Tidal and sediment dynamics in a fine-grained coastal region: A case study of the Jiangsu coast

-A

CASE STUDY OF THE

J

IANGSU COAST

Proefschrift

ter verkrijging van de graad van doctor aan de Technische Universiteit Delft,

op gezag van de Rector Magnificus prof. ir. K.C.A.M. Luyben, voorzitter van het College voor Promoties,

in het openbaar te verdedigen op Maandag 4 April 2016 om 15:00 uur

door

Peng Y

AO

Master of Science in Harbor, Coastal and Nearshore Engineering, Hohai Unversity, Nanjing, China,

promotors: Prof. dr. ir. M.J.F. Stive Prof. dr. ir. Z.B. Wang Samenstelling promotiecommissie:

Rector Magnificus voorzitter

Prof. dr. ir. M.J.F. Stive Technische Universiteit Delft, promotor Prof. dr. ir. Z.B. Wang Technische Universiteit Delft, promotor Onafhankelijke leden:

Prof. dr. ir. W.S.J. Uijttewaal Technische Universiteit Delft Prof. dr. ir. H.J. de Vriend Technische Universiteit Delft Prof. dr. ir. J.C. Winterwerp Technische Universiteit Delft Dr.ir. B.C. van Prooijen Technische Universiteit Delft Overige leden:

Prof. dr. Y.P. Chen Hohai University, China

Keywords: tides, tidal wave system, tidal current system, silt dynamics, sediment transport, morphordynamics

Printed by: GVO drukkers & vormgevers B.V.

Front & Back: Tidal flat of the Jiangsu coast, Photo by HelloRF Zcool/shutterstock.

Copyright © 2016 by Peng Yao

Email: p.yao@tudelft.nl; youngtoti@gmail.com ISBN 978-90-6464-994-3

An electronic version of this dissertation is available at

http://repository.tudelft.nl/.

All rights reserved. No part of the material protected by this copyright notice may be reproduced or utilized in any form or by any means, electronic or machanical, including photocopying, recording or by any information storage and retrieval system, without the prior permission of the author.

The Jiangsu coast is located in eastern China bordering the South Yellow Sea. It is strongly affected by the semi-diurnal tides. Both the tidal range and the tidal current vary greatly in space due to local tidal wave systems and morphologies. The radial tidal current pat-tern identified at the central coast is suggested to play a primary role in the evolution of a large-scale radial-shape sand ridge system. Another feature of the Jiangsu coast is the diversity of the bottom sediments with pronounced silt content. Inspired by the char-acteristics of both the hydrodynamics and sediment dynamics throughout the Jiangsu coast, this thesis focuses on advancing our understanding of the coastal tidal dynamics and the resulting sediment transport.

Regarding the radial tidal current pattern at the central Jiangsu coast, there have been plenty of studies exploring relevant formation mechanisms. A generally accepted inference is that the radial tidal current pattern is a consequence of the interaction be-tween the northern rotating tidal wave system and the southern progressive tidal wave. In this study, we examine the emergence of the radial tidal current in a schematized semi-enclosed tidal basin by introducing the tidal Current Amphidromic Point (CAP) and the tidal current inclination angles. After comprehensive numerical experiments, we find that the overall basin scale and the cross-basin phase difference play roles in the emergence of the radial tidal current. The radial tidal current only has an opportunity to emerge in a basin where the basin length (L) is larger than width (B) (i.e. L/B > 1), a lat-eral depth difference exists or the offshore incoming tidal wave has an oblique angle. The Yellow Sea is featured by these aforementioned prerequisites favouring the emergence of the radial tidal current. Furthermore, we discover that the radial tidal current is re-lated to the cross-basin CAP distribution pattern. When the radial tidal current emerges, the focal point is the CAP related to the velocity vectors rotating anti-cyclonically in the Northern Hemisphere. The CAP distribution deserves more attention for the identifica-tion of the radial tidal current pattern.

To understand the sediment dynamics in a silt-enriched environment in more detail, we have carried out a series of flume experiments under various wave and current con-ditions with field-collected silt-sand mixtures. According to the experiments, we find that the silt fraction has different features originating from both the sand fraction and the clay fraction. A high concentration layer is observed near the bottom together with ripples under pure wave conditions. Sediment concentrations inside the high concen-tration layer are quasi-stationary with the bulk Richardson number approaching a con-stant value. The thickness of the high concentration layer can be scaled with approxi-mately two times the damped wave boundary layer thickness. Thus, the wave motion in-duced turbulence is considered to be the main reason generating the high concentration layer. Moreover, suspensions inside the high concentration layer have a certain amount of sand content, which is different from the fluid mud in the cohesive muddy bed. For the vertical concentration profile, the silt fraction is also distributed differently from the

sand fraction, since the silt concentration decreases logarithmically within high concen-tration layer, while it is homogeneously distributed outside the high concenconcen-tration layer. Considering the specific features of the silt fraction, we recalibrated the formulations of van Rijn (2007a, b) based on our experiments and further developed a multi-fraction sediment transport model to predict the vertical concentration profile for silt and sand classes, and then tested the existing sediment formulations. The results show a promis-ing agreement with the measurements, for both wave-only and wave-with-current con-ditions.

Finally, the Jiangsu Regional Model is set up utilizing the aforementioned findings on tides and sediments. The Jiangsu Regional Model is used to examine whether our existing knowledge can be integrated for a relatively long-term (i.e. time scale of years) predictions on the sediment transport and the morphological changes of the Jiangsu coast. To this end, we first reasonably construct the bed composition throughout the model domain. Subsequently, the model is calibrated and validated against two inde-pendent measurements on water level, flow velocity and the sediment concentration. The results indicate that the present model can produce good results. The simulated annual-averaged SSCs depict a high value in the coastal region between the Old Yellow River Delta and the northern Radial Sand Ridge Field. The simulated morphological changes show a spatially distributed alternating-erosion-sedimentation pattern in the Old Yellow River Delta rather than pure erosion. Over the Radial Sand Ridge Field, the ridges are continuously growing and the adjacent tidal channels are deepening. The sim-ulated annual-averaged tide-induced sediment budget shows that the northern (i.e. the Old Yellow River Delta) and southern (i.e. the southern Radial Sand Ridge Field) Jiangsu coast are under erosion, while the central coast (i.e. the northern and central Radial Sand Ridge Field) is still in progradation. Furthermore, the simulated sediment bed in the Old Yellow River Delta shows a gradually coarsening trend while an overall fining trend is pronounced in the northern Radial Sand Ridge Field. All these long-term re-sults are in good agreement with observation-based estimations. The present modelling framework indeed has the ability for simulating sediment transport and morphological changes over a relatively long time span (i.e. time scale of years).

This thesis addresses series of findings on the radial tidal current pattern, character-istics of the silt-dominated sediments as well as the sediment transport and morpholog-ical changes along the Jiangsu coast. The proposed modelling approaches can serve as a basis and provide information on large-scale hydrodynamics and sediment dynamics for the management and planning of the Jiangsu coast. Future studies may be focused on (1) detailed investigation on the influencing factors on the emergence of the radial tidal current by the CAP system distribution; (2) the physics of the layered-bed system (i.e. the hard layer under ripples) for silt dominated mixtures; (3) improving the compu-tational efficiency of the Jiangsu Regional Model for longer time scale (i.e. tens of years).

De Jiangsu kust ligt in Oost-China aan de Zuid-Gele Zee. Deze kust wordt sterk beïnvloed door een dubbeldaags getij. Zowel het tijdverschil in het verticale getij als in de getijde stroming variëren sterk in ruimte als gevolg van het lokale getijdegolfklimaat en de lo-kale morfologie. Het stromingspatroon van het radiale getij aan de centrale Jiangsu kust is benoemd en wordt verondersteld een primaire rol spelen bij de ontwikkeling van de grootschalige, radiaal-vormige zandbanken. Een ander kenmerk van de Jiangsu kust is de grote diversiteit van de sedimenten met een uitgesproken slibfractie. Geïnspireerd door de kenmerken van zowel de sediment- als de hydrodynamica langs de gehele Ji-angsu kust, focust dit proefschrift zich op de bevordering van ons begrip van de getijde dynamiek en het resulterende sedimenttransport in het kustgebied.

