Michael Fettweis, Bouchra Nechad, Dries Van den Eynde, Frederic Francken, Vera Van Lancker

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 januari 2010 ‐ 30 juni 2010)  Michael Fettweis, Bouchra Nechad, Dries Van den Eynde, Frederic Francken, Vera Van Lancker 

MOMO/5/MF/201006/NL/AR/1 

Voorbereid voor Afdeling Maritieme Toegang, Departement Mobiliteit en Openbare Werken, Ministerie  van de Vlaamse Gemeenschap, contract MOMO  BMM 100 Gulledelle B–1200 Brussel België

Inhoudstafel  1. 1.1. 1.2. 1.3. 1.4.

Inleiding  Voorwerp van deze opdracht  Algemene Doelstellingen  Taken (januari 2010 ‐ december 2011)  Publicaties (januari 2010 – december 2011) 

4 4 4 5 7

2. 2.1. 2.1.1. 2.1.2. 2.1.3. 2.2. 2.2.1. 2.2.2. 2.2.3. 2.2.4. 2.2.5. 2.3. 2.3.1. 2.3.2. 2.3.3. 2.4.

Vergelijking  tussen  in  situ  en  remote  sensing  metingen  van  SPM  concentratie  8 Werkwijze  9 In‐situ tripode metingen  10 In‐situ 13‐uursmetingen en verticale profielen  11 Remote sensing data  13 Resultaten  14 Significante golfhoogte  14 SPM concentratie profielen  14 In situ en satelliet data op hetzelfde ogenblik  17 Frequentieverdeling van SPM concentratie  17 Oppervlaktecorrectie van tripode data  22 Discussie  22 Representativiteit van SPM concentratie data  22 Staalnamemethode  23 Meetonzekerheid  24 Conclusies  25

3.

Profielen saliniteit en temperatuur 

27

4. Referenties  29     Appendix 1: Fettweis M, Nechad B, Van Lancker V, Van den Eynde D. 2010. Evaluation of  in situ and remote sensing sampling methods of SPM concentration. AGU Ocean Scien‐ ce Meeting, 22‐26 February, Portland (USA).  Appendix 2: Chen P, Yu J, Fettweis M, Van den Eynde D, Maggi F. 2010. Flocculation in a  nutrient‐rich  coastal  area  (southern  North  Sea):  Measurements  and  modelling.  AGU  Ocean Science Meeting, 22‐26 February, Portland (USA).  Appendix 3: Fettweis M, Van den Eynde D, Francken F, Van Lancker V. 2010. SPM dynam‐ ics measured with an automated tripod in the Belgian nearshore area: natural dynam‐ ics and anthropogenic effects. Liège Colloquium, 26‐30 April.  Appendix 4:  Baeye M, Fettweis M, Van Lancker V, Francken F. 2010. Fluid mud dynamics  derived from ADV altimetry, Belgian Coastal Zone. Liège Colloquium, 26‐30 April.  Appendix  5:  Fettweis  M,  Francken  F,  Van  den  Eynde  D,  Verwaest  T,  Janssens  J,  Van  Lancker  V.  2010.  Storm  influence  on  SPM  concentrations  in  a  coastal  turbidity  maxi‐ mum area with high anthropogenic impact (southern North Sea). Continental Shelf Re‐ search. doi:10.1016/j.csr.2010.05.001  Appendix 6: Fettweis M, Baeye M, Francken F, Lauwaert B, Van den Eynde D, Van Lancker  V, Martens C, Michielsen T. Monitoring the effects of disposal of fine sediments from  maintenance dredging on SPM concentration Marine Pollution Bulletin (submitted) 

2

Appendix 7: SPM concentratieprofielen uit de 13‐uursmetingen te MOW1 en Kwintebank.  Appendix  8:  Saliniteits‐  en  temperatuursprofielen  uit  de  13‐uursmetingen  te  MOW1  en  Kwintebank. 

3

1. Inleiding  1.1. Voorwerp van deze opdracht   Het MOMO‐project (MOnitoring en MOdellering van het cohesieve sedimenttransport en  de evaluatie van de effecten op het mariene ecosysteem ten gevolge van bagger‐ en stor‐ toperatie) 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  2010  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 metin‐ gen. De combinatie van monitoring en modellering zal gegevens kunnen aanleveren over  de  transportprocessen  van  deze  fijne  fractie  en  is  daarom  fundamenteel  bij  het  beant‐ woorden 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 eco‐ systeem 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 (2002‐2004, 2004‐2006,  2006‐2008,  2008‐2009)  kan  gevonden  in  het  “Syntheserapport  over  de  effecten  op  het  mariene  milieu  van  baggerspeciestortingen”  (Lauwaert  et  al.,  2004;  2006;  2008,  2009a)  dat uitgevoerd werd conform art. 10 van het K.B. van 12 maart 2000 ter definiëring van de  procedure voor machtiging van het storten in de Noordzee van bepaalde stoffen en mate‐ rialen. Voor een uitgebreide beschrijving wordt verwezen naar de halfjaarlijkse rapporten. 

1.2. 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  sedi‐ mentatie te verminderen op de baggerplaatsen en anderzijds efficiënter te storten. Hier‐ voor  is  een  grondige  kennis  van  de  omgevingsfactoren  (Hydro‐meteorologische  en  sedi‐ mentologische condities, morfologie en geometrie van de vaargeulen en havens) essenti‐ eel, zie Pianc rapport (2008).   Aanslibbing  in  havens  en  vaargeulen  wordt  beïnvloed  door  de  gemiddelde  condities  (getijamplitude, SPM concentratie, sedimentoorsprong, saliniteit) als ook extreme gebeur‐ tenissen (vb storm). De efficiëntie van een stortplaats wordt bepaald door fysische (sedi‐ menttransport i.f.v. getij, doodtij‐springtij, wind, golven), economische en ecologische as‐

4

pecten. 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, re‐ circulatie), 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 bagger‐ specie, 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.  

1.3. Taken (januari 2010 ‐ december 2011)  In het bijzonder is bij het opstellen van de hieronder vermelde taken rekening gehouden  met  de  aanbevelingen  voor  de  minister  ter  ondersteuning  van  de  ontwikkeling  van  een  versterkt milieubeleid zoals geformuleerd in het Syntheserapport over de effecten op het  mariene  milieu  van  baggerspeciestortingen  (Lauwaert  et  al.,  2009b).  Tijdens  de  voorbije  twee jaren werd conform de aanbeveling van de ambtelijke werkgroep de samenwerking  met het WLH geïntensifieerd om zowel informatie, data en metingen uit te wisselen, als  ook het onderzoek op elkaar af te stemmen, zodat efficiënt aan de noden van de eindge‐ bruikers kan worden tegemoetgekomen.  

Taak 1: In situ 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 me‐ tingen  zullen  tijdsreeksen  en  verticale  profielen  worden  verzameld  van  de  stroming,  de  concentratie  aan  en  de  korrelgrootteverdeling  van  het  suspensiemateriaal,  de  tempera‐ tuur en de saliniteit. De optische metingen (transmissometer, Optical Backscatter Sensor)  zullen  gecalibreerd  worden  met  de  opgemeten  hoeveelheid  materie  in  suspensie  (gravi‐ metrische bepalingen na filtratie) om te komen tot concentraties. Stalen van suspensiema‐ teriaal 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 pe‐ riode. Het preferentieel station is MOW1, waar getracht zal worden om een quasi conti‐ nue meetreeks te verzamelen. Hiervoor zal gebruik gemaakt worden van 2 tripodes. Nadat  een meting beëindigd is zal deze tripode voor onderhoud aan wal gebracht worden terwijl  de tweede op de meetlocatie zal worden verankerd. Doel van deze metingen is het ver‐ zamelen  van  continue  metingen  in  het  gebied  zodat  een  verband  tussen  meteorologie,  seizoenen en getijden kan gelegd worden en de slibconcentratie gelegd kan worden.  

Taak 2: Haalbaarheidsstudie voor automatisch en online doorsturen data  gemeten met tripode  De data opgemeten met de tripode worden tot op heden enkel opgeslagen in het geheu‐ gen van de meettoestellen en zijn dus pas beschikbaar nadat de meting is afgelopen en de  tripode bovengehaald is. Met het oog op het efficiënter maken van de stortoperaties (zie  algemene  doelstellingen)  is  het  noodzakelijk  om  deze  informatie  on‐line  te  kunnen  heb‐

5

ben zodat de beste stortlocatie voor de gegeven hydro‐, meteo‐ en sedimentconcentratie  toestand kan gekozen worden.   De bedoeling van deze taak is om de verschillende bestaande systemen voor het au‐ tomatisch doorsturen van de meetdata te evalueren. Er zal hierbij contact genomen wor‐ den met MDK (G. Dumon), waar reeds heel wat ervaring hieromtrent aanwezig is. Na het  beëindigen van deze taak zal geëvalueerd kunnen worden op welke termijn een online da‐ tatransfersysteem operationeel kan zijn.  

Taak 3: Verwerking en interpretatie van metingen  De metingen vergaard tijdens de 13‐uursmetingen aan boord van de Belgica en met de tri‐ pode worden verwerkt en geïnterpreteerd. Hiervoor werd in het verleden reeds heel wat  software  ontwikkeld  (getijgemiddelde  waarden,  valsnelheid,..).  Naast  rapportage  van  de  data zal bijkomend aandacht geschonken worden aan:   • Bepaling van de SPM concentratie uit akoestische data (ADP, ADCP). Om de represen‐ tativiteit van deze methode te bepalen voor de data van het BCP zullen deze data ver‐ geleken worden met SPM concentraties afgeleid uit optische data (OBS).   • De verticale profielen (SPM concentratie, saliniteit, temperatuur) opgemeten tijdens  de 13‐uursmetingen zullen worden geanalyseerd. De SPM concentratie profielen leve‐ ren essentiële informatie voor het corrigeren van satellietdata. Verticale profielen van  saliniteit en temperatuur zijn belangrijk om de stratificatie ter hoogte van Zeebrugge  te kennen en de invloed op de wateruitwisseling van de haven.   • Bodemstalen en suspensiestalen (centrifuge) zullen worden geanalyseerd om de kor‐ relgrootteverdeling, het kalkgehalte en de organische fractie te bepalen.  

Taak 4: Evaluatie van in situ en remote sensing meetmethodes  De bedoeling van deze taak is de evaluatie van de temporele heterogeniteit van SPM con‐ centraties gemeten in de Belgische kustzone met in situ en met remote sensing technie‐ ken (satelliet). Omdat ‘match‐ups’ (satelliet meting is op hetzelfde ogenblik als in situ me‐ ting)  zelden  voorkomen,  zal  gebruik  gemaakt  worden  van  statistische  technieken  om  de  verschillen en overeenkomsten tussen de datasets te evalueren. Hierdoor kunnen verschil‐ lende data met elkaar vergeleken worden die niet op hetzelfde moment gemeten werden.  Verder zal aandacht geschonken worden aan de ‘sampling’ methode, aan de representati‐ viteit van de bestaande data sets en aan de gebruikte meetschema’s. Hieruit kan een sug‐ gestie worden gedaan om het huidige meetschema aan te passen en te optimaliseren.  

Taak 5: Verfijnen slibtransportmodel  Het gebruik van een numeriek sedimenttransportmodel vereist een regelmatige validatie  van de modelresultaten met meetgegevens en een verbetering van de beschrijving van de  processen in het model. Er zal verder gewerkt worden aan een calibratie van het floccula‐ tiemodel en de implementatie van het bodemmodel.  

Taak 6: Alternatieve stortschema’s   Onderzoek  naar  alternatieve  stortschema’s  en  stortlocaties  zal  voortgezet  worden  voor  stortplaats  B&W  Zeebrugge  Oost  conform  de  aanbevelingen  voor  de  minister  geformu‐ leerd in het syntheserapport 2009. Om de efficiëntie van mogelijke alternatieve locaties of  een andere bestaande stortplaats te testen wordt een terreinproef voorbereid in samen‐ werking met aMT. Tijdens de voorbereiding zal o.a. de duur ervan vastgelegd moeten wor‐ den, de monitoring moeten worden opgezet en hoe de proef zal moeten worden geëvalu‐ eerd. Hiervoor zal gebruik kunnen gemaakt worden van de resultaten uit taak 4 en de er‐

6

varing opgemaakt tijdens het topslib‐project.  

1.4. Publicaties (januari 2010 – december 2011)  Hieronder is een lijst met rapporten, publicaties en deelnamen aan workshops en confe‐ renties  waar  resultaten  en  data  verzameld  in  het  kader  van  het  MOMO  project  werden  voorgesteld:  Activiteits‐, Meet‐ en Syntheserapporten  Fettweis M, Nechad B, Van den Eynde D, Francken F, Van Lancker V. 2010. MOMO activi‐ teitsrapport  1  (1  januari  2010  –  30  juni  2010).  BMM‐rapport  MOMO/5/MF/201006/NL/AR/1, 32pp + app.  Conferenties/Workshops:  Fettweis M, Van den Eynde D, Francken F, Van Lancker V. 2010. SPM dynamics measured  with an automated tripod in the Belgian nearshore area: natural dynamics and anthro‐ pogenic effects. Liège Colloquium, 26‐30 April. (poster)  Baeye M, Fettweis M, Van Lancker V, Francken F. 2010. Monitoring morphological changes  using near‐bed ADV altimetry. Liège Colloquium, 26‐30 April. (poster)  Fettweis  M, Nechad  B, Van Lancker  V,  Van  den Eynde D.  2010. Evaluation  of  in  situ  and  remote sensing sampling methods of SPM concentration. AGU Ocean Science Meeting,  22‐26 February, Portland, USA. (presentatie).  Chen P, Yu J, Fettweis M, Van den Eynde D, Maggi F. 2010. Flocculation in a nutrient‐rich  coastal area (southern North Sea): Measurements and modelling. Poster at AGU Ocean  Science Meeting, 22‐26 February, Portland, USA. (poster).  Publicaties (tijdschriften, boeken)  Fettweis M, Baeye M, Francken F, Lauwaert B, Van den Eynde D, Van Lancker V, Martens  C, Michielsen T. Monitoring the effects of disposal of fine sediments from maintenance  dredging on SPM concentration Marine Pollution Bulletin (submitted)  Fettweis M, Nechad B. 2010. Evaluation of in situ and remote sensing sampling methods  for SPM concentrations on Belgian continental shelf (southern North Sea). Ocean Dy‐ namics (in revision)  Fettweis  M,  Francken F, Van  den  Eynde D,  Verwaest  T,  Janssens  J,  Van  Lancker  V.  2010.  Storm influence on SPM concentrations in a coastal turbidity maximum area with high  anthropogenic  impact  (southern  North  Sea).  Continental  Shelf  Research.  doi:10.1016/j.csr.2010.05.001   

