Activiteitsrapport 1 (1 april december 2008) Michael Fettweis, Dries Van den Eynde en Federico Maggi

Transcriptie

1 BEHEERSEENHEID VAN HET MATHEMATISCH MODEL VAN DE NOORDZEE SUMO GROEP MOnitoring en MOdellering van het cohesieve sedimenttransport en evaluatie van de effecten op het mariene ecosysteem ten gevolge van bagger- en stortoperatie (MOMO) Activiteitsrapport 1 (1 april december 28) Michael Fettweis, Dries Van den Eynde en Federico Maggi MOMO/4/MF/294/NL/AR/1 Voorbereid voor Afdeling Maritieme Toegang, Departement Mobiliteit en Openbare Werken, Ministerie van de Vlaamse Gemeenschap, contract MOMO BMM 1 Gulledelle B 12 Brussel België

2 Inhoudstafel 1. Inleiding Voorwerp van deze opdracht Algemene Doelstellingen ( ) Verminderen van de sedimentatie Efficiënter storten Taken (april 28 december 29) Publicaties (april 28 december 29) 6 2. Stormen en SPM concentratie ter hoogte van Zeebrugge: natuurlijke processen en menselijke impact 8 3. Flocculatiemodel BFLOC Beschrijving van 1D BFLOC Calibratie Gevoeligheid van BFLOC Implementatie in een 2D sedimenttransportmodel Uitbreiding naar 2D Resultaten 2D simulatie Conclusies Storminvloed op SPM concentratie Flocculatiemodel Referenties 22 Appendix 1: Fettweis, M., Francken, F. Van den Eynde, D. Persistent high SPM concentration layers in a coastal turbidity maximum area. 7 th Int. Conf. on Tidal Environments (Tidalites-28), september, Qingdao, China. Appendix 2: Fettweis, M. Uncertainty of excess density and settling velocity of mud flocs derived from in situ measurements. Particles in Europe, European workshop on the in situ measurements of particle size and concentration in the aquatic environment october 28, Bologna, Italië. Appendix 3: Doxaran, D., Babin, M., Ruddick, K., Park, Y., Fettweis, M. Tidal variations of particle size distribution and optical properties in turbid coastal waters. Perspectives for ocean colour remote sensing. Particles in Europe, European workshop on the in situ measurements of particle size and concentration in the aquatic environment october 28, Bologna, Italië. Appendix 4: Fettweis, M., Francken, F., Van den Eynde, D., Verwaest, T., Janssens, J., Van Lancker, V. Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea). submitted to Marine Geology 2

3 1. Inleiding 1.1. Voorwerp van deze opdracht MOMO staat voor monitoring en modellering van het cohesieve sedimenttransport en de evaluatie van de effecten op het mariene ecosysteem ten gevolge van bagger- en stortoperatie. Het MOMO-project maakt deel uit van de algemene en permanente verplichtingen van monitoring en evaluatie van de effecten van alle menselijke activiteiten op het mariene ecosysteem waaraan België gebonden is overeenkomstig het Verdrag inzake de bescherming van het mariene milieu van de noordoostelijke Atlantische Oceaan (1992, OSPAR-Verdrag). De OSPAR Commissie heeft de objectieven van haar huidig Joint Assessment and Monitoring Programme (JAMP) gedefinieerd tot 21 met de publicatie van een holistisch quality status report Noordzee en waarvoor de federale overheid en de gewesten technische en wetenschappelijke bijdragen moeten afleveren ten laste van hun eigen middelen. De menselijke activiteit die hier in het bijzonder wordt beoogd, is het storten in zee van baggerspecie waarvoor OSPAR een uitzondering heeft gemaakt op de algemene regel alle stortingen in zee zijn verboden (zie OSPAR-Verdrag, Bijlage II over de voorkoming en uitschakeling van verontreiniging door storting of verbranding). Het algemene doel van de opdracht is het bestuderen van de cohesieve sedimenten op het Belgisch Continentaal Plat (BCP) en dit met behulp van zowel numerieke modellen als het uitvoeren van metingen. De combinatie van monitoring en modellering zal gegevens kunnen aanleveren over de transportprocessen van deze fijne fractie en is daarom fundamenteel bij het beantwoorden van vragen over de samenstelling, de oorsprong en het verblijf ervan op het BCP, de veranderingen in de karakteristieken van dit sediment ten gevolge van de bagger- en stortoperaties, de effecten van de natuurlijke variabiliteit, de impact op het mariene ecosysteem in het bijzonder door de wijziging van habitats, de schatting van de netto input van gevaarlijke stoffen op het mariene milieu en de mogelijkheden om deze laatste twee te beperken. Een samenvatting van de resultaten uit de voorbije perioden (22-24, en 26-28) kan gevonden in het Syntheserapport over de effecten op het mariene milieu van baggerspeciestortingen (Lauwaert et al., 24; 26; 28) dat uitgevoerd werd conform art. 1 van het K.B. van 12 maart 2 ter definiëring van de procedure voor machtiging van het storten in de Noordzee van bepaalde stoffen en materialen. Voor een uitgebreide beschrijving wordt verwezen naar de halfjaarlijkse rapporten Algemene Doelstellingen ( ) Het onderzoek uitgevoerd in het MOMO project kadert in de algemene doelstelling om de baggerwerken op het BCP en in de kusthavens te verminderen, door enerzijds de sedimentatie te verminderen in de baggerplaatsen en anderzijds efficienter te storten. Een nauwe samenwerking tussen de BMM en het WLH is één van de vereisten om de doelstelling te kunnen realiseren. 3

4 Verminderen van de sedimentatie Vermindering van de sedimentatie zal kunnen bereikt worden door: een optimalisering van de vorm van de buitenmuur of een Current Deflecting Wall, zodat de wateruitwisseling tussen haven en zee vermindert. een aanpassing van de vorm van de toegangsgeul (bv verbreding, zachtere helling ) Efficiënter storten De efficiëntie van een stortplaats wordt bepaald door fysische (sedimenttransport i.f.v. getij, doodtij-springtij, wind, golven), economische en ecologische aspecten. Bij een efficiënte stortplaats is de recirculatie van het gestorte materiaal naar de baggerplaatsen zo klein mogelijk, is de afstand tussen bagger- en stortplaats minimaal en is de verstoring van het milieu verwaarloosbaar. Hieruit volgt dat er geen stortplaats kan bestaan die onder alle omstandigheden efficiënt is. Efficiënt storten zal kunnen betekenen dat in functie van de voorspelde fysische (wind, stroming, golven, sedimenttransport, recirculatie), economische (afstand, grootte baggerschip) en ecologische aspecten op korte termijn een stortlocatie zal worden gekozen. Om dit te bereiken is het volgende nodig: definiëren van een goede stortzones i.f.v. sedimenttransport, recirculatie baggerspecie, ecologie, economie, bathymetrie van de baggerplaatsen operationele voorspelling van de recirculatie van het gestorte materiaal door de operationele data uit hydrodynamische en sedimenttransportmodellen, real time meetstations, satellietbeelden, bathymetrie van de baggerplaatsen te integreren zodat een efficiënte stortlocatie kan bepaald worden Taken (april 28 december 29) Taak 1: Monitoring Taak 1.1: Slibconcentratie metingen: getijcyclus en langdurig Tijdens 4 meetcampagnes per jaar met de R/V Belgica zullen 13-uursmetingen uitgevoerd worden. De metingen zullen plaatsvinden in het kustgebied van het BCP. Tijdens de metingen zullen tijdsreeksen worden verzameld van de stromingen, de concentratie aan en de korrelgrootteverdeling van het suspensiemateriaal (LISST 1), de temperatuur en de saliniteit. De optische metingen (transmissometer, Optical Backscatter Sensor) zullen gecalibreerd worden met de opgemeten hoeveelheid materie in suspensie (gravimetrische bepalingen na filtratie) om te komen tot concentraties. Stalen van suspensiemateriaal zullen genomen worden met de centrifuge om de samenstelling ervan te bepalen. De tripode zal ingezet worden om stromingen, slibconcentratie, korrelgrootteverdeling van het suspensiemateriaal, saliniteit en temperatuur te meten gedurende een lange periode. Doel van deze metingen is het slibtransport in het gebied beter te begrijpen zodat een verband tussen de aanslibbing in de haven en de slibconcentratie erbuiten gelegd kan worden. Gelijkaardige metingen werden in de voorbije jaren reeds uitgevoerd ter hoogte van Blankenberge en de meetpaal MOW1; deze toonden aan dat zich regelmatig dicht tegen de bodem een laag met hoge slibconcentratie (>3g/l) heeft gevormd. Deze laag ontstaat door stormen en kan tot enkele dagen na de storm worden waargenomen. 4

5 Sub-taak 1.2: Monitoring voor proefproject Verlagen topslib in haven van Zeebrugge Een verankerd meetstation (tripode) werd reeds op 28 januari 28 geplaatst ter hoogte van het strand van Blankenberge in het kader van het proefproject verlagen van topslib in haven van Zeebrugge. De meting zal doorgaan tot minstens 14 dagen na het einde van het proefproject, dat voorzien is tegen 15 mei 28. Doel van deze verankering is het monitoren van de slibconcentratie, het onderzoeken of een verpomping van vloeibaar slib over de havendam een verhoging van de natuurlijke slibconcentratie dicht tegen het strand veroorzaakt en het slibtransport in het gebied beter te begrijpen. De invloed van een verpomping over de havendam zal worden geschat door de natuurlijke variabiliteit van de slibconcentratie (voor en na het proefproject) statistisch te vergelijken met de variabiliteit tijdens het verpompen. De aanwezigheid van een hooggeconcentreerde sliblaag dicht tegen de bodem, zoals werd waargenomen tijdens vorige tripode verankeringen, kan een significant effect hebben op de aanslibbing in de haven van Zeebrugge. Het is daarom belangrijk om te weten of tijdens het proefproject een dergelijk bodemlaag aanwezig is of zich heeft gevormd ten gevolge van een storm of van de verpomping zodat een verband tussen de aanslibbing in de haven en de slibconcentratie erbuiten gelegd kan worden. Deze informatie is nodig om een uitspraak over de efficiëntie van het verpompen over de westelijke havendam te formuleren. Sub-taak 1.3: Quasi permanent meetstation Uit vorig onderzoek bleek dat een verdubbeling van de tripode de meest haalbare optie is voor een (quasi) continu registrerend meetstation. Het aankoopdossier voor een tweede tripode werd opgestart, er wordt voorzien dat de tripode gedurende 28 operationeel zal zijn. Eerste testen zullen worden uitgevoerd en de meetdata zullen worden geëvalueerd. Sub-taak 1.4: Slibverdeling op de bodem Bodemstalen zullen worden geanalyseerd om de korrelgrootte, het kalkgehalte en de organische fractie te bepalen. Bij de box core bodemstalen zal ook de densiteit bepaald worden. Met deze data kan de slibverdeling in de kustzone verder verfijnd worden en zal onderscheid gemaakt kunnen worden tussen actief slib (i.e. slib dat in de cyclus van afzetting en resuspensie is betrokken) en inactieve slib (oude lagen die dagzomen en enkel eroderen tijdens extreme situaties). Een gedetailleerde kennis van de samenstelling van de zeebodem is belangrijk voor een nauwkeurige kwantificering van de erosiefluxen in sedimenttransportmodellen. Taak 2: Modellering Sub-taak 2.1: Verfijnen slibtransportmodel Het gebruik van een numeriek sedimenttransportmodel vereist een regelmatige validatie van de modelresultaten met meetgegevens en een aanpassing van parameterwaarden. Er zal verder gewerkt worden aan een integratie van de resultaten uit het flocculatie onderzoek uit 27 in het 2D sedimenttransportmodel. Sub-taak 2.2: Alternatieve stortschema s i.f.v. omgevingsfactoren Onderzoek naar alternatieve stortschema s (getijgebonden, enkel bij bepaalde 5