Met betrekking tot het huidige patroon van het radiale getij aan de kust van Jiangsu, zijn er tal van studies die de relevante mechanismen verkennen. Een algemeen aan-vaarde conclusie is dat het huidige patroon van het radiale getij een gevolg is van de in-teractie tussen het noordelijke, getijdegolf circulatie systeem en het zuidelijke, progres-sieve getijdegolfklimaat. In deze studie onderzoeken we het ontstaan van een radiaal-getij stroming in een geschematiseerd, semi-gesloten radiaal-getijde-bekken door de invoering van het Current Amphidromic Point (CAP) en het hellingsvlak van de getijde stroming. Na uitvoerige numerieke experimenten, concluderen we dat de omvang van het totale bekken en het verschil in de fase van het kruisbekken een rol speelt in de opkomst van de radiaal-getij stroming. De radiaal-getij stroming heeft alleen een kans te ontstaan in een waterbekken waar de (bekken) lengte (L) groter is dan de breedte (B) (d.w.z. L/B > 1). De radiaal-getij stroming zou zelfs kunnen ontstaan onder de voorwaarde van een vlakke bodem wanneer een getijdegolf onder een schuine hoek het bekken binnenkomt. We hebben ontdekt dat de radiaal-getij stroming is gerelateerd aan het kruisbekken pa-troon voor de distributie van de CAP. Het punt van aandacht tijdens het ontstaan van de radiaal-getij stroming, is de CAP gerelateerd aan de roterende snelheid van de anti-cyclonische vector op het noordelijk halfrond. De CAP distributie verdient meer aan-dacht voor de identificatie van het radiaal-getij stromingspatroon.

Om de dynamiek van sediment in een slib-omgeving in meer detail te begrijpen, hebben we een aantal experimenten uitgevoerd in een goot onder verschillende golf-en stromingsvoorwaardgolf-en met in het veld verzamelde slib-zand mgolf-engsels. Uit de ex-perimenten concluderen wij dat het slib-fracties verschillen van zowel de zand- en de klei-fracties. Een hoge concentratie laag is onder zuivere golfomstandigheden waarge-nomen nabij de bodem samen met rimpelingen. Sediment concentraties in de hoge concentratie laag zijn quasi-stationair met het bulk Richardson nummer met een bijna constante waarde. De dikte van de hoge concentratie laag verhoudt zich tot ongeveer twee keer de dikte van de gedempte golf buiten-laag. De door de golfbeweging

veroor-This samenvatting was translated from English to Dutch by Mariette van Tilburg and Marcel Stive.

zaakte turbulentie wordt, aldus, beschouwd als de belangrijkste oorzaak voor het gene-reren van de hoge concentratie laag. Bovendien, zweefdeeltjes in de hoge concentratie laag bevatten een bepaalde hoeveelheid zand, die verschilt van de vloeibare modder in het vaste-modder bed. In het verticale concentratie-profiel verspreidde de slibfractie zich ook anders dan de zandfractie, sinds de slib concentratie logaritmisch daalt bin-nen hoge concentratie laag, terwijl het zich homogeen distribueert buiten de hoge con-centratie laag. Gezien de specifieke kenmerken van de slibfractie, hebben wij de for-mules van Van Rijn (2007a, b) opnieuw gekalibreerd op basis van onze experimenten en hebben wij een multi-fractie sediment transportmodel ontwikkeld om het verticale concentratie-profiel voor slib en zand klassen te voorspellen, hebben wij vervolgens de bestaande sediment formules getest. De resultaten tonen een veelbelovende overeen-komst met de metingen, voor zowel alleen-golf als golf-met-stromingsvoorwaarden.

Tot slot, hebben wij met behulp van de bovengenoemde bevindingen over getijden en sedimenten het Jiangsu Regionale Model opgesteld. DeJiangsu Regionale Model wordt gebruikt om te onderzoeken of onze bestaande kennis met relatief lange termijn (d.w.z. op een tijdschaal van jaren) voorspellingen over het sedimenttransport en de morfolo-gische veranderingen van de Jiangsu kust kan worden geïntegreerd. Met dit doel heb-ben we eerst de samenstelling van het bed in het gehele domein van het model gecon-strueerd. Vervolgens is het model gekalibreerd en gevalideerd met twee onafhankelijke metingen van de waterstand, de stroomsnelheid en de concentratie van sediment. De resultaten wijzen erop dat het huidige model goede resultaten geeft. De gesimuleerde jaarlijkse-gemiddelde suspensieve sediment concentraties vertonen een hoge waarde in het kustgebied tussen de Oude Gele Rivier Delta en de noordelijke Radiale Zand Banken. De gesimuleerde morfologische veranderingen tonen een ruimtelijk gedistribueerd pa-troon van afwisselende erosie-sedimentatie in de Oude Gele Rivier Delta in plaats van alleen maar erosie. Over de Radiale Zand Banken, zijn de zandruggen voortdurend in aangroei en de aangrenzende getijdengeulen worden dieper. De gesimuleerde jaarlijks-gemiddelde, door getij-geïnduceerde, sediment kwantiteit toont aan dat de noordelijke (d.w.z. de Oude Gele Rivier Delta) en de zuidelijke (d.w.z. de Radiale Zand Banken) Ji-angsu kust eroderen, terwijl de centrale kust (d.w.z. de Noord- en Midden-Radiale Zand Banken) zich nog steeds in staat van progradatie bevindt. Bovendien toont het gesimu-leerde sediment bed in de Oude Gele Rivier Delta een geleidelijk verruwingstrend, terwijl er een algemene verfijningstrend plaats vindt in de noordelijke Radiale Zand Banken. Al deze lantermijn resultaten tonen een goede overeenkomst met de op observatie ge-baseerde ramingen. Het huidige model heeft inderdaad de mogelijkheid om sediment-transport en morfologische veranderingen over een relatief lange periode (d.w.z. tijds-bestek van jaren) te simuleren.

Dit proefschrift stelt een serie van bevindingen ter discussie met betrekking tot het radiaal getijde patroon, de kenmerken van de slib gedomineerde sedimenten, en het sedimenttransport en de morfologische veranderingen langs de kust van Jiangsu. De voorgestelde modellen en benaderingen kunnen de basis zijn en informatie verstrekken voor het beheer en de planning van de Jiangsu kust over de grootschalige hydrodynami-sche en sediment processen die hier plaats vinden. De focus van toekomstige studies kan zich richten op (1) gedetailleerd onderzoek naar de factoren van het ontstaan van de radiaal getijden stroming door CAP distributie; (2) de fysica van het gelaagde-bed (dat

wil zeggen de harde laag onder de rimpelingen) voor mengsels van merendeel slib; (3) de verbetering van de computationele efficiëntie van de JRM voor een langere tijdschaal (d.w.z. tientallen jaren).

A local tidal amplitude of a constituent A0 mean water level over a certain period

A_δ peak orbital excursion

c sediment concentration at certain elevation by mass or volume c depth-averaged sediment concentration by mass or volume ca reference concentration by mass

cave High concentration layer-averaged concentration by mass

c0

ave High concentration layer-averaged concentration by volume

cg el gelling concentration

cg el ,s dry bulk density of sand bed

D50 median grain size of the bed mixtures

Di mean grain size of the each sediment fraction

D_∗ dimensionless particle size F nodal amplitude factor

f Coriolis parameter

G local phase lag of a tidal constitent g gravitational acceleration

H wave height

h water depth

K amplitude of the equilibrium tide k wave number, k = 2π/L

ks,c current related bed roughness

ks,w wave related bed roughness

L wave length

Ric bulk Richardson number

Rif flux Richardson number

Sb bed-load transport rate

s relative density, s = ρs/ρw

s0 relative excess density, s0_{= s − 1}

T wave period

Te temperature

U_δ peak orbital velocity near bed, U_δ= ωAδ

um depth-averaged current velocity

u_∗,c current-related bed shear velocity V0+ u astronomical argument

ws settling velocity

za reference height

δHC L thickness of the high concentration layer

δs thickness of the near-bed sediment mixing layer

δw thickness of the wave boundary layer

ς(t) water level at time t

ε0 mean convective eddy viscosity in one wave cycle

εs sediment mixing coefficient

η height of wave ripple κ von Karman constant λ length of wave ripple

ν kinematic viscosity coefficient

ξ coefficient of the hiding-exposure effect T excess bed shear stress

τ current or wave related bed shear stress τcr critical bed shear stress

τ0 _{current or wave related effective bed shear stress}

ρs sediment density

ρw water density

φd damping coefficient due to sediment-induced stratification

φf l oc flocculation coefficient

φhs hinderer settling coefficient

φp porosity of sediment bed

φsl ope bed slope factor

χ astronomical argument of the equilibrium tide (relative to GMT) ψ wave mobility number