7

2. Vergelijking tussen in situ en remote sensing metin‐ gen van SPM concentratie  De  dynamica  van  gesuspendeerd  particulair  materiaal  (SPM)  controleert  het  sediment‐ transport, de afzetting en resuspensie van slib, de primaire productie en de samenstelling  van  benthische  gemeenschappen.  Kennis  van  SPM  concentratie  variabiliteit  in  het  Belgi‐ sche kustgebieden, waar de SPM concentratie bijzonder hoog is, is essentieel om tot een  efficiënter  stortbeleid  te  komen  en  om duurzame  socio‐economische  activiteiten  te  ont‐ wikkelen parallel met een bescherming van mariene gebieden. De SPM concentratie vari‐ eert sterk in de kustzone afhankelijk van seizoenen, getij en niet harmonische gebeurte‐ nissen zoals stormen. Een optimale meetstrategie is nodig om deze variabiliteit te meten.   In  situ  SPM  concentratie  metingen  gebeuren  vanuit  schepen  en  met  alleenstaande  structuren zoals tripoden en meetpalen. Met tripoden, meetpalen, boeien of andere plat‐ vormen  kunnen  continue  tijdseries  van  SPM  concentraties  op  specifieke  locaties  verza‐ meld worden en dit gedurende een langere tijdspanne (e.g. Cacchione et al., 1995; Li et  al., 1997; Blewett & Huntley, 1998; Ogston et al., 2000; Pepper & Stone, 2004; Ma et al.,  2008, Badewien et al., 2009). Schepen met ferry‐box systemen laten ook toe om data te  verzamelen over langere tijdsreeksen en dit op verschillende locaties (Buijsman & Ridde‐ rinkhof, 2007). Gewoonlijk wordt hier echter enkel de oppervlakte bemonsterd. Metingen  met onderzoeksschepen zijn beperkt in tijd en golfcondities, maar laten toe om verticale  profielen te verzamelen. Grote oppervlakten worden bestreken door satellieten, de data  zijn beperkt tot de oppervlakte (Bowers et al., 2002; Nechad et al., 2003; Zawada et al.,  2007; Eleveld et al., 2008; Doxaran et al., 2009). Voor de zuidelijke Noordzee zijn 40 tot 85  satellietbeelden  beschikbaar  per  jaar  (gemiddeld  63  beelden  gedurende  2003  tot  2008)  met een wolkenbedekking kleiner dan 50%.   Tijdseries van in situ SPM concentraties en satellietbeelden zijn nodig om het suspen‐ sietransport in kustgebieden te analyseren (Ruhl et al., 2001; Fettweis et al., 2007). Er zijn  echter een aantal nadelen verbonden met beide databronnen. Zo zijn satellietbeelden be‐ perkt door hun lage tijdsresolutie en omdat ze enkel de oppervlakte bemonsteren, terwijl  in situ metingen een lage horizontale resolutie hebben. Scheepstijd en budget zijn dikwijls  beperkt, het is daarom van groot belang om één of een combinatie van meetmethoden te  kiezen die een representatief substaal van de hele populatie kan geven in het tijds‐ en/of  ruimtedomein.  Indien  langdurige  variaties  veroorzaakt  door  natuurlijke  of  menselijke  in‐ grepen  geïdentificeerd  dienen  te  worden, dan  moeten de  overheersende  signalen  (getij,  doodtijspringtij en seizoen) eruit gefilterd worden. Dit vraagt een voldoende hoge meet‐ frequentie en dit gedurende een lange tijd. Voorbeelden van ontwerp en/of evaluatie van  meetschema’s zijn schaars, zeker in sedimenttransportstudies (Bograd et al., 1999; Shindo  & Otsuki, 1999; Caeiro et al., 2003; Hall & Davies, 2005; Werdell et al., 2009). Dikwijls kan  ook de beste meetstrategie niet worden toegepast omdat zij afhangt van een compromis  tussen kosten en baten. In praktijk worden daarom verschillende objectieven nagestreefd  die elk een aangepaste suboptimale strategie volgen. Autonome meetstations (tripoden)  met bijna continue metingen van SPM concentratie zijn relatief eenvoudig te ontwikkelen,  maar kunnen geen grote oppervlakte omvatten. Satellieten hebben een lage tijdsresolutie  en er treden hiaten op in de data door wolkenbedekking. Zo zijn er maar weinig satelliet‐ data beschikbaar tijdens stormen. Satellieten missen dus meestal de hoge SPM concentra‐

8

ties geassocieerd aan deze perioden. Satellieten kunnen daarenboven de heel hoge SPM  concentraties niet meten, omdat het zichtbare band (Bowers et al., 1998; Doxaran et al.,  2002; Nechad et al., 2010). Bij de interpretatie van de data is het nodig om de meetonze‐ kerheden van de gebruikte meetmethode en instrumentatie te kennen.   Het  doel  van deze  studie  (zie  taak  4 uit  §1.3)  is  om  de temporele  heterogeniteit  van  SPM concentratie te evalueren voor de Belgische kustzone, gebruikmakend van een groot  aantal data afkomstig van verschillende bronnen (MODIS satelliet en in situ metingen), om  een statistische analyse van deze data uit te voeren en om na te gaan hoe de verschillende  data  de  dynamica  van  kustzones  beschrijven.  Omdat  maar  enkele  match‐ups  (satelliet‐ beeld op hetzelfde ogenblik als de in situ meting) beschikbaar zijn, werden statistische me‐ thoden gebruikt om de data sets te evalueren. Deze benadering is nieuw en laat toe om  verschillende data sets te vergelijken die niet noodzakelijk op hetzelfde ogenblik werden  verzameld.  

  Figuur 2.1: Kaart van de zuidelijke Noordzee met de in situ SPM concentratie meetlocaties  MOW1 en Kwintebank. De achtergrond bestaat uit de gemiddelde oppervlakte SPM con‐ centratie (mg/l) afgeleid van MODIS beelden (2003‐2008). 

2.1. Werkwijze  SPM concentratie data werden verzameld met OBS instrumenten, die gebruikt werden op  de  R/V  Belgica  tijdens  13‐uursmetingen  en  gemonteerd  waren  op  de  tripode,  en  met  MODIS  satellietbeelden.  Om  de  oppervlakte  data  van  de  satellietbeelden  te  vergelijken  met de tripode data dicht tegen de bodem, warden de tripode data gecorrigeerd naar de  oppervlakte  toe.  Dit  werd  gedaan  met  behulp  van  de  13‐uursmetingen,  tijdens  dewelke  verticale  profielen  van  SPM  concentratie  werden  opgemeten.  De  gemeten  significante  golfhoogte te Bol van Heist (zie figuur 2.1) voor de periode 2003‐2008 werd gebruikt om  de meteorologische en golfinvloed te karakteriseren.  

9

  Figuur 2.2: Tripode op het achterdek van de Belgica.  Tabel 2.1: Tripode metingen te MOW1 en op de Kwintebank.   Locatie  Kwintebank  MOW1‐1  MOW1‐2  MOW1‐3  MOW1‐4  MOW1‐5  MOW1‐6  MOW1‐7  MOW1‐8 

Start datum en tijd  02/03/2004 15:10  18/10/2004 12:12  08/02/2005 08:10  04/04/2005 11:10  22/06/2005 08:15  22/11/2005 08:27  13/02/2006 11:25  27/03/2006 10:36  15/05/2006 12:24 

Eind datum en tijd  11/03/2004 09:40  08/11/2004 11:10  18/02/2005 10:00  15/04/2005 07:24  11/07/2005 11:35  05/12/2005 09:07  27/02/2006 10:28  18/04/2006 09:26  15/06/2006 10:22 

Duur (dagen)  8.83  20.92  10.08  10.83  19.13  13.04  13.96  19.08  30.96 

2.1.1. In‐situ tripode metingen  Een tripode werd gebruikt om o.a. SPM concentraties en stroomsnelheid te meten (zie fi‐ guur  2.2).  Hiervoor  werden  een  SonTek  3  MHz  Acoustic  Doppler  Profiler,  een  SonTek  5  MHz  Acoustic Doppler Velocimeter  Ocean, een  Sea‐Bird  SBE37  CT  systeem  en  twee  OBS  sensoren (ene op ongeveer 0.2 en de andere op ongeveer 2 m boven de bodem, verder  afgekort als mab = m above bottom) gemonteerd op de tripode. Enkel de OBS data wor‐ den  hierbij  gebruikt  omdat  door  onvoorspelbaarheden  in  de  sedimentdynamica  een  nauwkeurige schatting van de SPM concentratie uit akoestische backscatter signalen nog  steeds beperkt is (Hoitink & Hoekstra, 2005; Bartholomä et al., 2009). Tijdens de periode  2003‐2006 werd gedurende 147 dagen gemeten, waarvan 9 dagen op de Kwintebank en  138 dagen te MOW1 (tabel 2.1 en figuur 2.1). Respectievelijk 33%, 20%, 22% en 25% van  de data te MOW1 werden verzameld tijdens de lente, zomer, herfst en winter. De Kwinte‐ bank data werden alle verzameld tijdens de winter. De meetfrequentie was 2 minuten op  de Kwintebank en varieerde tussen 2 minuten en 20 minuten te MOW1. Voor analyse en  interpretatie werd een frequentie van 20 minuten gebruikt te MOW1, resulterend in on‐ geveer 10000 data te MOW1.  

10

Tabel  2.2:  13‐uursmetingen  te  MOW1  en  op  de  Kwintebank.  MT=Gemiddeld  tij,  NT=doodtij, ST=springtij.  Id‐Nr  2001/06‐A  2002/27‐B  2003/04‐A  2003/15  2003/17  2003/22  2003/25  2004/04  2004/05  2004/24  2004/25‐A  2005/02  2005/07‐A  2005/15‐A  2005/15‐B  2005/29  2006/06  2006/10‐A  2007/11‐A  2007/11‐B  2007/16  2007/25‐B  2008/02‐A  2008/02‐B   

Locatie  MOW1  MOW1  MOW1  Kwintebank  Kwintebank  MOW1  Kwintebank  Kwintebank  Kwintebank  MOW1  MOW1  MOW1  MOW1  Kwintebank  MOW1  MOW1  MOW1  MOW1  Kwintebank  MOW1  MOW1  MOW1  MOW1  Kwintebank 

Begin datum  07/03/2001 16:11  26/11/2002 18:34  19/02/2003 18:08  11/06/2003 18:20  26/06/2003 16:40  08/09/2003 18:00  09/10/2003 15:30  02/03/2004 20:00  11/03/2004 18:00  18/10/2004 13:20  08/11/2004 18:20  07/02/2005 14:00  04/04/2005 19:20  20/06/2005 20:20  21/06/2005 17:15  21/11/2005 17:00  28/03/2006 18:38  15/05/2006 17:28  02/05/2007 14:38  03/05/2007 06:43  10/07/2007 04:30  23/10/2007 01:30  06/02/2008 11:00  07/02/2008 17:00 

Eind datum  08/03/2001 09:38  27/11/2002 07:24  20/02/2003 05:01  12/06/2003 07:00  27/06/2003 05:00  09/09/2003 08:00  10/10/2003 04:20  03/03/2004 08:40  12/03/2004 07:10  19/10/2004 02:20  09/11/2004 07:00  08/02/2005 06:10  05/04/2005 08:20  21/06/2005 09:00  22/06/2005 06:10  22/11/2005 06:00  28/03/2006 06:00  16/05/2006 06:20  03/05/2007 03:00  03/05/2007 20:00  10/07/2007 17:00  23/10/2007 12:00  07/02/2008 01:00  08/02/2008 06:00 

  MT  NT  ST  MT  NT  MT  ST  NT  ST  ST  NT  MT  NT  MT  MT  NT  ST  ST  MT  MT  NT  MT  ST  ST 

2.1.2. In‐situ 13‐uursmetingen en verticale profielen  Gedurende de periode 2001‐2008 werden 16 13‐uursmetingen uitgevoerd te MOW1 en 8  op  de  Kwintebank  (tabel  2.2).  Het  Sea‐Bird  SBE09  SCTD  carrousel  meetsysteem  (figuur  2.3), dat tot 12 5 l Niskin flessen, een OBS, CTD en LISST 100 bevatte werd op ongeveer 3  m boven de bodem gehouden. Om de 20 minuten werd een Niskin fles gesloten en elk uur  werd de carrousel aan boord gebracht. Van elke Niskin fles werden drie substalen geno‐ men en aan boord gefiltreerd op vooraf gewogen GF/C filters. De filters werden later ge‐ droogd en opnieuw gewogen om de SPM concentratie te bepalen. Deze data werden line‐ air gecorreleerd met de OBS metingen om de OBS te kalibreren. De fouten op de filtratie‐ resultaten zijn relatief hoog bij lage concentraties aan suspensiemateriaal tengevolge van  de  relatief  hogere  systematische  fout  (Fettweis,  2008).  De  relatieve  standaard  deviatie  bedraagt respectievelijk 12% en 60%, rekening houdend met een systematische fout van  4.5  mg/l,  voor  alle  13‐uursmetingen  te  MOW1  en  op  de  Kwintebank.  De  relatieve  standaard deviatie op de SPM concentratie afgeleid van de OBS bedraagt minder dan 20%  (gemiddeld 10%) te MOW1 en tot 56% (gemiddeld 23%) op de Kwintebank. 

11

  Figuur 2.3: Sea‐Bird SBE09 SCTD carrousel meetsysteem  Bij het aan boord brengen van het carrousel meetsysteem wordt een verticaal profiel  opgemeten. Tijdens een getijcyclus worden aldus ongeveer 13 profielen opgemeten. In to‐ taal werden 198 verticale profielen verzameld te MOW1 en 103 op de Kwintebank. De 13‐ uursmetingen zijn  goed  verdeeld  over  dood‐,  gemiddeld  en  springtij.  Respectivelijk  25%,  19%, 31% en 25% van de data te MOW1 werden verzameld tijdens lente, zomer, herfst en  winter en 31%, 38% en 31% tijdens dood‐, gemiddeld en springtij. Voor de Kwintebank da‐ ta is de verdeling van de metingen over de seizoenen en de maancyclus als volgt: 37% len‐ te, 13% zomer, 13% herfst, 37% winter en 25% dood‐, 37% gemiddeld en 37% springtij.  De  gemeten  verticale  profielen  gaan  van  ongeveer  3  mab  naar  de  oppervlakte.  Het  ontbrekende onderste deel van het profiel werd met een lineaire regressie tussen de wa‐ terdiepte en het logaritme van de SPM concentratie berekend. Hiervoor werden de data  volgens stijgende diepte gesorteerd en uitgemiddeld over een dieptecel van 0.5 m: 

⎡1 ⎤ (2.1)  ⎢⎣ 2 (C i −1 + C i ) ( z i − z i −1 )⎥⎦   i=2 met C de gemeten SPM concentratie,  C  de per dieptecel gemiddelde SPM concentratie,  C=

1 Δz

n

∑

zi de diepte van het meetpunt i, Δz = (zn – z1) de verticale afstand tussen het eerste en het  laatste meetpunt per dieptecel, n het aantal meetdata per dieptecel, σi de standaard devi‐ atie op de meting i. De meetonzekerheid op de OBS kan dan worden geschreven als: 

σC =

∂C 1 1 σi = ∂C i 2 Δz

n −1

∑ (z i =2

i +1

− z i −1 ) σ i

  (2.2)  met  σ  de standaard fout op de gemiddelde SPM concentratie per dieptecel. Deze fout  C

wordt  opgeteld  bij  de  statistische  fout  tengevolge  van  (natuurlijke)  variabiliteit  van  de  meetdata. De systematische fout kan groter zijn dan de statistische fout op plaatsen met  weinig suspensiemateriaal (v.b. Kwintebank). De gefitte profielen werden gebruikt om de 

12

verhouding tussen de SPM concentratie aan de oppervlakte en op verschillende diepten te  berekenen om zodoende de SPM concentratie dicht tegen de bodem (tripode) te extrapo‐ leren naar de oppervlakte (Van den Eynde et al., 2007).  

2.1.3. Remote sensing data  De MODIS (MODerate resolution Imaging Spectroradiometer) level 1A data (L1A) voor de  periode 2003‐2008 werden verzameld (http://oceancolor.gsfc.nasa.gov, en figuur 2.4). De  L1A data bevatten de straling aan de top van de atmosfeer, die geometrisch gecorrigeerd  wordt  met  de  SeaDAS  software  (beschikbaar  op  dezelfde  NASA  website).  De  atmosferi‐ sche correctie voor troebel water (Ruddick et al., 2000) werd geïmplementeerd in SeaDAS  en gebruikt om de reflectie op het wateroppervlak af te leiden (water‐leaving reflectance).    