6 windrichting, ) zal uitgevoerd worden voor stortplaats B&W Zeebrugge Oost conform de aanbevelingen voor de minister geformuleerd in het syntheserapport 28. Het doel is om het effect van het getij (meteo) op de retourstroom naar de baggerplaatsen te bepalen en om de effecten van een verplaatsing van de stortplaats te evalueren. Overleg zal gebeuren met amt over mogelijke locaties van de stortplaats. Afhankelijk van de uitkomst kan een proefproject worden opgestart waarin monitoren en evalueren van de effecten van het storten van baggerspecie op de nieuwe stortplaats centraal staan. Sub-taak 2.2.1: 2D langdurige simulaties Simulatie van de effecten van getijgebonden storten op langdurige schaal zullen uitgevoerd worden met het 2D hydrodynamisch en sedimenttransportmodel van het BCP. Resultaten zullen per seizoen en doodtij/springtij worden geïnterpreteerd. Deze simulaties laten ook toe om eventuele effecten op de slibhuishouding ter hoogte van de Belgische kust tengevolge van een consequent getijgebonden storten of een verplaatsing van de stortplaats gedurende een langere periode (1 jaar) te analyseren. Sub-taak 2.2.2: 3D korte termijn simulaties 3D simulaties van getijgebonden storten zullen worden uitgevoerd met het detailmodel van het gebied rond Zeebrugge. Het hydrodynamische model zal hiervoor moeten worden uitgebreid met een slibtransportmodule, zodat de recirculatie van baggerspecie vanuit de stortplaats naar de buitenhaven van Zeebrugge kan worden berekend. Resultaten van deze simulaties zullen toelaten om een beter inzicht te bekomen in de recirculatie van het baggermateriaal en dus ook in de economische meerwaarde die een relocalisatie van de stortplaats zal opleveren Publicaties (april 28 december 29) Er werden volgende rapporten en publicaties opgesteld in het kader van het MOMO project: Activiteits-, Meet- en Syntheserapporten Fettweis, M., Van den Eynde, D., Maggi, F. 29. MOMO activiteitsrapport 1 (april 28 december 28). BMM-rapport MOMO/4/MF/293/NL/AR/1, 24pp + app. Conferenties/Workshops: Fettweis, M., Francken, F. Van den Eynde, D. 28. Persistent high SPM concentration layers in a coastal turbidity maximum area. 7 th International Conference on Tidal Environments, Tidalites- 28, september, Qingdao, China. (zie appendix 1) Fettweis, M. 28. Uncertainty of excess density and settling velocity of mud flocs derived from in situ measurements. Particles in Europe, European workshop on the in situ measurements of particle size and concentration in the aquatic environment october, Bologna, Italië. (zie appendix 2) Doxaran, D., Babin, M., Ruddick, K., Park, Y., Fettweis, M. 28. Tidal variations of particle size distribution and optical properties in turbid coastal waters. Perspectives for ocean colour remote sensing. Particles in Europe, European workshop on the in situ measurements of particle size and concentration in the aqua- 6

7 tic environment october, Bologna, Italië. (zie appendix 3) Publicaties (tijdschriften, boeken) Fettweis, M., Francken, F., Van den Eynde, D., Verwaest, T., Janssens, J., Van Lancker, V. Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea). Paper submitted to Marine Geology (zie appendix 4) 7

8 2. Stormen en SPM concentratie ter hoogte van Zeebrugge: natuurlijke processen en menselijke impact Sinds 23 wordt bij de BMM een verankerd meetstation (tripode) ingezet om de sedimentconcentratie en het sedimenttransport in de Belgische kustzone te meten. De resultaten tot februari 27 kunnen worden gevonden in Backers (24), Backers & Van den Brande (25, 26, 27a, 27b) en Fettweis et al. (27). De meeste meetdata werden verzameld in de nabijheid van Zeebrugge (MOW1, Blankenberge, MOW). De data werden geanalyseerd om de effecten van stormen op sedimenttransport en concentratie te identificeren. De resultaten werden voorgesteld op de Tidalites 28 conferentie en gebundeld in een paper (zie appendices 1 en 4). Hieronder volgt een korte samenvatting. Een multi-sensor tripode werd ingezet in het turbiditeitsmaximum ter hoogte van Zeebrugge om stormeffecten op de slibconcentratie (Suspended Particulate Matter, SPM) dicht tegen de bodem te bestuderen. De data hebben aangetoond dat tijdens of na een storm de SPM concentraties significant toenemen en dat hoog geconcentreerde slibsuspensies (High Concentrated Mud Suspensions, HCMS) gevormd worden. Er werd vastgesteld dat tijdens stormen ongeveer drie keer meer suspensiemateriaal in het water is dan gedurende kalme perioden. Verschillende bronnen van fijnkorrelige sedimenten die een invloed hebben op de SPM concentratie werden onderzocht. Dit zijn de windrichting en advectie van water; de recente gebeurtenissen (vb. een vorige storm, langdurige perioden met eenzelfde windrichting), het voorkomen van fluffy lagen, vers afgezet slib nabij de stortplaatsen, vaargeulen en aangrenzende gebieden en de erosie van medium geconsolideerd slib van Holoceen ouderdom. Erosieproeven werden door de Universiteit Stuttgart i.o. van het Waterbouwkundig Laboratorium uitgevoerd op verschillende sedimentstalen. Dit onderzoek gebeurde in het kader van het Belspo project QUEST4D. Hierdoor kon een schatting bekomen worden van de kritische erosieschuifspanning van de verschillende types van cohesieve sedimenten die in de Belgische kustzone voorkomen. De resultaten tonen aan dat de meeste slibafzettingen niet kunnen worden geërodeerd door getijstromingen, maar dat hogere schuifspanningen nodig zijn. Deze kunnen optreden tijdens stormen met hoge golven. De meetdata suggereren dat significante hoeveelheden fijn sediment in suspensie moet zijn om zeer hoge SPM concentraties dicht tegen de bodem te bekomen. De baggerstortplaatsen, de vaargeulen en de aangrenzende gebieden met vers afgezet slib zijn de belangrijkste bronnen voor het suspensiemateriaal tijdens stormen. Dit resultaat is belangrijk omdat aangetoond werd dat menselijke activiteiten (bouw van Zeebrugge, aanleg en verdieping van vaargeulen, storten van baggerspecie) significant hebben bijgedragen tot de vorming van hooggeconcentreerde sliblagen in de omgeving van Zeebrugge. 8

9 3. Flocculatiemodel BFLOC In een vorig MOMO rapport werd een overzicht gegeven van enkele bestaande flocculatiemodellen (Fettweis et al., 28a). De valsnelheid berekent met de modellen werd vergeleken met de valsnelheden afgeleid uit de in situ metingen (tripode data). De conclusie was dat een betere weergave van de fysische en biologischfysische processen nodig is in een flocculatiemodel om een betere overeenkomst te bekomen tussen metingen en modelresultaten. Er werd geopteerd om het 1D bioflocculatiemodel BFLOC (Maggi, subm) te implementeren. Het model is gebaseerd op het Winterwerp (1998) flocculatiemodel (zie ook Winterwerp et al., 26 voor een toepassing ervan in de Zeeschelde) en werd uitgebreid met een biologische componente. Het model werd door Maggi gecalibreerd met metingen van het BCP. Hieronder volgt een korte beschrijving van het model en de implementatie in het sedimenttransportmodel. De beschrijving werd overgenomen uit het manuscript van Maggi Beschrijving van 1D BFLOC Het suspensiemateriaal (SPM) kan worden opgedeeld in een minerale (1-Ω) en een organische fractie (Ω). Beide fracties bezitten cohesieve eigenschappen en beïnvloeden de groei (aggregatie) en het uiteenvallen van vlokken. Het volume aan vaste bestanddelen in een vlok (V) wordt geschreven als de som van het volume aan minerale (V m ) en organische bestanddelen (V b ): V = V + V = 1 ζ V + ζ (1) M B ( ) V met ζ=v b /V is de fractie aan organische bestanddelen. De verandering in de tijd van V kan geschreven worden als: dv dt dv dt dv dt M B = + (2) Indien we uitgaan van een fractale geometrie van de vlokken dan kan V geschreven worden als: nf nf 3 V = D f D (3) p met D f de vlokgrootte, D p de grootte van de primaire partikels en nf de fractale dimensie. De verandering van het minerale volume in een vlok in functie van de tijd is: dvm dvm dv dd f = M M (4) dt dv dd dt waarbij f dv M dv dv dd f = 1 ζ D nf 1 f = nf nf 3 D p (4a) (4b) 9

10 met dd dt k c ( 1 Ω) G k ( 1 ζ ) G 3 / 2 f a b 2 M M = D nf 4 nf 3 f (4c) D f ( D f D p ) ε ν = ν λ 2 G = een disspatieparameter (turbulente schuifspanning) en ε de dissipatiesnelheid per eenheid massa, ν de kinematische viscositeit en λ de Kolmogorov turbulentieschaal. Vergelijking 4c komt overeen met het flocculatiemodel van Winterwerp (1998). De termen (1-Ω) en (1-ζ) werden toegevoegd om rekening te kunnen houden met de aggregatie van de minerale fractie van het SPM en het verlies aan minerale fractie tijdens het uiteenvallen van de vlok. De parameters k a en k b beschrijven aggregatie en het uiteenbreken van de vlokken, ze kunnen als volgt worden weergegeven: k k ' 1 a = ( + ζ ) 3 p 1 ka (5a) nf D nf ρ ' 1 b = ( 1+ ) kb 3 nf D p ' ' k a k b.5 μ ζ (5b) nf Fy met en niet-dimensionale calibratieparameters, μ de dynamische viscositeit van het water, F y de vloksterkte en ρ = ζ ρ B + (1 + ζ) ρ M de gemiddelde densiteit van de vaste bestanddelen in een vlok (densiteit van de organische: ρ B, en minerale: ρ M bestanddelen). De term (1+ζ) werd toegevoegd om rekening te kunnen houden met de levende organismen in de vlokken die organische kleverige substanties (EPS en TEP) aanmaken en zowel de cohesie van de vlok vergroten als ook de kans tot aggregatie verhogen. De verandering in biomassa in functie van de tijd wordt beschreven met behulp van een vrij eenvoudig model als dvb dvb dv dd f dvb = M B + MGD (6) dt dv dd dt dt met dv B dv dv dd f dd f dt dv dt B = ζ D f nf 1 f = nf nf 3 D p 3 / 2 ka c ΩG kb ζ G 2 B = D nf 4 nf 3 f D f ( D f D p ) (6a) (6b) M (6c) V = B M GD ηvb 1 (6d) K De laatste term beschrijft de verandering van de biomassa tengevolge van de groei en het afsterven van cellen. Hierin is η de groeisnelheid van biomassa en K de biomassacapaciteit van de vlok. De groeisnelheid wordt uitgedrukt als C η = ˆ η (7) K C' m + 1

11 met ηˆ de maximale specifieke groeisnelheid, K m de half-verzadigingsconcentratie en C de concentratie aan nutriënten. K wordt verondersteld proportioneel te zijn met het poriënvolume, V p, van de vlok: 3 K = β V = β D V (8) p ( ) f waarbij β de proportionaliteitsfactor is. De differentiaalvergelijkingen 4 en 6 kunnen numeriek worden opgelost voor de vlokgrootte D f met behulp van een expliciete eindige-differentiemethode. Met behulp van de Winterwerp (1998) formulering voor de effectieve densiteit van vlokken: 3 nf D p Δρ = ( ρ ρ w ) (9) D f kan de valsnelheid van de vlokken, w s, worden berekend: nf 1 a ( ρ ρ ) D w 3 nf f ws = g D p (1) b μ Re met g de gravitatieversnelling, Re het vlok Reynolds getal (= w s D f ν -1 ) en a en b vormparameters van de vlok Calibratie Het model werd in een 1D versie door Maggi (submitted) afgeijkt op vlokgrootte. Hierbij werden metingen van SPM concentratie en partikelgrootte ter hoogte van Zeebrugge (MOW1: 13-uursmeting 8-9/9/23; 51 N 22.48, 3 E 7.99 ) en op de Kwintebank (tripode, 2-11/3/24; 51 N 18.19, 2 E 4.3 ) gebruikt, zie Fettweis et al. (26, 27) en figuren 3.1 en 3.2. De gebruikte parameterwaarden kunnen in tabel 3.1 gevonden worden. De turbulente schuifspanning werd berekend met een toepassing van het 3D hydrodynamisch model COHERENS v1 op het BCP (OPTOS-BCP). OPTOS-BCP omvat het gebied tussen 51 N en N in latitude en tussen 2.8 E en 4.2 E in longitude. De horizontale resolutie bedraagt.71 (longitude) en.42 (latitude), wat ongeveer overeenstemt met een roosterafstand van 8 m. In het verticaal vlak wordt gerekend met 2 sigma lagen. Randvoorwaarden zijn waterstand en dieptegemiddelde stroomsnelheid, deze worden voorzien door het OPTOS-NOS model, dat ook gebruik maakt van de COHERENS code, maar de hele Noordzee en een deel van het Kanaal omvat. Meteorologische data zijn afkomstig van het UK Meteorological Office. In OPTOS-BCS wordt turbulente kinetische energie berekend met een transportvergelijking (k-ε turbulentiemodel) en de mixing length met een algebraïsche formulering (Mellor en Yamada, 1974). De met behulp van vergelijkingen (9) en (1) berekende valsnelheid en effectieve densiteit wordt weergeven in figuren 3.5 en 3.6. De resultaten van valsnelheid tonen zeer duidelijk aan dat de gebruikte methode heel gevoelig is aan de berekende (gekozen) fractale dimensie. Uit de metingen werd in Fettweis et al. (27) een fractale dimensie van 2.59 afgeleid voor de Kwintebank tijdserie. Door deze waarde te gebruiken i.p.v. een fractale dimensie van 2, bekomt men valsnelheden die ongeveer een factor 1 groter zijn. Dit toont dat kleine variaties in gemeten parameterwaarden kunnen leiden tot grote verschillen in valsnelheid en onderstreept het belang van een nauwkeurige inschatting van de fractale dimensie. 11