ω angular frequency of waves,ω = 2π/T

Abbreviations:

CAP tidal Current Amphidromic Point EAP tidal Elevation Amphidromic Point SSC Suspended Sediment Concentration

Summary vii

Samenvatting ix

Notations xiii

1 Introduction 1

1.1 Background. . . 2

1.1.1 Coastal region and the relevant physical processes. . . 2

1.1.2 Tides and tide-induced morphology in the coast ocean . . . 2

1.1.3 Sediment transport in the coastal region. . . 4

1.2 The Bohai Sea, Yellow Sea and East China Sea. . . 5

1.3 The Jiangsu coast. . . 5

1.3.1 Geography. . . 5

1.3.2 Drive forcing. . . 7

1.3.3 Sediment characteristics. . . 8

1.4 Problem definition and Research objectives . . . 8

1.5 Outline . . . 10

2 Tidal wave systems in the Yellow Sea 13 2.1 Introduction . . . 14

2.2 Model set up and performance . . . 16

2.2.1 Model set up. . . 16

2.2.2 Model performance . . . 20

2.3 Results and influence factor analyses. . . 24

2.3.1 Co-tidal charts and tidal current field . . . 24

2.3.2 Analyses of influencing factors for tidal motions. . . 26

2.4 Discussion on the factors of influence on the tidal wave and tidal current pattern . . . 29

2.4.1 Regional bathymetry. . . 29

2.4.2 The role of the geographic position of the Shandong Peninsula . . . 31

2.5 Conclusion . . . 36

3 Tidal current system in the South Yellow Sea 39 3.1 Introduction . . . 40

3.2 Regional Setting. . . 43

3.3 Method. . . 44

3.3.1 Simplified tidal model . . . 44

3.3.2 Case configurations . . . 46

3.3.3 Illustration of a tidal current system . . . 48

3.3.4 Definition of the radial tidal current . . . 49 xv

3.4 Results . . . 49

3.4.1 Tidal regimes of the reference case. . . 49

3.4.2 Shape of the basin . . . 52

3.4.3 Lateral depth difference . . . 55

3.5 Further explorations on the influence of the lateral phase difference . . . . 57

3.5.1 Idealized model setting . . . 57

3.5.2 Pre-allocating lateral phase difference along the open boundary . . 59

3.5.3 Self-generating lateral phase difference . . . 62

3.6 Discussions. . . 63

3.6.1 Influencing factors on tidal current system and the radial tidal cur-rent . . . 63

3.6.2 Application in the South Yellow Sea . . . 64

3.6.3 Application of a CAP system . . . 66

3.7 Conclusions. . . 67

4 Experiment inspired numerical modeling of sediment concentration over sand– silt mixtures 69 4.1 Introduction . . . 70

4.2 Flume experiment . . . 72

4.3 Experiment Results and Discussions . . . 76

4.3.1 Bed forms . . . 76

4.3.2 High concentration layer under waves. . . 80

4.3.3 Vertical profile of Suspended Sediment Concentration. . . 85

4.4 Numerical study on predicting the suspended sediment concentration. . . 87

4.4.1 Reference concentration. . . 87

4.4.2 Multi-fraction Approach. . . 90

4.4.3 Boundary conditions. . . 91

4.4.4 Comparison of model results with measurements. . . 92

4.5 Conclusions. . . 95

5 Sediment transport modeling in a sand–silt mixed environment: A case study of the Jiangsu coast 97 5.1 Introduction . . . 98

5.2 General information on the study area . . . 101

5.3 Methodology . . . 103

5.3.1 Sediment transport model. . . 103

5.3.2 Bed level and bed composition update. . . 106

5.3.3 General configurations of the Jiangsu Regional Model (JRM) . . . . 107

5.3.4 Initial bed composition . . . 109

5.3.5 Open boundary conditions for sediment. . . 111

5.3.6 Model skill assessment. . . 112

5.4 Model calibration and validation on short-term scales . . . 114

5.4.1 Model calibration . . . 114

5.4.2 Model verification . . . 115

5.5 Model results and long-term validation. . . 121

5.5.1 Tidal regimes. . . 121

5.5.2 Sediment dynamics . . . 122

5.6 Discussions. . . 130

5.6.1 Sensitivity of the southernmost SSC boundary conditions (influ-ence of the Yangtze-derived sediment). . . 130

5.6.2 Limitations and remarks for further research . . . 131

5.7 Conclusions. . . 132

6 Conclusions and recommendations 135 6.1 General. . . 136

6.2 Conclusions. . . 136

6.3 Recommendations for future work . . . 139

A Definitions of the characteristic tidal parameters 141 B Modified van Rijn (2007a, 2007b) sediment transport formulations 143

References 149 List of Figures 167 List of Tables 171 Acknowledgements 173 Curriculum Vitae 175 List of Publications 177

1

I

NTRODUCTION

Learn from yesterday, live for today, hope for tomorrow. The important thing is not to stop questioning. — Albert Einstein

1

1.1.

B

ACKGROUND

1.1.1.

C

OASTAL REGION AND THE RELEVANT PHYSICAL PROCESSES

T

HEcoastal region is broadly defined as a region that bridges the terrestrial land and the ocean, forming one of the most dynamic interfaces on earth (McLean et al.,

2001;Dronkers,2005). It includes adjacent lands, nearshore and offshore waters ( Davidson-Arnott,2010). Estuaries, beaches, intertidal and nearshore zones support a diverse range of marine and terrestrial species. Humans have been utilizing the coastal region through-out the history, for the purposes of naval, commercial, recreation and tourism. General features associated with the coastal region used in this study are illustrated in Figure1.1.

Figure 1.1: Schematized coastal region with spatial boundaries.

The coastal region is generally a result of various processes in different spatial and temporal scales (Davis,1994): geological processes, e.g. mountain formation and ero-sion, require millions of years; climatic processes, e.g. sea level changes, require thou-sands of years; the aforementioned two processes are superimposed over the cumula-tive combinations of the shorter time scale processes, e.g. tides, currents, waves and sediment transport (De Vriend et al.,1993;Stive et al.,2002) . In this study we limit our-selves to these small time scale processes and the present-day conditions, excluding the geological and climatic processes.

1.1.2.

T

IDES AND TIDE

-

INDUCED MORPHOLOGY IN THE COAST OCEAN

A

STRONOMICALtides (herein after referred to tides) are one of the major driving forces in the coastal regionPugh(1996). Tides are generated by regular movements of the moon–earth and earth–sun systems, causing periodical sea level change (rise and fall) and periodically reversing currents (flood and ebb) at coast boundaries. The tide itself is a long wave with a period of about 12 hours and 25 minutes (semi-diurnal tide) in most coasts, while it has a dominant period of about 24 hours (diurnal tide) at some

lo-1

cations. Tidal wave propagation in shallow coastal waters can be influenced by effects of shoaling (due to the change of the basin geometries and bathymetries), damping (due to bottom friction), reflection (due to land boundaries) and deformation (due to differ-ences in propagation velocities at low and high water).

Tidal current ridges were first introduced byOff (1963) to describe the significant features of series of rhythmic linear sand bodies on the tidal-dominated continental shelves, such as the North Sea and the east China marginal seas (i.e. the Yellow Sea and the East China Sea). Figure1.2presents an example of the distribution of the sand ridges in the North Sea. Strong tidal currents (i.e. 0.5–2.5 m/s) are considered to play a main role in shaping the ridge patterns in an environment with sufficient sediment supply (Off,1963;Liu et al.,1998;Dyer & Huntley,1999). These sand ridges are commonly ob-served in a parallel manner (Dyer & Huntley,1999), such as the shore-faced connected ridges along the Dutch coast (van de Meene & van Rijn,2000). An unique example of the tidal current ridges is shown in the south western Yellow Sea (Liu et al.,1989; see Fig-ure1.3). Series of individual ridges spread toward sea in a radial pattern with the apex zone connected with the shore-face of the central Jiangsu coast. This study concerns the hydrodynamics and sediment dynamics over such unique geomorphological features.

