Dean, R. G. (1987). Coastal Sediment Processes, Toward Engineering Solutions. Pro-
ceedings of the Specialty Conference on Advances in Understanding of Coastal
Sediment Processes, 1, 1–24.
Dean, R. G. (2002). Beach Nourishment: Theory and Practice. (Vol. 18: Advanced Series
on Ocean Engineering). World Scientific Publishing.
Dean, R. G., Berek, E. P., Gable, C. G. & Seymour, R. J. (1982). Longshore transport
determined by an efficient trap. Proceedings of the 18th International Conference
on Coastal Engineering, 954–968.
Dean, R. G., Chen, R. & Browder, A. E. (1997). Full scale monitoring study of a sub-
merged breakwater, Palm Beach, Florida, USA. Coastal Engineering, 29(3-4),
291–315.
Dean, R. G. & Dalrymple, R. A. (2004). Coastal processes with engineering applications.
Cambridge University Press.
Dibajnia, M. & Watanabe, A. (1993). Sheet flow under nonlinear waves and currents.
Proceedings of the 23rd International Conference on Coastal Engineering, 2015–
2028.

554 Bibliography
Doering, J. C. & Bowen, A. J. (1995). Parametrization of orbital velocity asymmetries

of shoaling and breaking waves using bispectral analysis. Coastal Engineering,
26(1–2), 15–33.
Dohmen-Janssen, C. M., Hassan, W. N. & Ribberink, J. S. (2001). Mobile-bed effects in
oscillatory sheet flow. Journal of Geophysical Research: Oceans, 106(C11), 27103–
27115.
Dronkers, J. (1986). Tidal asymmetry and estuarine morphology. Netherlands Journal
of Sea Research, 20(2-3), 117–131.
Dronkers, J. (1998). Morphodynamics of the Dutch delta. 8th International Biennial Con-
ference on Physics of Estuaries and Coastal Seas, 297–304.
Dronkers, J. (2005). Dynamics of Coastal Systems (Vol. 25). World Scientific.
Duran-Matute, M., Gerkema, T. & Sassi, M. G. (2016). Quantifying the residual volume
transport through a multiple-inlet system in response to wind forcing: The
case of the western Dutch Wadden Sea. Journal of Geophysical Research: Oceans,
121(12), 8888–8903. https://doi.org/https://doi.org/10.1002/2016JC011807
ECMWF. (n.d.). https://www.ecmwf.int/en/forecasts/datasets/reanalysis- datasets/
era5
Egbert, G. D. & Erofeeva, S. Y. (2002). Efficient inverse modeling of barotropic ocean
tides. Journal of Atmospheric and Oceanic Technology, 19(2), 183–204. https://
doi.org/10.1175/1520-0426(2002)019<0183:EIMOBO>2.0.CO;2
Ehlers, J. (1988). The morphodynamics of the Wadden Sea. Balkema.
Einstein, H. A. (1942). Formulas of the transportation of bed load. Transactions of the
American Society of Civil Engineers, 107(1), 561–577.
Einstein, H. A. (1950). The bed-load function for sediment transportation in open chan-
nel flows. (Vol. 1026). US Department of Agriculture Soil Conservation Service,
Washington, DC.
Elias, E. P. L. (2006). Morphodynamics of Texel Inlet. (Doctoral dissertation). Delft Uni-
versity of Technology. IOS Press, The Netherlands.
Elias, E. P. L., Stive, M. J. F., Bonekamp, J. G. & Cleveringa, J. (2003). Tidal inlet dynamics
in response to human interventions. Journal of Coastal Engineering, 45(4), 629–
658.
Elias, E. P. L., Van der Spek, A. J. F., Wang, Z. B. & De Ronde, J. (2012). Morphody-
namic development and sediment budget of the Dutch Wadden Sea over the
last century. Netherlands Journal of Geosciences, 91(3), 293–310.
ENW. (2007). Technisch rapport duinafslag (tech. rep.). Expertisenetwerk waterveiligheid.
Erofeeva, S., Padman, L. & Howard, S. L. (2020). Tide Model Driver (TMD) version 2.5,
Toolbox for Matlab.
Escoffier, F. F. (1940). The stability of tidal inlets. Shore and Beach, 8, 111–114.
Eysink, W. D. (1991). Morphologic Response of Tidal Basins to Changes: The Dutch
Coast. Paper No. 8. 22nd International Conference on Coastal Engineering, 1948–
1961.

Bibliography 555
Eysink, W. D. & Biegel, E. J. (1992). Impact of sea level rise on the morphology of the
Wadden Sea in the scope of its ecological function. ISOS*2 Project, phase 2 (tech.
rep. Report H1300). Delft Hydraulics, Delft.
Farr, T. G. & Kobrick, M. (2000). Shuttle radar topography mission produces a wealth
of data. Eos Transactions AGU, 81(48), 583–585. https://doi.org/https://doi.org/
10.1029/EO081i048p00583
FitzGerald, D. M. & Penland, S. (1987). Backbarrier dynamics of the east Friesian Islands.
Journal of Sedimentary Research, 57(4), 746–754.
Fredsøe, J. & Deigaard, R. (1992). Mechanics of Coastal Sediment Transport. (Vol. 3).
World Scientific, Singapore.
Friedrichs, C. T. & Aubrey, D. G. (1988). Non-linear tidal distortion in shallow well-
mixed estuaries: A synthesis. Estuarine, Coastal and Shelf Science, 27(5), 521–
545.
Galloway, W. E. (1975). Process framework for describing the morphologic and strati-
graphic evolution of deltaic depositional systems. In M. L. Broussard (Ed.), Deltas:
Models for exploration (pp. 87–98). Houston Geological Society.
Gornitz, V. & Lebedeff, S. (1987). Global sea-level changes during the past century. Sea
Level Fluctuation and coastal evolution, SEPM Special Publication No. 41, 2–16.
Grant, W. D. & Madsen, O. S. (1979). Combined wave and current interaction with a
rough bottom. Journal of Geophysical Research: Oceans, 84(C4), 1797–1808.
Gyr, A. & Hoyer, K. (2006). Sediment Transport: A Geophysical Phenomenon. In R.
Moreau (Ed.). Springer.
Hageman, B. P. (1969). Development of the western part of the Netherlands during the
Holocene. Geologie en Mijnbouw, 48, 373–388.
Hallermeier, R. J. (1978). Uses for a calculated limit depth to beach erosion. Proceedings
of the 16th Coastal Engineering Conference, 1493–1512.
Hallermeier, R. J. (1981). A profile zonation for seasonal sand beaches from wave cli-
mate. Coastal Engineering, 4, 253–277.
Hayes, M. O. (1967). Relationship between coastal climate and bottom sediment type
on the inner continental shelf. Marine Geology, 5(2), 111–132.
Hayes, M. O. (1979). Barrier island morphology as a function of tidal and wave regime.
In S. P. Leatherman (Ed.), Barrier Islands (pp. 1–27).
Holman, R. A. & Bowen, A. J. (1982). Bars, bumps, and holes: Models for the generation
of complex beach topography. Journal of Geophysical Research: Oceans, 87(C1),
457–468.
Holman, R. A., Symonds, G., Thornton, E. & Ranasinghe, R. (2006). Rip spacing and
persistence on an embayed beach. Journal of Geophysical Research, 111(C01006).
Holthuijsen, L. H. (2007). Waves in Oceanic and Coastal Waters. Cambridge University
Press.
Hudson, R. Y. (1952). Wave forces on breakwaters. Proceedings of the American Society
of Civil Engineers, 78(1), 1–22.