12 14 12 Particle size Kolmogorov scale Particle size (micro m) Kolmogorov turbulence scale (micro m) SPM concentration (g/l) Currents Water depth Current velocity (m/s) Water depth (m) Time (hours) 6 Figuur 3.1: MOW1 (8-9/9/23, 23-22)). Boven: gemiddelde partikelgrootte en Kolmogorov turbulentie schaal (3D model). Midden: SPM concentratie. Onder: Waterdiepte en verticaal gemiddelde stroomsnelheid (3D model). Nul op de x-as komt overeen met de start van de meting op 8/9/23 18h. 12

13 6 Particle size Kolmogorov scale Particle size (micro m) Kolmogorov turbulence scale (micro m) SPM concentration (g/l) Currents Water depth Current velocity (m/s) Water depth (m) Time (hours) 12 Figuur 3.2: Kwintebank (2-11/3/24). Boven: gemiddelde partikelgrootte en Kolmogorov turbulentieschaal (3D model). Midden: SPM concentratie. Onder: Waterdiepte en verticaal gemiddelde stroomsnelheid (3D model). Nul op de x-as komt overeen met de start van de meting op 2/3/23 15h1. 13

14 125 Model biomass-free Model biomass affected Measurement 1 Particle size (micro m) Time (hours since start) Figuur 3.3: MOW1 (8-9/9/23, 23-22). Vergelijking tussen de gemeten en gesimuleerde (1D BFLOC) vlokgrootte (zonder en met biomassa). Nul op de x-as komt overeen met de start van de meting op 8/9/23 18h Model biomass-free Model biomass affected Measurement 35 Particle size (micro m) Time (hours since start) Figuur 3.4: Kwintebank (2-11/3/24). Vergelijking tussen de gemeten en gesimuleerde (1D BFLOC) vlokgrootte (zonder en met biomassa). Nul op de x-as komt overeen met de start van de meting op 2/3/23 15h1. 14

15 .15 Model biomass-free Model biomass-affected Derived from measurement Fall velocity (mm/s) Model biomass-free Model biomass-affected Derived from measurement 6 Excess density (kg/m**3) Time (hours since start) Figuur 3.5: MOW1 (8-9/9/23, 23-22). Valsnelheid (boven) en effectieve densiteit (onder) berekend met het 1D BFLOC model en afgeleid uit de metingen (methode zie Mikkelsen & Pejrup, 21; Fettweis, 28). De berekening van de valsnelheid en de effectieve densiteit uit de metingen werd met een fractale dimensie van 2 i.p.v (Fettweis et al., 27) of 1.46 (Fettweis, 28) uitgevoerd. Nul op de x-as komt overeen met de start van de meting op 8/9/23 18h. 15

16 4.5 4 Model biomass-free Model biomass-affected Derived from measurement (nf=2.59) Derived from measurement (nf=2) Fall velocity (mm/s) Model biomass-free Model biomass-affected Derived from measurement 15 Excess density (kg/m**3) Time (hours since start) Figuur 3.6: Kwintebank (2-11/3/24). Valsnelheid (boven) en effectieve densiteit (onder) berekend met het 1D BFLOC model en afgeleid uit de metingen (methode zie Mikkelsen & Pejrup, 21; Fettweis, 28). Voor de berekening van de valsnelheid en de effectieve densiteit uit metingen werd een fractale dimensie van 2 verondersteld. De resultaten voor een fractale dimensie van 2.59 (zie Fettweis et al., 27) worden tevens getoond. Nul op de x-as komt overeen met de start van de meting op 2/3/23 15h1. 16

17 Een belangrijke conclusie uit de calibraties is dat de parameters in BFLOC plaatsafhankelijk zijn (zie tabel 3.1). Dit komt ook tot uiting in de geografische variabiliteit op het BCP van de beschikbaarheid aan organisch materiaal en de vlokgrootte Fettweis et al. (26, 27). Voor Ω werd een waarde van 4% genomen te MOW1 en 7% op de Kwintebank. Dit komt overeen met het gemiddelde percentage POC en PON in het suspensiemateriaal (Fettweis et al., 27). De nutriëntenconcentratie C is afkomstig uit Lacroix et al. (27). In deze berekening werd een gemiddelde fractale dimensie van 2 veronderstelt. Dit komt overeen met een gemiddelde waarde van de fractale dimensie, zie de waarden tussen 1.46 en 2.59 op het BCP (Fettweis, 28) en de waarden tussen 1.7 en 2.25 op andere locaties (Lick et al., 1993; Ten Brinke & Dronkers, 1993; Kranenburg, 1994; Winterwerp et al. 26). De parameters k a en k b voor aggregatie en uiteenbreken van vlokken zijn van dezelfde grootte orde dan deze vermeld in Winterwerp (1998) en Maggi (27). Verder verschillend de waarden K m en ηˆ met minder dan een factor 2 van deze uit Ploug & Grossart (1999) en Kiorboe (23). Maggi (submitted) besluit hieruit dat het model voldoende robuust is en dat de SPM dynamica voor en belangrijk deel gecontroleerd wordt door de concentratie aan organisch materiaal en de biomassa dynamica van levende organismen (bacteriën). Tabel 3.1: Parameterwaarden voor het flocculatiemodel zoals gebruikt door Maggi (submitted) voor Zeebrugge (MOW1) en de Kwintebank. Ω, C, Dp, ρm, ρb, Fy en nf werden uit de literatuur gehaald. De overige parameters werden zodanig afgeijkt, dat het verschil tussen gesimuleerde en gemeten vlokgrootte minimaal is. Zeebrugge Kwintebank (MOW1) Ω (gehalte OM in SPM) [-].4.7 C (nutriënten concentratie) [mol l -1] D p (primaire partikelgrootte) [µm] 2 6 ρ M (densiteit minerale fractie) [kg m -3 ] ρ B (densiteit organische fractie) [kg m -3 ] F y (vloksterkte) [N] nf (fractale dimensie) [-] 2 2 k [-] ' a ' k b [-] ηˆ (max specifieke groeisnelheid) [s -1 ] K m (half verzadigingsconcentratie) [mol l -1 ] β [-] V M () [mm³] V B () [mm³] Gevoeligheid van BFLOC Het effect van SPM concentratie en turbulente schuifspanning op de vlokgrootte werd door Berlamont et al. (1993) beschreven en kan worden teruggevonden in de modelresultaten. Een verhoging van de SPM concentratie vergroot de vlokgrootte, terwijl een hogere turbulente schuifspanning de vlokgrootte doet afnemen. 17

18 Maggi (submitted) heeft de invloed van de biomassa op de vlokgrootte nagegaan door het BFLOC model te calibreren voor een situatie zonder biomassa (dus Ω=, ηˆ =). In figuren worden de resultaten voor deze situaties getoond. De vergelijking met de resultaten met biomassa laten duidelijk zien dat de biomassa in het SPM een significante invloed heeft op de vlokdynamica Implementatie in een 2D sedimenttransportmodel Uitbreiding naar 2D Om het bioflocculatiemodel in een 2D (en 3D) numeriek sedimenttransportmodel te implementeren moet een geografische spreiding van parameters opgenomen worden. 1 8 (POC+PON)/SPM (%) SPM concentration (mg/l) Figuur 3.7: De verhouding (POC+PON)/SPM concentratie in functie van de SPM concentratie, de trendline is (POC+PON)/SPM (%) = SPM (R²=.49). De data geven de meetresultaten van zestien 13-uursmetingen weer die werden uitgevoerd tussen op het BCP. Blauwe en rode kruis: turbiditeitsmaximum (locaties MOW1 en 13); groene driehoek: rand van het turbisiteitsmaximum (B&W S1) en roze vierkant: offshore (Kwintebank en Hinderbank). Uit voorgaand onderzoek blijkt dat parameters zoals het gehalte aan organisch materiaal en de primaire partikelgrootte een geografische spreiding hebben. Belangrijkste organische parameter in het model is het gehalte aan organisch materiaal, Ω. In de suspensiestalen is het percentage hiervan gelegen tussen 5-12% (gemiddeld 7.5%), zie Fettweis et al. (27). De POC en PON concentratie wordt sinds 24 gemeten in de suspensiestalen tijdens 13-uursmetingen. Uit deze data blijkt dat de POC concentratie varieert tussen.1 en 25 mg/l, deze van PON tussen.5 en 3.3 mg/l en de SPM concentratie tussen 2 en 758 mg/l. Indien we de som van POC en PON concentratie als representatief voor het gehalte aan organisch materiaal beschouwen dan kan de parameter Ω berekend worden uit het verband tussen SPM en de organische fractie in het SPM, zie figuur 3.7. De som van POC en PON concentratie is niet gelijk aan het gehalte aan organisch materiaal, maar kan be- 18

19 schouwd worden als een goede indicator ervan. De data uit het turbisiteitsmaximum tonen dat het POC en PON percentage niet lineair verbonden is met de SPMconcentratie. Dit wijst op het feit dat tijdens depositie relatief meer minerale deeltjes (of vlokken met minerale bestanddelen) bezinken en afgezet worden dan vlokken met minder minerale bestanddelen of biologische deeltjes. Of anders gesteld dat de niet aan vlokken gebonden biomassa (plankton) of de lichtere vlokken met meer organisch materiaal niet (of veel minder) bezinken en afgezet worden. Daarenboven kan uit de figuur ook een geografisch gebonden variatie (turbisiteitsmaximum offshore) worden vastgesteld die ook het niet lineaire verband tussen de verhouding (POC+PON)/SPM t.o.v. SPM concentratie verklaart. De primaire partikelgrootte (D p ) werd bepaald op 5 suspensiestalen genomen op 3 verschillende locaties (Fettweis, 28). De gemiddelde partikelgrootte op de Hinderbank bedraagt 7.2±3. µm, op de Kwintebank 2.1±1.5 µm en ter hoogte van Oostende 1.1±3.7 µm. Uit de korrelgroottedata blijkt ook de minerale fractie van het SPM op de Kwintebank een iets hoger gehalte aan silicaten en calciet vertoont in vergelijking met deze uit het turbiditeitsmaximum. Er zijn op dit moment nog niet genoeg data beschikbaar om het bestaan van een geografische spreiding voldoende te onderbouwen. ' ' De parameters die via calibratie werden bepaald werden ( k a, kb, ηˆ, K m, B) worden in het 2D model berekend in functie van de SPM concentratie: onder 2 mg/l wordt de Kwintebank en boven 1 mg/l de MOW1 waarde genomen. Daartussen wordt een lineaire interpolatie toegepast. Een uitgebreide analyse van al de bestaande meetdata zal een meer nauwkeurige inschatting van deze parameters opleveren Resultaten 2D simulatie De hydrodynamische berekeningen werden uitgevoerd met een implementatie van het 3D COHERENS (Luyten et al., 1999) en het MU-STM sedimenttransport model op het Het MU-BCS rooster. Het gebruikte rooster heeft een resolutie van 42,86 in lengtegraad ( m) en 25 in breedtegraad (772 m) en omvat het Belgische Continentaal Plat. We verwijzen naar het vorige MOMO rapport (Fettweis et al., 28b) en naar voor een beschrijving van de modellen. Het MU-STM model met BFLOC gebruikt de diepte gemiddelde stroming, de waterstand en de turbulente schuifspanning G als input. De periode februari-maart 24 werd gesimuleerd. Uit figuur 3.8 blijkt dat metingen, 1D en 2D resultaten een gelijkaardig verloop kennen. De resultaten tonen dat de SPM concentratie onderschat wordt in het 2D model. Dit is de belangrijkste oorzaak voor de verschillen in partikelgrootte en valsnelheid. De onderschatting van de SPM concentratie is waarschijnlijk te wijten aan enerzijds het feit dat de metingen op 1.4 m boven de bodem werden uitgevoerd, terwijl het 2D model dieptegemiddeld rekent en anderzijds dat het 2D model een seizoenaal gemiddelde SPM concentratie als randvoorwaarde heeft. Hierdoor kunnen de op het moment van de meting heersende SPM concentratie niet goed worden voorgesteld. De tripode heeft op de Kwintebank vrij hoge SPM concentraties gemeten (tot 1 mg/l), waarden die bij latere verankeringen of 13-uursmetingen nooit werden gemeten. 19

20 Measurements 2D model.14 SPM concentration (g/l) Measurements 1D Model 2D model 25 Particle size (micro m) From measurements (nf=2) 1D model 2D model.8 Fall velocity (mm/s) Time (hours since start) Figuur 3.8: Vergelijking tussen metingen, 1D Bfloc model en het 2D sedimenttransportmodel met Bfloc. Boven: Gemeten en gemodelleerde SPM concentratie. Midden: Gemeten en berekende partikelgrootte. Onder: Berekende en uit metingen afgeleide valsnelheid. Nul op de x-as komt overeen met de start van de meting op 2/3/23 15h1. 2