1

1.1.3.

S

EDIMENT TRANSPORT IN THE COASTAL REGION

T

HEcoastal region consists of soft materials, such as, sand, silt and clay (Fredsøe & Deigaard,1992). The sediment transport itself is a complex and multi-dimensional process that closely relates to the interactions involving various external forcing agents (e.g. tide, wave and wind), turbulence and sediment grains (van Rijn,1993;Soulsby,

1997;Winterwerp & van Kesteren,2004). In the coastal region, both the magnitude and direction of sediment transport vary in different locations: erosion may occur at some places, whereas deposition is found at other locations. Consequently, the coastal mor-phologies are continuously changing. In turn, changing of the coastal morphology af-fects the tides, waves and currents. Hence, the water motion and the morphology evolve interdependently linked by the sediment transport. Thus, sediment transport plays a vital role in morphological processes in the shallow coastal waters (e.g. estuaries, tidal inlets and lagoons, coasts) (Dyer,1986;Le Hir et al.,2001;Uncles,2002;Collins & Balson,

2007;Fagherazzi & Overeem,2007).

Over the years, many efforts have been made to understand the sediment transport mechanisms and the coastal morphological evolution, by developing and updating nu-merical/analytical models and field instruments (Mehta et al.,1989;van Rijn,1993;Jay et al.,1997;Ouillon et al.,2003;Winterwerp & van Kesteren,2004;Collins & Balson,2007;

Fagherazzi & Overeem,2007;Papanicolaou et al.,2008;Amoudry & Souza,2011;Erikson et al.,2013;van Rijn et al.,2013). Regarding sediment transport modeling, there are many studies to formulate the sediment transport for pure non-cohesive sediments (Nielsen,

1992;van Rijn,1993;Soulsby,1997etc.), and cohesive sediments (seeWinterwerp & van Kesteren,2004for an overview). In these theories, the non-cohesive sediment is defined as the partial size larger than 63µm; while the cohesive sediment refers to sediment sizes smaller than 63µm (also named as mud, i.e. sum of the silt and clay fractions).

Natural bottom sediments in the coastal region are rarely composed of grains with one size but rather of different sizes. Different sediment fractions behave different re-sulting in a selective sediment transport (Greenwood & Xu,2001). In particular, when the bottom sediments are mixed by both cohesive and non-cohesive materials, things become rather complicated. According to several experiments, it is known that the crit-ical bed shear stress can dramatcrit-ically increase with the increase of the clay component in the mixture (Mitchener & Torfs,1996;Jacobs et al.,2011). Hence, cautions should be taken when dealing with sediment transport over a mixed bed.

1

1.2.

T

HE

B

OHAI

S

EA

, Y

ELLOW

S

EA AND

E

AST

C

HINA

S

EA

T

HEeast part of the Chinese marginal seas includes the Bohai Sea, the Yellow Sea

and the East China Sea, bounded by the Chinese main land, the Korean Penin-sula, Kyushu Island and Ryukyu Island (Figure1.3). Over the continental shelf of the Bohai Sea, the Yellow Sea and the East China Sea, tide is the primary driving force for the morphological processes, while waves and storms are secondary effects (Liu et al.,1998;

Zhang et al.,1999;Wang et al.,2012b). The Bohai Sea, the Yellow Sea and the East China Sea is a wide and semi-enclosed basin, composed of several bays, peninsulas, estuaries and other natural geophysical settings. The bathymetry of the Bohai Sea, the Yellow Sea and the East China Sea shows significant variations as well. The water depth changes from ~2000 m in the south (Okinawa Trough) to less than 50 m in the north, with a step-like characteristic in the longitudinal direction. In the South Yellow Sea, a north–south trough is situated in the centre with a depth of 60–80 m. Separated by the trough, the water depth is slowly decreasing with a gentle slope toward the Chinese coast, but more rapidly decreasing toward Korean Peninsula.

1.3.

T

HE

J

IANGSU COAST

1.3.1.

G

EOGRAPHY

GENERAL INFORMATION

T

HEJiangsu coast is located in eastern China facing the South Yellow Sea (Figure1.3). The coastline of Jiangsu starting from the Xiuzhen River Estuary in the north, ex-tending southward to the Yangtze River Estuary, with a total length of ~954 km. The silty coast dominates ~93% of the total shoreline (Ren,1986). A large-scale and well-developed tidal flat system is mainly located in the central Jiangsu coast, sheltered be-hind a radial sand ridge system. The width of the tidal flats is about 10–13 km and the maximum width can reach 36 km (Wang & Zhu,1990;Healy et al.,2002).

In this study, the region starting from the Jiangsu coastline toward a depth of around 50 m (i.e. approximately the western edge of central Yellow Sea trough; see Figure1.3) is defined as the inner shelf of the Jiangsu coast. The averaged water depth of the inner shelf region is around 30 m with a gentle bottom slope. Two distinct geomorphological units can be identified in the inner shelf, viz. the Old Yellow River Delta in the north and the fan-shaped Radial Sand Ridge Field in the south (Figure1.3) (Zhang,1984;Liu et al.,

1

Figure 1.3: Map of the Bohai Sea, Yellow Sea and East China Sea. The locations of the Jiangsu coast as well as the Old Yellow River Delta and the Radial Sand Ridge Field are labelled in the map. OYRD=Old Yellow River Delta; RSRF=Radial Sand Ridge Field; Pen.=Peninsula; Riv.=River. The black rectangular box denotes the inner shelf zone defined in this study. The upper right map is the zoom-in of the Radial Sand Ridge Field.

OLDYELLOWRIVERDELTA

T

HEYellow River discharged into the South Yellow sea via the northern Jiangsu coast during 1128–1855 AD. In this period, the Yellow River brought considerable amount of fine silts into the Jiangsu coast, making the shoreline continuously progradating sea-ward and forming the Old Yellow River Delta and a large-scale tidal flat system in the central Jiangsu coast (Gao,2009). After 1855, the lower reach of the Yellow River shifted

1

northward discharging into the Bohai Sea. Consequently, the Old Yellow River Delta (both subaerial and submarine delta area) has been subject to severe erosion since 1855 AD. However, the erosion is only active in the area surround the Old Yellow River Delta, and it does not affect the central Jiangsu coast, where the tidal flats still accumulate. Thus, the Jiangsu coast can be roughly divided in two parts: an erosion-dominant open coast in the north (i.e. Old Yellow River Delta) and an accretion-dominant ridge-sheltered coast in the south (Ren,1986;Liu et al.,2011).

RADIALSANDRIDGEFIELD

I

Nthe South–Western region of the Yellow Sea, an enormous fan-shaped morphologi-cal feature exists, made up of approximately 70 individual ridges of various size, cov-ering an area of about 22470 km² (Ren,1986). These ridges are centred at the central Jiangsu coast and spread seaward up to 25 m water depth in a radial pattern (Liu et al. 1989;Wang 2002; see Figure1.3). They are named as Radial Sand Ridge Field. According to both field observations and numerical simulations, the local flood/ebb currents flow landward/seaward in the same direction as the ridge orientation(Wang,2002). Such a tidal current pattern is commonly referred to as radial tidal current in previous studies (e.g.Zhu & Chang 2001a;Uehara et al. 2002). The radial tidal current is considered to be independent from local geomorphology, which has been demonstrated byZhu & Chang

(2001a) through several numerical experiments (e.g. using a flat and a linear sloped bathymetry replacing the Radial Sand Ridge Field, respectively). Besides, the simula-tion of the Paleo-tidal current field reveals that this special tidal current pattern existed since 7000 years ago and therefore is expected to play a crucial role in the formation and development of the Radial Sand Ridge Field (Uehara et al.,2002;Uehara & Saito,2003;

Zhu & Chen,2005).

1.3.2.