    Figuur 2.4: Twee voorbeelden van MODIS satellietbeelden en de afgeleide SPM concentra‐ tie (mg/l) voor 29/05/2009 (boven) en 31/05/2009 (onder).  SPM concentratie werd berekend uit deze ‘water‐leaving reflectance’ met behulp van  een algoritme gekalibreerd voor troebel water (Nechad et al., 2010). De nauwkeurigheid  van de SPM concentratie werd bepaald voor de fouten afkomstig van het optische model  dat gebruikt werd om de reflectie om te rekenen naar SPM concentraties. De optische ei‐ genschappen  van  de  partikels  in  suspensie  werden  geparameteriseerd.  Fouten  ontstaan  indien  de  grootte  en  samenstelling  van  de  partikels  significant  veranderen  en  daarmee  gepaard  ook  de  optische  eigenschappen.  Dit  treed  op  tijdens  een  getij  en  door  windin‐ vloed (Nechad et al., 2010). De onzekerheid op SPM concentratie tengevolge van fouten in  de ‘water‐leaving reflectance’ werd bepaald op basis van 29 match‐ups in weinig tot mid‐ delmatig  troebel  water  (3‐80  mg  l‐1).  De totale  fout  op  SPM  concentratie  bedroeg  onge‐ veer 37% voor de MODIS beelden. Voor SPM concentraties >10 mg/l is de relatieve fout  op  de  SPM  concentratie  afgeleid  uit  MODIS  beelden  significant  lager  dan  bij  lagere  con‐ centraties. Dit wordt verklaard door de relatief hogere fouten in de ‘water‐leaving reflec‐ tance’ bij lage SPM concentraties en ook doordat het algoritme voor SPM concentratie af‐

13

geijkt werd voor troebel water. Verder moet ook vermeld worden dat het zichtbare spec‐ trum van satellieten gewoonlijk verzadigd geraakt bij zeer hoge turbiditeit (Doxaran et al.,  2002). De band gecentreerd op de golflengte van 667 nm, die hier gebruikt werd om SPM  concentraties  te  bepalen,  bereikt  een  detectielimiet  bij  ongeveer  200  mg/l.  Banden  met  langere golflengte kunnen deze beperking omzeilen als de sensor/ruis verhouding (Wang  et al., 2009) of de fouten bij de atmosferische correctie aanvaardbaar zijn.  Op het BCP beschikken we over ongeveer 60 gedeeltelijk wolkenvrije beelden per jaar.  en dus 460 data te MOW1 en 502 op de Kwintebank voor de beschouwde periode. 64%  van de beelden zijn genomen in de lente en zomer en 36% in de herfst en winter. Tidens  deze seizoenen zijn SPM concentraties het hoogst. De MODIS satelliet vliegt over het BCP  tussen 12:40‐13:40 UTC. In situ metingen die binnen het uur van een satellietbeeld liggen,  werden gemiddeld en vergeleken met deze uit het satellietbeeld. De relatieve standaard  deviatie van de in situ data gedurende dit uur bedraagt gemiddeld 27%.  

2.2. Resultaten  2.2.1. Significante golfhoogte  De  gemiddelde  significante  golfhoogte  (Hs)  tijdens  2003‐2008  bedroeg  0.55  m.  Lagere  waarden waren er in de lente (0.48 m) en zomer (0.53 m) en hogere in de herfst (0.62 m)  en winter (0.60 m), zie figuur 2.5. De gemiddelde Hs tijdens tripode metingen te MOW1  was 0.58 m en op de Kwintebank 0.47 m; deze waarden zijn iets hoger dan het gemiddel‐ de  over  heel  de  periode  2003‐2008.  Significante  golfhoogten  groter  dan  1.5  m  werden  minder frequent opgemeten tijdens de tripode metingen (7% te MOW1 t.o.v. 10% gedu‐ rende de periode 2003‐2008).   Bij 13‐uursmetingen was de gemiddelde Hs = 0.49 m; deze metingen zijn beperkt tot  goede  weersomstandigheden  met  Hs  gewoonlijk  kleiner  dan  1.5  m.  Het  geometrisch  ge‐ middelde  van  Hs  tijdens  de  MODIS  overvlucht  over  het  BCP  vliegt  is  0.60  m.  Gedurende  wolkenvrije  condities  wanneer  SPM  concentratie  data  beschikbaar  zijn,  verkleint  het  ge‐ middelde  tot  0.44  m.  Dit  toont  aan  dat  de  SPM  concentraties  uit  satellietdata  een  naar  goede weersomstandigheden (lage golven) vertekent beeld geven en dat een significante  golfhoogte kleiner dan 0.44 m als proxy gebruikt kan worden voor wolkenvrije condities.  

2.2.2. SPM concentratie profielen  Enkele voorbeelden van de gemeten en gefitte profielen tijdens 13‐uursmetingen worden  getoond in figuur 2.6 (zie appendix 7 voor alle profielen). Er kunnen twee types worden  onderscheiden. Beiden kunnen door een logaritmische functie worden benaderd. Het eer‐ ste  type  bestaat  uit  goed  gemengde  profielen  met  een  geringe  verticale  stratificatie  (de  verhouding  tussen  verticaal  gemiddelde  en  oppervlakte  concentratie  is  kleiner  dan  2),  terwijl het tweede type gekenmerkt wordt door grotere gradiënten (de verhouding is gro‐ ter dan 2). Dit laatste type komt vooral voor rond hoog‐ en laagwater, wanneer de toene‐ mende  stroomsnelheid  een  kritische  waarde  overschreden  heeft  en  het  sediment  gere‐ suspendeerd wordt. Het maximum in stroomsnelheid treedt ongeveer 1 uur voor HW op.  Het eerste type profiel wordt waargenomen bij een lage erosieflux, omdat er ofwel geen  erodeerbaar  materiaal  meer  aanwezig  is  of  doordat  de  bodemschuifspanning  onder  een  drempelwaarde voor erosie valt. Zij vertegenwoordigen aldus perioden van verticale men‐ ging  of  een  relaxatie  periode  tijdens  kentering.  Op  de  Kwintebank  behoren  88%  en  te  MOW1 74% van de profielen tot dit type. De correlatiecoëfficiënt tussen de gefitte en de  gemeten data is hoog (MOW1: R² = 0.77; Kwintebank: R² = 0.98).  

14

Probability density

Probability density

spring x*=48 cm s*=1.79 0.2

0.1

summer x*=53 cm s*=1.82 0.2

0.1

0.0

0.0 0

50

100 150 significant wave height (cm)

200

0

250

0.3

50

100 150 significant wave height (cm)

200

250

0.3

autumn x*=62 cm s*=1.81

Probability density

Probability density

 

0.3

0.3

0.2

0.1

0.0

winter x*=60 cm s*=1.95 0.2

0.1

0.0

0

50

100 150 significant wave height (cm)

200

250

0

50

100 150 significant wave height (cm)

200

250

 

Figuur  2.5:  Waarschijnlijkheidsverdeling  van  de  significante  golfhoogte  te  Bol  van  Heist  voor 2003‐2008 en de overeenstemmende lognormale functie. 

  Figuur 2.6: Enkele voorbeelden van gemeten en berekende SPM concentratie profielen uit  de getijcyclus metingen. De diepte is relatief t.o.v. de totale waterdiepte. KB=Kwintebank,  de andere profielen zijn van MOW1 (zie ook appendix 7). 

15

De gefitte profielen werden gebruikt om een correlatie tussen de log‐getransformeer‐ de  SPM concentraties  aan de  oppervlakte,  op  2  mab  en op  0.2  mab,  en  de  verticaal ge‐ middelde waarde te berekenen (figuur 2.7 en tabel 2.3). De correlatie is hoog voor de type  1  profielen  op  beide  locaties  en  voor  de  type  2  profielen  op  de  Kwintebank  (R²  >  0.8),  maar is minder goed voor de type 2 profielen te MOW1 (R² = 0.4 tot 0.6). De laagste corre‐ laties  treden  op  bij  de  data  op  0.2  mab  en  tonen  aan  dat  een  extrapolatie  van  de  data  dicht tegen de bodem naar de oppervlakte toe, hoogstwaarschijnlijk niet nauwkeurig is bij  type 2 profielen.  Een derde type profiel werd waargenomen in de tripode data (<2 mab). Het wordt ge‐ karakteriseerd door zeer hoge SPM concentraties dicht tegen de bodem, die mogelijks het  gevolg zijn van de aanwezigheid van tijdelijke vloeibare sliblagen of hooggeconcentreerde  slibsuspensies  en  die  als  gevolg  hebben  dat  er  een  slechte  correlatie  optreed  tussen  de  waarden dicht tegen de bodem (0.2 mab) en de 2 mab SPM concentraties. Zij zijn typisch  geassocieerd met stormen (Fettweis et al., 2010). Deze HCBS lagen kunnen niet benaderd  worden door de boven gebruikte profielen, omdat de dynamica dicht tegen de bodem tij‐ dens zulke condities losgekoppeld is van de dynamica hogerop in de waterkolom.     100

1000 MOW1

surface SPM concentration (mg/l)

surface SPM concentration (mg/l)

profile 1 profile 2

100

Kwintebank

profile 1 profile 2

10

1

10 10

100 vertical averaged SPM concentration (mg/l)

1000

1

100

10 vertical averaged SPM concentration (mg/l)

 

Figuur 2.7: Relatie tussen oppervlakte en de verticaal gemiddelde SPM concentratie voor  type 1 en 2 profielen afgeleid uit de gefitte profielen (13‐uursmetingen). De correlatie werd  berekend na log‐transformatie (profiel 1: R²=0.94, profiel 2: R²=0.61), zie Tabel 2.3.  Tabel 2.3: Lineaire correlatie tussen de log‐getransformeerde oppervlakte SPM concentra‐ tie (x) en de verticaal gemiddelde SPM concentratie, deze op 2 mab en deze op 0.2 mab (y).  De relatie log(y)=a+b log(x) werd opgesteld met de gefitte profielen. (mab=m above bed)  vert gem  Type 1 profiel

Type 2 profiel

Alle profielen

a=-0.01 b=0.95 R²=0.94 a=0.11 b=0.69 R²=0.63 a=0.30 b=0.72 R²=0.63

MOW1  2 mab  a=0.05 b=0.89 R²=0.88 a=0.27 b=0.55 R²=0.47 a=0.54 b=0.56 R²=0.46

0.2 mab  a=0.10 b=0.84 R²=0.83 a=0.45 b=0.45 R²=0.38 a=0.73 b=0.44 R²=0.35

Kwintebank  vert gem  2 mab  0.2 mab  a=-0.04 b=0.99 R²=0.98 a=-0.14 b=0.95 R²=0.94 a=-0.01 b=0.95 R²=0.96

a=-0.05 b=0.96 R²=0.96 a=0.18 b=0.89 R²=0.89 a=-0.01 b=0.88 R²=0.91

a=-0.04 b=0.95 R²=0.95 a=-0.18 b=0.87 R²=0.86 a=0.02 b=0.86 R²=0.89

   

16

De  relatie  tussen  de  getij‐gemiddelde  data  op  2  mab  en  0.2  mab  werd  berekend  na  log‐transformatie van de data, zie figuur 2.8. Voor SPM concentraties tussen 100 tot 1000  mg/l is de waarde dicht tegen de bodem (0.2 mab) ongeveer 1.5‐1.7 keer groter dan op 2  mab (R²=0.69). Tijdens twee metingen (MOW1‐4, MOW1‐8) was de 0.2 mab SPM concen‐ traties  1.4‐5.6  keer  groter  dan  deze  op  2  mab;  de  correlatiecoëfficiënt  is  hierbij  laag  (R²=0.33).  100

1000

Kwintebank

1.4 mab SPM concentration (mg/l)

2 mab SPM concentration (mg/l)

MOW1-1, 2, 3, 5, 6, 7 MOW1-4, 8

100

10

10 10

100 1000 0.2 mab SPM concentration (mg/l)

10000

10

0.2 mab SPM concentration (mg/l)

100

 

Figuur 2.8: Relatie tussen de getij‐gemiddelde SPM concentratie op 0.2 m boven de bodem  (mab) (x) en op 2 mab (y) voor de tripode metingen. De correlatie werd berekend na log‐ transformatie van de data (R²=0.69 voor MOW1‐1, 2, 3, 5, 6, 7 en R²=0.33 voor MOW1‐4,  8). (mab=m above bed) 

2.2.3. In situ en satelliet data op hetzelfde ogenblik  Te  MOW1  zijn  er  19  keer  op  hetzelfde  ogenblik  tripode  en  de  satelliet  data  beschikbaar  (match‐ups). De correlatiecoëfficiënt tussen– na log‐transformatie – de oppervlakte SPM  concentratie afkomstig van MODIS en de SPM concentratie van de tripode is R²=0.8 voor  de 2 mab en R²=0.7 voor de 0.2 mab data. Merk op dat het verschil tussen oppervlakte en  bodem SPM concentratie heel groot is (15‐30 keer lager), zie figuur 2.9a. De relatie tussen  oppervlakte en bodem SPM concentratie, zoals berekend voor de gefitte profielen van de  13‐uursmetingen,  werd  gebruikt  om  een  diepte  correctie  toe  te  passen.  Bij  het  gebruik  van de correctiefactoren voor alle profielen (tabel 2.3) wordt een relatief goede overeen‐ komst  tussen  MODIS  en  2  mab  tripode  data  (R²=0.5)  bekomen  (figuur  2.9),  niettegen‐ staande dat de gecorrigeerde MODIS data de SPM concentratie onderschatten (zie § 2.3).  

2.2.4. Frequentieverdeling van SPM concentratie  Door frequentieverdelingen van de verschillende data sets op te stellen, kunnen we met  behulp  van  standaard  statistische  tests  (Χ²  test,  Kolmogorov‐Smirnov  test)  bepalen  of  twee  verdelingen  afkomstig  zijn  van  eenzelfde  populatie.  Indien  de  data  verzameld  met  verschillende methoden een gelijkaardige log‐normale verdeling, geometrisch gemiddelde  en standaard afwijking kennen, dan kunnen we besluiten dat – binnen een gegeven onze‐ kerheidmarge – de methoden gelijkaardige substalen van de gehele populatie opleveren.  De drie verschillende SPM concentratie data (getijcyclus, tripode en satelliet) werden ge‐ bruikt om frequentieverdelingen op te stellen. De verdelingen komen goed overeen met  log‐normale verdelingen, zie figuren 2.10‐2.12. Een variabele x is log‐normaal verdeeld in‐ dien de log(x) normaal verdeeld is. De PDF (Probability Density Function) van zulk een va‐ riabele kan als volgt worden geschreven:  

17

f (x ) =

1 ⎛ (log(x ) − μ )2 ⎞⎟   exp⎜ − 2 xσ 2π ⎝ 2σ ⎠ 1

(2.3) 

waarbij μ de mediaan en σ de standaardafwijking is. Een log‐normale verdeling wordt ge‐ karakteriseerd door de mediaan en de standaard deviatie van log(x). Door de waarden te‐ rug  om  te  zetten  naar  de  gemeten  grootheid  kan  de  verdeling  gekarakteriseerd  worden  door het geometrisch gemiddelde (x*) en de multiplicatieve standaardafwijking (s*). Het  interval x* gedeeld door s* tot x* vermenigvuldigd met s* omvat 68.3% van de populatie  en x*/2s* tot x* × 2s* 95.5%. Volgende benadering werd gebruikt:   1

⎞n ⎛1 n ⎞ ⎛ n x* = exp⎜ ∑ log( xi )⎟ = ⎜⎜ ∏ xi ⎟⎟   ⎝ n i =1 ⎠ ⎝ i =1 ⎠ 1 ⎛ ⎞ ⎜ ⎡ 1 n ⎡ ⎛ x ⎞⎤ 2 ⎤ 2 ⎟ i s* = exp⎜ ⎢ ∑ ⎢log⎜ x * ⎟⎥ ⎥ ⎟   n 1 − ⎜⎜ ⎢⎣ ⎝ ⎠⎦ ⎥⎦ ⎟⎟ i =1 ⎣ ⎝ ⎠  

(2.4) 

(2.5) 

1000 surface SPM conc. MODIS (mg/l)

(a)

100

10 0.2 mab 2 mab 2 mab (R²=0.80) 0.2 mab (R²=0.66)

1 10

100 1000 SPM concentration tripod (mg/l)

10000

1000 corrected SPM conc.MODIS (mg/l)

(b)

100

0.2 mab 2 mab 2 mab (R²=0.52) 0.2 mab (R²=0.23)

10 10

100 1000 SPM concentration tripod (mg/l)

10000

 

Figuur 2.9: Correlatie tussen SPM concentratie op 2 en 0.2 mab en de oppervlakte (a) en  dieptegecorrigeerde (b) MODIS SPM concentratie tijdens een match‐up. De correlatie wer‐ den berekend na log‐transformatie en rekening houdend met meetfouten, zie § 2.1.3.  