21 4. Conclusies 4.1. Storminvloed op SPM concentratie Tijdens en na stormen werden heel hoge concentraties aan suspensiemateriaal gemeten ter hoogte van Zeebrugge. De baggerstortplaatsen, de vaargeulen en de aangrenzende gebieden met vers afgezet slib zijn de belangrijkste bronnen voor het suspensiemateriaal tijdens deze gebeurtenissen. Menselijke activiteiten (bouw van Zeebrugge, aanleg en verdieping van vaargeulen, storten van baggerspecie) dragen significant bij tot de vorming van hooggeconcentreerde sliblagen in de omgeving van Zeebrugge Flocculatiemodel De simulatie toont aan dat met het BFLOC flocculatiemodel, dat rekening houdt met biologisch-fysische processen, een goede overeenkomst kan bekomen worden tussen gemeten en gemodellerde vlokgrootte. Verdere inspanningen zijn nodig om de parameterwaarden beter af te ijken zodoende optimaler met seizoenale (maandelijkse) en geografische variaties rekening gehouden kan worden. Toekomstige ontwikkelingen zijn enerzijds een implementatie in een 3D sedimenttransportmodel op basis van COHERENS v2 en een koppeling met het 3D biochemisch model MIRO (Lacroix et al., 27) dat voor een tijdsafhankelijke input (seizoenaal) van de nutriëntconcentratie kan zorgen. 21

22 5. Referenties Backer, J. 24. RV Belgica Meetcampagnes MOMO 23-24, Monitoring Sediment Transport. BMM-MDO/BELGICA/MOMO-I/Rapport/ 24. Backers, J. & Van den Brande, R. 25. RV Belgica meetcampagnes MOMO BMM-MDO/BELGICA/MOMO/Rapport pp. Backers, J. & Van den Brande, R. 26. RV Belgica meetcampagnes MOMO BMM-MDO/BELGICA/MOMO/Rapport 26. Backers, J., Van den Branden, R. 27a. RV Belgica meetcampagnes MOMO Deel I. BMM-MDO/27-16/MOMO/26-27, 65pp + CD. Backers, J., Van den Branden, R. 27b. RV Belgica meetcampagnes MOMO Deel II. BMM-MDO/27-49/MOMO/26-27, 46pp + CD. Berlamont, J, Ockenden, M., Toorman, E., Winterwerp, J.C The characterisation of cohesive sediment properties. Coastal Engineering, 21, Fettweis M., Francken, F., Pison, V., Van den Eynde D. 26. Suspended particulate matter dynamics and aggregate sizes in a high turbidity area. Marine Geology, 235, Fettweis, M., Francken, F., Van den Eynde, D. 27. MOMO activiteitsrapport (april 27 september 27). BMM-rapport MOMO/3/MF/277/NL/AR/2, 62pp + app. Fettweis, M. 28. Uncertainty of excess density and settling velocity of mud flocs derived from in situ measurements. Estuarine Coastal and Shelf Science, 78, Fettweis, M., Francken, F., Van den Eynde, D. 28a. MOMO activiteitsrapport (april 27 september 27). BMM-rapport MOMO/3/MF/281/NL/AR/3, 34pp + app. Fettweis, M., Van den Eynde, D., Francken, F., Nechad, B. 28b. MOMO activiteitsrapport (october 27 - maart 28). BMM-rapport MOMO/3/MF/285/ NL/AR/4, 57pp + app. Kiorboe, T. 23. Marine snow microbial communities: scaling of abundances with aggregate size, Aquatic Microbial Ecology, 33, Kranenburg, C On the fractal structure of cohesive sediment aggregates. Estuarine, Coastal and Shelf Science, 39, Lacroix, G., Ruddick, Y., Park, Y., Gypens, N., Lancelot, C. 27. Validation of the 3D biogeochemical model MIRO&CO with field nutrient and phytoplankton data and MERIS-derived surface chlorophyll a images. Journal of Marine Systems, 64, Lauwaert, B., Fettweis, M., Cooreman, K., Hillewaert, H., Moulaert, I., Raemaekers, M., Mergaert, K. & De Brauwer, D. 24. Syntheserapport over de effecten op het mariene milieu van baggerspeciestortingen. BMM, DVZ & amt rapport, BL/24/1, 52pp. Lauwaert, B., De Brauwer, D., Fettweis, M., Hillewaert, H., Hostens, K., Mergaert, K., Moulaert, I., Parmentier, K. & Verstraeten, J. 26. Syntheserapport over de effecten op het mariene milieu van baggerspeciestortingen (vergunningsperiode 24-26). BMM, ILVO & amt rapport, BL/26/1, 87pp+ app. 22

23 Lauwaert, B., Bekaert, K., Berteloot, M., De Brauwer, D., Fettweis, M., Hillewaert H., Hoffman, S., Hostens, K., Mergaert, K., Moulaert, I., Parmentier, K., Vanhoey, G., Verstraeten, J. 28. Syntheserapport over de effecten op het mariene milieu van baggerspeciestortingen (vergunningsperiode 26-28). BMM, ILVO, ak & amt rapport, BL/28/1, 128pp. Lick, W., Huang, H., Jepsen, R Flocculation of fine-grained sediments due to differential settling. Journal of Geophysical Research, 98 (C6), Maggi, F. 27. Variable fractal dimension: a major control for floc structure and flocculation kinematics of suspended cohesive sediment. Journal of Geophysical Research. doi:1.129/26jc3951. Maggi, F. Bioflocculation of suspended particles in nutrient-rich aqueous ecosystems. Journal of Hydrology (submitted). Mellor, G.L., Yamada, T A hierarchy of turbulence closure models for planetary boundary layers. Journal of Atmospheric Science, 31, Mikkelsen, O.A., Pejrup, M. 21. The use of a LISST-1 laser particle sizer for insitu estimates of floc size, density and settling velocity. Geo-Marine Letters, 2, Ploug, H., Grossart, H.-P Bacterial production and respiration in suspended aggregates - a matter of the incubation method, Aquatic Microbial Ecology, 2, Ten Brinke, W.B.M., Dronkers, J Physical and biotic aspects of fine-sediment import in the Oosterschelde tidal basin (The Netherlands). Netherlands Journal of Sea Research, 31, Winterwerp, J.C A simple model for turbulence induced flocculation of cohesive sediment. Journal of Hydraulic Research, 36, Winterwerp, J., Manning, A.J., Martens, C., De Mulder, T., Vanlede, J. 26. A heuristic formula for turbulence-induced flocculation of cohesive sediment. Estuarine, Coastal and Shelf Science, 68,

24 ƒ COLOPHON Dit rapport werd voorbereid door de BMM in april 29 Zijn referentiecode is MOMO/4/MF/293/NL/AR/1. Status draft finale versie herziene versie vertrouwelijk Beschikbaar in het Engels Nederlands Frans Indien u vragen hebt of bijkomende copies van dit document wenst te verkrijgen, gelieve een te zenden naar M.Fettweis@mumm.ac.be, met vermelding van de referentie, of te schrijven naar: BMM 1 Gulledelle B 12 Brussel België Tel: Fax: BEHEERSEENHEID VAN HET MATHEMATISCH MODEL VAN DE NOORDZEE SUMO GROEP De lettertypes gebruikt in dit zijn Gudrun Zapf-von Hesse s Carmina Medium 1/14 voor de tekst en Frederic Goudy s Goudy Sans Medium voor titels en onderschriften. 24

25 APPENDIX 1 Fettweis, M., Francken, F. Van den Eynde, D. 28. Persistent high SPM concentration layers in a coastal turbidity maximum area. 7 th Int. Conf. on Tidal Environments (Tidalites-28), September, Qingdao, China

26 TIDALITES 28, Qingdao (China), September Persistent high SPM concentration layers in a coastal turbidity maximum area Michael FETTWEIS*, Frederic FRANCKEN, Dries VAN DEN EYNDE Royal Belgian Institute of Natural Science, Management Unit of the North Sea Mathematical Models, Gulledelle 1, 12 Brussels, Belgium * m.fettweis@mumm.ac.be Suspended Particulate Matter (SPM) transport has a major influence on coastal ecosystems and on the siltation of ports and naviagtion channels. The distribution of SPM is also affected by human activities such as port development, dredging and dumping and aggregate extraction. These activities have further major impacts on the environment, due to shoreline erosion and bathymetrical, morphological and hydrodynamical changes and eventually degradation or total loss of natural habitats. In a dynamic coastal area, such as the southern North Sea, the natural variability of cohesive sediment processes due to tides and meteorological effects is very high as also the human impact. A major question to be answered is to which extent is the physical environment (hydrodynamics, sediment transport, sedimentology and morphology) the result of natural changes and/or anthropogenic activities? Are the human impacts local or do they have a wider influence? Detailed knowledge on SPM concentration variability and on the processes causing these variations are necessary to asses the effects of human impact. The SPM concentration variations in the turbidity maximum in the Belgian-Dutch coastal zone (southern North Sea) are complex and the cause of in- or decrease of SPM concentration is except for the tidal and spring-neap tidal signal often difficult to identify. Long-term measurements using a benthic lander (tripode) have therefore been carried out between 23 and 27 on three locations. In total 24 days of SPM concentration at 2m and.2m above the bed have been measured using OBS. Further, data on salinity, temperature, current velocity (ADV and ADP) and particle size (LISST 1C) have been collected. Regularly very high SPM concentrations have been observed near the bottom in areas where current velocities are high (>1.2 m/s during spring tide) and deposition of mud limited to the formation of fluffy layers, which are quickly resuspended. The High Concentration Bed Layers (HCBL) are limited in height to a few tens of cm and are not detected at 2 m above the bed (see figure below) and occur at the beginning of a storm when SPM concentration is high due to mainly wave erosion. The fact that the mixing capacity of waves is lower than the erosion capacity, can result in saturated conditions and thus in the formation of a fluid mud layer (Winterwerp, 21). HCBL have regularly been observed in the area during a few tides, which proves either that no residual transport occurs or that erosion of cohesive sediments from the bed layer is higher than the advection of it. The figure below shows that during the period with high SPM concentration the significant wave height and the wave induced turbulence is high, resulting in a high erosion capacity. The SPM concentration variations are affected by tides, waves and wind direction. The major wind direction in Western Europe is from the SW, which results in a residual inflow of water and suspended sediments from the English Channel towards the North Sea (Fettweis et al., 27a). The measurements show that an increase of SPM concentration during storms is strongly dependent on the direction of the wind and on the availability of erodable cohesive sediments in the coastal zone. This makes that the effect of every storm is different, because the meteorological and sedimentological history determines the SPM concentration. The increase in SPM concentration can therefore not always be linked in time with the onset of a storm. Storms that bring offshore water with low SPM concentration towards the coastal zone will e.g. first result in a slight decrease of SPM concentration followed by an increase due to wave action. HCBL have probably an important influence on the SPM transport in whole the area. It is argued that local mud sources, such as mud in sandy beds or consolidated mud layers, have to exist besides the well known and

27 generally accepted English Channel as major source of SPM in the southern North Sea (Fettweis et al., 27b), in order to obtain sufficient amount of SPM to develop HCBL. Local sources are probably only effective during extreme conditions, their impact is therefore difficult to quantify. In the coastal area consolidated mud layers exist that cannot erode under normal conditions (critical erosion stress >1Pa). Wave forcing may result in failure of the consolidated bed and in the formation of mud pebbles (Jacinto & Le Hir, 21). Mud pebbles have been found frequently in the area (Fettweis et al. 27b). 3 SPM 1 SPM 2 25 SPM concentration (mg/l) Time (days) Wave height TKE Sign. wave height (cm) TKE per unit volume (kg/ms**2) Time (days) November 26, Belgian coastal zone (southern North Sea). Above: SPM concentration (SPM1:.2 and SPM2: 2. m above bed); below: Significant wave height and turbulent kinetic energy per unit volume, derived from ADV current velocity data at.2 m above the bed. References Fettweis, M., Nechad, B., Van den Eynde, D. 27a. An estimate of the suspended particulate matter (SPM) transport in the southern North Sea using SeaWiFS images, in-situ measurements and numerical model results. Continental Shelf Res, 27, Fettweis, M., Du Four, I., Zeelmaekers, E., Baeteman, C., Francken, F., Houziaux, J.-S., Mathys, M., Nechad, B., Pison, V., Vandenberghe, N., Van den Eynde, D., Van Lancker, V., Wartel, S. 27b. Mud Origin, Characterisation and Human Activities (MOCHA). Final Scientific Report, Belgian Science Policy, 59pp. Silva-Jacinto, R. Le Hir, P. 21. Response of stratified muddy beds to water waves. In: Coastal and Estuarine Fine Sediment Processes (McAnally, W.H., Mahta, A.J., eds.). Proceedings in Marine Science, 3, Winterwerp, J. 21. Stratification effects by cohesive and non-cohesive sediment. Journal of Geophysical Research, 16 (C1),

28 APPENDIX 2 Fettweis, M. Uncertainty of excess density and settling velocity of mud flocs derived from in situ measurements. Particles in Europe, European workshop on the in situ measurements of particle size and concentration in the aquatic environment October 28, Bologna, Italy.