D

RIVE FORCING

T

IDE, especially the semi-diurnal tide, is the most dominant forcing in the Yellow Sea (Choi,1984;Fang,1986). The mean tidal range along the Jiangsu coast is between 2 and 4 m. The largest observed tidal range is 9.28 m in the Huangshayang tidal channel (Wang et al.,2012b). The mean tidal flow velocity is weak in the Haizhou Bay (between 0.3–0.5 m/s) and the Old Yellow River Delta (between 0.6 m/s–1 m/s), and it becomes strong in the Radial Sand Ridge Field. The mean tidal flow velocity can be larger than 2 m/s in several major tidal channels (e.g. Xiyang channel, Huangsha Yang channel) of the Radial Sand Ridge Field during spring tide. There are several small rivers discharg-ing into the sea along the Jiangsu coast. Most of the rivers are channelized with small discharges. The Yangtze River discharges into the sea in the southernmost of the study

1

area. Approximately 10% of the total discharge from the Yangtze River may flow

north-ward, diluting the salinity in the near-Estuary region (i.e. southernmost of the Radial Sand Ridge Field).

1.3.3.

S

EDIMENT CHARACTERISTICS

T

HEbottom sediment greatly varies throughout the inner shelf of the Jiangsu coast. The sediment is mainly composed of fine sand (> 125µm) in the north Jiangsu coast (i.e. the Haizhou Bay). In the Old Yellow River Delta and in the tidal channels of the Radial Sand Ridge Field, the contents of fine silt and clay (< 30µm) increase, and the main sediment type is clayey/fine silt (8–30µm). On the ridges of the Radial Sand Ridge Field, the contents of coarse silt and fine sand (> 30µm) increase, and the main types of sediment are sandy silt and silty sand (30–125_{µm). In the offshore area where the depth} is more than 30 m, the sediment becomes finer seaward as sand, sandy silt and clayey silt, respectively. Overall, the grain sizes of the bottom sediment grains nowadays range from 8–250µm with pronounced silt contents. The diversity of the bottom sediment may be attributed to the energetic tidal currents and the historical fluvial sediment supplies from the two largest sediment-carrying rivers in the past, i.e. the Yangtze River and the Yellow River. Nowadays, the fluvial sediment inputs are less important for the sediment suspensions in the Jiangsu coastal waters, whereas seabed erosion is the major sediment source (Fu & Zhu,1986;Milliman et al.,1986;Qin et al.,1989).

1.4.

P

ROBLEM DEFINITION AND

R

ESEARCH OBJECTIVES

T

WOcharacteristics can be identified for the morphological evolution of the Jiangsu coast: the special tidal wave system and the associated tidal current pattern; the silt-dominated sedimentary environment. Inspired by the Jiangsu coast, we aim to in-vestigate (1) the underlying physics behind the special tidal wave and tidal current pat-tern and (2) the sediment transport with respect to a silt-dominated mixed environment. Specific research topics are briefly discussed as follows.

(1) Tidal wave systems in the Bohai Sea, the Yellow Sea and the East China Sea

The first co-tidal chart regarding the Bohai Sea, the Yellow Sea and the East China Sea was drawn byOgura(1933). Since the 1970s, numerical modeling has been commonly employed to investigate the tidal regimes in the Bohai Sea, the Yellow Sea and the East China Sea (Choi,1984;Fang,1986;Yanagi & Inoue,1994;Kang et al.,1998). These models were constructed with different focuses and model settings (e.g. mesh resolution, area of domain, etc.) with fruitful results. By reviewing the previous models, we notice that

1

there is limited understanding on the role of the river discharge, tidal generating force and the land reclamation to the tidal dynamics of the Bohai Sea, the Yellow Sea and the East China Sea, especially for the Jiangsu coastal region. Furthermore, regarding the special tidal wave in the Radial Sand Ridge Field, the local geometry and the basin shape are suggested to be important aspects of influence but there is no consensus yet (Zhang,

1991;Zhang et al.,1999;Wang et al.,2012b). To address these questions, we setup a large-scale model covering the domain of Bohai Sea, the Yellow Sea and the East China Sea, with special interests in the South Yellow Sea. Based on the large-scale tidal wave model, we investigate the tidal wave propagation in the Bohai Sea, the Yellow Sea and the East China Sea addressing the aforementioned concerns.

(2) Physics underlying the formation of the radial tidal current off the Jiangsu coast

Based on previous studies (e.g.Zhang et al.,1999), the special radial tidal current is hy-pothesized to be formed by the interaction of the progressive wave from the East China Sea and the rotating tidal wave system in the south Yellow Sea. The hypothesis is pro-posed based on the simulated tidal elevation systems (represented by a co-tidal chart) and the tidal current pattern (represented by either tidal vector fields or a tidal ellipse field). However, compared with the tidal elevation system, the tidal current system is more sensitive to the changes of the external factors (i.e. basin shape, bottom friction, bathymetry, open boundary conditions) (Xia et al.,1995;Carbajal,1997). To deepen the existing understandings and seek possible explanations on the radial tidal current, we apply a schematized tidal model and conduct relevant numerical experiments from the perspective of the tidal current amphidromic system.

(3) Sediment transport mechanisms over a silt-enriched bed under currents and waves

Previous studies commonly adopt the cohesive formulations to deal with the silt-sized mud assuming that the silt behaves similarly to clay materials. However, the cohesive characteristic of the mud is determined by the content of the clay and the very fine silt (of which the grain size is less than 8µm) (van Rijn,2006). The critical clay content of a mixture is found between 5% and 10%, controlling the cohesion of the mixtures (van Ledden et al.,2004). Furthermore, the mineral compositions of silt consist mainly of quartz and feldspar, which are basically non-cohesive (Lambe & Whitman,1979). Sev-eral studies show that the silt-enriched mixtures exhibit cohesive-like behavior during erosion (Roberts et al.,1998), but flocculation has not been observed based on the set-tling experiments on silt (te Slaa et al.,2015,2013). Thus, it is uncertain whether the previous formulations are suitable for the silt-dominated mixtures. In order to extend the knowledge of sediment transport in the silt range, we carried out a series of flume

1

experiments on two kinds of sand-silt mixtures under various wave and current

condi-tions. Subsequently, we aim to improve the existing sediment modeling approach for sand-silt mixtures based on experimental data.

(4) Can we predict the sediment transport over a relative long-term period in a silt-en-riched coastal region in large-scale?

As mentioned before, bottom sediments in the Jiangsu coastal region are mainly mix-tures of sand, silt and clay with pronounced silt contents. Due to the large domain and the complex morphologies of the Jiangsu coastal region, the overall sediment transport and morphological changes are still insufficiently understood using only the local-scale measurements. Numerical modeling can serve as an alternative tool to investigate the sediment dynamics in this region. However, we note that the existing numerical mod-eling on sediment transports along the Jiangsu coast is rather scarce. Furthermore, all these previous models were developed for a short-time period (i.e. several tidal cycles) and applied cohesive sediment transport formulations (e.g. Partheniades-Krone formu-lations; seePartheniades,1965) to compute the suspended sediment transport over a spatially homogenous cohesive muddy bed setting. Such setting may result in reason-able results for shorter time scale but it is questionreason-able for long-term simulations (i.e. time scale of years). To obtain a better understanding on sediment dynamics in a rel-atively long-term period, we develop a sediment transport model integrating previous experimental findings on silt class rather than simply using the formulations for cohe-sive sediment. The model ability on the Jiangsu coastal sediment transport is further examined by a series of existing observation-based understandings.

1.5.

O

UTLINE

F

OLLOWINGthe above objectives, this thesis is organized as follows.

In chapter2, the tidal wave systems in the Bohai Sea, the Yellow Sea and the East China Sea are investigated thoroughly through a large-scale tide wave model with rela-tively high resolution. In chapter3, we focus on deepening the present understanding on the formation mechanisms of the radial tidal current pattern off the central Jiangsu coast. Chapter4introduces a series of flume experiment, which are designed to extend the knowledge of sediment transport in the silt range. The experimental data are utilized to examine the existing sediment theories on cohesive and non-cohesive sediment and improve the present method. Chapter5integrates the findings of chapter4into a model-ing framework (i.e. Delft3D), which is further used to set up the Jiangsu Regional Model. The ability of the present sediment transport model over relatively long-term simulation

1

(i.e. time scale of years) is examined. Chapter2,3,4and5, which make up the main contents of this thesis, are arranged as standalone research manuscripts with separated introductory sections and conclusions. Chapter6synthesizes the main findings of this study and presents the recommendations for future studies.