18

  Figuur 2.10: Frequentieverdeling van tripode SPM concentratie data op 2 mab en 0.2 mab.  De data te MOW1 zijn gegroepeerd in klassen van 50 mg/l en deze van de Kwintebank in  klassen  van  5  mg/l.  De  overeenkomstige  log‐normale  verdeling  heeft  een  Χ²  test  waar‐ schijnlijkheid van p>0.9. Ook aangeduid zijn het geometrisch gemiddelde x* vermenigvul‐ digd/gedeeld door de multiplicatieve standaard deviatie s*. (mab=m above bed)  Het geometrisch gemiddelde, de multiplicatieve standaard deviatie en de X² test resul‐ taten van de SPM concentratieverdelingen kunnen gevonden worden in Tabel 2.4. De data  van  de  13‐uursmetingen  worden  getoond  voor  drie  diepten  die  overeenkomen  met  de  meetdiepte  van  satellieten  (oppervlakte)  en  van  de  tripode  (2  en  0.2  mab).  De  Χ²  test  waarschijnlijk  werd  berekend  in  de  veronderstelling  dat  SPM  concentraties  gefit  kunnen  worden met een log‐normale functie. Over het algemeen varieert s* tussen 1.5 en 2.8 en  valt dus in de range van 1.4 tot 3 die ook in de andere domeinen binnen de natuurweten‐ schappen werd gevonden (Limpert et al., 2001). Indien de test waarschijnlijkheid klein is (p  < 0.05), dan zou de nul hypothese verworpen moeten worden. Lage waarschijnlijkheden  treden op bij de 13‐uursmetingen, in het bijzonder in de oppervlakte data op de Kwinte‐ bank en in de 0.2 mab te MOW1. De afwijking van de log‐normale verdeling van de opper‐ vlakte data van de Kwintebank (figuur 2.11) zijn mogelijks te wijten aan de hoge meetfou‐ ten van SPM concentratie (zie §2.1.2). Voor de MOW1 data dicht tegen de bodem kan de  afwijking het gevolg zijn van de soms onrealistisch hoge waarden veroorzaakt door de log‐ extrapolatie van de verticale profielen naar 0.2 mab (zie appendix 7). We kunnen daarom  argumenteren dat een type I fout optreedt en dat deze SPM concentraties dus wel dege‐ lijk een log‐normale verdeling hebben.    

19

  Figuur 2.11: idem Figuur 2.10 maar nu voor de 13‐uursmetingen aan de oppervlakte, op 2  mab en op 0.2 mab afgeleid uit de gefitte profielen. De data te MOW1 zijn opgedeeld in  klassen van 20 mg/l en deze van de Kwintebank in klassen van 5 mg/l. (mab=m above bed) 

  Figuur 2.12: idem Figuur 2.10 maar nu voor de MODIS oppervlakte SPM concentratie data.  De data zijn opgedeeld in klassen van 5 mg/l . (mab=m above bed) 

20

Tabel  2.4:  Geometrisch  gemiddelde  van  de  SPM  concentratie  (x*)  afkomstig  van  13‐ uursmetingen (13h), tripode metingen en MODIS satelliet te MOW1 en Kwintebank. p is de  Χ² test waarschijnlijkheid t.o.v. een log‐normale verdeling, s* de multiplicatieve standaard  deviatie; de x* aan de oppervlakte in de tripode data werd met behulp van de relatie voor  ‘alle profielen’ uit Tabel 2.3 berekend. (mab=m above bed)  Locatie 

MOW1 

Kwintebank 

Methode  MODIS  13h  13h  13h  tripode 

Diepte  oppervlakte  oppervlakte  2 mab  0.2 mab  oppervlakte 

tripode  tripode  MODIS  13h  13h  13h  tripode 

2 mab  0.2 mab  oppervlakte  oppervlakte  2 mab  0.2 mab  oppervlakte 

tripode  tripode 

2 mab  0.2 mab 

x* (mg/l)  23  39  81  93  59 (2 mab)  68 (0.2 mab)  174  326  6  9  12  13  19 (2 mab)  23 (0.2 mab)  29  36 

s*  2.4  2.1  2.6  2.8  1.7 (2 mab)  1.6 (0.2 mab)  2.5  2.8  2.5  2.1  2.2  2.3  1.4 (beiden) 

p  0.51  0.81  0.09  0.02  0.97 (2 mab)  0.11 (0.2 mab)  0.99  0.99  0.28  0.02  0.56  0.79  0.00 (beiden) 

1.5  1.5 

0.93  0.93 

Tabel  2.5:  Geometrisch  gemiddelde  SPM  concentratie  (x*)  te  MOW1  voor  verschillende  golfcondities. De verdelingen komen overeen met bepaalde orbitale golfsnelheden aan de  bodem (Uw); p is de Χ² test waarschijnlijkheid t.o.v. een log‐normale verdeling; s* de multi‐ plicatieve standaard afwijking; oppervlakte x* de met behulp van de relatie voor ‘alle pro‐ fielen’ uit Tabel 2.3 naar de oppervlakte gecorrigeerde x*. (mab=m above bed)  ws (m) 

>0.5 

>0.3 

>0.03 

alle 

<0.3 

<0.03 

  2 mab  x* (mg/l)  338  278  176  178  155  145  s*  1.6  1.8  2.5  2.0  2.5  2.4  p  0.73  0.95  0.99  0.91  0.99  0.99  oppervlakte x* (mg/l)  91  82  63  59  58  56    0.2 mab  x* (mg/l)  617  513  390  313  308  263  s*  1.9  2.0  2.8  2.8  2.8  2.8  p  0.99  1.00  1.00  0.93  0.91  0.96  oppervlakte x* (mg/l)  92  87  75  68  67  63    MOW1 is gelegen in een ondiep gebied waar golfeffecten belangrijk zijn. Door data te  selecteren waar de orbitale golfsnelheid aan de bodem, Uw, hoger (respectievelijk lager) is  dan een zekere waarde kan het gemiddelde en de multiplicatieve standaard deviatie van  tijdens  een  storm  (respectievelijk  goed  weer)  worden  berekend  (tabel  2.5).  De  orbitale  golfsnelheid aan  de bodem werd  berekend  uit  de  significante golfhoogte  (Hs), de  water‐ diepte  en het JONSWAP  golfspectrum (Soulsby,  1997). Een  Uw  van  0.03  m/s,  0.3  m/s en 

21

0.5 m/s komt overeen met een significante golfhoogte van ongeveer 0.5 m, 1.5 m en 2.5  m in een waterdiepte van 10 m. De resultaten te MOW1 tonen dat de frequentieverdelin‐ gen  gelijkaardig  zijn  voor  data  met  Uw  >  0.03  m/s  en  Uw  <  0.3  m/s.  De  gemiddelde  SPM  concentratie te MOW1 neemt toe van 145 mg/l (Uw  < 0.03 m/s) naar 338 mg/l (Uw > 0.5  m/s) op 2 mab en van 263 mg/l (Uw < 0.03 m/s) naar 617 mg/l (Uw > 0.5 m/s) op 0.2 mab;  dit bevestigt het niet lineaire gedrag van het systeem. De SPM concentratie verdeling uit  de 13‐uursmetingen op 2 mab komen goed overeen – maar toch nog met significante ver‐ schillen – met de tripode data tijdens goede weersomstandigheden (Uw < 0.03 m/s;  x*=81  mg/l  versus  145  mg/l).  Gelijkaardige  resultaten  werden  gevonden  voor  de  0.2  mab  data  van de 13‐uurmetingen en van de tripode data tijdens goede weersomstandigheden (93  mg/l versus 263 mg/l).  

2.2.5. Oppervlaktecorrectie van tripode data  Door de tripode data van de Kwintebank en MOW1 te extrapoleren naar de oppervlakte  kunnen  de  drie  datasets  (tripode,  13‐uurmeting  en  MODSI)  met  elkaar  vergeleken  wor‐ den. De oppervlakte correctie werd uitgevoerd voor zowel de 2 mab als de 0.2 mab tripo‐ de  data  door  de  relatie  voor  ‘alle  profielen’  uit  tabel  2.3  toe  te  passen.  De  oppervlakte  waarden voor 0.2 mab zijn iets hoger dan voor 2 mab (tabel 2.4). De verschillen tussen sa‐ telliet,  13‐uursmeting  en  tripode  oppervlakte  data  zijn  echter  nog  steeds  significant.  De  laagste gemiddelde SPM concentratie wordt gevonden bij de satelliet data (Kwintebank: 6  mg l‐1, MOW1: 23 mg l‐1) en de hoogste bij de tripode data (Kwintebank 19 mg l‐1, MOW1:  59 mg l‐1 beiden voor de 2 mab extrapolatie).   De resultaten tonen dat voor de Kwintebank gemiddelde SPM concentratie uit satelliet  data  binnen  een  standaard  afwijking  gelegen  is  van  de  data  afkomstig  van  de  13‐ uursmetingen en vice versa. Dit geldt niet voor de tripode data, die niet binnen een stan‐ daard  afwijking  van  de  13‐uurmetingen  en  satelliet  data  vallen.  Dit  is  mogelijks  veroor‐ zaakt door het feit dat de tripode data beperkt zijn tot maart 2004 en dus niet representa‐ tief zijn voor een heel jaar (tabel 2.1). Indien enkel de winter satelliet data worden gese‐ lecteerd dan wordt een gemiddelde SPM concentratie bekomen van 12 mg l‐1, die echter  nog steeds laag is vergeleken met deze afgeleid uit de tripode data. Te MOW1 is de ge‐ middelde SPM concentratie uit 13‐uursmetingen gelegen binnen een standaard afwijking  van deze van de tripode en de satelliet data. De gemiddelde SPM concentratie afkomstig  van de tripode en de satelliet zijn niet binnen een standaard afwijking gelegen.  

2.3. Discussie  De  resultaten  van  standaard  statistische  testen  tonen  aan  dat  de  data  van  13‐ uursmetingen, tripode en MODIS verschillende verdelingen hebben en dat zij dus andere  subpopulaties  van  de  hele  SPM  concentratie  populatie  vertegenwoordigen.  Onderaan  worden de verschillen tussen de staalnamemethode, de representativiteit van de metin‐ gen en de meetfouten meer in detail besproken. Deze analyse zal enkel voor de MOW1  data  worden  uitgevoerd,  omdat  deze  representatief  zijn  voor  seizoenen  en  extreme  ge‐ beurtenissen.  

2.3.1. Representativiteit van SPM concentratie data  Een  eerste  selectie  van  SPM  concentraties  werd  gedaan  door  deze  volgens  de  orbitale  golfsnelheid op te delen (zie §2.2.4). De naar de oppervlakte toe gecorrigeerde tripode da‐ ta (tabel 2.5) berekend met de 2 mab data variëren tussen 91 mg/l (Uw > 0.5 m/s), 59 mg/l  (alle data) en 56 mg/l (Uw < 0.03 m/s) en zijn nog steeds significant hoger dan deze van de 

22

13‐uursmetingen  en  satelliet  data  (tabel  2.4).  Dit  resultaat  onderstreept  dat  de SPM  dy‐ namica dicht tegen de bodem en de vorming van vloeibaar slib of hooggeconcentreerde  suspensielagen los gekoppeld is van de processen die zich hogerop in de waterkolom af‐ spelen en wijst mogelijks op het feit dat 18 13‐uursmetingen en 460 satellietbeelden niet  representatief zijn voor de SPM concentratie populatie bij goede weersomstandigheden te  MOW1.  

2.3.2. Staalnamemethode  Staalname wordt aanzien als een statistische handeling om individuele SPM concentraties  te selecteren, die toelaten conclusies te trekken over de hele SPM concentratie populatie  op  een  locatie  of  in  een  groter  gebied.  Een  gemeenschappelijk  doel  van  deze  metingen  zou  moeten  zijn  om  een  representatieve  subpopulatie  te  verzamelen,  waaruit  bevindin‐ gen, binnen de foutenmarges, over de hele populatie kunnen worden afgeleid. Indien we  met tijdsafhankelijke en meer specifiek harmonisch variërende processen te maken heb‐ ben, zoals SPM concentratie, dan zou men op de hoogte moeten zijn van het aantal data  dat nodig is vooraleer een subpopulatie als representatief kan beschouwd worden. Het is  daarom  essentieel  om  te  weten  hoe  representatief  een  bepaalde  staalnamemethode  is.  Verschillen tussen datasets kunnen bv. het gevolg zijn van het feit dat in situ en and remo‐ te sensing technieken gebruik maken van andere staalnamemethoden.   Satellieten kunnen beschouwd worden als een toevallige staalname (random sampler)  die een vertekend beeld geven, omdat enkel tijdens wolkenvrije momenten data kunnen  verzameld worden, en omdat satellieten verzadigd raken bij hoge SPM concentraties. 13‐ uursmetingen vanuit een schip en langdurige metingen met tripoden kunnen worden om‐ schreven als een ‘event based’ staalnamemethode, die gekenmerkt wordt door een wille‐ keurig startmoment en het verder verzamelen van data gedurende minstens een getijcy‐ clus.  Door  de  staalname  te  groeperen  gedurende  een  getijcyclus  wordt  de  belangrijkste  SPM concentratie variatie bemonsterd.   Een  meer  objectieve  benadering  is  om  verschillende  meetschema’s  toe  te  passen  op  de volledige dataset en om de analyse uit te voeren op de genomen subpopulatie (Schlep‐ pi et al., 2006). Indien we aannemen dat de tripode data te MOW1 een goede representa‐ tie zijnn van de natuurlijke variabiliteit gedurende een jaar, dan kunnen we deze willekeu‐ rig  bemonsteren  met  analoge  meetschema’s  als  van  satellieten  of  13‐uursmetingen.  Zo‐ doende kunnen we nagaan hoe een laagfrequente staalname het gemiddelde en de stan‐ daard deviatie beïnvloeden. Satellietdata komen meestal voor tijdens periodes met weinig  wind (Fettweis et al., 2007), daarom werd een significante golfhoogte van 0.44 m gebruikt  als  proxy  voor  wolkenvrije  condities.  Drie  meetschema’s  werden  onderzocht;  schema  1  bestaat uit 60 willekeurig genomen stalen; schema 2 uit 60 willekeurige stalen bij een Hs <  0.44 m en schema 3 uit 60 willekeurige stalen genomen tussen 13h en 14h en bij een Hs <  0.44 m. Een significante golfhoogte van 0.44 m komt overeen met de mediaan van de sig‐ nificante golfhoogte tijdens passage van de satelliet. Het schema voor de 13‐uursmetingen  bestaat uit 6 willekeurige staalname momenten bij Hs <  1.5 m met telkens meetreeksen  van 13 opeenvolgende uren (totaal 78 data). Elk meetschema werd 10 keer herhaald om  de variabiliteit tengevolge van de ‘random number generator’ te bepalen (tabel 2.6).  Het  satelliet  meetschema  1  geeft  zeer  gelijkaardige  gemiddelde  SPM  concentratie  (2  mab 182 mg l‐1 t.o.v. 178 mg l‐1: 0.2 mab: 347 mg l‐1 t.o.v. 313 mg l‐1) en multiplicatieve  standaard afwijking dan voor alle tripode data, terwijl satelliet meetschema 2 een vermin‐ dering  van  de  gemiddelde  SPM  concentratie  (2  mab:158  mg  l‐1;  0.2  mab:  258  mg  l‐1)  en 