29 Particles in Europe (PiE) workshop, October 28, Bologna, Italy Extended Abstract Uncertainty Of Excess Density And Settling Velocity Of Mud Flocs Derived From In Situ Measurements. Michael Fettweis Royal Belgian Institute of Natural Science, Management Unit of the North Sea Mathematical Models, Gulledelle 1, 12 Brussels, Belgium, 1. Introduction Flocculation is the process of floc formation and break-up and has a direct impact on settling velocity. The settling velocity is a function of the particle size and effective density, and can be described by Stokes law under the assumption that the particle Reynolds number is smaller than one and that the particles are spherical and non porous. However, because the Suspended Particulate Matter (SPM) consists of a population of flocs with heterogeneous sizes, density, and shape, the settling velocity of mud flocs in real environments may vary. Experimental observations are always subject to uncertainties that can be typically attributable to random measurement errors (lack of precision), systematic errors (lack of accuracy), human error, and intrinsic variable stochasticity. Within the field of flocculation of cohesive sediment dynamics, stochastic uncertainty is of primary importance as recently recognised in the works by Lee and Matsoukas (2), Jackson (25), Khelifa and Hill (26) and Maggi (27) who studied the fluctuations of the average and median floc size over time. Measurements of effective density and settling velocity are inherently associated with uncertainties (errors). These are linked with a lack of accuracy of the measuring instruments, a lack of precision of the observations and the statistical nature of the variables (floc size, primary particle size and primary particle density). When using observations some understanding of the uncertainties is needed. Based on the theory of error propagation the error of the effective density and the settling velocity of mud flocs has been estimated using the measurement data of OBS, SPM filtration, LISST1C, CTD and Sedigraph. The measurements have been carried out between 23 and 25 in the southern North Sea during 8 tidal cycles, (Figure 1). An extended presentation of the results can be found in Fettweis (28). Figure 1: Yearly averages of vertically averaged SPM concentration in the southern North Sea derived from 362 SeaWiFS images ( ), see Fettweis et al. (27). Also shown are the locations of the tidal measurement stations. The coordinates are latitude ( N) and longitude ( E). 2. Calculation of fractal dimension, effective density and settling velocity By describing mud flocs with the fractal theory (Meakin, 1991, Van Leussen, 1994), the floc effective density can be written as (Kranenburg, 1994):

30 nf 3 D f ρ = ρ f ρ w = ( ρ p ρ w ) (1) D p where ρ is the effective density; ρ f, ρ w, and ρ p are the floc, water, and primary particle densities, respectively; D f and D p are the floc and primary particle sizes, respectively, and nf is the floc fractal dimension. The primary particle is defined as the first-order constituent of a floc and may consists of clay or other silicate minerals, carbonate and organic particles. D p can be represented by the median diameter of mineral grains in the flocs. Because ρ p, ρ w and D p can be considered independent variables, Eq. 1 can be reduced to nf 3 ρ = k a D f (2) with k a a correlation parameter. If ρ and D f are known then the fractal dimension can be derived using a linear regression on a log-log plot. By doing so, it is assumed that nf is constant, this assumption has recently been questioned (e.g. Khelifa and Hill, 26; Maggi et al., 27 and Maggi, 27). The effective floc density can be calculated if the floc and water density are known (see Mikkelsen and Pejrup, 21). The water density has been derived from conductivity, temperature, and pressure measured by the CTD. The floc density can be expressed as: M f ρ f = (3) V f with V f the floc volume and M f the floc mass concentration per unit volume. The water and primary particle mass can be written as M w =ρ w V w and M p =ρ p V p, respectively, with V w and V p the water and primary particle volumes in the floc. The floc density (Eq. 3) can eventually be calculated with M f written as: ( ) M p M f = M p + M w = M p + ρ w V f V p = M p + ρ w V f (4) ρ p The fall velocity, w s, for flocs with fractal structure can be written as (Winterwerp, 1998): w s ( ρ ρ ) nf 1 f α p w D 3 nf = g Dp 18 β η Re.687 where Re is the floc-reynolds number, g is the gravitational acceleration, η is the molecular viscosity of water ( kg/ms), and α and β are shape factors. M p has been measured with an OBS and through filtration; V f and D f have been measured with a LISST 1C and D p with a Sedigrapg. In Figure 2 typical distributions of D f measured with the LISST 1C are shown. The density of primary particles, ρ p, has been calculated, based on an analysis of the floc constituents. The density was obtained from the primary particle size distribution and the CaCO 3 and total organic content (TOC) content. In figure 3 primary particle size distribution measured of the mineral fraction by the Sedigraph are presented. The fractal dimension was derived from a linear regression on a log-log plot of effective density and floc size. (5) 2 23/22 MOW1 4 23/25 Kwintebank Volume concentration ( l/l) 12 8 Volume concentration ( l/l) Particle size ( m) 1 1 Particle size ( m) Figure2: Particle (floc) size distribution of the SPM measured by the LISST 1C as a function of volume concentration. Only the distributions with a transmission greater then 2% are shown.

31 Mass fraction (%) Particle size ( m) Figure 3: Primary particle size distribution of the mineral fraction of the SPM measured with a Sedigraph 51 and sieving. The rising tail at 62.5 µm in 4 out of 5 spectra comprises the sand fraction without further detail. 3. Estimates and origin of uncertainties The errors have been estimated based on the theory of error propagation (see e.g. chapter 14.2 of Numerical Recipes, Press et al., 1989). The results indicate that major sources of uncertainties are from the primary particle and floc sizes, the primary particle density, and the SPM concentration from filtration and from OBS. The relative standard deviations for effective density, fractal dimension and settling velocity are about 1%, 2.5% and 1%, respectively. These uncertainties should rather be regarded as lower limits of the real error, because the errors due to lack of accuracy of the OBS, LISST and Sedigraph have been omitted as they are unknown. The origin of the error on the settling velocity is mainly due to uncertainties on the primary particle size D p and the floc size D f. These results are complementary to those of Khelifa and Hill (26), who have underlined the dominant effect that primary particle size has on the effective density and thus on the settling velocity. The results from the error analysis have shown that the statistical uncertainties on the settling velocity will always be high when dealing with natural flocs or particles and that they cannot be reduced by increasing the accuracy of the instruments, the way of measuring or the method to calculate the settling velocity. In other words, they are always the dominating ones. These actions will however increase the reliability of the settling velocity in the sense that the value corresponds better with reality - as systematic errors are reduced or precision is increased. 4. Conclusions Measurements of settling velocity are inherently associated with uncertainties due to lack of accuracy of the measuring instruments and due to the statistical nature of particle size distributions (and effective densities) in the suspended matter. Theses errors occur when using direct or indirect method to obtain settling velocities, the conclusions are: (1) The relative standard deviation on settling velocity due to statistical uncertainties is minimum 1%. The origin of the error is mainly from uncertainties of the primary particle and the floc sizes, respectively. These statistical uncertainties will always be high when dealing with natural flocs or particles and cannot be reduced by increasing the accuracy of the instruments; (2) The statistical error on the effective density is mainly due to uncertainties of the SPM concentration and of the primary particle density; (3) More reliable values of settling velocity can be obtained by e.g. increasing the precision of the measurements, the accuracy of the instruments and not assuming self-similarity of floc structures; (4) It is crucial to have data on primary particle size and density at the same moment as floc size and SPM concentration data, as these parameters are of major importance to calculate the effective density and the settling velocity. Measurements of suspended matter should include an analysis of its major constituents (organic matter, CaCO 3 and silicate minerals) and the grain size. An important part of our understanding of flocculation and cohesive sediment dynamics (deposition and erosion) is based on measurements. The uncertainties associated with indirect (or direct) settling velocity measurements are very high due to their statistical nature; the total error will be even higher because systematic errors due to a lack of accuracy of the measuring instruments are

32 not included. References Fettweis, M., Nechad, B., Van den Eynde, D. 27b. An estimate of the suspended particulate matter (SPM) transport in the southern North Sea using SeaWiFS images, in situ measurements and numerical model results. Cont. Shelf Res. 27: Fettweis, M. 28. Uncertainty of excess density and settling velocity of mud derived from in situ measurements. Est., Coast. Shelf Sci., 78, Jackson, G. A. 25. Coagulation theory and models of oceanic plankton aggregation. In: Flocculation in Natural and Engineered Environmental Systems. Droppo, I., Leppard, G., Liss, S., Milligan, T. (eds.). CRC Press, Khelifa, A., Hill, P.S. 26. Models for effective density and settling velocity of flocs. J. Hydr. Res. 44: Kranenburg, C On the fractal structure of cohesive sediment aggregates. Est. Coast. Shelf Sci. 39: Lee, K., Matsoukas, T. 2. Simultaneous coagulation and break-up using constant-n Monte Carlo. Powder Techn. 11, Mikkelsen, O.A., Pejrup, M. 21. The use of a LISST-1 laser particle sizer for in-situ estimates of floc size, density and settling velocity. Geo-Mar. Let. 2: Maggi, F. 27. Variable fractal dimension: a major control for floc structure and flocculation kinematics of suspended cohesive sediments. J. Geoph. Res. 112, C712. Maggi, F., Mietta, F., Winterwerp, J.C. 27. Effect of variable fractal dimension on the floc size distribution of suspended cohesive sediment. J. Hydr. 343: Meakin, P Fractal aggregates in geophysics. Rev. Geophys. 29: Press, W.H., Flannery, B.P., Teukolsky, S.A., Vetterling, W.T Numerical recipes: The art of scientific computing (Fortran Version). Press Syndicate of the University of Cambridge, 72pp. van Leussen, W Estuarine macroflocs and their role in fine-grained sediment transport. Ph. D. Thesis, Univ. Utrecht, The Netherlands. Winterwerp, J A simple model for turbulence induced flocculation of cohesive sediments. J. Hydr. Res. 36:

33 APPENDIX 3 Doxaran, D., Babin, M., Ruddick, K., Park, Y., Fettweis, M. Tidal variations of particle size distribution and optical properties in turbid coastal waters. Perspectives for ocean colour remote sensing. Particles in Europe, European workshop on the in situ measurements of particle size and concentration in the aquatic environment October 28, Bologna, Italy.

34 Particles in Europe (PiE), Bologna (Italy), October 13-14, 28 Tidal variations of particle size distribution and optical properties in turbid coastal waters. Perspectives for ocean colour remote sensing. David Doxaran 1, Marcel Babin 1, Kevin Ruddick 2, Younge Park 2 and Michael Fettweis 2 1 Laboratoire d'océanographie de Villefranche, UMR CNRS / UPMC, Villefranchesur-Mer, France doxaran@obs-vlfr.fr 2 Management Unit of the North Sea Mathematical Models (MUMM, BMM, UGMM, RBINS, Brussels, BELGIUM Max. 25 words Measurements of surface water turbidity, inherent optical properties (IOPs), suspended particle concentration and size distribution were carried out in Belgian coastal waters during a 12 hours tidal cycle. The IOPs (namely the light absorption, attenuation and scattering coefficients) were measured using a Wetlabs field spectrophometer designed with three visible and six near-infrared spectral channels, and a short pathlength appropriate for turbid coastal waters. Particle size distributions were measured using a LISST-1X (Sequoia Scientific). Additional measurements provided by field (temperature, salinity, water depth, current velocities and seawater reflectance) and ocean colour satellite (MODIS 3 and SEVIRI 4 ) data are used to describe the dynamics of the study area Results show a clear 6-hour cycle in parameters related to suspended matter concentrations, i.e. water turbidity, light absorption, attenuation and scattering coefficients. Moreover, the spectral slopes of the attenuation and scattering coefficients, expected to be particle size-related, also exhibit variations indicating particle sorting over the tidal cycle, i.e. settling of coarse particles at slack tide and resuspension at mid-tides. We first analyze how the spectral variations of the IOPs reflect changes in the shape of the particle size distribution. Based on bio-optical modelling and field measurements, we then determine how changes in the shape of the particle size distribution can be detected from the seawater reflectance. We finally discuss the potential of ocean colour remote sensing data in order to retrieve information on the size of particles suspended in turbid coastal waters. 3 Moderate Resolution Imaging Spectroradiometer 4 Spinning Enhanced Visible and InfraRed

35 APPENDIX 4 Fettweis, M., Francken, F., Van den Eynde, D., Verwaest, T., Janssens, J., Van Lancker, V. Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea). Paper submitted to Marine Geology