2

T

IDAL WAVE SYSTEMS IN THE

Y

ELLOW

S

EA

In this chapter, a two-dimensional tidal wave model which covers the Bohai Sea, the Yel-low Sea and the East China Sea is developed and validated to investigate the near-field and far-field hydrodynamic conditions along the Jiangsu coast in large scale. After a brief review of previous models, some influencing factors, such as tide generating force, river discharges as well as shoreline changes due to land reclamation, are examined in this study. We suggest that whether these factors should be considered in the model depends on the different purposes and geographical regions of interest. Then, a series of experiments are designed to further investigate the previously proposed important factors of influence (i.e. local bathymetry and the role of Shandong Peninsula) on the tidal motions in the Yel-low Sea. The numerical experimental results show that the interaction between the tidal wave system in the north Yellow Sea and the incoming tidal wave plays a main role on the formation of the rotating tidal wave system in the south Yellow Sea, whereas the role of ge-ometric position of the Shandong Peninsula is secondary. With respect to the tidal current, it is found that the radial shape of it is independent of the local bathymetry, but may be generated by the convergent tidal wave formed by the meeting of the rotating tidal wave in the south Yellow Sea and the incoming tidal wave from the East China Sea. Furthermore, a certain water depth is crucial for the intensity of current velocity to generate the special topography of radial sand ridges.

Parts of this chapter have been published in:

Su, M., Yao, P., Wang, Z. B., Zhang, C. K., & Stive, M. J. F. (2015). Tidal Wave Propagation in the Yellow Sea. Coast.

Eng. J., 57(3), 1550008-1-155008-29, doi: 10.1142/S0578563415500084.

2

2.1.

I

NTRODUCTION

T

HEeast part of the Chinese marginal seas includes the Bohai Sea, the Yellow Sea and the East China Sea (Figure1.3). The southwest of the Yellow Sea, where the Jiangsu coast is located, is famous for abundant tidal flats which account for 25% of the tidal flat resources in China (Tao et al.,2011a; see Chapter1for more detail). Near the Jiangsu coast, the hydrodynamic condition is very special because of its radial tidal current pat-tern (Zhang et al.,2013). In order to gain a better understanding of this hydrodynamic feature, it is necessary to advance the knowledge about the underlying mechanism of tidal wave propagation in the whole Bohai Sea, the Yellow Sea and the East China Sea first.

The mechanism of the tidal wave propagation in the Bohai Sea, the Yellow Sea and the East China Sea has been a research focus for many years. The first study can be traced back to 1933, which was conducted byOgura(1933). Several studies have been performed to establish a first impression of the tidal wave system in this region by using observed data (e.g. Fang,1986). Since the late 1970’s, numerical modeling has become the most commonly used method (see Table2.1), besides analytical modeling (Kang 1984;Shen & Huang 1993). Many meaningful results have been generated based on these models with different focuses and model settings (e.g. mesh resolution, area of domain). Analytical models are always used to explore internal tidal dynamics (e.g. tidal amphidromic points formation, tidal wave propagation pattern) based on idealized do-mains. 2D and 3D models mainly focus on practical issues, such as tidal wave propaga-tion and tidal current in the Bohai Sea, the Yellow Sea and the East China Sea.

The consideration of influencing factors, such as the tide generating force and river discharges, varies in different models (Table2.1). Although the influence of tide generat-ing force on the tidal amplitude in the Yellow Sea and in the East China Sea is commented to be several percent (An,1977;Kang et al.,1991,1998), some research still suggests that the contribution of tide generating force should not be neglected in the Bohai Sea, the Yellow Sea and the East China Sea in order to generate an accurate tidal motion (e.g.Yu et al. 2006;Song et al. 2013a). As many studies have ignored tide generating force, it is necessary to analyze and discuss where the most affected region is and to what extent the accuracy can be affected by ignoring tide generating force in this large-scale model. Similarly, as the Bohai Sea, the Yellow Sea and the East China Sea is surrounded by many rivers (Figure1.3) which are always ignored by previous studies, it is also unclear to what extent the river discharge can influence the tidal wave pattern. In addition to the influ-ences of these natural physical processes, the human intervention on the coastal line is also important in the Bohai Sea, the Yellow Sea and the East China Sea. For example, a

2

Table 2.1: Summary of the some tidal models about the Bohai Sea, the Yellow Sea and the East China Sea.

Authors

Domain

Resolution

TGF

Choi

(

1980

)

BYE

12’ × 15’ × 1 layer

No

Kang et al.

(

1991

)

BYE*

7.5’× 10’× 1 layer

Yes

Zhao et al.

(

1994

)

BYE

15’× 15’× 1 layer

Yes

Kang et al.

(

1998

)

BYE

3.75’× 5’ × 1 layer

Yes

Guo & Yanagi

(

1998

)

BYE

12.5 km× 20 layers

No

Wan et al.

(

1998

)

BYE

5’× 5’× 5 layers

No

Wang et al.

(

1999a

)

BYE

10’× 10’× 6 layers

Yes

Lin et al.

(

2000

)

BYE

6’× 6’ × 1 layer

Yes

Zhu & Chang

(

2000

)

BYE

10’× 10’ × 1 layer

Yes

Bao et al.

(

2001

)

BYE

5’× 5’× 15 layers

No

Li

(

2003

)

BYE

8~32 km× 20 layers

No

Chen

(

2008

)

BYE

4’× 4’ × 1 layer

Yes

Zhu & Liu

(

2012

)

BYE

1.5~50 km× 10 layers

–

Zhu & Liu

(

2012

)

BYE

2’× 2’× 22 layers

–

Note: BYE= Bohai Sea, Yellow Sea and East China Sea; BYE*=Bohai Sea, Yellow Sea and East China Sea+East/Japan Sea; TGF=Tide Generating Force. “Yes” and “No” denote TGF is considered and irgnored in the model, respectively.

large-scale reclamation (>1800 km² between 2010 and 2020) is implemented by the Chi-nese government near the Jiangsu coast. It is necessary to identify to what extent such abrupt shoreline changes can influence the tidal wave motion on both near- and far-field. Therefore, further studies are required to focus on these questions and to provide suggestions on the setup of large-scale models.

with regard to the near-field hydrodynamics of the Jiangsu coast, many studies have focused on the radial tidal current and its formation mechanism. There are three main different understandings to interpret its formation. The first is the bathymetry control assumption, which considers that radial tidal current is generated by the local radial sand ridges (Ren,1986;Zhang,1991). The second theory suggests that radial tidal cur-rent is independent of the local bathymetry but controlled by the special local tidal wave systems (Zhu & Chang,2001a). The local tidal wave pattern is convergent-like, which is hypothesized to be formed by the meeting of the progressive wave from the East China Sea and the rotating tidal wave system in the south Yellow Sea. And the geometric

posi-2

tion of the Shandong Peninsula is thought to play a significant role on the formation of such tidal wave and tidal current pattern (Huang & Wang,1987;Zhang & Zhang,1996;

Wang et al.,1998;Zhang et al.,1999). The third opinion, considering the effects of both bathymetry and tidal wave system, suggests that the formation of radial tidal current is due to the interaction between tidal waves and topography/local shorelines (e.g. Lin et al. 2000;Wang et al. 2012b;Ye 2012). Comparing these theories, there are two points, local bathymetry and the role of Shandong Peninsula on the local tidal wave system and the tidal current pattern, requiring further study to determine their effects.

Therefore, in this study, a large-scale model covering the domain of the Bohai Sea, the Yellow Sea and the East China Sea is set up, with special interests in the south Yellow Sea, to investigate the overall tidal motion in the near-field of the Jiangsu coast. After the verification of the model, the effects of influencing factors, such as tide generating force, river discharge as well as shoreline changes due to land reclamation, on the performance of the model are examined using numerical experiments. Then, a series of experiments are carried out to explore the effect of the local bathymetry and the role of Shandong Peninsula on the tidal wave and tidal current pattern.

2.2.

M

ODEL SET UP AND PERFORMANCE

2.2.1.