23

multiplicatieve standaard afwijking als gevolg heeft (tabel 2.6). De resultaten van het sa‐ telliet meetschema 2 zijn gelijkaardig met de gemiddelde SPM concentratie van de tripode  data bij Uw  < 0.03 m s‐1 (2 mab: 158 mg l‐1 t.o.v. 145 mg l‐1; 0.2 mab: 258 mg l‐1 t.o.v. 263  mg l‐1). Uit deze resultaten zou men kunnen concluderen dat 60 stalen representatief zijn  voor de hele populatie en dat toepassen van satelliet meetschema 2 de gemiddelde SPM  concentratie verlaagd. Dit laatste wordt dan vooral beïnvloed door de beperking tot goede  weersomstandigheden,  maar  ook  de  staalname  tijd  blijkt  een  invloed  te  hebben  zoals  wordt  aangetoond  door  het  satelliet  meetschema  3,  waarmee  de  gemiddelde  SPM  con‐ centratie verder daalt. (tabel 2.6). Deze conclusie wordt verder versterkt door het feit dat  herhaling  van  hetzelfde  meetschema  niet  resulteert  in  een  hogere  standaard  afwijking.  Door de over het algemeen betere weersomstandigheden tijdens de lente en zomer be‐ komt men meer lente en zomer (60%) dan herfst en winter (40%) data bij gebruik van sa‐ telliet schema’s 2 en 3. Dit komt goed overeen met de verdeling van de beschikbare satel‐ lietdata over de seizoenen.   Het  13‐uurmeetschema  resulteert  in  een  gelijkaardige  gemiddelde  SPM  concentratie  dan bij de tripode data met Hs < 0.3 m s‐1 (2 mab: 185 mg l‐1 t.o.v. 155 mg l‐1; 0.2 mab: 284  mg l‐1 t.o.v. 308 mg l‐1). De relatieve standaardafwijking tengevolge van de ‘random num‐ ber generator’ bedraagt 33%, dit is hoger dan bij de satelliet meetschema’s, hoewel het  aantal stalen niet veel verschilt (78 t.o.v. 60). De hogere variabiliteit tijdens een getijcyclus  wordt behouden, terwijl dit verdwijnt bij de satelliet meetschema’s (Fettweis et al., 2007).  Tabel 2.6: Geometrisch gemiddelde SPM concentratie (x*) te MOW1 bekomen door gelijk‐ aardige meetschema’s toe te passen op de tripode data dan deze van satellieten (SAT) en  13‐uursmetingen  (13h),  zie  discussie.  p  is  de  Χ²  test  waarschijnlijkheid  t.o.v.  een  log‐ normale verdeling, s* de multiplicatieve standaard deviatie; de x* aan de oppervlakte in de  tripode data werd met behulp van de relatie voor ‘alle profielen’ uit tabel 2.3 berekend.  Meetschema 

Sat 1 

Sat 2 

x*(mg/l) ±stdv   s*  p  oppervlakte x* (mg/l) 

182±12  2.6  0.49  64±2 

158±13  2.2  0.62  59±3 

x*(mg/l) ±stdv   s*  p  oppervlakte x* (mg/l) 

347±42  2.7  0.78  71±3 

258±23  2.4  0.64  62±3 

Sat 3  2 mab  153±19  2.2  0.45  58±4  0.2 mab  236±31  2.5  0.59  59±3 

TC  183±63  2.1  0.23  64±12  284±105  2.3  0.35  64±11 

2.3.3. Meetonzekerheid  Oppervlaktecorrectie van een subpopulatie uit de tripode data bekomen door toepassing  van satelliet meetschema 3 en 13‐uursmeetschema geeft nog steeds verschillende waar‐ den voor de SPM concentratie dan deze afkomstig van de satelliet (58 mg l‐1 t.o.v. 23 mg l‐ 1 ) en de 13‐uursmetingen (64 mg l‐1 t.o.v. 39 mg l‐1). Correctie voor type 2 profielen moet  toegepast worden om een overeenkomst binnen een standaardafwijking te bekomen tus‐ sen satelliet meetschema 3 (x* = 30 mg l‐1, s* = 1.5), 13‐uursmeetschema (x* = 32 mg l‐1,  s* = 1.5) en de satelliet data (x* = 23 mg l‐1, s* = 2.4). Type 2 profiele komen slechts in 25%  van de metingen te MOW1 voor. Dit wijst mogelijks op een onderschatting van type 2 pro‐

24

fielen tijdens 13‐uursmetingen, maar wordt waarschijnlijk ook veroorzaakt doordat remo‐ te  sensing  data  beperkt  zijn  tot  lagere  SPM  concentraties  tengevolge  van  verzadiging.  Daarenboven is de benadering van verticale profielen met een eenvoudige logaritmische  functie, die geen rekening houdt met bezinking en verticale menging – niettegenstaande  de hoge lineaire correlatie coëfficiënt tussen de gefitte en de gemeten waarden (MOW1:  R²=0.77; Kwintebank: R²=0.98) – waarschijnlijk niet nauwkeurig voor type 2 profielen dicht  tegen de zeebodem. De zeer hoge SPM concentraties (>3.6 g l‐1 op 0.2 mab en 2.6 g l‐1 op  2  mab)  te  MOW1  in de  tripode  data,  die  echter nooit gemeten  werd  door  satellieten  of  tijdens  13‐uursmetingen  (maximum  1  g  l‐1  tijdens  2006/06),  bevestigd  dat  de  dynamica  dicht tegen de bodem sterk beïnvloed wordt door golfinwerking. De vorming van hoogge‐ concentreerde sliblagen in golfgedomineerde gebieden en de verschillend met de dynami‐ ca in de rest van de waterkolom wordt beschreven in de literatuur (de Wit & Kranenburg,  1997; Li & Mehta, 2000; Winterwerp, 2006).  

2.4. Conclusies  Er werd aangetoond dat de SPM concentratie afgeleid uit drie verschillende meetmetho‐ den (tripode op vaste plaats, 13‐uursmeting en satelliet) verschillende frequentieverdelin‐ gen  opleveren.  De  analysen  werden  uitgevoerd  voor  MOW1,  gelegen  in  het  turbiditeits‐ maximum, en voor de Kwintebank, gelegen in een gebied met lage turbiditeit. Om de SPM  concentraties  dicht  tegen  de  bodem  te  kunnen  vergelijken  met  de  satellietdata  werden  correcties  toegepast.  Deze  werden  opgesteld  uit  de  SPM  concentratieprofielen  opgeme‐ ten  tijdens  de  13‐uursmetingen.  De  verschillen  tussen  de  datasets  kunnen  worden  ver‐ klaard door de verschillende meteorologische condities gedurende de metingen; doordat  de SPM dynamica dicht tegen de bodem deels losgekoppeld is van de processen die zich  hoger in de waterkolom afspelen; door de staalnamemethode; door de gebruikte correc‐ tie naar de oppervlakte die een logaritmisch profiel verondersteld en door meetfouten. De  belangrijkste conclusies zijn:   • Tengevolge van de hoge tijdelijke en ruimtelijke variabiliteit van SPM concentratie in  het  turbiditeitsmaximum,  zijn  intensievere  meetschema’s  nodig  dan  in  de  meer  off‐ shore gelegen gebieden met lage concentraties.   • SPM concentraties afkomstig van satelliet, 13‐uursmeting en tripode zijn zeer gelijk‐ aardig op de Kwintebank rekening houdend met de grotere meetonzekerheid in ge‐ bieden met lage turbiditeit.   • Satellieten of enkele 13‐uursmetingen per jaar kunnen geen langdurige continue me‐ tingen  vervangen  in  gebieden  met  hoge  turbiditeit.  Deze  metingen geven  enkel  een  deel  van  de  hele  populatie  weer  dat  bovendien  vertekend  is  naar  goede  weersom‐ standigheden en lente‐zomer perioden (satelliet). Sediment transport berekend met  deze data zal altijd een onderschatting geven van de werkelijkheid.   • De gemiddelde SPM concentratie berekend uit 60 willekeurige staalnamen per jaar is  representatief voor de gemiddelde SPM concentratie van de hele populatie. Waneer  een satelliet meetschema wordt gebruikt met een selectiecriterium voor significante  golfhoogte (Hs  < 0.44 m) en staalnametijd, dan stellen 60 stalen de SPM concentratie  tijdens  goede  weersomstandigheden  voor.  De  gemiddelde  SPM  concentratie  bere‐ kend  uit  6  willekeurige  13‐uursmetingen  per  jaar  met  een  Hs  <  1.5  m  als  criterium,  geeft gelijkaardige gemiddelde SPM concentraties als de hele populatie. De hoge SPM  concentratie variabiliteit gedurende een getijcyclus wordt behouden, terwijl deze ver‐ loren gaat bij een satelliet meetschema.  

25

•

De  gemiddelde  SPM  concentratie  afgeleid  uit  satelliet  data  is  gelegen  binnen  één  standaard afwijking van de gemiddelde uit de 13‐uursmetingen op beide onderzochte  locaties. Dit duidt op een goede overeenkomst tussen de verdelingen uit beide data‐ sets. Indien op de SPM concentratie gemeten door de tripode en geëxtrapoleerd naar  de  oppervlakte  een  meetschema  gelijkaardig  als  bij  satellieten  wordt  toegepast,  wordt  een  significant  hoger  gemiddelde  bekomen  dan  dit  afkomstig  uit  de  satelliet  SPM concentraties. De reden hiervoor zijn gelegen in de onzekerheden geassocieerd  met de gebruikte extrapolatie en het feit dat hogere SPM concentraties (>200 mg l‐1)  uitgefilterd worden door verzadiging van de satelliet.  

26

3. Profielen saliniteit en temperatuur  Verticale  salinteits‐  en  temperatuursstratificatie  werd  opgesteld  met  de  data  van  de  13‐ uursmetingen, zie §2.1.2. De figuren worden in appendix 8 gegroepeerd, hierbij werd het  ontbrekende  onderste  deel  van  het  profiel  met  een  lineaire  regressie  tussen  de  water‐ diepte en de saliniteit/temperatuur berekend. In tabel 3.1 wordt het maximaal gemeten  verschil  tussen  de  onderste  en  bovenste  waarneming  in  een  profiel  tijdens  de  13‐ uursmeting weergegeven samen met het maximaal verschil tijdens een getij (horizontale  gradiënt). De verticale verschillen zijn over het algemeen kleiner dan 1 (0.2 °C). Enkele uit‐ schieters in saliniteit werden waargenomen te MOW1 (1.7) en Kwintebank (5.5).   Tabel 3.1: Maximale verticale en horizontale saliniteits‐ en temperatuursgradiënt per getij.  Id‐Nr      2001/06‐A  2002/27‐B  2003/04‐A  2003/15  2003/17  2003/22  2003/25  2004/04  2004/05  2004/24  2004/25‐A  2005/02  2005/07‐A  2005/15‐A  2005/15‐B  2005/29  2006/06  2006/10‐A  2007/11‐A  2007/11‐B  2007/16  2007/25‐B  2008/02‐A  2008/02‐B   

Locatie      MOW1  MOW1  MOW1  Kwintebank  Kwintebank  MOW1  Kwintebank  Kwintebank  Kwintebank  MOW1  MOW1  MOW1  MOW1  Kwintebank  MOW1  MOW1  MOW1  MOW1  Kwintebank  MOW1  MOW1  MOW1  MOW1  Kwintebank 

Verticaal  Saliniteit  Temperatuur   (°C)  0.35  0.07  1.68  0.12  1.71  0.05  0.05  0.11  0.41  0.19  0.10  0.03  0.54  0.16  1.46  0.68  0.32  0.13  0.28  0.06  0.08  0.06  0.11  0.09  0.11  0.06  0.09  0.09  0.42  0.13  0.16  0.19  0.11  0.04  0.03  0.02  5.48  0.12  0.18  0.04  0.46  0.11  0.08  0.15  0.27  0.22  0.03  0.02 

Horizontaal  Saliniteit  Temperatuur (°C)  SBE19  SBE09  SBE19  SBE09  0.78  0.66  0.13  0.06  1.00  0.93  0.19  0.60  1.79  1.36  0.44  0.24  0.19  0.16  0.94  0.94  ‐  0.82  ‐  0.44  0.42  0.40  0.30  0.38  0.99  0.99  0.65  0.71  1.27  2.68  0.77  0.76  1.15  1.23  0.63  0.27  0.45  0.41  0.23  0.13  0.36  0.42  0.35  0.34  0.30  0.31  0.26  0.26  0.38  0.44  0.22  0.22  0.79  0.79  0.48  0.50  1.26  0.78  0.40  0.22  0.60  0.59  0.24  0.23  1.21  1.28  0.45  0.49  0.60  ‐  0.47  0.37  0.85  0.54  0.19  0.32  1.55  0.96  0.72  0.45  1.11  0.79  0.42  0.45  ‐  0.81  ‐  0.82  0.90  0.80  0.76  0.70  0.15  0.36  0.29  0.29 

In tabel 3.2 worden de horizontale gradiënten per getij tijdens de tripodemetingen ge‐ toond. De resultaten tonen dat het maximaal verschil in saliniteit en temperatuur tijdens  een getij kan oplopen tot 4.6 (1.6 °C) in de kustzone ter hoogte van Zeebrugge. De medi‐ aan van deze horizontale gradiënten bedraagt 0.75 (0.35 °C) en 0.73 (0.35 °C) te MOW1 en 

27

Blankenberge  respectievelijk.  De  cummulatieve  frequentieverdeling  hiervan  worden  ge‐ toond in figuur 3.1. Uit de figuur blijkt dat een salinteitsverschil van 2 of meer tijdens een  getij in 10% van de data teruggevonden wordt.  Tabel  3.2 Maximale  en  mediane  saliniteits‐  en  temperatuursgradiënt  per  getij  tijdens tri‐ pode metingen te MOW1 en Blankenberge  Locatie    MOW1  MOW1  MOW1  MOW1  MOW1  MOW1  MOW1  MOW1  Blank  Blank  Blank  Blank  Blank  Blank 

Periode    oct‐nov 2004  feb 2005  apr 2005  jun‐jul 2005  nov‐dec 2005  feb 2006  mar‐apr 2006  may‐jun 2006  nov‐dec 2006  dec 2006 – feb 2007 jan‐feb 2008  mar‐apr 2008  apr‐jun 2008  may‐jun 2009 

Saliniteit  Max  D50  1.89  0.53  4.03  1.10  1.77  0.75  2.18  0.78  1.76  0.90  1.02  0.77  4.63  1.05  4.04  0.59  1.98  1.11  2.86  0.62  2.50  0.60  2.77  1.08  2.12  0.36  3.08  0.51 

Temperatuur  Max  D50  0.92  0.20  0.94  0.32  ‐  ‐  1.02  0.46  086  0.43  0.46  0.18  1.18  0.35  0.96  0.42  1.13  0.46  1.64  0.34  0.69  0.32  0.79  0.30  0.82  0.34  0.65  0.32 

 

  Figuur 3.1: Cumulatieve frequentieverdeling van de maximale saliniteits‐ en temperatuurs‐ verschillen tijdens een getij. Data afkomstig van tripode metingen te MOW1 en Blanken‐ berge.  