556 Bibliography
Hughes, S. A. & Chiu, T. Y. (1978). The variations in beach profiles when approximated by
a theoretical curve. (tech. rep. TR/039). Coastal and Oceanographic Engineering
Department, Univ. Florida, Gaines- ville, Florida.
Inman, D. L. & Bagnold, R. A. (1963). The Sea. In M. N. Hill (Ed.). Interscience, New
York, N.Y.
Inman, D. L. & Nordstrom, C. E. (1971). On the Tectonic and Morphologic Classification
of Coasts. Journal of Geology, 79(1), 1–21.
IPCC. (1992). Global climate change and the rising challengeof the sea. Response Strategies
WorkingGroup, CZM Subgroup, Ministry of Transport, Public Works andWater
Management, DG Rijkswaterstaat.
IPCC. (2001). Climate Change 2001: Synthesis Report. A Contribution of Working Groups I,
II, and III to the Third Assessment Report of the Integovernmental Panel on Climate
Change. (R. T. Watson & the Core Writing Team, Eds.). Cambridge University
Press, Cambridge, United Kingdom, and New York, NY, USA.
IPCC. (2007). Climate Change 2007: Synthesis Report. Contribution of Working Groups I, II
and III to the Fourth Assessment Report of the Intergovernmental Panel on Climate
Change (R. K. Pachauri & A. Reisinger, Eds.). IPCC, Geneva, Switzerland.
IPCC. (2014). Climate Change 2014: Synthesis Report. Contribution of Working Groups I,
II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate
Change (R. K. Pachauri & l. A. Meyer, Eds.). IPCC, Geneva, Switzerland.
IPCC. (2021). Climate Change 2021: The Physical Science Basis. Contribution of Working
Group I to the Sixth Assessment Report of the Intergovernmental Panel on Cli-
mate Change (V. Masson-Delmotte, P. Zhai, A. Pirani, S. L. Connors, C. Péan,
S. Berger, N. Caud, Y. Chen, L. Goldfarb, M. I. Gomis, M. Huang, M. Leitzell,
E. Lonnoy, J. B. R. Matthews, T. K. Maycock, T. Waterfield, O. Yelekçi, R. Yu &
B. Zhou, Eds.). Cambridge University Press. In press.
Iribarren, C. R. (1938). Una fórmula para el cálculo de los diques de escollera. M. Bermejillo-
Pasajes, Madrid, Spain.
JARKUS. (n.d.). https://publicwiki.deltares.nl/display/OET/Dataset+documentation+
JarKus
Jarrett, J. T. (1976). Tidal prism-inlet area relationships (General Investigation of Tidal
Inlets Report No. 3). Vicksburg, MS, US Army Engineer Waterways Experiment
Station.
Jonsson, I. G. (1967). Wave boundary layers and friction factors. Proceedings of the 10th
Internation Conference on Coastal Engineering, 2, 127–148.
Kamphuis, J. W. (1991). Alongshore sediment transport rate. Journal of Waterway, Port,
Coastal and Ocean Engineering, 117(6), 624–641.
Kamphuis, J. W. (2000). Introduction to Coastal Engineering and Management (P. L.-F.
Liu, Ed.; Vol. 16). World Scientific.
Katopodi, I. & Ribberink, J. S. (1992). Quasi-3D modelling of suspended sediment trans-
port by currents and waves. Coastal Engineering, 18(1-2), 83–110.
Kinsman, B. (1965). Wind waves: their generation and propagation on the ocean surface.
Courier Corporation.

Bibliography 557
Klein, A. H. F. (2003). Morphodynamics of Headland-bay Beaches: Examples from the

coast of Santa Catarina State, Brazil. (Doctoral dissertation). University of Al-
garve. Algarve, Portugal.
Klein, A. H. F., Da Silva, G. M., Ferreira, Ó. & Dias, J. A. (2005). Beach sediment distri-
bution for a headland bay coast. Journal of Coastal Research, 285–293.
Komar, P. D. (1998). Beach Processes and Sedimentation (Second). Prentice-Hall, Upper
Saddle River, New Jersey.
Komar, P. D. & Inman, D. L. (1970). Longshore sand transport on Beaches. Journal of
Geophysical Research, 75(30), 5914–5927.
Le Méhauté, B. (1976). An Introduction to Hydrodynamics and Water Waves. Springer,
Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-85567-2
LeConte, L. J. (1905). Discussion on the paper, “Notes on the improvement of river and
harbor outlets in the United States” by D. A. Watt, paper no. 1009. Trans. ASCE,
55, 306–308.
Lippmann, T. & Holman, R. A. (1990). The spatial and temporal variability of sand bar
morphology. Journal of Geophysical Research, 95(C7), 11575–11590.
List, J. H., Farris, A. & Sullivan, C. (2005). Reversing storm hotspots on sandy beaches:
Spatial and temporal characteristics. Marine Geology, 226, 261–279.
List, J. H. & Farris, A. S. (1999). Large-scale shoreline response to storms and fair
weather. Coastal Sediments ’99: Proceedings of the 4th International Symposium
on Coastal Engineering and Science of Coastal Sediment Processes., 1324–1338.
Lodder, Q. J., Wang, Z. B., Elias, E. P. L., Van der Spek, A. J. F., De Looff, H. & Townend,
I. H. (2019). Future Response of the Wadden Sea Tidal Basins to Relative Sea-
Level rise–An Aggregated Modelling Approach. Water, 11(10), 2198.
Longuet-Higgins, M. S. (1953). Mass transport in water waves. Philosophical Transac-
tions of the Royal Society of London. Series A, Mathematical and Physical Sciences,
245, 535–581.
Longuet-Higgins, M. S. & Stewart, R. W. (1962). Radiation stress and mass transport
in gravity waves, with application to ‘surf beats’. Journal of Fluid Mechanics,
13(4), 481–504.
Longuet-Higgins, M. S. & Stewart, R. W. (1964). Radiation stresses in water waves: a
physical discussion with applications. Deep-Sea Research, 11, 529–562.
Luijendijk, A., De Vroeg, H., Swinkels, C. & Walstra, D.-J. (2011). Coastal response
on multiple scales: a pilot study on the IJmuiden Port. The Proceedings of the
Coastal Sediments 2011. https://doi.org/DOI:10.1142/9789814355537_0046
Luijendijk, A., Hagenaars, G., Ranasinghe, R., Baart, F., Donchyts, G. & Aarninkhof,
S. G. J. (2018). The State of the World’s Beaches. Scientific Reports, 8(5).
Mangor, K. (2004). Shoreline Management Guidelines (Ed. 3). DHI, Denmark.
Masselink, G. & Hughes, M. G. (2003). Introduction to Coastal Processes and Geomorpho-
logy. Hodder Arnold, London.
Masselink, G. & Short, A. D. (1993). The Effect of Tide Range on Beach Morphodynam-
ics and Morphology: A Conceptual Beach Model. Journal of Coastal Research,
9, 785–800.