M

ODEL SET UP

T

HISstudy focuses on the tidal wave propagation and the tidal current pattern in the Yellow Sea, where the tidal current changes are predominantly horizontal (Xing et al.,2012). A 2DH modeling method is considered to be adequate for this study, since it requires less computation time compared with the 3D method especially for large-scale domain. Therefore, a 2DH tidal wave model is setup based on the Delft3D modeling system (Deltares,2011), focusing on the tidal dynamics in the Bohai Sea, the Yellow Sea and the East China Sea with a special interest in the south Yellow Sea. The Delft3D–FLOW module is developed by resolving the Navier-Stokes equations for incompressible fluids under shallow water conditions and under the Boussinesq assumptions. More details of the numerical schemes can be found in the Delft 3D–FLOW user’s manual (Deltares,

2011).

The model domain covers an area which is bounded by latitudes of 24° and 41°N and by longitudes of 117° and 131°E (Figure2.1). To fit the coastal lines, orthogonal curvilin-ear grids are applied with a resolution from approximately 0.7’× 0.7’ to 2.8’× 2.8’, under spherical coordinates. A relatively higher grid resolution is used in the area of interest, e.g. the Jiangsu coast, the Yellow Sea, while the grid is coarser in the area near open boundaries. The enlarged grids near the Jiangsu coast are shown in Figure2.2a.

Previ-2

Figure 2.1: Map of the the Bohai Sea, the Yellow Sea and the East China Sea, and distribution of the main rivers. Bathymetry contours are given in m. Circle dots show the location of the water level observation stations, which are divided into 10 zones (A–J) represented by different colours. Red crosses and stars show the location of tidal current observation stations.

ous numerical studies were able to satisfy the basic requirement of the hydrodynamic analysis in the Bohai Sea, the Yellow Sea and the East China Sea; however, finer grids are necessary for a more accurate simulation because the small scale changes of the topog-raphy and shoreline can only be considered by high resolution.

The bathymetry data (Figure2.1) are obtained, first, from the General Bathymetric Chart of the Oceans (GEBCO) bathymetry database (IOC, IHO and BODC,2003) with 30

2

arc-second grid resolution. Other bathymetry data near the China coast, e.g. Jiangsu area, are obtained by digitalizing the official marine charts published by the Maritime Safety Administration of the P.R. of China, to improve the simulation accuracy near the coastal region.

Figure 2.2: Grid (a) and distribution of water level observation stations (b) near the Jiangsu coast.

The open boundaries of this large-scale model are set at the deep-sea areas, suffi-ciently far away from the Jiangsu coast. In total, there are three open boundaries in-cluded in this model (Figure2.1). At these boundaries, the water level is prescribed by 13 astronomical tidal components (M2, S2, N2, K2, K1, O1, P1, Q1, MF, MM, M4, MS4 and MN4), obtained from a global scale ocean model — TPXO7.2 model (available at

http://volkov.oce.orst.edu/tides/global.html). Harmonic components of the tidal gauges near open boundaries (e.g. Ryukyu Islands), collected from the Interna-tional Hydrographic Organization (IHO) tidal dataset, are used to correct the data ob-tained from the TPXO7.2 model. Then, the water levels at open boundaries are com-puted by: ς(t) = A0+ k X i =1 AiFicos(ωit + (V0+ u)i−Gi), (2.1)

where,ς(t) is the water level at time t; A0is the mean water level over a certain period; k

is the number of relevant constituents; i is the index of a constituent; Ai is the local tidal

amplitude of a constituent; Fiis the nodal amplitude factor;ωiis the frequency; (V0+u)i

is the astronomical argument; Giis the local phase lag.

2

relatively deep, tide generating force is included in this model and calculated based on the tidal potential of the equilibrium tide and the earth tide (Schwiderski,1980;Deltares,

2011). For the equilibrium tide, the tidal potential is decomposed into the series: ϕ = X

ν=0,1,2ϕv

(λ,φ,t), (2.2)

where,ν = 0,1,2 refers to long-period species, diurnal species and semi-diurnal species, respectively:

ϕ0= Ki(1.5 cos2φ − 1)cos(ωit + 2λ + χi), (2.3)

ϕ1= Kisin(2φ)cos(ωit + λ + χi), (2.4)

ϕ2= Kicos2φcos(ωit + 2λ + χi), (2.5)

where,λ, φ are geographical co-ordinates, Kiis the amplitude of the equilibrium tide;ωi

is the frequency of the equilibrium tide;χi is the astronomical argument of the

equilib-rium tide (relative to GMT); t is the universal standard time. Then, the equilibequilib-rium tidal potentialϕ is corrected by the theory of the earth tide to obtain the net tidal potential. In the model, the tide generating force of 10 tidal constituents (M2, S2, N2, K2, K1, O1, P1, Q1, MF, and MM) is considered, because all other constituents fall below 4% of the dominating semi-diurnal (M2) tide (Dietrich,1963;Schwiderski,1980).

Since there are many rivers flowing into the domain of the Bohai Sea, the Yellow Sea and the East China Sea (Figure2.1), the 14 rivers along the coastline of China and Ko-rea are considered in this model. Multi-year averaged monthly discharges of these rivers are collected from database of Global River Discharge (Vorosmarty et al.,1998). These monthly discharges of different rivers are prescribed in the model as the river bound-aries.

For the shallow water areas (e.g. tidal flats) inside the domain of the Bohai Sea, the Yellow Sea and the East China Sea, the drying and flooding processes are considered in this model. The main scheme considering the drying and flooding processes is to conduct a series of checks on both water level points and velocity points (due to the use of staggered grids) before computation in each time step. For water level points, we used the maximum value (dmax) of four surrounding depth points as the water depth.

At a velocity point, the bed level is the arithmetic average of the value specified in the vertices of the cell (dmean). For the water levels at a velocity point, a so-called upwind

method is applied (see the Delft 3D–FLOW Manual,Deltares,2011, for more details). Then, whether computational grids are wet or dry is determined by comparing the total water depth at the water level point and the velocity point with the threshold depthδ (in

2

this study_{δ=0.01 m).}

The model was started at zero initial conditions (cold start), running three months (from 1 July to 1 October 2006), which cover the field investigation periods near the Jiangsu coast. The first 7 days are considered as a spin-up period and omitted in the analysis. The simulation time frame is referred to as the Universal Time Coordinated (UTC) with the time step of 1 minute.

2.2.2.

M

ODEL PERFORMANCE

VERIFICATION OF THE WATER LEVEL

I

Nthe Bohai Sea, the Yellow Sea and the East China Sea model, we setup 190 water level observation stations along the land boundaries (colored circles in Figure2.1). The computed water levels at these stations are verified by the measured values in the form of harmonic constants. The harmonic constants of water level stations are col-lected from the International Hydrographic Organization (IHO) tidal dataset, the Ad-miralty Tidal Tables (Hydrographer of the Navy,1979) and published papers byZhang

(2005) andTao et al.(2011b). These points are numbered (from 1 to 190) and divided into 10 groups (A–J) in a clockwise direction for a clear illustration. Note that the phase lag in this study refers to UTC.

To graphically illustrate the results, we present the amplitude ratio (the simulated amplitude divided by the measurement) and the phase difference (the simulated phase subtracted by the measurement) into a circular bar chart. The amplitude ratio, which is dimensionless, can show how much the amplitude is overestimated or underestimated in the model. It is worth to emphasize that the amplitude ratio is more accurate to eval-uate the model performance than the absolute amplitude error, because the same ab-solute differences may represent different agreements with the observations. Due to the limited space, only the simulated results of the M2 tide are shown in this paper (Figure

2.3). In order to estimate the deviation between observations and simulations, the Root Mean Square (RMS) values for the four major tidal components (M2, S2, K1, and O1), in terms of amplitude ratio differences and phase differences, are obtained by Eq.(2.6), respectively. R M S = s PN 1(xi− ¯x)2 N , (2.6)

where, xiis the amplitude ratio Ac/Ao(or calculated phase gc) of a tidal constituent; Ao

and Acare the observed and calculated amplitude, respectively; ¯x is 1 when calculating

2

Figure 2.3: Harmonic constants comparison of the M2 tidal constituent. The outer, middle and inner circles present the number of the observation points, amplitude ratio and phase difference, respectively. The letters ’A–J’ indicate the 10 regions of the observation points. Different colors indicate the different level of agreement between the measurements and the simulations. The color green, orange, and red indicate an error less than 0.1, 0.2 and over 0.2 for amplitude ratio, respectively, and less than 10°, 20°, and over 20° for phase difference, respectively.

phase; N is the number of data. In general, the RMS values in terms of the amplitude ratio difference and the absolute phase difference are 0.12 and 8.35°, 0.14 and 9.98°, 0.20 and 12.94°, 0.22 and 8.99° for the M2, K1, S2, O1, respectively. Given that the number of tidal gauges is almost 200, the differences can be considered relatively small. Especially, the accuracy of the M2 constituent is satisfactory. In conclusion, this model can satis-factorily reproduce the tidal wave system in the Bohai Sea, the Yellow Sea, and the East China Sea.