28

4. Referenties  Badewien  TH,  Zimmer  E,  Bartholomä  A,  Reuter  R  (2009)  Towards  continuous  long‐term  measurements of suspended particulate matter (SPM) in turbid coastal waters. Ocean  Dynamics, 59:227‐238. doi:10.1007/s10236‐009‐0183‐8  Bartholomä A, Kubicki A, Badewien TH, Flemming BW (2009) Suspended sediment trans‐ port in the German Wadden Sea—seasonal variations and extreme events. Ocean Dy‐ namics, 59:213‐255. doi:10.1007/s10236‐009‐0193‐6  Blewett J, Huntley DA (1998) Measurement of suspended sediment transport processes in  shallow water off the Holderness Coast, UK. Marine Pollution Bulletin, 37:134‐143  Bograd SJ, Rabinovich AB, Thomson RE, Eert AJ (1999). On sampling strategies and interpo‐ lation schemes for satellite‐tracked drifters. Journal of Atmospheric and  Oceanic Tech‐ nology, 16:893‐904  Bowers DG, Boudjelas S, Harker GEL (1998) The distribution of fine suspended sediments  in  the  surface  waters  of  the  Irish  Sea  and  its  relation  to  tidal  stirring.  International  Journal of Remote Sensing, 19:2789‐2805  Bowers DG, Gaffney S, White M, Bowyer P (2002) Turbidity in the southern Irish Sea. Con‐ tinental Shelf Research, 22:2115‐2126  Buijsman MC, Ridderinkhof H (2007). Long‐term ferry‐ADCP observations of tidal currents  in  the  Marsdiep  inlet.  Journal  of  Sea  Research,  57:  237‐256.  doi:10.1016/j.seares.2006.11.004   Cacchione DA, Drake DE, Kayen RW, Sternberg RW, Kineke GC, Tate GB (1995) Measure‐ ments in the bottom boundary subaqueous delta. Marine Geology, 125:235‐257  Caeiro S, Painho M, Goovaerts P, Costa H, Sousa S (2003). Spatial sampling design for se‐ diment quality assessment in estuaries. Environmental Monitoring & Software, 18:853‐ 859. doi:10.1016/S1364‐8152(03)00103‐8  de Wit PJ, Kranenburg C (1997) The wave‐induced liquefaction of cohesive sediment beds.  Estuarine Coastal and Shelf Science, 45:261‐271  Downing  J  (2006)  Twenty‐five  years  with  OBS  sensors:  The  good,  the  bad,  and  the  ugly.  Continental Shelf Research, 26: 2299‐2318. doi:10.1016/j.csr.2006.07.018.  Doxaran  D,  Froidefond  J‐M,  Lavender  S,  Castaing  P  (2002)  Spectral  signatures  of  highly  turbid  waters.  Application  with  SPOT  data  to  quantify  suspended  particulate  matter  concentrations. Remote Sensing of Environment, 81:149‐161  Doxaran,  D,  Froidefond  J‐M, Castaing  P,  Babin  M  (2009) Dynamics  of  the  turbidity  maxi‐ mum zone in a macrotidal estuary (the Gironde, France): Observations from field and  MODIS  satellite  data.  Estuarine  Coastal  and  Shelf  Science,  81:321‐332.  doi:10.1016/j.ecss.2008.11.013  Eleveld MA, Pasterkamp R, van der Woerd HJ, Pietrzak JD (2008) Remotely sensed seaso‐ nality  in  the  spatial  distribution  of  sea‐surface  suspended  particulate  matter  in  the  southern  North  Sea.  Estuarine  Coastal  and  Shelf  Science,  80:103‐113.  doi:10.1016/j.ecss.2008.07.015  Fettweis M (2008) Uncertainty of excess density and settling velocity of mud derived from  in  situ  measurements.  Estuarine  Coastal  and  Shelf  Science,  78:426‐436.  doi:10.1016/j.ecss.2008.01.007  Fettweis M, Nechad B, Van den Eynde D (2007) An estimate of the suspended particulate  matter (SPM) transport in the southern North Sea using SeaWiFS images, in situ meas‐ urements  and  numerical  model  results.  Continental  Shelf  Research,  27:1568‐1583.  doi:10.1016/j.csr.2007.01.017  Fettweis M, Francken F, Van den Eynde D, Verwaest T, Janssens J, Van Lancker V. (2010)  Storm influence on SPM concentrations in a coastal turbidity maximum area with high  anthropogenic  impact  (southern  North  Sea).  Continental  Shelf  Research. 

29

doi:10.1016/j.csr.2010.05.001  Hall P, Davies AM (2005) The influence of sampling frequency, non‐linear interaction, and  frictional effects upon the accuracy of the harmonic analysis of tidal simulations. Ap‐ plied Mathematical Modelling, 29: 533‐552. doi:10.1016/j.apm.2004.09.015  Hoitink AJF, Hoekstra P (2005) Observations of suspended sediment from ADCP and OBS  measurements  in  a  mud‐dominated  environment.  Coastal  Engineering,  52:103‐118.  doi:10.1016/j.coastaleng.2004.09.005  Lauwaert B, Fettweis M, Cooreman K, Hillewaert H, Moulaert I, Raemaekers M, Mergaert  K, De Brauwer D (2004) Syntheserapport over de effecten op het mariene milieu van  baggerspeciestortingen. BMM, DVZ & aMT rapport, BL/2004/01, 52pp.  Lauwaert B, De Brauwer D, Fettweis M, Hillewaert H, Hostens K, Mergaert K, Moulaert I,  Parmentier K, Verstraeten J (2006) Syntheserapport over de effecten op het mariene  milieu  van  baggerspeciestortingen  (vergunningsperiode  2004‐2006).  BMM,  ILVO  &  aMT rapport, BL/2006/01, 87pp+ app.  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 (2008) Syn‐ theserapport over de effecten op het mariene milieu van baggerspeciestortingen (ver‐ gunningsperiode 2006‐2008). BMM, ILVO, aK & aMT rapport, BL/2008/01, 128pp.  Lauwaert B, Bekaert K, Berteloot M, De Backer A, Derweduwen J, Dujardin A, Fettweis M,  Hillewaert  H, Hoffman S, Hostens K, Ides S, Janssens J, Martens C, Michielsen T, Par‐ mentier K, Van Hoey G, Verwaest T (2009a) Synthesis report on the effects of dredged  material disposal on the marine environment (licensing period 2008‐2009). Report by  BMM, ILVO, CD, aMT and WL BL/2009/01. 73pp.  Lauwaert B, Bekaert K, Berteloot M, De Backer A, Derweduwen J, Dujardin A, Fettweis M,  Hillewaert  H, Hoffman S, Hostens K, Ides S, Janssens J, Martens C, Michielsen T, Par‐ mentier K, Van Hoey G, Verwaest T (2009b) Syntheserapport over de effecten op het  mariene  milieu  van  baggerspeciestortingen  (vergunningsperiode  2008‐2009).  Report  by BMM, ILVO, CD, aMT and WL BL/2009/01. 18pp.  Li MZ, Amos CL, Heffler DE (1997) Boundary layer dynamics and sediment transport under  storm and non‐storm conditions on the Scotian Shelf. Marine Geology, 141:157‐181  Li Y, Mehta AJ (2000) Fluid mud in the wave‐dominated environment revisited. In: Coastal  and Estuarine Fine Sediment Dynamics. In: McAnally WH, Mehta AJ (eds) Coastal and  Estuarine Fine Sediment Dynamics, Proceedings in Marine Science 3, Elsevier, Amster‐ dam, 79‐93  Limpert E, Stahel W, Abbt M (2001) Log‐normal distributions across the sciences: Keys and  clues. BioScience, 51:341‐352  Ma  Y,  Wright  LD,  Friedrichs  CT  (2008)  Observations  of  sediment  transport  on  the  conti‐ nental  shelf  off  the  mouth  of  the  Waiapu  River,  New  Zealand:  Evidence  for  current‐ supported  gravity  flows.  Continental  Shelf  Research,  28:516‐532.  doi:10.1016/j.csr.2007.11.001  Nechad B, De Cauwer V, Park Y, Ruddick KG (2003) Suspended Particulate Matter (SPM)  mapping  from  MERIS  imagery.  Calibration  of  a  regional  algorithm  for  the  Belgian  coastal waters. ESA MERIS user workshop, 10‐13th November 2003  Nechad  B,  Ruddick  KG,  Park  Y  (2010)  Calibration  and  validation  of  a  generic  multisensor  algorithm for mapping of Total Suspended Matter in turbid waters. Remote Sensing of  Environment, 114:854‐866. doi:10.1016/j.rse.2009.11.022  Ogston  AS,  Cacchione  DA,  Sternberg  RW,  Kineke  GC  (2000)  Observations  of  storm  and  river‐flood  driven  sediment  transport  on  the  northern  California  continental  shelf.  Continental Shelf Research, 20:2141‐2162  Pepper DA,  Stone  GW  (2004)  Hydrodynamic  and  sedimentary  responses  to  two contras‐ ting winter storms on the inner shelf of the northern Gulf of Mexico. Marine Geology,  210:43‐62. doi:10.1016/j.margeo.2004.05.004  PIANC (2008) Minimising harbour siltation. PIANC report No 102, 88pp. 

30

Ruddick KG, Ovidio F, Rijkeboer M (2000) Atmospheric correction of SeaWiFS imagery for  turbid coastal and inland waters. Applied Optics, 39:897‐912  Ruhl CA, Schoelhammer DH, Stumpf RP, Lindsay CL (2001) Combined use of remote sen‐ sing and continuous monitoring to analyse the variability of suspended‐sediment con‐ centrations  in  San  Francisco  Bay,  California.  Estuarine  Coastal  and  Shelf  Science,  53:801‐812. doi:10.1006/ecss.2000.0730  Schleppi  P,  Waldner  PA,  Fritschi  B  (2006)  Accuracy  and  precision  of  different  sampling  strategies and flux integration methods for runoff water: comparisons based on meas‐ urements  of  the  electrical  conductivity.  Hydrological  Processes,  20:395‐410.  doi:10.1002/hyp.6057  Shindo K, Otsuki A (1999). Establishment of a sampling strategy for the use of blue mussels  as an indicator of organotin contamination in the coastal environment. Journal of Envi‐ ronmental Monitoring, 1:243‐250  Soulsby R (1997) Dynamics of marine sands. Thomas Telford Publications, London. 249pp  Van den Eynde D, Nechad B, Fettweis M, Francken F (2007) SPM dynamics in the southern  North Sea derived from SeaWifs imagery, in situ measurements and numerical model‐ ling. In: Maa JP‐Y, Sanford LP, Schoelhammer DH (eds) Estuarine and coastal fine sedi‐ ment dynamics. Proceedings in Marine Science 8. Elsevier, Amsterdam, pp 299‐311  Wang M, Son S, Shi W (2009) Evaluation of MODIS SWIR and NIR‐SWIR atmospheric cor‐ rection algorithms using SeaBASS data. Remote Sensing of Environment 113:635‐644.  doi:10.1016/j.rse.2008.11.005  Werdell PJ, Bailey SW, Franz BA, Harding LW, Feldman GC, McClain CR (2009) Regional and  seasonal  variability  of  chlorophyll‐a  in  Chesapeake  Bay  as  observed  by  SeaWiFS  and  MODIS‐Aqua.  Remote  Sensing  of  Environment,  113:1319‐1330.  doi:  10.1016/j.rse.2009.02.012  Winterwerp J (2006) Stratification effects by fine suspended sediment at low, medium and  very  high  concentrations.  Journal  of  Geophysical  Research,  111:C05012.  doi:10.1029/2005JC003019  Zawada DG, Hu C, Clayton T, Chen Z, Brock JC, Muller‐Karger FE (2007) Remote sensing of  particle backscattering in Chesapeake Bay: A 6‐year SeaWiFS retrospective view. Estua‐ rine, Coastal and Shelf Science, 73:792‐806. doi:10.1016/j.ecss.2007.03.005 

31

COLOPHON                                 

                            BEHEERSEENHEID VAN HET MATHEMATISCH MODEL NOORDZEE

VAN

Dit rapport werd voorbereid door de BMM in juni 2010  Zijn referentiecode is MOMO/5/MF/201006/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 e‐mail te zenden naar [email protected], met vermelding van de referentie,  of te schrijven naar:  BMM  100 Gulledelle  B–1200 Brussel  België  Tel:   +32 2 773 2111  Fax:  +32 2 770 6972  http://www.mumm.ac.be/ 

     

DE

SUMO GROEP

 

 

 

  

32

APPENDIX 1      Fettweis, M., Nechad, B., Van Lancker, V., Van den Eynde, D. 2010. Evaluation of in  situ  and  remote  sensing  sampling  methods  of  SPM  concentration.  AGU  Ocean  Science Meeting, 22‐26 February, Portland (USA) 

Evaluation of in situ and remote sensing sampling methods of SPM concentration Michael Fettweis, Bouchra Nechad, Vera Van Lancker, Dries Van den Eynde Royal Belgian Institute of Natural Science, Management Unit of the North Sea Mathematical Models (MUMM), Gulledelle 100, Brussels, Belgium (Email: [email protected]) Time series of in situ SPM concentration and satellite imagery are valuable data sources for the analysis of suspended-sediment transport in coastal or estuarine areas. Still, shortcomings remain, with satellite imagery suffering from a low temporal resolution and only related to surface data, whilst in situ measurements have a limited spatial resolution. Ship time and budget are often limited; it is thus of primary importance to choose a sampling method or a combination of methods, providing a representative sub-sample of the population in the time- and space domain. If long-term variations, induced by natural changes or anthropogenic effects, need resolving, overprinting tidal, neap-spring as also seasonal signals need filtering. This requires sufficiently dense sampling in time and long data series. To our knowledge only few efforts are being made to design or to evaluate existing sampling schemes. Often the best sampling strategy cannot be chosen, as it depends also on the availability of remote sensing and in situ data. Autonomous stations (tripods) with an almost continuous measurement of SPM concentration are relatively easy to design; however optical remote sensing images, available at lower time resolution and fair weather conditions only, often miss the occurring high ranges of SPM concentration during storms. Knowledge on the uncertainty, introduced by the sampling method and instrumentation, is therefore important in data interpretation, as well as data assimilation. The aim of the presentation is the evaluation of the temporal SPM heterogeneity in the Belgian nearshore, using a large set of SPM concentration data from MODIS (MODerate resolution Imaging Spectroradiometer) ocean colour satellite and from in situ measurements (tidal cycle, tripod). As match-ups (satellite picture at the same time as in situ measurements) are scarce, statistical methods are used to evaluate the differences and similarities in the data sets. This approach is new and allows comparing different data sets, not necessarily sampled at the same moment in time. Further, the sampling strategy is analysed and the representativeness of the different data sets is discussed.

APPENDIX 2     

Chen P, Yu J, Fettweis M, Van den Eynde D, Maggi F. 2010. Flocculation  in a nutrient‐rich coastal area (southern North Sea): Measurements and  modelling. Poster at AGU Ocean Science Meeting, 22‐26 February,  Portland (USA). 