558 Bibliography
Miche, R. (1944). Mouvements ondulatoires des mers en profundeur constante on decrois-

ante. Annales des Ponts et Chaussees.
Mil-Homens, J., Ranasinghe, R., van Thiel de Vries, J. S. M. & Stive, M. J. F. (2013). Re-
evaluation and improvement of three commonly used bulk longshore sediment
transport formulas. Coastal Engineering, 75, 29–39. https://doi.org/https://doi.
org/10.1016/j.coastaleng.2013.01.004
Mitrovica, J. X. & Milne, G. (2002). On the origin of late Holocene sea-level high-
stands within equatorial ocean basins. Quaternary Science Reviews, 21, 2179–
2190. https://doi.org/10.1016/S0277-3791(02)00080-X
Mol, A. C. S. (2007). Schematisation of boundary conditions for morphological simula-
tions. Deltares.
Moore, B. D. (1982). Beach profile evolution in response to changes in water level and
wave height (Master’s thesis). University of Delaware, Newark, DE.
Morton, R. A. (1977). Historical Shoreline Changes and Their Causes, Texas Gulf Coast.
Gulf Coast Association of Geological Societies Transactions, 27, 352–364.
Munk, W. H. (1950). Origin and generation of waves. Coastal Engineering Proceedings,
1(1), 1–4. https://doi.org/10.9753/icce.v1.1
Nairn, R. B., Roelvink, J. A. & Southgate, H. N. (1990). Transition Zone Width and Im-
plications for Modelling Surfzone Hydrodynamics. 22nd International Confer-
ence on Coastal Engineering (pp. 68–81). https://doi.org/10.1061/9780872627765.
007
NC OneMap Geospatial Portal. (2020). North Carolina Department of Information Tech-
nology, Government Data Analytics Center, Center for Geographic Informa-
tion and Analysis. https://www.nconemap.gov
Nielsen, P. (1992). Coastal Bottom Boundary Layers and Sediment Transport. World Sci-
entific. https://doi.org/10.1142/1269
Nielsen, P. (2009). Coastal and Estuarine Processes. World Scientific. https://doi.org/10.
1142/7114
O’Brien, M. P. (1931). Estuary and Tidal Prisms Related to Entrance Areas. Civil Engin-
eering, 1(8), 738–739.
O’Brien, M. P. (1969). Equilibrium flow areas of inlets on sandy coasts. Journal of the
Waterways and Harbors Division, ASCE, 95(1), 43–53.
Oertel, G. F. (1988). Hydrodynamics and Sediment Dynamics of Tidal Inlets. Lecture
Notes on Coastal and Estuarine Studies, vol 29. In D. G. Aubrey & L. Weishar
(Eds.). Springer, New York, NY. https://doi.org/10.1007/978-1-4757-4057-8_17
Pelnard-Considère, R. (1956). Essai de theorie de l’èvolution des formes de ravage en
plages de sable et de galets. (tech. rep. No. 3).
Ranasinghe, R. & Stive, M. J. F. (2009). Rising Seas and Retreating Coastlines. Climate
Change, 97, 465–468.
Ranasinghe, R., Symonds, G., Black, K. & Holman, R. A. (2004). Morphodynamics of
intermediate beaches: a video imaging and numerical modelling study. Coastal
Engineering, 51, 629–655.

Bibliography 559
Ranasinghe, R., Symonds, G. & Holman, R. A. (1999). Quantitative characterisation of

rip currents via video imaging. Coastal Sediments ’99, 987–1002.
Ranasinghe, R. & Turner, I. L. (2006). Shoreline response to submerged structures: a
review. Coastal Engineering, 53(1), 65–79.
Renger, E. & Partenscky, H. W. (1974). Stability criteria for tidal basins. Coastal Engin-
eering Proceedings, 1605–1618.
Reniers, A. J. H. M., Roelvink, J. A. & Thornton, E. B. (2004). Morphodynamic mod-
eling of an embayed beach under wave group forcing. Journal of Geophysical
Research, 109(C01030).
Ribberink, J. S. (1998). Bed load transport for steady and unsteady oscillatory flow.
Coastal Engineering, 34, 59–82.
Ribberink, J. S., De Vroeg, J. H. & Van Overeem, J. (1992). Kustverdediging Eierland
(Texel) - hydraulische morfologische effektstudie (fase ii, deel iii) : morfologische
berekeningen. (tech. rep.). Delft Hydraulics.
Rijkswaterstaat. (1949). Getijtafels voor Nederland.
Roelvink, J. A., Reniers, A. J. H. M., Van Dongeren, A. R., Van Thiel de Vries, J., McCall,
R. T. & Lescinski, J. (2009). Modelling storm impacts on beaches, dunes and
barrier islands. Coastal Engineering, 56(11-12), 1133–1152.
Roelvink, J. A. & Stive, M. J. F. (1989). Bar-generating cross-shore flow mechanisms on
a beach. Journal of Geophysical Research, 94, 4785–4800.
Rouse, H. (1937). Modern conceptions of the mechanics of turbulence. Transactions of
the American Society of Civil Engineers, 102, 463–505.
Schoonees, J. S. & Theron, A. K. (1993). Review of the field-data base for the longshore
sediment transport. Coastal Engineering, 19, 1–25.
Schoonees, J. S. & Theron, A. K. (1996). Improvement of the most accurate longshore
transport formula. Proceedings of the 25th International Conference on Coastal
Engineering, 3, 3652–3665.
Sha, L. P. & Van den Berg, J. H. (1993). Variation in ebb-tidal delta geometry along the
coast of the Netherlands and the German Bight. Journal of Coastal Research,
9(3), 730–746.
Shields, A. (1936). Anwendung der Ahnlichkeitsmechanik und der Turbulenzforschung
auf die Geschiebebewegung. (Doctoral dissertation). Mitt. Der Preuss. Versuch-
sanst. Fur Wasserbau und Schiffbau, Berlin, Germany.
Shore Protection Manual. (1984). U.S. Army Corps of Engineers.
Short, A. D. (2005). Sandy Coasts. In M. L. Schwartz (Ed.), Encyclopedia of coastal science
(pp. 821–825). Springer Netherlands. https : / / doi . org / 10 . 1007 / 1 - 4020 - 3880 -
1_267
Sleath, J. F. A. (1978). Measurements of bed load in oscillatory flow. Journal of Water-
way, Port, Coastal and Ocean Engineering, ASCE, 104(3), 291–307.
Smit, M. (2010). Formation and evolution of nearshore sandbar patterns. (Doctoral dis-
sertation). TU Delft.