A detailed comparison of the harmonic constants of the M2 tidal constituent near the Jiangsu coast (Figure2.2b) is given in Table2.2. Generally, the results are in reason-able agreement with the observations. However, the relatively large deviation in several

2

points (i.e. Lüsi port) illustrates that the grid may still be too coarse to accurately simu-late the hydrodynamic conditions when the regional bathymetry is complex on smaller scales. To simulate the tide near the Jiangsu coast with more accuracy, local model with re-fined grid is necessary (see Chapter5).

Table 2.2: Comparison of the harmonic constant of the M2 tidal constituent near the Jiangsu coast.

Station Observed Calculated gc−go ac/ao ao(m) go(°) ac(m) gc(°) (°) (–) Lianyungang 1.70 310.16 1.45 309.56 −0.60 0.85 Yanwei 1.53 320.16 1.32 324.36 4.20 0.86 Sheyang estuary 0.90 58.16 0.92 53.67 −4.51 1.02 Yangkou port 2.54 137.16 2.28 143.73 6.57 0.90 Lüsi port 1.75 124.00 1.38 129.50 5.50 0.79 Lianxing 1.31 108.16 1.06 101.78 −6.38 0.81

Note: ao and go represent the observed amplitude and phase, respectively; ac and gc

represent the calculated amplitude and phase, respectively; gc−gois the phase-lag

dif-ference and ac/aois the amplitude ratio.

VERIFICATION OF THE TIDAL CURRENT

I

Nthis study, tidal currents have been verified at 22 observation stations, including 18 stations in the middle of the sea and 4 stations near the Jiangsu coast (Figure2.1). As to the area near the Jiangsu coast, due to the shallow water depth and complicated bathymetry, long-term measurements are rare. Hence, short-term field measurements (Zhang,2012) are used to validate the model. The field survey was conducted on 24 Au-gust 2006 and 31 AuAu-gust 2006, with an interval of one hour. With respect to the offshore regions, three methods are applied to evaluate the agreement of long-term simulation. First, asChen(2008) suggested, we validate computed tidal current at several tidal cur-rent observation stations in the middle of the sea (red crosses in Figure2.1) by the pre-scribed time-series values in the Tidal Tables (National Marine Data and Information Service,2005). Second, we validate some other observation stations (M5, MS, CM7, B, D, I, M4, ADCPa, and ADCPc, see Figure2.1) with tidal current harmonic constants. The data are based on quasi-long-term tidal current measurements (Larsen et al.,1985;Guo & Yanagi,1998;Teague et al.,1998). In addition, field measurements carried out by the Ocean University of China at station NY (see Figure2.1) are selected to validate the con-tinuous time-series results of our model (for more details on the measurements refer to

Song et al.(2013a) andZhu(2009).

Model validation against time series data is shown in Figure2.4. Figure2.4a shows good agreement between the simulations and the predicted values from the Tidal

Ta-2

Figure 2.4: Validation of the tidal currents velocity.

bles in the Lianqingshi Fishing Port (123.51°E, 33.50°N). For stations near the Jiangsu coast (short-term validation), the simulated result (Figure2.4b) at R11 station (120.99°E, 33.69°N) is a little different from the measured values, which is difficult to simulate in detail because the local complex bathymetry cannot be well represented by a large-scale model. Figure2.4c and2.4d show the comparisons between the time-series observation data at station NY (123.50°E, 38.00°N) with the model simulations in two directions. The results reveal that our model can capture the main feature of the tidal current reasonably

2

in both directions. As to the harmonic constants of tidal current, comparisons are listed in Table2.3(taken the M2 and S2 constituents as examples). At most stations, the ratio of the velocity magnitude is almost 1, indicating a high degree of satisfaction. The rela-tively large deviation of phase at some stations (e.g. station D, MS and ADCPc) may be due to the large water depth, where tide induced current is only a part of the local cur-rent. In general, our model results are still reasonable. The present model can be used to simulate the tidal wave and tidal current in this large area.

Table 2.3: Comparisions between the simulated and observed M2 and S2 constituents of the tidal currents.

Station LON (°) LAT (°) M2 S2 U V U V Hc/Ho (–) ∆G (°) Hc/Ho (–) ∆G (°) Hc/Ho (–) ∆G (°) Hc/Ho (–) ∆G (°) M5 124.50 32.00 1.23 5.16 1.09 8.84 1.50 13.13 1.24 10.96 MS 124.80 30.52 1.10 14.26 0.96 18.53 1.67 8.36 1.65 0.23 CM7 125.45 28.65 1.09 1.96 0.98 3.77 0.88 2.73 1.00 5.55 B 124.08 36.95 1.07 0.85 0.93 0.99 1.00 2.14 0.88 3.94 D 124.58 36.00 1.21 16.13 1.03 20.79 0.99 17.97 1.02 18.72 I 124.69 34.3 0.89 2.07 1.01 10.39 0.85 4.24 1.07 9.72 M4 122.82 31.25 0.72 6.98 0.99 0.62 0.86 8.32 0.97 13.40 ADCPa 125.25 33.27 1.00 2.13 0.93 −2.07 1.10 13.52 1.00 8.96 ADCPc 124.03 35.97 1.23 10.23 0.99 −13.3 1.20 2.51 1.14 −1.59 Note: Hoand Gorepresent the observed magnitude and phase of tidal current,

respec-tively; Hc and Gc represent the calculated magnitude and phase of the tidal current,

respectively;∆G=Gc−Go ,is the phase-lag difference and Hc/Hois the magnitude ratio.

LON denotes the longitude; LAT means the latitude.

2.3.

R

ESULTS AND INFLUENCE FACTOR ANALYSES

2.3.1.

C

O

-

TIDAL CHARTS AND TIDAL CURRENT FIELD

T

IDALwave motions are usually presented in the form of co-tidal charts. The sim-ulated co-tidal charts of eight primary tidal constituents by the model are in good agreement with the co-tidal charts in the Marine Atlas of the Bohai Sea, Yellow Sea and East China Sea (Atlas of the Oceans Editorial Board,1992) and those in the references

2

(e.g. Fang 1986;Kang et al. 1998). Figure2.5shows the co-tidal charts of the M2 and K1 tidal constituents simulated by the model. Additionally, the simulated position of the amphidromic point of the M2 tide in the south Yellow Sea (121.55°E, 34.80°N) is very close to the location (121.68°E, 34.60°N) suggested by previous numerical results (Xing et al.,2012;Chen et al.,2013).

Figure 2.5: Simulated co-tidal chart of the M2 (a) and K1 (b) tidal constituents. Black lines denote co-phase lines refering to the UTC. The same notations are used in the following figures.

Figure2.6a shows the simulated tidal ellipses of the M2 tidal constituent with a ro-tational direction. It is clear that a radial tidal current field exists at the central Jiangsu coast, and Jianggang is the focal point of the converging and diverging currents. Note that a clockwise flow pattern occurs in the radial sand ridge field, whereas a counter-clockwise flow pattern appears in the north of the Jiangsu coast. As tide-induced resid-ual currents always play an important role in the marine processes, especially in the shallow coastal regions (Lie,1999;Lee et al.,2011;Sanay et al.,2007), the Eulerian tidal-induced residual current of the M2 tide is also computed and shown in Figure2.6b. It can be found that the direction of the residual current forms a clockwise circulation pat-tern over the sand ridges. The direction of the residual currents is mainly toward the ridges (landward) in several tidal channels, which indicates the accretion of the radial sand ridges.