Flocculation in a nutrient-rich coastal area (southern North Sea): Measurements and modeling Peihung Chen, Jason Yu National Sun Yat-sen University, Department of Marine Engineering, 80424 Kaohsiung, Taiwan Michael Fettweis, Dries Van den Eynde Royal Belgian Institute of Natural Science, Management Unit of the North Sea Mathematical Models (MUMM), Gulledelle 100, Brussels, Belgium Federico Maggi The University of Sydney, School of Civil Engineering J05, Sydney NSW 2006, Australia Knowledge on cohesive sediment transport processes is required to predict the distribution of suspended and deposited cohesive sediments in natural or anthropogenically created environments such as navigation channels and harbours. Settling of mud flocs is controlled by flocculation and hence also determines the transport of cohesive sediments. Flocculation is the process of floc formation and break-up which has a direct impact on settling velocity. The settling velocity is a function of the particle size and excess density and varies strongly in natural environments because Suspended Particulate Matter (SPM) consists of a population of flocs with heterogeneous sizes, densities, shapes and constituents (e.g. Eisma and Kalf, 1987). Natural SPM comprises many different substances with concentrations that are generally site specific and time varying. Although an accurate taxonomy is currently lacking, the SPM can be divided by inorganic and organic fractions. The inorganic fraction mainly consists of clay minerals, carbonates, quartz and other silicates. The organic fraction of the SPM is prevalently made of a variety of micro-organisms, their metabolic products, residuals from dead organisms, and fecal pellets (e.g. Mehta, 1989; Droppo et al., 1997; Grossart et al., 2003; Bhaskar et al., 2005;). The two fractions of the SPM are intimately related by physical, biological and chemical processes which make the SPM a complex, reactive biomaterial distributed in the water body. Maggi (2009) presented a flocculation model where a coupling between the mineral and micro-organism dynamics was implemented. The model was calibrated using in situ measurements of SPM concentration, turbulent shear rate and average floc size collected in the Belgian North Sea (Fettweis et al., 2006). Using the tuned coefficients, the model described the particle size observations well. The significant influence of micro-organism and organic matter on the particle size as well as the well-known variation of particle size with turbulent shear stress could be reproduced. The aim of the presentation is to show results of calibration of the model against a large set of SPM concentration and particle size measurements from different location on the Belgian continental shelf, extended with turbulence shear rate modeled with COHERENS-3D. Refernces Bhaskar, P.V., Grossart, H.P., Bhosle, N.B., Simon, M., 2005. Production of macroaggregates from dissolved exopolymeric substances (EPS) of bacterial and diatom origin. FEMS Microbiology Ecology 53, 255–264. Droppo, I.G., Leppard, G.G., Flanning, D.T., Liss, S.N., 1997. The freshwater floc: a functional relationship of water and organic and inorganic floc constituents affecting suspended sediment properties. Water, Air and Soil Pollution 99, 43–54. Eisma, D., Kalf, J. 1987. Distribution, organic content and particle size of suspended matter in the North Sea. Netherlands Journal of Sea Research, 21, 265-285. Maggi, F. 2009. Biological flocculation of suspended particles in nutrient-rich aqueous ecosystems. Journal of Hydrolog, 376, 116-125. Fettweis, M., Francken, F., Pison, V., Ven den Eynde, D., 2006. Suspended particulate matter dynamics and aggregate sizes in a high turbidity area. Marine Geology, 235, 63–74. Grossart, H.P., Kiorboe, T., Tang, K., Ploug, H., 2003b. Bacterial colonization of particles: growth and interactions. Applied and Environmental Microbiology, 69, 3500–3509. Mehta, A.J., 1989. On estuarine cohesive sediment suspension behavior. Journal of Geophysical Research – Oceans 94 (c10), 14303–14314.

Paper Number: G035B-08

Flocculation in a nutrient-rich coastal area (southern North Sea): Measurements and modeling Peihung Chen1, Jason C-S Yu1, Michael Fettweis2, Dries Van den Eynde2 and Federico Maggi3 Sun Yat-sen University, Department of Marine Engineering, 80424 Kaohsiung, Taiwan 2 Royal Belgian Institute of Natural Science, Management Unit of the North Sea Mathematical Models (MUMM), Gulledelle 100, Brussels, Belgium 3 The University of Sydney, School of Civil Engineering J05, Sydney NSW 2006, Australia E-mail: [email protected] 1National

Introduction

The Belgian and southern Dutch coastal waters are an effective trap for fine-grained cohesive sediments. Most of these suspended sediments originate from the English Channel transported into the North Sea through the Dove Strait. Continuous dredging and dumping activities in the Belgian coastal water and harbors add about 10 millions tons of dry matter annually, from which 70% is silt and clay (Fettweis et al., 2006 ). These are the main sources of recent fine grained sediments in the southern North Sea (Fig.1 ). Natural Suspended Particulate Matter (SPM) comprises many different substances with concentrations that are generally site specific and time varying. The SPM can be divided by inorganic and organic fractions. The inorganic fraction mainly consists of clay minerals, carbonates, quartz and other silicates. The organic fraction of the SPM is prevalently made of a variety of micro-organisms, their metabolic products, detritus, and fecal pellets (e.g. Mehta, 1989; Droppo et al., 1997; Grossart et al., 2003; Bhaskar et al., 2005). Settling of mud flocs is controlled by flocculation and hence also determines the transport of cohesive sediments. Flocculation is the process of floc formation and break-up which has a direct impact on settling velocity. The settling velocity is a function of the particle size and excess density and varies strongly in natural environments because SPM consists of a population of flocs with densities, heterogeneous sizes, shapes and constituents (e.g. Eisma and Kalf, 1987; van Leussen, 1994). A flocculation model (BFLOC) proposed by Maggi (2009) taking the coupling effects between the mineral and micro-organism

LISST Sampling bottles OBS

Fig. 1 SPM distribution in the North Sea

Fig. 2 (a,left) Instrument setup aboard; (b,right) Bottom sediment sample.

dynamics was calibrated using two sets of in situ data collected in the Belgian North Sea (Fettweis et al., 2006). The model has shown significant influence of micro-organism and organic matter on the particle size and the variation of particle size with turbulence. The aim of this poster is to show the results of calibration of the model against a larger set of SPM measurements from different location on the Belgian continental shelf, extended with turbulence shear rate modeled with COHERENS-3D. Flocs observed from the time series through out a complete tidal cycle during several years and seasons are analyzed and compared with various variables measured at the same time, i.e. SPM, POC. Sensitivity tests are carried out for understanding the influences of each parameter. Modeled floc sizes with the tuned parameters during tidal cycles have shown good agreement for some sets of data. The possible causes of those model results not fitted with the measurements are also discussed.

Data and analysis

Table 2 Parameters used for sensitivity test Date Set: 2004-16, 2004-25-A (ref: 2003-22 from Maggi (2009))

Data collected during 2003 and 2004 (see Table 1) on the Belgian continental shelf are discussed here (see figure 1 for the sites). These are ship board measurements for a complete tidal cycle (13 hours) which consist of in situ observations of SPM using OBS and floc sizes with LISST 100 optical sensors (figure 2). Water samples were taken for lab experiments on SPM for OBS-SPM calibration and the POC/PON for organic fractions in the SPM . Continuous filtration of the water was also carried out for analyzing the primary particles. Velocities were measured by current meter or shipboard ADCP. The currents, surface elevation and turbulent kinetic energy have been computed for these measuring periods using a 3D hydrodynamic operational model for the Belgian continental shelf (OPTOSBCS). The turbulence closure scheme used in OPTOS-BCS describes the turbulent energy dissipation as the product of a velocity and a length scale Mellor and Yamada (1974). Parameters were first analyzed in order to find their correlations. OBS backscatter data are compared with sampled SPM have shown strong correlation (see Table 2 and Figure 3), and used for calibrating the SPM time series observed from OBS. The biomass concentrations are calculated using the sampled POC/PON fraction of the SPM. Higher organic fractions can be observed with lower SPM concentrations (Fiure 4). The relation between Kolmogrov micro-scale(λ) and current velocity is also obtained for computing the turbulent shear rates in the BFLOC model (Figure 5).

Table 1 Tidal cycle measurements, further the linear regression coefficients between the OBS signal and the SPM concentration from filtrations are shown (TM= coastal turbidity maximum)

10

(POC+PON)/SPM (%)

8

Nr.

Data

Location

Area

SPM=A+B×OBS A

B

2003-15

11-12/06/2003

Kwinebank

Offshore

0.31

2.253

2003-25

09-10/10/2003

Kwinebank

Offshore

1.30

1.472

2

2004-16 2004-25-A

15-16/07/2004 08-09/11/2004

B&W Oostende MOW1

TM TM

5.60 7.95

1.643 1.542

0

2004-25-B

09-10/11/2004

Hinderbank

Offshore

3.06

1.953

Min

Max

(σ )

(p-3σ )

(p+3σ )

0.189

0.0378

0.0756

0.3024

2

kb '

[−]×10-6

11.41

2.2820

4.5640

18.2560

3

η max

[s-1]×10-6

6.586

1.3172

2.6344

10.5376

4

Km

[M]×10-6

1.159

0.2318

0.4636

1.8544

5

β

[−]

0.226

0.0452

0.0904

0.3616

Date Set: 2003-15, 2003-25, 2004-25-B (ref: 2004-04-05 from Maggi (2009)) No.

Parameter

Original value

20% of the value

Min

Max

(p-3σ )

(p)

(σ )

1

ka '

[−]

0.156

0.0312

0.0624

0.2496

2

kb '

[−]×10-6

4.261

0.8522

1.7044

6.8176

3

η max

[s-1]×10-6

1.2252

2.4504

9.8016

4

Km

[M]×10-6

1.968

0.3936

0.7572

3.1488

5

β

[−]

0.285

0.0570

0.1140

0.4560

Nr.

100

200

300 400 500 SPM concentration (mg/l)

600

700

800

6.126

ref: Michael Fettweis, 2008

Tuned parameter

SPM ± stdv

Df ±stdv

d ±stdv

003-15

4.5±1.1

160±38

2.06±0.02

kb ' = 2.025 × 10

003-25

27±12

75±20

2.08±0.04

ka ' = 0.148

−6

004-16

32±14

81±22

3.23±0.06

kb ' = 9.934 × 10

004-25-A

89±54

88±25

1.72±0.03

ka ' = 0.076

004-25-B

3.6±1.3

115±34

3.25±0.03

−6

kb ' = 18.202 × 10

2004-16(B&W Oostende) 200 Floc size(um)

20% of the value

(p)

[−]

4

0

160

Original value

Parameter

ka '

(p+3σ )

Table 3 RMSE between the BFLOC model results and the measured floc sizes using calibrated parameters.

6

Fig. 3 OBS-SPM Calibration. Fig. 4 Organic fractions in SPM. Fig. 5 Current-Turbulence.

Results of the Biological Flocculation Model

No. 1

in-situ BFLOC(org)

Floc size(um)

BFLOC(ka') Maggi (2009) presented a flocculation model where a coupling between the mineral and micro-organism dynamics 120 80 was implemented. The model was calibrated using two sets of in situ measurements of SPM concentration, turbulent 40 shear rate and average floc size collected in the Belgian North Sea (Fettweis et al., 2006). Using the coefficients tuned by 0 Maggi (2009), the model described the particle sizes well comparing with the observations. The significant influence of 15-Jul 15:45 15-Jul 17:38 15-Jul 19:31 15-Jul 21:25 15-Jul 23:18 16-Jul 01:12 16-Jul 03:05 16-Jul 04:59 2003-15(Kwintebank) biomass on the particle size can clearly observed. A sensitivity test has been carried out for the 5 parameters used in the 300 in-situ BFLOC model, i.e. the aggregation parameter; breakup parameter; the biomass growth rate; half-saturation concentration 240 BFLOC(org) BFLOC(kb') 180 and carrying capacity coefficient. A similar Monte Carlo analysis was carried out to characterize the parameters. We 120 applied five independent normally-distributed probability density functions to the parameters with the averages obtained 60 from standard deviation equal to 20% of each parameters (σ) according to the parameters of Maggi (2009). The range of 0 11-Jun 18:25 11-Jun 20:13 11-Jun 22:01 11-Jun 23:49 12-Jun 01:37 12-Jun 03:25 12-Jun 05:13 12-Jun 07:02 parameter values was limited to ±3 times the standard deviation in 300 replicate test (see Table 2). Root mean square Fig. 6 Modeled time series from (a, up) the better fitted errors (RMSE) of each test set are calculated. The best parameters for each measurement are listed in Table 3. results; (b,low) not fitted results. The RMSE can be improved from 65.363µm down to 26.988µm for the measurement 2004-25-A. There are no significant improvements but well fitted with the observations for the sets 2003-25 and 2004-16 (Figure 6a). The worst predictions are for the 2003-15 (Fig 6b) and 2004-25-B, though RMSE could decrease (See Table3). The floc sizes of these two data sets are mostly biger than 100µm (160±38µm, 115±34µm). The floc sizes of these two measurements are not well distributed from the LISST records (see Figure 7). This may due to the real particles are mostly larger than 500µm which were exceeded the valid LISST record range. And, thus caused the modeled results based on the biological flocculation depart from the observations. Further study will be required, e.g. correction of the floc size distribution patterns will be necessary for verifying the influences.

Fig. 7 Particle (floc) size distribution of the SPM measured by the LISST as function of volume concentration

−6

kb ' = 1.938 × 10

−6

RMSE(µm) Original Tuned

O−T

85.059

68.516

16.543

16.222

16.156

0.066

24.481

23.414

1.067

65.363

26.988

38.375

58.316

39.188

19.128

Conclusions

Five data sets of SPM measurements in Belgian coastal area are further studied and the floc dynamics are modeled after Maggi (2009). Data sets with fine particles are well modeled using calibrated parameters. Two sets of measurements were not well predicted by the flocculation model are investigated. This may cause by the uncertainties of the LISST records since the particles are exceeding the valid instrument ranges. Further study on this issue will be necessary.

References Bhaskar, P.V., Grossart, H.P., Bhosle, N.B., Simon, M., 2005. Production of macroaggregates from dissolved exopolymeric substances (EPS) of bacterial and diatom origin. FEMS Microbiology Ecology 53, 255–264. Droppo, I.G., Leppard, G.G., Flanning, D.T., Liss, S.N., 1997. The freshwater floc: a functional relationship of water and organic and inorganic floc constituents affecting suspended sediment properties. Water, Air and Soil Pollution 99, 43–54. Eisma, D., Kalf, J., 1987. Distribution, organic content and particle size of suspended matter in the North Sea. Netherlands Journal of Sea Research, 21, 265-285. Fettweis, M., 2008.Unvertainty of excess density and settling velocity of mud flocs derived from in situ measurements. Estuarine, coast and shelf science, 78, 426–436. Fettweis, M., Francken, F., Pison, V., Ven den Eynde, D., 2006. Suspended particulate matter dynamics and aggregate sizes in a high turbidity area. Marine Geology, 235, 63–74. Grossart, H.P., Kiorboe, T., Tang, K., Ploug, H., 2003b. Bacterial colonization of particles: growth and interactions. Applied and Environmental Microbiology, 69, 3500–3509. Maggi, F., 2009. Biological flocculation of suspended particles in nutrient-rich aqueous ecosystems. Journal of Hydrolog, 376, 116-125. Mehta, A.J., 1989. On estuarine cohesive sediment suspension behavior. Journal of Geophysical Research – Oceans 94 (c10), 14303–14314. Van Leussen, W., 1994. Estuarine macroflocs and their role in fine-grained sediment transport. Ph. D. thesis, University of Utrecht, The Netherland.

APPENDIX 3     

Fettweis M, Van den Eynde D, Francken F, Van Lancker V. SPM dynamics  measured  with  an  automated  tripod  in  the  Belgian  nearshore  area:  natural  dynamics  and  anthropogenic  effects.  Liège  Colloquium,  26‐30  April 2010 

SPM dynamics measured with an automated tripod in the Belgian nearshore area: natural dynamics and anthropogenic effects FETTWEIS M1, VAN DEN EYNDE D1, FRANCKEN F1, VAN LANCKER V1 1

Royal Belgian Institute of Natural Sciences, MUMM, Belgium

Large amounts of sediments are dredged in the Belgian nearshore area to maintain ships’ access to ports and harbours. The EU Marine Water Framework Directive provides a framework that embodies the principles of environmental protection, improvement and restoration on an integrated basis. In addition, the WFD specifies that there must be no temporal deterioration in chemical and biological status for many water bodies, and identifies (Annex VIII) ‘material in suspension’ as one of the main pollutants. When human activities occur in habitats characterised by cohesive seabed sediments or by high turbidities, resuspension of material or dredging and dumping can result in higher concentrations of suspended particulate matter (SPM), which can spread over large areas. The manner in which the system reacts to engineering works needs to be understood to ensure cost-effective operations at sea, to better gauge the human footprint, and to develop environmental policies aiming at a more sustainable management of the marine environment. Reference situations are rarely available in the marine environment and, therefore, true impacts are difficult to be assessed unambiguously. In situ data have been collected near Zeebrugge in the framework of a scientific experiment that was set up by the Ministry of the Flemish Community during May 2009 in order to evaluate an alternative dredging method for the Albert II dock (port of Zeebrugge). The dredged matter was directly pumped outside the port and the SPM concentration and other parameter were measured using a tripod at about 2 km from the dumping site. Data have been collected before, during and after the dredging works. Mounted instruments on the tripod include a SonTek 3 MHz ADP, a SonTek 5 MHz ADV Ocean, a Sea-Bird SBE37 CT system, two OBS (one at 0.2 and another at 2 m above bed (mab), a LISST 100 and two SonTek Hydra systems for data storage and batteries. In order to measure changes of a natural highly dynamic quantity it is of primary importance to choose a sampling method, providing a representative sub-sample of the population in the time- and space domain. If long-term variations, induced by natural changes or anthropogenic effects, need resolving, overprinting tidal, neap-spring as also seasonal signals need filtering. This requires sufficiently dense sampling in time and long data series together with the uncertainty, introduced by the sampling method and instrumentation. The aim of the measurements was to investigate the effects of dumping works on the SPM concentration and to evaluate the temporal SPM heterogeneity in the Belgian nearshore, using a large set of SPM concentration. Statistical methods, based on probability density and auto correlation functions are used to evaluate the data sets.