36 Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea) Michael Fettweis 1, *, Frederic Francken 1, Dries Van den Eynde 1, Toon Verwaest 2, Job Janssens 2, Vera Van Lancker 1, 3 1 Royal Belgian Institute of Natural Science, Management Unit of the North Sea Mathematical Models, Gulledelle 1, 12 Brussels, Belgium 2 Flanders Hydraulics Research, Berchemlei 115, 214 Antwerp, Belgium 3 Ghent University, Renard Centre of Marine Geology (RCMG), Krijgslaan 281, S8, 9 Gent, Belgium * m.fettweis@mumm.ac.be Abstract Multi-sensor tripod measurements in the high-turbidity area of the Belgian nearshore zone (southern North Sea) allowed investigating storm effects on near bed suspended particulate matter (SPM) concentrations. The data have shown that during or after a storm the SPM concentration increases significantly and that high concentrated mud suspensions (HCMS) are formed. Under these conditions, about 3 times more mass of SPM was observed in the water column, as compared to calm weather condtions. The following different sources of fine-grained sediments, influencing the SPM concentration signal, have been investigated: wind direction and the advection of water masses; the previous history and occurrence of fluffy layers; freshly deposited mud near the disposal grounds of dredged material, navigation channels and adjacent areas; and the erosion of medium-consolidated mud of Holocene age. Based on erosion behaviour measurements of in-situ samples, the critical erosion shear stresses have been estimated for different cohesive sediment samples outcropping in the study area. The results have shown that most the mud deposits cannot be eroded by tidal currents alone, but higher shear stresses, as induced by storms with high waves, are needed. Erosion can however occur during storms with high waves. The data suggest that in order to obtain very high SPM concentrations near the bed, significant amounts of fine-grained sediments have to be resuspended and/or eroded. The Disposal grounds of dredged material, navigation channels and adjacent areas with freshly deposited mud have been found to be the major source of the fine-grained sediments during storms. This result is important, as it suggests that dredging and the associated disposal of sediments have made available fine-grained matter that contributes significantly to the formation of high SPM concentrations and high concentrated mud suspensions. 1

37 Introduction Improving our understanding of suspended particulate matter (SPM) variability in nearshore areas is essential for a more sustainable exploitation of the marine environment, to better understand the human footprint and to develop marine policies matching international agreements aiming at sustainable development. Tides, waves, medium-scale meteorological events, biological processes, flocculation and deposition affect SPM concentration. The SPM concentration is further influenced by the remote or local availability of fine sediments and by human activities such as dredging and disposal of dredged material or port construction. Fine sediments from remote sources are transported mainly by advection caused by wind-driven or tidal currents. Local sources consist of sediments having a significant amount of fines; wave action and tidal currents can resuspend or erode these deposits. The aim of this paper is to present and discuss SPM concentration variations in a shallow coastal turbidity maximum area generated by medium-scale meteorological events. Waves have an important impact on cohesive sediment processes on continental shelves (e.g. Green et al., 1995; Cacchione et al., 1999; Traykovski et al., 27; Shi et al., 28). During these events high concentrated mud suspensions (HCMS) have been measured, which can be formed by settling of suspended matter or fluidization of cohesive sediment beds (Maa and Mehta, 1987; van Kessel and Kranenburg, 1998; Li and Mehta, 2). Fluidization or liquefaction of mud layers occurs if the stress caused by wave pressure exceeds the yield stress of the sediments. The high concentrated mud suspensions (HCMS) and fluid mud deposits are easily transported and may result in very high sediment fluxes. In the vicinity of Zeebrugge (southern North Sea), long-term SPM concentration time series have been collected using a benthic boundary layer tripod: these form a unique dataset for the investigations of storm influences on SPM concentration. The measuring locations are situated in a very energetic area that is amongst the most turbid in the North Sea. The origin of the suspended matter in the southern North Sea is mainly from the inflow of fine-grained sediments through the Dover Strait, generated by a flood-dominated and wind-induced residual water transport towards the North Sea (Eisma, 1981; van Alphen, 199; Lafite et al., 2). Fettweis et al (27a) point out that this SPM flux cannot solely be responsible for the often very high SPM concentration in the area. They suggest that erosion of the nearshore medium consolidated Holocene mud deposits contributes to the sediment balance, but comment that other sources of fine-grained sediments might exist. These sources could be related to human activities as most of the nearshore areas in the southern North Sea have a long history of human impact (e.g. coastal defence works, port construction, dredging and disposal of sediments). Verwaest (27) stresses the impact of dredging and disposal activities on SPM concentrations. The cohesive sediment processes associated with storms in a coastal high turbidity zone are not well documented; the present paper is therefore an attempt to assess, based on field measurements and numerical model results, the hydrodynamics, sediment dynamics and bed erosion processes during extreme meteorological events. These data will help to better understand the variability of SPM concentration and its relation to erosion, resuspension and transport. 2. Regional settings The study area is situated in the southern North Sea, more specifically in the Belgian-Dutch nearshore zone. The depth is generally between -2 m below MLLWS (Mean Lowest Low Water Spring), except in the mouth of the Westerschelde estuary where depths can reach more than 2 m below MLLWS (Fig. 1). The tidal regime is semi-diurnal, with tidal ranges that diminish towards the northeast. The mean tidal range at Zeebrugge is 4.3 m and 2.8 m at spring and neap tide, respectively. The tidal currents are generally flood-dominant dominant (towards the northeast), as also the residual water 2

38 transport. The maximum current velocities are more than 1 ms -1. The winds are dominantly from the southwest and the highest waves occur during north-western winds (Fettweis and Van den Eynde, 23). SPM forms a turbidity maximum between Oostende and the mouth of the Westerschelde. SPM concentration in the coastal zone varies between minimum 2-7 mg l -1 and maximum 1-1 mg l -1 during calm meteorological conditions at 3 m above the sea bed; lower values (<1 mg l -1 ) occur in the offshore area (Fettweis et al., 27a). The sea bed consists mainly of Quaternary sandy deposits. Most of them have been deposited during the Holocene transgression; they are grouped today in sand banks, which form the thickest accumulation of Quaternary deposits and the most prominent sea bed feature in the area (Le Bot et al. 25). The nearshore deposits consist of mainly fine sands with variable mud content. The cohesive sediments (mud, clay) occur mainly in the eastern nearshore part of the Belgian Continental Shelf and are characterised by a particular rheological and/or consolidation state. Four different types are distinguished, namely Eocene clay, Holocene mud (±3 yr BP), modern mud (<1 yr BP) and freshly deposited mud (Fettweis et al, 27b). Generally, the freshly deposited mud occurs as thin fluffy layers or locally as gradually soft consolidated thicker packages (±.2-1 m). These thicker soft mud layers consisting also of modern mud deposits, are limited to the area around the disposal place of dredged material near Oostende and the port of Zeebrugge. The Holocene deposits, extending over most of the foreshore area, consist of medium-consolidated mud with intercalation of thin more sandy horizons; they are often covered by sand layers (order of cm to dm) or fluffy layers of a few cm thick. In the offshore swales, the thickness of the Quaternary cover is locally less than 2.5 m; in these areas Eocene outcrops (clay) are to be expected (Le Bot et al., 25). The port of Zeebrugge and its connection to the open sea (Pas van het Zand, Fig. 1) as well as the navigation channel towards the Westerschelde estuary (Scheur, Fig. 1) are efficient sinks for cohesive sediments. To conserve the maritime access to the coastal harbours and to the Scheldt estuary dredging is needed. Maintenance dredging amounts today to about 8.6 million tons of dry matter yearly and capital dredging to 2.8 million tons of dry matter (averages over ), see Lauwaert et al. (28). About 45% of the maintenance dredging is carried out in the navigation channels and 55% in the harbours. 93% of the dredging in the harbours is from the port of Zeebrugge. Half of the matter from the navigation channels is dredged in the access channel towards Zeebrugge (Pas van het Zand). The dredged matter from the navigation channels consist of 7-85% of mud; from the harbours this amounts to more than 95%. 3. Materials and Methods 3.1 Instrumentation and measuring sites Data collection was conducted between April 25 and February 27 using a tripod at two sites (Fig. 1). Both sites are characterised by the occurrence of near-bed Holocene medium-consolidated mud, albeit covered with an ephemeral slightly muddy fine sand layer with a median grain size of about 17 µm. The thickness of the sand layer increases towards the shore. The water depth is 9 m MLLWS at the MOW1 site and about 5 m MLLWS at the Blankenberge site (Fig. 1). The tripod measuring system was developed to monitor SPM concentration and current velocity. Mounted instruments include a SonTek 3 MHz ADP, a SonTek 5 MHz ADV Ocean, a Sea-Bird SBE37 CT system, two OBS, two SonTek Hydra systems for data storage and batteries and a LISST 1C (Table 1). Three periods have been selected for further analysis to assess storm influences (Table 2). Field calibration of the OBS sensors have been carried out during 5 tidal cycles between April 25 and May 26 at site MOW1. A Niskin bottle was closed every 2 minutes, thus resulting in about 4 samples per tidal cycle. Three sub 3

39 samples were filtered on board of the vessel from each water sample, using pre-weighted filters (Whatman GF/C). After filtration, the filters were rinsed once with Milli-Q water (±5 ml) to remove the salt, and dried and weighted to obtain the SPM concentration. A linear regression between all OBS signals and SPM concentrations from filtration (about 24) was assumed. 3.2 Calculation of wave- and current induced shear stress from ADV The high frequency ADV measurements (25 Hz) permit to decompose the velocity in terms of a mean and a fluctuating part. Several studies reported on the possibility to estimate shear stress from the second moment (turbulence) statistics (Pope et al., 26; Verney et al., 27; Andersen et al., 27). The estimation is based on the calculation of the turbulent kinetic energy (TKE), which can be obtained from the variance of the velocity fluctuations. The shear stress has found to be proportional to the TKE through τ = C TKE, where C=19 was adopted as proposed by Stapleton and Huntley (1995) and Thompson et al. (23). Although some potential disadvantages can be credited to the TKE-method, such as the ADV s inherent Doppler noise which may lead to high-biased estimates of the TKE from an ADV (Voulgaris and Trowbridge, 1998; Nikora and Goring, 1998), recent studies have pointed out that the TKE-based calculation of shear stresses is probably the most reliable method (Pope et al., 26; Andersen et al., 27). 3.3 Erosion behaviour measurements Erodibility measurements have been performed on mud samples taken at different locations near the navigation channels Pas van het Zand and Scheur and near the harbour of Oostende (Fig. 1). Boxcores were subsampled using cylindrical perspex tubes (diameter 13.5 cm) allowing to retrieve relatively undisturbed mud samples of the first 4 cm (on average) of the sea bed. Using a gamma-ray densitometer, bulk density profiles of these samples were determined in a nondestructive way. The measurements were performed at the University of Stuttgart using the SETEG-flume (Kern et al., 1999; Witt and Westrich, 23): a straight pressure duct with rectangular cross section. The top of the perspex tubes can be attached to a circular hole in the bottom of the flume, and by pushing the sediment upwards until its surface is level with the bottom surface of the flume the erosion behaviour of the top layer of the sediment can be investigated. After erosion of the top layer, the sediment can be pushed further upwards and the underlying layers can be studied. The critical shear stress for erosion is determined by visual observation of the onset of erosion by a gradual increase of the discharge. Shear stresses are calculated from the measured discharge via the Darcy-Weisbach equation, using the Colebrook formula to determine the roughness coefficient (Streeter, 1996). In addition, the flume is equipped with the socalled SEDCIA-system, enabling to measure the erosion rate. This system consists of a camera which observes the timedependent shift of a series of parallel laser lines projected under a certain inclination angle on the sediment s surface, from which the sediment s volume change and hence also the mass change, since the bulk density is known can be calculated in function of time. 3.4 Hydrodynamic numerical model The currents, surface elevation and turbulent kinetic energy have been modelled using an implementation of the COHERENS hydrodynamic model to the Belgian Continental Shelf, termed hereafter OPTOS-BCS. The 3D model solves the continuity and momentum equations on a staggered sigma coordinate grid with an explicit mode-splitting treatment of the barotropic and baroclinic modes. A description of the COHERENS model can be found in Luyten et al. (1999): in this application only 2D results of bottom shear stress have been used. OPTOS-BCS covers an area between 51 N and N in latitude and between 2.8 E and 4.2 E in longitude. The horizontal resolution is.24 (longitude) 4