560 Bibliography
Smith, S. D. & Banke, E. G. (1975). Variation of the sea surface drag coefficient with
wind speed. Quarterly Journal of the Royal Meteorological Society, 101(429), 665–
673. https://doi.org/https://doi.org/10.1002/qj.49710142920
Soulsby, R. L. (1994). Manual of Marine Sands. Thomas Telford, London.
Soulsby, R. L. (1997). Dynamics of marine sands: a manual for practical applications.
Thomas Telford, London.
Soulsby, R. L., Hamm, L., Klopman, G., Myrhaug, D., Simons, R. R. & Thomas, G. P.
(1993). Wave-current interaction within and outside the boundary layer. Coastal
Engineering, 21(1–3), 41–69.
Spanhoff, R. & Van de Graaff, J. (2007). Towards a better understanding and design
of shoreface nourishments. Proceedings of the 30th International Conference on
Coastal Engineering, 4141–4153.
Steetzel, H. J. (1993). Cross-shore transport during storm-surges. (Doctoral dissertation).
Delft University of Technology.
Steijn, R. C., Van Banning, G. K. F. M. & Roelvink, J. A. (1998). Gevoeligheidsberekenin-
gen Eijerland en ZW-Texel – fase 2: ZW Texel. Rapport van het Samenwerkings-
verband Alkyon / WL|Delft Hydraulics. (A266 / Z2430). June 1998. (Sensitivity
computations Eijerland and Southwest Texel - phase 2: Southwest Texel). (tech.
rep.).
Stewart, R. H. (2008). Introduction to Physical Oceanography. http://hdl.handle.net/
1969.1/160216
Stive, M. J. F. (1985). A scale comparison of waves breaking on a beach. Coastal Engin-
eering, 9(2), 151–158.
Stive, M. J. F., De Schipper, M. A., Luijendijk, A. P., Aarninkhof, S. G. J., Van Gelder-
Maas, C., Van Thiel de Vries, J. S. M., De Vries, S., Henriquez, M., Marx, S. &
Ranasinghe, R. (2013). A new alternative to saving our beaches from local sea-
level rise: the sand engine. Journal of Coastal Research, 29(5).
Stive, M. J. F. & De Vriend, H. J. (1995). Modeling shoreface profile evolution. Marine
Geology, 126, 235–248.
Stive, M. J. F. & Reniers, A. J. H. M. (2003). Sandbars in motion. Science, 299(5614), 1855–
1856.
Stive, M. J. F., Van de Kreeke, J., Lam, N. T., Tung, T. T. & Ranasinghe, R. (2009). Em-
pirical relationships between inlet cross-section and tidal prism: A review. Pro-
ceedings of Coastal Dynamics, Tokyo, Japan, 2009.
Svendsen, I. A. (1984). Wave heights and set-up in a surf zone. Coastal Engineering, 8,
303–329. https://doi.org/10.1016/0378-3839(84)90028-0
Swart, D. H. (1974). Offshore sediment transport and equilibrium beach profiles (Doctoral
dissertation). Delft University of Technology.
Symbols and abbreviations used on admiralty charts (np5011). (2020). United Kingdom
Hydrographic Office.
Szmytkiewicz, M., Biegowski, J., Kaczmarek, L. M., Okrój, T., Ostrowski, R., Pruszak, Z.,
Óżyńsky, G. & Skaja, M. (2000). Coastline changes nearby harbour structures:

Bibliography 561
comparative analysis of one-line models versus field data. Coastal Engineering,

40(2), 119–139. https://doi.org/https://doi.org/10.1016/S0378-3839(00)00008-9
Tamisiea, M. E., Mitrovica, J. X., Davis, J. L. & Milne, G. A. (2003). II: Solid Earth Physics:
Long Wavelength Sea Level and Solid Surface Perturbations Driven by Polar Ice
Mass Variations: Fingerprinting Greenland and Antarctic Ice Sheet Flux. Space
Science Reviews, 108, 81–93.
Thorne, C. R., Abt, S. R., Barends, F., Maynord, S. T. & Pilarczyk, K. W. (1995). River,
coastal, and shoreline protection : Erosion control using riprap and armourstone.
Torfs, H. (1995). Erosion of mud/sand mixtures (Doctoral dissertation). Katholieke Uni-
versiteit Leuven.
Turner, I. L., Whyte, D., Ruessink, B. & Ranasinghe, R. (2007). Observations of rip spa-
cing, persistence and mobility at a long, straight coastline. Marine Geology, 236,
209–221.
Valentin, H. (1952). Die Küsten der Erde: Beiträge zur allgemeinen und regionalen Küst-
enmorphologie (Vol. 246). J. P. Gotha.
Van de Graaff, J. (1977). Dune erosion during a storm surge. Coastal Engineering, 1, 99–
134.
Van de Kreeke, J. (1992). Stability of tidal inlets; Escoffier’s analysis. Shore and Beach,
60(1).
Van de Kreeke, J. (2004). Equilibrium and cross-sectional stability of tidal inlets: applic-
ation to the Frisian Inlet before and after basin reduction. Coastal Engineering,
51, 337–350.
Van de Kreeke, J. & Robaczewska, K. (1993). Tide-induced residual transport of coarse
sediment: Application to the Ems estuary. Netherlands Journal of Sea Research,
31(3), 209–220.
Van de Meene, J. W. H. & Van Rijn, L. C. (2000). The shoreface-connected ridges along
the central Dutch coast - Part 2: Morphological modelling. Continental Shelf
Research, 20, 2325–2345. https://doi.org/10.1016/S0278-4343(00)00049-2
Van der Graaff, J., Niemeyer, H. D. & Van Overeem, J. (Eds.). (1991). Coastal Engineering
[Artificial Beach Nourishments], 16(1).
Van der Meer, J. W. (1988). Stability of Cubes, Tetrapods and Accropode. Proceedings
of the Breakwaters Conference ‘88, Thomas Telford, London, United Kingdom.
Van der Salm, G. L. S. (2013). Coastline modelling with UNIBEST:Areas close to structures
(Master’s thesis). Delft University of Technology.
Van der Spek, B.-J., Bijl, E., Van de Sande, B., Poortman, S., Heijboer, D. & Bliek, B. W.
(2020). Sandbar Breakwater: An Innovative Nature-Based Port Solution. Water,
12(5). https://doi.org/10.3390/w12051446
Van Ledden, M. & Wang, Z. B. (2001). Sand-mud morphodynamics in a former estuary.
IAHR Proceedings of the River, Coastal and Estuarine Morphodynamics Confer-
ence, Obihiro, Japan, 505–514.
Van Rijn, L. C. (1984a). Sediment Transport, Part I: Bed Load Transport. Journal of
Hydraulic Engineering, 110(10).

562 Bibliography
Van Rijn, L. C. (1984b). Sediment Transport, Part II: Suspended Load Transport. Journal
of Hydraulic Engineering, 110(11).
Van Rijn, L. C. (1989). Handbook of sediment transport by currents and waves. Delft
Hydraulics, Delft, The Netherlands.
Van Rijn, L. C. (1993). Principles of sediment transport in rivers, estuaries and coastal
seas. Aqua Publications, Amsterdam, The Netherlands.
Van Rijn, L. C. (1999). Principles of coastal morphology. Aqua Publications, Amsterdam,
the Netherlands.
Van Rijn, L. C. (2000). General view on sand transport by currents and waves. (tech. rep.)
[Report Z2899.30]. Delft Hydraulics, Delft, The Netherlands.
Van Rijn, L. C. (2010). Coastal erosion control based on the concept of sediment cells.
(tech. rep.). EU Conscience project: Concepts and Science for Coastal Erosion
Management.
Van Thiel de Vries, J. S. M. (2009). Dune Erosion During Storm Surges (Doctoral disser-
tation). Delft University of Technology.
Van Veen, J. (1950). Eb-en vloedschaarsystemen in de Nederlandse getijwateren. Tijd-
schrift koninklijk Nederlands Aardrijkskundig Genootschap, 67(2), 303–352.
Van Veen, J., Van der Spek, A. J. F., Stive, M. J. F. & Zitman, T. (2002). Ebb and flood
channel systems in the Netherlands tidal waters. VSSD, Delft, The Netherlands.
Van Veen, J., Van der Spek, A. J. F., Stive, M. J. F. & Zitman, T. (2005). Ebb and flood
channel systems in the Netherlands tidal waters. Journal of Coastal Research,
21(6).
Vanoni, V. A. (1975). Sedimentation Engineering. ASCE, New York.
Vellinga, P. (1986). Beach and dune erosion during storm surges. (Doctoral dissertation).
Vellinga, P. & Leatherman, S. P. (1989). Sea level rise, consequences and policies. Cli-
mate Change, 15, 175–189.
Visser, P. (2014). Short Term Morphological Impact of the Eierlandsedam. (Master’s thesis).
Vollmers, H. J. (1989). The state of art of physical modelling of sediment transport. Pro-
ceedings of the International Symposium on Sediment Transport Modeling, ASCE,
New York, 7–12.
Walton, T. L. & Adams, W. D. (1976). Capacity of inlet outer bars to store sand. Proceed-
ings of the 15th International Conference on Coastal Engineering, ASCE, 1919–
1937. https://doi.org/10.1061/9780872620834.112
Wang, Z. B., De Ronde, J. G., Van der Spek, A. J. F. & Elias, E. P. L. Responses of the
Dutch coastal system to the (semi-) closures of tidal basins. In: International
Conference on Estuaries and Coasts, Sendai, Japan. 2009.
Wang, Z. B., Louters, T. & de Vriend, H. J. (1995). Morphodynamic modelling of a tidal
inlet in the Wadden Sea. Journal of Marine Geology, 126, 289–300.
Wang, Z. B. & Ribberink, J. S. (1986). The validity of a depth-integrated model for sus-
pended sediment transport. Journal of Hydraulic Research, 24(1), 53–67. https:
//doi.org/10.1080/00221688609499332