40 and.14 (latitude), corresponding both to a grid size of about 25 m. Boundary conditions of water elevation and depthaveraged currents for this model have been provided by the operational models OPTOS-NOS (covering the North Sea) and OPTOS-CSM (covering the North-West European Continental Shelf Model) of the Management Unit of North Sea Mathematical Models (see Pison and Ozer (23) have described the validation of the current velocities of OPTOS-BCS using ADCP measurements. The bottom shear stress under the combined effect of currents and waves is calculated in OPTOS-BCS with Bijker s formulae of 1966 (Koutitas, 1988). The currents are from the hydrodynamic model, whereas the wave data are from wave buoy measurements (Coastal Service of the Ministry of the Flemish Community). 4. Results 4.1 Tripod data The data of SPM concentration, water depth, ADV current velocity and bottom shear stress collected at MOW1 in autumn 25 are presented in Fig 2. Spring tide occurred around December 1 (day 1), neap tide around November 26 (day 5). The measuring period was characterised by 2 days of calm weather followed by a WNW storm with wind velocities of more than 16 m/s (7 Bf) and significant wave heights of up to 3.5 m in the coastal zone. The water was pushed up against the coast and low water levels raised with almost 2 m. The SPM concentrations follow a quarterdiurnal (ebb-flood) signal. Although the currents are flood dominated, no significant difference in magnitude between ebb and flood SPM concentration peak occurs. A significant increase in SPM concentration occurred almost immediately after the beginning of the WNW storm (day 3-4). The SPM concentration measured by OBS1 (.3 meter above bottom, mab) increases to 1-3 g l -1, whereas at 2 mab (OBS2) the SPM concentrations remain less than.5 g l -1. The sudden increase in SPM concentration at the onset of the storm was induced by the exceptional meteorological conditions and partly because the storm occurred after a calm period and around neap tide. Neap tidal conditions and calm periods favour the deposition of fluffy layers, as has been observed from bottom samples and from model simulations (Fettweis and Van den Eynde, 23). After the WNW storm the SPM concentration decreased in magnitude, however both OBS s still measured very high peak concentrations (OBS1: up to 3 g l -1, OBS2: up to 1 g l -1 ) until the end of the deployment (1 week after the storm). The high minima in SPM concentration measured by the OBS1 (.5-1 g l -1 ), indicates that a HCMS or fluid mud layer was formed. It is only days after the storm that the SPM concentrations, measured by both OBS s, show again similar minima and that the near bed high SPM concentrations have disappeared. 77 days of data were collected at the Blankenberge site between November 8, 26 and February 7, 27. The data from 7-18 November 26 and 24 December 26 8 January 27 are presented in Fig During both periods different storms have occurred. On November, a NW storm generated significant wave heights of about 2.7 m. The second half of December was characterized by low wind speeds from mainly W-SW. On December 31, wave heights of nearly 3 m were registered (Fig. 4). The beginning of January 27 was characterised by storms from mainly a NE direction. The highest SPM concentrations during the November 12 storm have been registered only about one day after the storm by OBS1 and about two days after by OBS2 (Fig. 3). The OBS1 data are characterised by very high minima in SPM concentrations (>.8 g l -1 ). The OBS2 has measured only during a short period after the storm an increase in SPM concentration. This indicates that vertical mixing was limited. Similar data have been collected during the storm of December 31. The period before the storm was characterised by low differences in SPM concentration between OBS1 and OBS2. Stratification in SPM concentration has been observed only at the end of the ebbing tide and during slack 5

41 water and is due to settling of suspended particles. The increase in SPM concentration after the December 31 storm is similar as observed during the November 12 storm - only detected one day after the storm in both OBS. This increase in SPM concentration occurred during ebb indicating that the suspended matter has mainly been transported from the NE, i.e. from the mouth of the Westerschelde estuary and in the direction of the wind-driven and the ebb current. Local resuspension of mud layers was at that time of minor importance. A few conclusions can be drawn from the three measurement periods: 1. There is a dominant quarter-diurnal (ebb-flood) signal in the SPM concentration time-series. The spring-neap tidal signal can be identified clearly during calm meteorological conditions; 2. The data from LISST 1C prove that the suspended particles are flocs: the biggest particles have been measured during slack water and low current velocities and the smallest during peak current velocities; 3. Considerable variations in SPM concentrations exist during a tidal cycle: maximum concentrations were sometimes up to 5 times higher than the minimum concentrations; 4. The very high SPM concentrations measured near the bed are related to storm periods; our data prove that HCMS occur near the bottom in the coastal turbidity maximum of the Belgian-Dutch nearshore zone; and 5. Wind-driven advection can have a significant influence on SPM concentration. 4.2 Erosion behaviour measurements In Fig. 5 the measured depth profile of the bulk density, ρ, and the critical shear stress for erosion, τ ce, of two cores with Holocene mud and one with freshly deposited mud are shown (location is indicated in Fig. 1). The profiles show that a high variability of τ ce can be observed in muddy sediments. The cores containing Holocene mud are both characterised by a 5-1cm thick surface layer of freshly deposited mud above soft to medium-consolidated mud. The Holocene mud deposits are characterised by intercalations of thin sandy layers. The top layer has a τ ce ranging from.5 to 2.5 Pa. The Holocene mud layers are very strong, with a τ ce up to 13 Pa, while the intermediate sandy layers exhibit much lower values (~1 Pa). The core with freshly deposited mud has been collected in the navigation channel. The upper 45 cm are not consolidated and have a τ ce of 1-4 Pa. A total of 35 sediment samples have been taken in the nearshore zone from which the following observations can be made: Generally, the top layer has low values for both τ ce (.5 to 1 Pa) and ρ (fluffy layers); The sediment stability of freshly deposited mud strongly increases in the first few cm. A value of ~4 Pa is reached at a depth of a few cm; Only in some cores a positive correlation between τ ce and ρ exist; The presence of shells has a negative effect on the stability of muddy sediments; and Sandy fractions always exhibit low to medium stability (maximum τ ce : ~4 Pa). 4.3 Hydrodynamic model results The numerical model has been used to simulate the bottom shear stress during autumn 25 (22 November 5 December). The maximum bottom shear stress is shown in Fig. 6 with and without wind and wave effects. The figure shows that the bottom shear stress increases significantly to values of more than 7 Pa. Without meteorological effects the maxima, occurring during spring tide in the channel towards the Westerschelde (Scheur), reaches about 4 Pa. The difference between model results and shear stress results from ADV can be ascribed to inaccuracies in calculating nearbed shear stress from ADV data and to the fact that vertically-averaged model results are used to calculate the bed shear 6

42 stress. 5. Discussion The formation of HCMS in wave-dominated areas is well documented in the scientific literature (de Wit and Kranenburg, 1997; Winterwerp, 1999; Li and Mehta, 2). The occurrence of fluid mud or HCMS on many continental shelves is associated with wave or current-driven sediment gravity flows off high-load rivers (Wright and Friedrichs, 26). However, the origin of the suspended matter in the southern North Sea and in the Belgian-Dutch nearshore zone has been mainly ascribed to the inflow of fine-grained sediments through the Dover Strait (Gerritsen et al., 2), as no high-load rivers exist in the area (Fettweis et al., 27a). This SPM flux amounts to about t per year, from which about 5% is transported along the continental side (i.e. France-Belgian-Dutch nearshore area), see Fettweis et al. (27a). Due to the high tidal energy permanent layers of freshly deposited to very soft consolidated cohesive sediments occur in the Belgian-Dutch nearshore area only in some protected hydrodynamic environment such as ports, navigation channels and disposal grounds of dredged material. However, thin fluffy layers, temporarily present, do occur over large areas. The fluctuation of SPM concentration with time is complex and it is not always straightforward to identify the origin of some of the variations. The relation between tidal amplitude and tide-averaged SPM concentration for the MOW1 data is shown in Fig. 7. The low correlation points to the fact that the neap-spring tidal signal is overprinted by other processes, such as high wave conditions and wind-induced long- and cross-shore advection. Using SeaWiFS images the average mass of SPM over the period in the turbidity maximum area has been calculated as about t. Variations of the order of 3% occur between spring and neap tides, as also of the order of 4-6% in-between seasons. Based on the tripod measurements, the SPM mass during storm conditions has been estimated as t in the same area. An important amount of fine-grained matter has thus to be resuspended, eroded or transported during a storm. Below, some points are discussed in more detail to better identify the possible sources of fine-grained sediments and the processes that result in the significant increase in SPM concentration during storms. 5.1 SPM transport and wind-driven advection The effect of winds on SPM concentration is variable and depends also on the wind direction and the availability of muddy sediments. Along- or cross-shore advection, enhanced during winds from the SW/NE or NW/SE, respectively, transports water masses with low SPM concentration and higher salinity to the measurement location. During periods with high salinity variations during a tide, a negative correlation between salinity and SPM concentration exists, see Fig. 8. High salinities and low(er) SPM concentrations are associated with flooding and low salinities and high(er) SPM concentrations with ebbing tide. The data show that during the first 3 weeks of January 27, SW winds prevailed resulting in advection of high salinity and low turbid water from the English Channel towards the southern North Sea. At the end of January, the wind direction changed towards S-SE and low salinity, high turbidity water originating from the Schelde estuary dominated the signal. 5.2 Holocene mud as SPM source The largest reservoir of fine-grained sediments in the nearshore area consists of the medium-consolidated Holocene mud (bulk density >15 kg m -3 ). The area where these mud deposits occur in the first meter of the seabed is ~744 km² (Fettweis et al, 27b). Erosion behaviour measurements (see above) confirm that consolidated cohesive bed layers are difficult to erode by only fluid-transmitted stress. The maximum bottom shear stress during calm periods and during spring tide is about 4 Pa (see Fig. 2-4 and 6). The sediments of the mud fields can thus not be eroded under calm 7

43 meteorological conditions. Near bed shear stresses derived from the ADV data amount up to 4 Pa (average is 12 Pa) during the storm of November 25 at MOW1 (Fig 2), end of December 26 (Fig. 3) and January 27 (Fig. 4). These values indicate that erosion of Holocene mud by fluid-transmitted shear stress can occur and that SPM could have been released into the water column from the Holocene mud fields under storm conditions. Using the Ariathurai-Partheniades formulation (Ariathurai, 1974) for erosion of cohesive sediments, a mean bed shear stress of 12 Pa, a critical erosion -2 shear stress of 1 Pa and an erosion rate of.5 g m s -1 the mass of Holocene mud that can be eroded amounts to 864 t km -2 day -1. If all the Holocene mud would outcrop then about t per day could have been eroded by horizontal tidally- and wave-induced shear stress during the storm. This is a maximum estimate as large parts of the Holocene mud fields are probably covered by thin layers of sandy sediments. Most probably, erosion of Holocene mud by entrainment represents only a minor source of suspended sediments during storms. Still, in literature, it is well documented that a cohesive mud bed may be eroded and fluidised by waves (Maa and Mehta, 1987; De Wit and Kranenburg, 1997; Li and Mehta, 2; Silva-Jacinto and Le Hir, 21). The effect of water waves on a cohesive, deformable bed can be described by a pressure wave, inducing normal and shear stresses in the bed. These stresses modify the strength of the bed and thus also the erodibility of the sediments. Silva Jacinto and Le Hir (21) explain the observed liquefaction and failure of consolidated mud layers along sandy layers by waves due to pumping effects. Mud pebbles are an indication of this type of erosion, they have been observed regularly in the area of investigation (Fettweis et al., 27b). However, due to insufficient data availability, the mass of mud pebbles or of suspended matter originating from wave induced bed failure cannot be quantified yet. 5.3 Mixed sediments as SPM source Other important erosion mechanisms are due to the mutual interaction of cohesive and non-cohesive sediments (Le Hir et al., 27): one can distinguish between sand grains moving on a cohesive substrate and erosion of mixed sediments. Thin sand layers on top of Holocene mud layers and mixed sediments have often been observed in the turbidity maximum area. Williamson and Torfs (1996) were the first to show that the addition of mud to a sandy bed increases the sediment shear strength. Van Ledden et al. (24) argue that a transition in erosion behaviour can be expected when the bed changes between cohesive and non-cohesive properties, and when the network structure is formed by another sediment fraction. Mixed sediments could therefore act as a reservoir for fine-grained material that is only resuspendable under more extreme meteorological conditions. However, on the basis of existing data, we cannot quantify the amount of fine sediments that is released as SPM due to the erosion of mixed sediments. 5.4 Freshly deposited mud as SPM source The movement of recently deposited sediment in the wave boundary layer exposes fine-grained sediment to resuspension when shear stresses become sufficiently high. These resuspended sediments are then transported further by waves and currents. The numerical model results indicate that, under normal conditions, the bed shear stress in the navigation channels and at location MOW1 or Blankenberge is lower than 4 Pa (Fig. 6). The critical erosion shear stresses of in-situ samples in the navigation channels and around Zeebrugge and the disposal ground of Oostende are, below the fluffy surface layer, generally higher than 4 Pa. The deposits of fresh mud below the fluffy layer in these areas forms thus a reservoir of SPM that will only be resuspended during periods with high shear stresses e.g. caused by storms. The fluffy surface layer having thicknesses of a few centimetres and a τ ce of <1 up to ~4 Pa, will be fully resuspended during 8

44 periods with higher stresses (storms, spring tides). Generation of fluffy layers can occur during periods with low stresses (neap tides). The total surface of the area where freshly deposited mud is found is not precisely known. The surface of the navigation channels, where mud is dredged, equals to ±15 km². Mud with similar erosion behaviour has also been sampled near the disposal ground of dredged material near Oostende and around Zeebrugge. Based on bed samples, the surface of these deposits has been estimated as 3 km². The bulk density of these sediments amounts to kg/m³. If we assume a thickness of 2 cm very soft mud in these areas, then the total mass of mud available for resuspension equals t. This is of the same order of magnitude as what has been estimated to be in suspension during storms (see above). The MOW1 and the Blankenberge site are both situated in the vicinity of the navigation channel towards Zeebrugge (Pas van het Zand) and the Westerschelde (Scheur). The data suggest that an important part of the HCMS, measured at both sites, could have been resuspended from the very soft mud deposits in the navigation channels and adjacent areas. 6. Conclusion Measurements have been collected at two locations in the vicinity of the port of Zeebrugge and its navigation channels. Between autumn 25 and winter 27, three stormy periods have been selected with similar wave conditions. The data have shown that during or after a storm, the SPM concentration increases significantly and that HCMS have been formed. SPM concentration is clearly related to high waves and winds. The wind direction and the advection of water masses, the previous history and the availability of fine-grained sediments in fluffy layers, the very soft mud deposits around navigation channels, and the erosion of medium-consolidated mud of Holocene age influence the SPM signal. The data suggest that for the generation of very high SPM concentrations near the bed, significant amounts of finegrained sediments have to be resuspended and/or eroded. The navigation channels and other areas with soft mud have been found to be the major source of the fine-grained sediments during storms. This result is important as it suggest that the deepening of the navigation channels has made available fine-grained matter that contributes significantly to the formation of high concentration mud suspensions. This suggests that HCMS were probably less frequent in the past when anthropogenic activities where limited. Acknowledgement This study was partly funded by the by the Belgian Science Policy within the framework of the QUEST4D project (SD/NS/6A) and by the Maritime Access Division of the Ministry of the Flemish Community in the framework of the MOMO project. G. Dumon (Coastal Service, Ministry of the Flemish Community) made available wave measurement data. We want to acknowledge the crew of the RV Belgica, Zeearend and Zeehond for their skilful mooring and recuperation of the tripod. The assistance of marine divers during recuperation of the tripod on February 7 27 is also acknowledged. The measurements would not have been possible without technical assistance of A. Pollentier, J.-P. De Blauwe and J. Backers (measuring service of MUMM, Oostende). 7. References Andersen, T. J., Fredsoe, J., Pejrup, M. 27. In-situ estimation of erosion and deposition thresholds by Acoustic Doppler Velocimeter (ADV). Estuarine, Coastal and Shelf Science, 75, Ariathurai, C. R A finite element model for sediment transport in estuaries. University of California, Davis. 9