Bibliography 563
Wegener, A. (1912). Die Entstehung der Kontinente. Geologische Rundschau, 3, 276–

292.
Wegener, A. (1929). Die Entstehung der Kontinente und Ozeane. 4th edition (1st edition
1915). [(rev. eds. 1920,1922,1929; in German, from 1922 also in English)]. Vieweg
& Sohn, Braunschweig.
Whitehouse, R. J. S. (1998). Scour at marine structures: a manual for practical applica-
tions. Thomas Telford, London.
World Coast Conference. Preparing to meet the coastal challenges of the 21st century.
In: (Noordwijk, The Netherlands). 1993.
Wright, L. D. & Short, A. D. (1984). Morphodynamic variability of surf zones and
beaches: A synthesis. Marine Geology, 56, 93–118. https://doi.org/10.1016/0025-
3227(84)90008-2
Zhang, K., Douglas, B. C. & Leatherman, S. P. (2004). Global Warming and Coastal
Erosion. Climate Change, 64(41).

565
Subject index
wave number, 92 bed armouring, 282

bed form, 292
accommodate, see management
bed friction, see bed shear stress
strategy
bed load transport, 284
accretion, 10
waves and currents, 291
accretion length, 387
bed shear stress, see also damping, 172,
active zone, 323, 330
195, 196, 222, 246, 279
actual transport, 362
aeolian transport, 318 Bijker coefficient, 296
Airy wave theory, 111, 191, 194, 206 borrow pit, 502
amphidromic point, 137, 144 bottom friction, see bed shear stress
amphidromic system, 137 bound long wave, 270, 351
angle of repose, 275 boundary layer, 194, 246, 306, 356, 429
angular frequency, 92 Boussinesq model, 191
asymmetry, see wave asymmetry or Bowen, 311, 352
tidal asymmetry braided channel, 422
branched channel, 422
back-barrier system, 435
breaker bar, see bar
backshore, 321
breaker index, 181
Bagnold, 290, 310, 355, 369, 451
breaker point, 181, 216
bar, 168, 228, 333, 340, 365
breaker types, 183
channel margin-, 420
breaker zone, see surf zone
storm-, 162
breaking, 171, 210, 235
swash-, 420
bar cycles, 341 depth-induced, 181
bar system, 235 steepnes-induced, see
basalt, 66 white-capping
basin, 248 breakwater, 2, 386, 470
bay, 411 emerged detached-, 179, 487
bay barrier, 405 submerged detached-, 399, 489,
Bayram, 373 510
beach, 333 Bruun, 59, 328, 347
beach erosion, 203 bulk density, 275
beach slope, 372 bulk longshore transport formula, 365
beach state, 334 bulk transport, 362
beach widening, 509 buoy, 99
Beaufort scale, 110 bypassing, 389, 395, 469, 481

566 Subject index
canyon, 379 continental shelf, see shelf

capillary wave, 93 continuity equation, see mass balance
carbonate sediment, 50, 66 convergence
catchment, 419 plate-, 39
CERC formula, 367 wave energy-, 254, 392
Chézy coefficient, 197 conveyor belt, see great conveyor belt
channel margin linear bars, 420 coral reef, 71
channel volume, 447 Coriolis, 95, 117, 153
channeling, 425 Coriolis acceleration, 97, 141
Chart Datum, 146 Coriolis parameter, 97
chenier plain, 63 Cornu spiral, 180
chutes, 445 crest angle, 180
circulation, 444 critical velocity, 281
circulation pattern, 390, 430, 434 cross-shore profile, 5
classical wave equation, 136 cross-shore sediment transport, 391
clay, 68 Curray’s diagram, 61
cliff, 49 curvature-induced flow, 267
closure curve, see Escoffier cusp, 334, 336
closure depth, 19, 330, 380 cyclone, see hurricane
cnoidal wave theory, 191
co-phase line, 137 daily inequality, 127, 165
co-range line, 137 damping, 46, 253
co-tidal line, 137 preferential-, 257
coarse sediment, 451 Dean, 372
coast, 19 Dean’s approach, 329
Coastal Engineering, 1 decomposition
coastal plain, 19, 409 transport rate-, 350
coastal profile, 321 deep water, 94, 113
coastal protection, 468 deflocculation, 315
coastal waters, 19 degenerate amphidromic point, 137
coastal zone, 19 delta formation, 64
coastlands, 19 deltaic coastline, 80, 405
coastline, 1 density
coastline disturbance, 392, 403 bulk-, 275
collapsing breaker, 183 dry bulk-, 275
collision coast, see leading-edge coast grain-, 274
continental-, 43 relative-, 275
island arc-, 45 saturated bulk-, 275
compensation current, 236 deposited volume, 11
computer model, 379 design conditions, 344, 498, 508
continental drift, 36 differential pull, 124
continental levering, 55 differential warming, 151
continental sediments, 66 diffraction, 170, 178

Subject index 567
diffraction zone, 178 ebb spit, 421

diffusion, 301 ebb-tidal sediment transport, 442
dimensionless fall velocity, 335 eddy, 232
dispersion tidal-, 419
direction-, 116 eddy viscosity, 226, 231, 307
frequency-, 113, 116 sediment concentration dependent,
of momentum, see turbulence 306
wave-, 111 edge waves, 338
dispersion relation, 113 efficiency factor, 292
dissipation, 171, 221 Einstein, 291
dissipation model, 222, 225, 329 emergence, 62
dissipative beaches, 335 end moraines, 50
distributary channel, 85 energetics approach, 290, 310, 369
diurnal tide, 129, 131, 164 energy balance, 116, 170
divergence energy conservation, see energy
plate-, 39 balance
wave energy-, 254, 392 energy dissipation rate, 311
doldrums, 155 energy penetration, 179
downdrift coast, 386 eolianite, 50
downward crossing, 102 episodic event, 336
drag coefficient, 277 equilibrium, 24
drag force, 280 equilibrium concentration, 314, 457
dredging, 502 equilibrium concept, 25
dry bulk density, 275 equilibrium flow curve, see stability
dune erosion, see erosion curve
dune reinforcement, 508 equilibrium profile
dune retreat, 347 suspended load, 356
dune vegetation, 78 equilibrium theory of tides, 121
DUROS-plus, 347 equinoctal tides, 130
Dutch coast, 15, 117, 147, 237, 238, 243, equinoxes, 129
251, 270, 325, 330, 333, 341, erosion, 10, 21
381, 413, 498 dune-, 344, 474
dynamic coupling, 462 permanent-, 471
dynamic equilibrium, 324, 327 structural-, 471, 504
dynamic pressure, 193 Escoffier, 438
dynamic viscosity, 277 estuarine circulation, 412
estuary, 82, 411
earth’s rotation vector, 96 estuary entrance, 415
ebb, 238 estuary mouth, 415
ebb channel, 420, 445 eustasy
ebb current, 238 geoidal-, 54
ebb dominance, 243, 259, 261, 450 glacio-, 54
ebb shield, 420 eustatic change, 54

568 Subject index
extreme values, 119 gradient-type transport, 307

grading, 274
fall velocity, 275, 315, 330
grain density, 274, 275
dimensionless-, 335
grain size, 330
feldspar, 66
granite, 66
fetch, 110
gravitational acceleration, 123
fine sediment, 456
gravitational pull, 122
fjord, 50, 60
gravity force, 280
flats area, 448
great circle, 96, 117
flocculation, 69, 315
great conveyor belt, 157
flood, 238
greenhouse effect, 150
flood channel, 420, 445
gross transport, 377
flood current, 238
group speed, 114, 116
flood dominance, 243, 260, 261, 449
groyne, 2, 438, 470, 483, 492
flood ramp, 420
gyres, 426
flow separation, 280
forced behaviour, 338 Hallermeier, see closure depth
forcing templates, 337 hard measures, 2
form factor (tide), 164 hard methods, 470
free behaviour, 338 harmonic analysis, 145
free stream velocity, 195 havoc, 269
frequency, see wave frequency headland, 233
friction heat imbalance, 150
bottom-, see bed shear stress higher harmonics, 262
linear-, 240, 249 hindered settling, 278
quadratic-, 250 Holocene, 39
friction factor, 197 horizontal diffusivity, see eddy
current-only, 297 viscosity
currents and waves, 295 horizontal orbital velocity, see
Jonsson, 197, 198 horizontal particle velocity
Soulsby, 198 horizontal particle excursion, 193
waves-only, 297 horizontal particle velocity, 200
friction-dominated flow, 250 horizontal tide, see tide
funnel, 15 hurricane, 155, 163
hydrostatic pressure, 193
Galloway, 83
hypsometric curve, 449, 464
Genesis, 486
geographical variation, 69 ice age, 39
geological timescale, 36 immersed mass, 284
geology, 36 immersed weight, 284
glacial, 39 inertia, 265
glacier, 39, 50 inertia dominated flow, 251
global warming, 150 inertial frame of reference, 95
gorge, 410, 415 infra-gravity wave, 93

Subject index 569
initiation of motion, 279, 290 wave-induced, 247, 363

inlet, 5, 12, 95, 415, 438, 475 longshore current profile, 227
Inman and Bagnold, 369 Longshore Drift Rate, 395
interference, 114 longshore transport, 10, 364, 392, 432,
interglacial, 39 476
intertidal area, 95 wave-induced, 362
intertidal flat, 422 longshore transport gradients, 392
intertidal zone, 166, 333 Longuet-Higgins streaming, see
Intertropical Convergence Zone, 153 streaming
IPCC, 65 lower shoreface, 321, 350
Iribarren parameter, 334, 370
irregular wave, 228 management strategy, 468
irregular waves, 100 mangrove, 77, 167
Irribarren parameter, 182 marginal flood channel, 420
isostasy marginal sea coast, 42, 45
glacio-, 55 marine feeding, 39, 413
hydro-, 55 Marram, 79
mass balance, 21
jetty, 2, 386, 478 cross-shore, 213
JONSWAP spectrum, 106, 111 tidal basin, 249
mass flux, 201, 230, 265
Kamphuis, 372
material derivative, 314
Kelvin wave, 142, 240
mean current, 351
kinematic viscosity, 277
Mean Sea Level, 7, 92, 145
kwelders, see salt marsh
median particle diameter, 274
lagoon, 410 memory effect, 313, 457
landfall, 163 Miche breaking criterion, 180
Large Oscillating Water Tunnel, 288 mixing length, 226
Le Méhauté, 191 momentum, 200, 203
leading-edge coast, 42, 43 momentum balance
lift force, 280 cross-shore, 213
linear wave theory, see Airy wave geostrophic-, 143
theory tidal basin, 249
lithification, 50 momentum flux, 204
lithosphere, 39 monsoon, 155
littoral drift, see longshore transport Moore’s grain size, 330
littoral transport, see longshore morphodynamic regime, 334
transport morphodynamics, 21
littoral zone, see also surf zone, 225 morphological timescale, 25
log-normal distribution, 119 morphology, 2
long waves, see shallow water mud, 68, 315
longshore current, 212, 225, 233, 295 muddy, 73
tidal, 247 multiple line theory, 380, 390

570 Subject index
n, 115 progradation, 61
NAP, 7 prograding coast, 81
neap tide, 127 progressive wave, 417
negative feedback, 21 propagation velocity, see phase speed
neo-trailing-edge coast, 43, 49 protect, see management strategy
net transport, 377 pumping mode, 255
Nikuradse roughness, 197
quadratic friction law, 197, 222
no-slip condition, 195
quartz, 66, 274
non-dispersive, 113
quasi-steady approach, 289
non-linear effects, 184
North Sea, 144 radiation stress, 203, 205, 230
nourishment, 330, 469, 478, 500 advective part, 208
nourishment lifetime, 506 pressure part, 208
random waves, see irregular waves
ocean conveyor belt, see great
Rayleigh distribution, 102, 107, 108
conveyor belt
reference concentration, 289
ocean waves, 92
reflective beach, 334
classification, 93
refraction, 170, 175, 338, 392
oceanic ridge, see ridge
current-, 178, 429
one-line theory, 380
depth-, 178, 233
orbital path, 191
refraction coefficent, 177
outbreaking, 390
regression, 60
overshoot, 445
relative density, 275
overtide, 262, 452
relative tidal range, 167
particle excursion, see horizontal relaxation timescale, 457
particle excursion reset event, 336
passive margins, see trailing-edge coast residual current, 265, 266, 426, 450
perched beach, 510 bathymetry-induced-, 266
phase speed, 93, 113, 114, 252 Coriolis-induced-, 267
phase-locked, 185 residual transport, 450
pier, 490 resonance, 47, 165, 253, 417
Pierson-Moskowitz spectrum, 111 retreat, see management strategy
platen, see intertidal flat return current, 202
Pleistocene, 39 revetment, 2, 470, 497
plunging breaker, 163, 183 reworked sediments, 66, 413
pole, 99 Reynolds averaging, 305
pororoca, 259 Reynolds number
porosity, 275, 283 grain-, 277, 281
positive feedback, 22 ria, 60
potential flow, 265 ridge, 39, 333
preferential damping, 257 rip channel, 336
pressure gauge, 99 rip currents, 233
pressure gradient, 230, 268 ripple factor, 292