45 Cacchione, D.A., Wiberg, P.L., Lynch, J.F., Irish, J.D., Traykovski, P Estimates of suspended-sediment flux and bedform activity on the inner portion of the Eel continental shelf. Marine Geology, 154, de Wit, P.J., Kranenburg, C The wave-induced liquefaction of cohesive sediment beds. Estuarine, Coastal and Shelf Science, 45, Eisma, D Supply and deposition of suspended matter in the North Sea. In: Holocene Marine Sedimentation in the North Sea Basin. (Nio, D.D., Shuttenhelm, R.T.E., van Weering, T.C.E., eds.). IAS Special Publication, 5, Blackwell Scientific, Fettweis, M., Van den Eynde, D. 23. The mud deposits and the high turbidity in the Belgian-Dutch coastal zone, Southern bight of the North Sea. Continental Shelf Research, 23, Fettweis, M., Nechad, B., Van den Eynde, D. 27a. An estimate of the suspended particulate matter (SPM) transport in the southern North Sea using SeaWiFS images, in-situ measurements and numerical model results. Continental Shelf Research, 27, Fettweis, M., Du Four, I., Zeelmaekers, E., Baeteman, C., Francken, F., Houziaux, J.-S., Mathys, M., Nechad, B., Pison, V., Vandenberghe, N., Van den Eynde, D., Van Lancker, V., Wartel, S. 27b. Mud Origin, Characterisation and Human Activities (MOCHA). Belgian Science Policy, EV/35, 59pp. Gerritsen, H., Boon, J.G., van der Kaaij, T., Vos, R.J. 21. Integrated modelling of suspended matter in the North Sea. Estuarine, Coastal and Shelf Science, 53, Green, M.O., Vincent, C.E., McCave, I.N., Dickson, R.R., Rees, J.M., Pearson, N.D Storm sediment transport: observations from the British North Sea shelf. Continental Shelf Research, 15, Günther, H., Rosenthal, W., The hybrid parametrical (HYPAS) wave model. In: Ocean wave modelling, Swamp group. Plenum Press, New York, Kern, U., Haag, I., Schürlein, V. Holzwarth, M., Westrich, B Ein Strömungskanal zur Ermittlung der tiefenabhängigen Erosionsstablität von Gewässersedimenten. Wasserwirtschaft, DWA, 89(2), Koutitas, C.G., Mathematical Models in Coastal Engineering. Pentech Press, London, UK, 156pp. Lafite, R., Shimwell, S., Grochowski, N., Dupont, J.-P., Nash, L., Salomon, J.-C., Cabioch, L., Collins, M., Gao, S. 2. Suspended particulate matter fluxes through the Strait of Dover, English Channel: observations and modelling. Oceanologica Acta, 23, Lauwaert, B., Bekaert, K., Berteloot, M., De Brauwer, D., Fettweis, M., Hillewaert, H., Hoffman, S., Hostens, K., Mergaert, K., Moulaert, I., Parmentier, K., Vanhoey, G., Verstraeten, J. 28. Synthesis report on the effects of dredged material disposal on the marine environment. BMM, ILVO, AK & AMT rapport, BL/28/1, 19pp. Le Bot, S., Van Lancker, V., Deleu, S., de Batist, M., Henriet, J.-P., Haegeman, W. 25. Geological and geotechnical properties of Eocene and Quarternary deposits on the Belgian continental shelf: synthesis in the context of offshore wind farming. Netherlands Journal of Geosiences, 84, Le Hir, P., Monbet, Y., Orvain, F. 27. Sediment erodability in sediment transport modelling: Can we account for biota effects? Continental Shelf Research, 27, Li, Y., Mehta, A.J. 2. Fluid mud in the wave-dominated environment revisited. In: Coastal and Estuarine Fine Sediment Dynamics. (McAnally, W.H., Mehta, A.J., eds.), Proceedings in Marine Science, 3,

46 Luyten, P. J., Jones, J. E., Proctor, R., Tabor, A., Tett, P., Wild-Allen, K COHERENS, a coupled hydrodynamicalecological model for regional and shelf seas: User Documentation. MUMM report, Brussels, Belgium. 911pp. Maa, P.-Y., Mehta, A. J Mud erosion by waves: a laboratory study. Continental Shelf Research, 7, Nikora, V.I., Goring, D.G., ADV measurements of turbulence: can we improve their interpretation? Journal of Hydraulic Engineering, 124, Pison, V., Ozer, J. 23. Operational products and services for the Belgian coastal waters. In: Building the European capacity in operational modeling, Proc. 3 rd Int. Conf. on EuroGOOS (Dahlin, H., Flemming, N.C., Nittis, K., Petersson, S.E., eds.). Elsevier Oceanography Series 69, Pope, N. D., Widdows, J., Brinsley, M. D. 26. Estimation of bed shear stress using the turbulent kinetic energy approach - A comparison of annular flume and field data. Continental Shelf Research, 26, Shi, J.Z., Gu, W.-J., Wang, D.-Z. 28. Wind wave-forced fine sediment erosion during slack water periods in Hangzhou Bay, China. Environmental Geology, 55, Silva-Jacinto, R. Le Hir, P. 21. Response of stratified muddy beds to water waves. In: Coastal and Estuarine Fine Sediment Processes (McAnally, W.H., Mehta, A.J., eds.). Proceedings in Marine Science, 3, Stapleton, R.L., Huntley, D.A., Seabed stress determination using the inertia dissipation method and the turbulent kinetic energy method. Earth Surface Processes and Landforms 2, Streeter, V.L Fluid Mechanics. McGraw-Hill Book Company, New York, 75pp. Thompson, C.E.L., Amos, C.L., Jones, T.E.R., Chaplin, J. 23. The Manifestation of Fluid-transmitted Bed Shear Stress in a Smooth Annular Flume - A Comparison of Methods. Journal of Coastal Research, 19, Traykovski, P., Wiberg, P.L., Geyer, W.R. 27. Observations and modeling of wave-supported sediment gravity flows on the Po prodelta and comparison to prior observations from the Eel shelf. Continental Shelf Research, 27, Van Alphen, J.S.L.J A mud balance for Belgian-Dutch coastal waters between 1969 and Netherlands Journal of Sea Research, 25, Van den Eynde, D., MU-WAVE: An operational wave forecasting system for the Belgian coast. 3rd Int Workshop on Wave Hindcasting and Forecasting, Montréal, Canada, van Ledden, M., van Kesteren, W.G.M., Winterwerp, J. 24. A conceptual framework for the erosion behaviour of sand-mud mixtures. Continental Shelf Research, 24, van Kessel, T., Kranenburg, C Wave-induced liquefaction and flow of subaqueous mud layers. Coastal Engineering, 34, Verney, R., Deloffre, J., Brun-Cottan, J. C., Lafite, R. 27. The effect of wave-induced turbulence on intertidal mudflats: Impact of boat traffic and wind. Continental Shelf Research, 27, Verwaest, T. 27. Using simple semi-empirical models for integrated assessment of scenarios for a navigation channel. The case of the port of Ostend, Belgium. Coastal Sediments 27, New Orleans, USA. Voulgaris, G., Trowbridge, J.H., Evaluation of the Acoustic Doppler Velocimeter (ADV) for Turbulence Measurements. Journal of Atmospheric and Oceanic Technology, 15, Williamson, H., Torfs, H Erosion of mud/sand mixtures. Coastal Engineering, 29,

47 Winterwerp, J On the dynamics of high-concentrated mud suspensions. PhD thesis, TU Delft, The Netherlands, 172pp. Witt, O., Westrich, B. 23. Quantification of erosion rates for undisturbed contaminated cohesive sediment cores by image analysis. Hydrobiologia, 494, Wright, L.D., Friedrichs, C.T. 26. Gravity-driven sediment transport on continental shelves: A status report. Continental Shelf Research, 26,

48 Table 1: Oceanographic instruments mounted on the tripod and distance in m above bed (mab). The saturation concentration of the OBS s was 3.3 g l -1. Survey Nr ADV ADP OBS1 OBS2 SBE37 LISST1 X 25/ / / Table 2: Tripod deployments at MOW1 and Blankenberge. Survey Nr Location Start - End date (dd/mm/yyyy hh:mm) Duration (days) 25/29 MOW1 22/11/25 8:11-5/12/25 9: /23 Blankenberge 8/11/26 14:29 27/11/26 9: /27 Blankenberge 18/12/26 1:44-7/2/27 13:

49 Figure 1: (A) Yearly and depth-averaged SPM concentration (mg/l) in the southern North Sea, derived from 37 SeaWiFS images ( ), see Fettweis et al (27a). (B) Bathymetry (m below mean sea level) of the Belgian coastal area. Indicated are the tripod measuring stations (black dots: MOW1, Blankenberge), the navigation channel (Pas van het Zand and Scheur) and the location of the box-core samples for erosion behaviour measurements (white dots). Coordinates are in latitude ( N) and longitude ( E). 14

50 SPM conc (mg/l) Depth below surface (m) Velocity (m/s) Shear stress (Pa) SPM1 SPM Time (days) 4 2 Wave height (m) Figure 2: MOW1 site (survey 25/29), tripod measurements of 22 November 5 December 25 (survey 25/29). From up to down: depth below water surface (m) and significant wave heights; ADV current velocity (m/s); shear stress (Pa) derived from the ADV and SPM concentration at.3 mab (SPM1) and 1.9 mab (SPM2). 15

51 SPM conc (mg/l) Depth below surface (m) Velocity (m/s) Shear stress (Pa) SPM1 SPM Time (days) Wave height (m) Figure 3: Blankenberge site (survey 26/23), tripod measurements of 7-2 November 26. From up to down: depth below water surface (m) and significant wave heights; ADV current velocity (m/s); shear stress (Pa) derived from the ADV; and SPM concentration at.2 mab (SPM1) and 2.2 mab (SPM2). 16

52 SPM conc (mg/l) Depth below surface (m) Velocity (m/s) Shear stress (Pa) SPM1 SPM Time (days) Wave height (m) Particle size (µm) Figure 4: Blankenberge site (survey 26/27), tripod measurements of 27 December 26 1 January 27. From up to down: depth below water surface (m) and significant wave heights; ADV current velocity (m/s) and mean particle size (µm); shear stress (Pa) derived from the ADV; and SPM concentration at.2 mab (SPM1) and 2.2 mab (SPM2). 17

53 A B C τce (Pa) τce (Pa) τce (Pa) Tau-crit Bulk density depth below bed [cm] 2 3 depth below bed (cm) 2 3 depth below bed (cm) Tau-crit bulk density 5 1,1 1,2 1,3 1,4 1,5 1,6 1,7 1,8 bulk density [g cm - ³] 5 tau-crit bulk density 1,1 1,2 1,3 1,4 1,5 1,6 1,7 1,8 bulk density (g cm - ³) 5 1,1 1,2 1,3 1,4 1,5 1,6 1,7 1,8 bulk density (g cm - ³) Figure 5: Depth profile of critical erosion shear stress (τ ce ) and bulk density of box cores. (A) 7 cm of soft mud above 4 cm of medium-consolidated mud (Holocene) with intercalations of sand, muddy sand and shell layers (location: near Zeebrugge). (B) 5 cm of very soft mud with fine sand above 4 cm of alternating medium-consolidated mud and thin muddy sand layers with shells (Holocene). (C) 4 cm of freshly deposited mud from the navigation channel ( Pas van het Zand) on top of medium-consolidated mud (Holocene). The sandy and shelly layers have a lower τ ce and a higher bulk density. 18

54 Figure 6: Hydrodynamic model results of maximum bottom shear stress (Pa) during a spring tide without wind (A) and during the November 25 storm (B). Coordinates are in latitude ( N) and longitude ( E). 19