Subject index 571
river bend, 267 wind-, 236

river mouth, 404 set-up difference, 389, 398
roller, 184 settling velocity, see fall velocity
roller balance equation, 184 𝑆, 𝜑-curve, 373
roller energy, 201 shadow zone, see also diffraction zone,
roller momentum, 228 232, 388
Rossby number, 97 shallow water, 94, 113
rotational, 221 shallow-water equations, 135
Rouse number, 309 shallow-water approximation, 216
runnel, 333 shear stress
bed-, see bed shear stress, 449
salient, 396, 488 grains, 279
salt marsh, 75, 167, 422 time-averaged, 296
sand, 67 waves and currents, 292
sawtooth, 187 shear velocity, 279
sawtooth asymmetry, 259, 260, 456, sheet flow, 279, 285
458 shelf, 46, 321
scale, 23 sheltering, 170, 385, 388
scale analysis, 249 Shields parameter, 281
schorren, see salt marsh shoal, 22, 234, 426, 482
scour, 493, 498 bypass-, 396
sea, 93, 116 shoaling, 170, 172, 211, 254, 351
sea dike, 498 shoaling factor, 174
sea level rise, 36, 65, 464 shoreface, 321
relative-, 348 short basin, 255
seasonality, 8 short waves, see deep water
seawall, 2, 470, 493 silt, 68, 457, 502
secondary current pattern, 398 simple dissipation model, see
secondary flow, 265, 267 dissipation model
curvature-induced, 426 single line theory, see one-line theory
secondary waves, 395 skewed waves, 185
sediment availability, 61, 70 skewness, 184, 186, 258, 261, 351
sediment concentration, 275 skin-friction factor, 295
sediment continuity, 304 slack, 238, 244, 260
sediment supply, 48, 348 slikken, see intertidal flat
seiche, 269 Slufter, 317
self-organisation, 336 Snell’s law, 175, 221
semi-diurnal tide, 131, 164 soft measures, 2, 469, 500
set-down solitary wave theory, 181, 191
wave-, 204, 211, 212, 216 spectral analysis, 101, 104
wind-, 236 spilling breaker, 163
set-up, 230, 232, 428 spilling breakers, 183
wave-, 204, 211, 212 spit, 402, 476, 482

572 Subject index
spring tide, 127 tidal amplitude, 135

stability, 462 tidal analysis, 144
stability curve, 440 tidal asymmetry, 243, 250, 256, 261,
standing wave, 252, 417 449
stationary, 100 tidal bore, 257
steric changes, 54 tidal bulge, 125
stirring, 351, 364 tidal character, 164
Stokes drift, 201, 203, 265, 450, 461 tidal components, see tidal constituents
Stokes series expansion, 190 tidal constituents, 130, 131, 263
Stokes wave, 185, 187 tidal current, 95, 245
Stokes wave theory, 180, 185 cross-shore distribution, 245
storm profile, 8 tidal curve, 164
storm surge, 15, 46, 98, 167, 237, 344, tidal delta, 419
474 ebb-, 435
strand plain, 63 tidal flat, 166
streaming, 196, 202, 295, 357 tidal force, 125
structure, 2, 246 tidal inlet, see inlet
submerged-, see breakwater tidal jet, 265
submerged breakwater, see breakwater tidal levels, 146
submergence, 62 tidal prism, 255, 416, 432, 435, 447, 462
subtidal zone, 166 tidal propagation
summer profile, 340, 352 in basin, 257
supratidal flat, see salt marsh tidal velocity, see tidal current or tide
supratidal zone, 167 tide, 94, 175
surf beat, 271 horizontal-, 238, 244
surf zone, 2, 162, 217, 225, 271, 333, 363 vertical-, 238
surface elevation, 92 tide gauge, 99
surface roller, see roller tide generation, 121
surge, 15 tide-dominated environment, 166, 409
surging breakers, 183 time series analysis, 102
suspended load transport, 285, 289, tombolo, 396, 488
314, 355 total load formula, 290
suspended sediment transport, 301 trade winds, 117, 154, 155
current-related, 302 trailing-edge coast, 42, 43, 48
wave-related, 302 Afro-, 48
swash bar, 420 Amero-, 48
swash platform, 420 neo-, 49
swell, 93, 116, 163 transgression, 60
transgressive coast, 81
Taylor expansion, 351 transport capacity, 362
terminal lobe, 420 transport pattern, 430
thermohaline ocean circulation, 157 trestle, 490
throat, see gorge tropical easterlies, see trade winds

Subject index 573
tsunami, 98, 114, 175 wave energy, 102

tunneling, 175 wave force, 212, 222
turbidity current, 64 wave frequency, 91, 92, 111
turbulence, 195, 223, 226, 231 wave generation, 110
turbulent diffusivity, 306, 307 wave group, 114, 270
turbulent force, 227 wave height, 92
typhoon, see hurricane maximum-, 108
root-mean-square-, 103, 109
undertow, 202, 230, 295, 352 significant-, 103, 108
undulation, 233 wave measurements, 98
Unibest, 384, 486 wave momentum, 200
uniform energy dissipation, 328 wave pattern, 424
updrift coast, 386 wave period, 92
upper shoreface, 321, 322, 350 mean-, 103
upward crossing, see downward peak spectral-, 106, 109
crossing significant-, 103, 109
Ursell parameter, 189 zero-crossing-, 103, 109
wave ray, 172, 176
variance, 102
wave statistics
variance density spectrum, 105
long-term, 119
Vellinga erosion profile, 346
short-term, 100
velocity distribution, 231
wave steepness, 92, 180, 372
velocity moment, 351
wave tunnel, 285
vertical tide, see tide
wave zone, see diffraction zone
viscosity, 195, 277
wave-by-wave analysis, see time series
vorticity, 194, 221
analysis
wadden, see intertidal flat wave-dominated environment, 162,
wantide, see watershed 409
wash load, 284 wave-induced current, 428
water level gradient wave-induced pressure, 193, 204
alongshore, 389 wavelength, 92
watershed, 48, 418 wavenumber, 111
wave action, 170 Weibull distribution, 108, 119, 121
wave asymmetry, 184, 187 well-graded, 274
wave celerity, see phase speed well-sorted, 274
wave climate, 119, 160, 376 westerlies, 154
east coast swell, 162 white-capping, 171, 180
protected, 162 wind set-up, see set-up
storm-, 160 wind shear stress, 236
west coast swell, 161 wind waves, 93
wave condition, 376 wind-induced current, 428
wave crest, 180 winter profile, 8, 340, 352
wave direction, 109 zonal wind systems, 157

575
Credits
Lecture notes
In the 1960s, the topic of coastal engineering at the faculty of Civil Engineering of TU
Delft was not really taught as a separate graduate course. It was rather an integral
part of the hydraulic engineering chairs and courses of Professors Van Bendegom and
Jansen with more focus on rivers than coasts. In 1972 Professor Bijker was appointed
the first chair focusing on coastal engineering alone. He asked Associate Professor
Massie to compose the first coastal engineering lecture notes, which still serve as in-
spiration for our current notes. In 1988 Professor D’Angremond succeeded Professor
Bijker, who invited both MSc graduate Van de Velden and Associate Professor Van de
Graaff to develop two new courses and lecture notes, viz. Introduction to Coastal En-
gineering (CT4300) and Coastal Morphology and Protection (CT5309) respectively. In
1994, the topic of Coastal Inlets and Tidal Basins was added as a separate third course
(CT5303) developed by Professor Stive, newly appointed in a part-time position. With
the dedicated assistance of MScs Elias and Hibma, he composed lecture notes by in-
tegrating existing material of Professors De Vriend, Dronkersand Wang and Drs Eij-
sink, Van der Spek and Van Dongeren. Per the academic year 2009-2010, the courses
CT4300, CT5309 and CT5303 were replaced with two larger courses Coastal Dynamics
I (CIE4305) and Coastal Dynamics II (CIE4309), for which the lecture notes were de-
veloped that formed the basis of this book. The development of these two new courses
was driven by our wish to update and streamline our coastal engineering curriculum
and make room for new developments in coastal and morphodynamic process know-
ledge and modelling. We wish to acknowledge the efforts of all our colleagues who
developed these earlier courses.
Photos
Figure Source
Fig. 2.5 c b ea Matthijs Buijs
Fig. 2.8 c be a Marcel Stive
Fig. 2.13 c be a Alejandra Gijón Mancheño
Fig. 2.15 c be a Blenda Gomes Rocha
Fig. 2.16 c be a Ascha Simons

576 Credits
Fig. 2.31 c b ea Erik Mosselman

Fig. 3.9 beeldbank.rws.nl, Rijkswaterstaat
Fig. 3.12 c be a Stefanie Ross
Fig. 5.8 Roberto Lo Savio / Shutterstock.com
Fig. 5.69 Richard P Long / Shutterstock.com
Fig. 7.9 p Jurriaan Brobbel / Rijkswaterstaat
Fig. 8.20 Photo by Carolina Piccoli / map by OpenStreetMap contributors
Fig. 8.21 Leonid Andronov / Shutterstock.com
Fig. 8.27 p U.S. Geological Survey, Department of the Interior / USGS
Fig. 9.6 p NC OneMap Geospatial Portal (2020)
Fig. 9.14a beeldbank.rws.nl, Rijkswaterstaat
Fig. 9.14b Rens Jacobs / beeldbank.rws.nl, Rijkswaterstaat
Fig. 9.14c Rens Jacobs / beeldbank.rws.nl, Rijkswaterstaat
Fig. 9.14d Joop van Houdt / beeldbank.rws.nl, Rijkswaterstaat
Fig. 9.19 Ehlers (1988)
Fig. 10.8 c b Institut Cartogràfic i Geològic de Catalunya (ICGC)
Fig. 10.9 Joop van Houdt / beeldbank.rws.nl, Rijkswaterstaat
Fig. 10.10 Courtesy of CDR International B.V., from Van der Spek et al. (2020)
Fig. 10.13 Rens Jacobs / beeldbank.rws.nl, Rijkswaterstaat
Fig. 10.14 c be a Ad Reniers
Fig. 10.17 UK Environmental Agency under an Open Government Licence
Fig. 10.18 p U.S. Geological Survey, Department of the Interior / USGS
Fig. 10.22a Wayland Smith / Shutterstock.com
Fig. 10.22b Carol Blaker / Shutterstock.com
Fig. 10.26a Evelyn Simak / c ba Wikipedia Commons
Fig. 10.26b Evelyn Simak / c ba Wikipedia Commons
Fig. 10.30 c ba Ceinturion Wikipedia Commons
Fig. 10.40 Ni_Giri / Shutterstock.com
Fig. 10.41 c b OrtoPNOA 2019 scne.es
Fig. 10.42a Joop van Houdt / beeldbank.rws.nl, Rijkswaterstaat

Credits 577
Fig. 10.42b Joop van Houdt / Rijkswaterstaat

Fig. 10.42c Joop van Houdt / Rijkswaterstaat

579
Errata and improvements
Version history
This section provides a record of changes made to this textbook since its initial public-
ation as an open textbook (January 2021). If the edits are minor, the version number
increases by 0.1. If the changes involve substantial updates, the edition number in-
creases to the next whole number. The e-book always reflects the most recent version
and is available for online use and free download from the TU Delft Open Textbook
repository at textbooks.open.tudelft.nl.
Version Date Change

1.0 January, 2021 Original
1.1 January, 2022 Corrected misspellings, typographical errors and punc-
tuation errors in text
Added missing symbols in text
Corrected misspellings and other minor errors in figures
Corrected typographical errors in equations
Added some brief clarifications
Updated or corrected a few references
Added ‘Errata and improvements’ on page 579
1.2 January, 2023 Corrected misspellings, typographical errors and punc-
tuation errors in text
Updated Figs. 1.14, 2.18, 2.19 and 3.8
Corrected typographical errors in equations
Added some brief clarifications
List of changes
When the version number increases by 0.1, students (and other readers) can continue
to use their softback of the previous version, as long as they copy the most important
changes in their hardcopy. To this end, an e-book with markup, serving as an overview
of errata and improvements, is available via the TU Delft Open Textbook repository
by selecting Errata and improvements. These corrections are considered important for
understanding the material. Other, less crucial changes are not tracked.

580 Errata and improvements
Errata and improvements that are suggested, but not yet part of a new version, are
recorded on online errata and improvements lists, respectively, to which we invite our
readers to contribute. In order to suggest a correction or improvement, please fill out
the suggestion form. These lists as well as the suggestion form are also accessible by
selecting Errata and improvements on the TU Delft Open Textbook repository.

Coastal Dynamics

Uploaded by

Copyright:

Available Formats

Coastal Dynamics

Uploaded by

Document Information

Original Title

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

Coastal Dynamics

Uploaded by

Copyright:

Available Formats

Coastal Dynamics

Version 1.2 (2023)

Judith Bosboom and Marcel J.F. Stive

Judith Bosboom and Marcel J.F. Stive

This textbook on Coastal Dynamics focuses on the

Dr. Judith Bosboom is a senior lecturer Coastal

Marcel J.F. Stive Cover image:

Prof. Marcel Stive is an emeritus professor

Judith Bosboom & Marcel J.F. Stive

Bosboom, J. and Stive, M. J. F. (2023). Coastal Dynamics. Delft University of Technology,

Publisher TU Delft Open

The softback is printed by Holland Ridderkerk (HollandRidderkerk.nl) on behalf of the

Last change date: 2023-01-11

2 Large-scale geographical variation of coasts 35

Last change date: 2023-01-11

2.3 Tectonic control of coasts . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Last change date: 2023-01-11

3.7.4 Spring and neap tide . . . . . . . . . . . . . . . . . . . . . . . . 127

4 Global wave and tidal environments 149

5 Coastal hydrodynamics 169

Last change date: 2023-01-11

5.5.7 3D effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231

6 Sediment transport 273

Last change date: 2023-01-11

6.9.1 Choice of models . . . . . . . . . . . . . . . . . . . . . . . . . . 318

7 Cross-shore transport and profile development 321

8 Longshore transport and coastline changes 361

Last change date: 2023-01-11

9 Coastal inlets and tidal basins 409

10 Coastal protection 467

Last change date: 2023-01-11

10.4 Modification of longshore transport processes . . . . . . . . . . . . . . 476

A Linear wave theory 515

B Waves breaking on a beach 521

C Hydrographic charts 525

D Stability of structures 533

E Responses to the closures of Dutch tidal basins 537

Last change date: 2023-01-11

E.2.1 Closure of Zuiderzee . . . . . . . . . . . . . . . . . . . . . . . . 539

Acronyms and abbreviations 549

Subject index 565

Errata and improvements 579

Last change date: 2023-01-11

Last change date: 2023-01-11

Last change date: 2023-01-11

Last change date: 2023-01-11

1.2. Coastal dynamics for coastal engineers

Last change date: 2023-01-11

The purposes of these ‘soft’ and ‘hard’ works are diverse:

• Control of shoreline erosion;

• Marine forces (waves, tides, currents and other oceanographic phenomena);

Last change date: 2023-01-11

1.2.2. Position of Coastal Dynamics in the TU Delft curriculum

Coastal Dynamics I and II

Last change date: 2023-01-11

Relevant prior knowledge

1.3. Study goals

Last change date: 2023-01-11

1.4. Examples of engineering applications