1

Royal Belgian Institute of Natural Sciences, Management Unit of the North Sea Mathematical Models, Gulledelle 100, 1200 Brussels, Belgium

SPM dynamics measured with an automated tripod in the Belgian nearshore area: natural dynamics and anthropogenic effects M. FETTWEIS, D. VAN DEN EYNDE, F. FRANCKEN, V. VAN LANCKER Royal Belgian Institute of Natural Science (RBINS), Management Unit of the North Sea Mathematical Models (MUMM), Gulledelle 100, 1200 Brussels, Belgium, E‐mail: [email protected]

INTRODUCTION Many access channels and harbours suffer from sedimentation of fine sediments and formation of fluid mud layers. Conventional dredging methods with trailer suction dredgers, and disposal of the dredged material at  designated locations, incur substantial costs.  In May 2009 an experimental study was carried out in the port of Zeebrugge (Albert II dock) to investigate whether the thickness of the mud layer with a density lower than 1200 kg/m³, could be reduced by pumping  using a cutter dredger. The dredged matter was pumped over the harbour breakwater into the sea at a location closer to the shore and the port as compared to the existing disposal sites (Lauwaert et al., 2009). The  investigations performed during the disposal experiment consist of measurements in the water column (e.g. temperature, salinity, SPM concentration, particle size and current velocity) at a fixed location situated about 3  km west of the disposal site to evaluate far field effects. Although the evaluation of the project was negative in terms of efficiently reducing the thickness of the fluid mud layer insight was gained in the monitoring efforts  needed to assess the impact on SPM concentration outside the port. 

Fluid mud from the port of Zeebrugge

The SPM concentration variability at Blankenberge is the  result of various processes related to tides, storms and  seasonal changes. In order to identify the effects from  dredged material disposal, statistical methods are used.  By constructing frequency distributions of different data  sets we can calculate using standard statistic tests if two  distributions are drawn from the same distribution  function. If disposal of dredged material over the western  breakwater of the port has an impact on SPM  concentration then this should be detected in variations of  statistical parameters during the data collected during the  field experiment (see Fettweis & Nechad, 2010). Further,  changes in SPM concentration due to disposal could results  in a shift of the SPM concentration maxima during the  tides.  

Probability density

0.1

0.0 0.3

2 (Dec 2006 - Feb 2007) x*=309 mg/l s*=2.97 p=0.57

0.2

0.1

2 (Dec 2006 - Feb 2007) x*=149 mg/l s*=2.32 p=0.93

0.2

0.1

0.0

0.3 3 (Jan - Feb 2008) x*=183 mg/l s*=2.36 p=0.77

0.2

Probability density

Probability density

0.1

3 (Jan - Feb 2008) x*=105 mg/l s*=2.46 p=0.59

0.2

0.1

0.0 0.3 4 (Mar - Apr 2008) x*=290 mg/l s*=3.01 p=0.96

0.2

Probability density

Probability density

Right: Bathymetry in the  southern North Sea and  the tripod location at  Blankenberge

0.1

0.0

0.0 0.3

0.1

4 (Mar - Apr 2008) x*=150 mg/l s*=2.46 p=0.99

0.2

0.1

0.0 0.4

0.0 0.3

5 (Apr - Jun 2008) x*=258 mg/l s*=2.71 p=0.53

0.2

Probability density

Probability density

Above: Albert II dock  (port of Zeebrugge) and  the dredging locations.

Tripod deployment at Blankenberge

1 (Nov-Dec 2006) x*=144 mg/l s*=2.26 p=0.93

0.2

0.3

0.0 0.3

0.1

5 (Apr - Jun 2008) x*=102 mg/l s*=2.41 p=0.81

0.3

0.2

0.1

Numerous morphological, and sedimentological effects result from dredging and disposal works in nearshore areas. The current study focuses on the effects of a continuously disposal of fine‐grained  dredged material on the increase of SPM concentration and on the possible formation of fluid mud close to the shoreline. The major conclusion from the data are:

0.0 0.3

0.0 0.3 6 (May - Jun 2009) x*=584 mg/l s*=2.65 p=1.00

0.2

Probability density

Probability density

Tripod location

0.3 1 (Nov-Dec 2006) x*=340 mg/l s*=2.93 p=1.00

0.2

Probability density

Probability density

Probability density

0.3

Disposal location of  dredged material

METHOD A tripod has been deployed for 240 days during 6  measuring periods before, during and after the  experiment. SPM concentration was measured at 0.2 m  and 2 m above bed.

0.1

6 (May - Jun 2009) x*=149 mg/l s*=2.24 p=0.98

0.2

0.1

0.0

0.0 0

500

1000 1500 2000 2500 0.2 mab SPM concentration (mg/l)

3000

0

250

500 750 1000 1250 2 mab SPM concentration (mg/l)

Probability density distribution of the SPM  concentration data at 0.2 mab (left) and 2 mab (right) during the 6 measuring periods and the  corresponding log‐normal probability density  functions and Χ² test probability p. The data are  binned in classes of 50 mg l‐1, the dashed lines  correspond to the median x* times/over the  multiplicative standard deviation s* . The field experiment took place duirng period 6

1500

SPM concentration at 0.2 mab (SPM1) and 2 mab (SPM2) and tide‐averaged  data during periods 1, 5 and 6 (field experiment)

• There is a dominant quarter‐diurnal signal in the time‐series. The spring‐neap tidal signal is clearly visible during calm weather;  • SPM concentrations maxima were sometimes up to 50 times higher than the minima;  • The very high SPM concentrations near the bed during winter  and  autumn are caused by storms and suggest that high concentrated  mud layers or fluid layers are formed (Fettweis et al., 2010);  • During the field experiment (May – June 2009) the near‐bed  dynamics was mainly influenced by HCBS layers, whereas at 2 mab a spring‐neap tidal signal could be identified

CONCLUSION Harbour authorities world‐wide are obliged to dredge their major shipping channels, and subsequently to dispose of the dredge spoil offshore. This study provided a statistical technique to evaluate the effects of disposal operations in a highly turbid area with high natural variability. The following arguments show that the disposal of dredged matter have caused an increase in SPM concentration: • The SPM concentration near the bed (0.2 mab) was exceptionnally high during the field experiment. The median SPM concentration was 240 mg/l higher than  during measuring period 1 (high autumn SPM concentration) and about 330 mg/l higher than during the same period in 2008.  • The port and the disposal site are situated in ebb‐direction of the measuring location. During the experiment a generally higher SPM concentration near the bed  during ebb and at 2 mab during flood was observed, suggesting that disposal of dredged material is mainly transported in the benthic layer.  • The time lag between high wave and high SPM concentration proves that the SPM has been advected towards the measuring location rather than eroded locally. Further readings: Fettweis M, Nechad B (2010) Evaluation of in situ and remote sensing sampling methods for SPM concentrations, Belgian continental shelf (southern North Sea). Ocean Dynamics (in revision) Fettweis M, Francken F, Van den Eynde D, Verwaest T, Janssens J, Van Lancker V (2010) Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea).  Continental Shelf Research (in revision) Lauwaert B, Bekaert K, Berteloot M, De Backer A, Derweduwen J, Dujardin A, Fettweis M, Hillewaert H, Hoffman S, Hostens K, Ides S, Janssens J, Martens C, Michielsen T, Parmentier K, Van Hoey G, Verwaest T (2009)  Synthesis report on the effects of dredged material disposal on the marine environment (licensing period 2008‐2009). Report by MUMM, ILVO, CD, aMT and WL BL/2009/01. 73pp.

This study was funded by the Maritime Access Division of the Ministry of the Flemish Community, within the framework of the MOMO project and partly by the Belgian Science Policy projects QUEST4D (SD/NS/06A).

APPENDIX 4     

Baeye  M,  Fettweis  M,  Van  Lancker  V,  Francken  F.  2010.  Monitoring  morphological  changes  using  near‐bed  ADV  altimetry. Liège Colloquium, 26‐30 April 2010 

Monitoring Morphological Changes using near-bed ADV Altimetry. BAEYE, M.1, FETTWEIS, M.2, FRANCKEN, F.2, VAN LANCKER, V.2 1

Ghent University, Belgium Management Unit of the North Sea Mathematical Models, Belgium

2

Morphological changes of the bottom is of interest to many coastal and offshore applications such as engineering works (windmills, harbour construction, dredging), aggregate extraction, the underwater heritage protection and the object burial in general. The objective of this study is understanding seabed morphodynamics, comprising several spring/neap tidal cycles. Continuous time series were made available from a Sontek Acoustic Doppler Velocimeter (ADV) mounted on a multisensor tripod, at different locations on the Belgian Continental Shelf (RBINSMUMM). Most measurements are conducted in near coastal shallow waters, though also offshore locations have been targeted. Data from the nearshore moorings clearly show dynamic trends in accretion and erosion. The results are statistically related to time-series of tidal forcing, wind and wave stresses using cross- and multi-spectral analyses and wavelet transforms. Furthermore, the data are coupled to time series of an optical backscatter sensor (OBS) to detect occurrences of near-bed fluffy layers and fluid mud sheets.

Please send your abstract by e-mail to [email protected]

Please indicate hereafter if you need special technical equipment for your presentation(s) x

1 2

Renard Centre of Marine Geology, Krijgslaan 281 B-9000 Ghent MUMM, Gulledelle 100, B-1200 Brussels (St Lambrechts-Woluwe)

Fluid Mud Dynamics Derived from ADV Altimetry, Belgian Coastal Zone Matthias Baeye1, Michael Fettweis2, Frederic Francken², Vera Van Lancker2

1

2

Department of Geology and Soil Science, Renard Centre of Marine Geology, Ghent University, Krijgslaan 281, B-9000, Gent, Belgium. [email protected]. Department VI of the Belgian Royal Institute of Natural Sciences, Management Unit of Mathematical Models North Sea, Gulledelle 100, B-1200 Brussels (St Lambrechts-Woluwe), Belgium. [email protected], [email protected], [email protected].

Cohesive sediment in coastal systems

eg. fluid mud/HCMS (high-concentration mud suspensions) Depending on sediment properties, meteo-hydrodynamic conditions, availability of the sediment.

In the southern North Sea

A tripod frame mooring during winter of 2007 (28 days), see Figure 3

Belgian Continental Shelf Characteristics macrotidal regime (tidal amplitude maximum of 4-5 m), occurrence of moderate wave conditions (0.5-2 m of significant wave height).

Study area? near-shore/west off Zeebrugge harbor

Aim of study?

shallowness, maximum current velocities up to 1 m/s,

evaluating the probability for detecting fluid mud formation by means of an ADV (acoustic Doppler velocimeter) mounted on a tripod frame (Figure 1)

highly energetic hydrodynamic conditions, Fettweis and Van den Eynde (2003): highly turbid with mean SPM concentrations (50 - 1000 mg/l), occurrence of turbidity maxima and mud fields.

correlating observed sea bed level changes with hydro-meteo conditions

-

ADV, OBS (optical backscatter sensor), LISST (Laser In-Situ Scattering and Transmissometry) SonTek ADV/Ocean (5MHz) (besides flow measurements)

distance between probe tip and nearby physical boundary within range  “detecting the spike in signal strength corresponding to the reflection of the acoustic pulse from that boundary” (Velasco and Huhta 2005), see Figure 2. IMPORTANT: Pitch and roll variations of probe are taken into account settling of the tripod frame causes biased sediment levels

Conclusions:

ADV altimetry reveals depositional and erosional events Storms (significant wave height > 2m) + spring tide (doy 1-9)

Delivery of sediment in suspension

Bed level accretion during neap tide = Long-term occurrence of fluid mud (doy 9-22)

Erosion as result of remobilization during accelerating tidal flow (doy 19-22) and/or storm passage

and so on … Observation of short-term accretional events during slack waters = Rapid siltation from saturated mud suspension conditions (Winterwerp et al 2001)

Entrainment during accelerating tide Observation of erosional events dependent on meteorological conditions

Ebb currents more erosive than flood (doy 10-11) for for westerly wind

Resuspension of all sediments remaining a ’harder’ sea bed surface (doy 23-28) for easterly winds

Figure 1

Figure 2

Figure 3 References: Fettweis, M. and Van den Eynde D. 2003. “The mud deposits and the high turbidity in the Belgian-Dutch coastal zone, southern bight of the North Sea.” Continental Shelf Research, 23, 669-691 Velasco D.W. and Huhta C.A. “Experimental verification of acoustic Doppler velocimeter (ADV) performance in fine-grained, high sediment concentration fluids.” SonTek/YSI report. 23 pgs. Winterwerp J.C., Uittenbogaard, R.E. and de Kok, J.M. 2001. “Rapid siltation from saturated mud suspensions.” Intercoh Conf. ’98. 22 pgs. This study is partly financed through a grant of the Flemish Institute to stimulate scientific-technological research in industry (IWT) and partly financed by the projects MOMO (Flemish Government) and QUEST4D (Belgian Science Policy)

APPENDIX 5      Fettweis M, Francken F, Van den Eynde D, Verwaest T, Janssens J, Van  Lancker  V.  2010.  Storm  influence  on  SPM  concentrations  in  a  coastal  turbidity  maximum  area  with  high  anthropogenic  impact  (southern  North Sea). Continental Shelf Research. doi:10.1016/j.csr.2010.05.001   

Continental Shelf Research ] (]]]]) ]]]–]]]

Contents lists available at ScienceDirect

Continental Shelf Research journal homepage: www.elsevier.com/locate/csr

Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea) Michael Fettweis a,n, Frederic Francken a, Dries Van den Eynde a, Toon Verwaest b, Job Janssens b, Vera Van Lancker a a b

Royal Belgian Institute for Natural Science (RBINS), Management Unit of the North Sea Mathematical Models (MUMM), Gulledelle 100, 1200 Brussels, Belgium Flanders Hydraulics Research, Berchemlei 115, 2140 Antwerp, Belgium

a r t i c l e in fo

abstract

Article history: Received 28 May 2009 Received in revised form 17 March 2010 Accepted 6 May 2010

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 conditions. 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 of 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. 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. & 2010 Published by Elsevier Ltd.

Keywords: Suspended particulate matter Storm influence Southern North Sea Anthropogenic impact

1. 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 complying with international agreements aiming at sustainable development. Tides, waves, medium-scale meteorological events, and biological processes affect SPM concentration in nearshore areas. Coastal beds are usually composed of heterogeneous particles often consisting of sand- and clay-sized material. The mutual interaction of cohesive and non-cohesive sediments during erosion influences SPM concentration (Le Hir et al., 2007). The SPM concentration is further controlled by the seasonal

n

Corresponding author. Tel.: + 32 2 7732132; fax: +32 2 7706972. E-mail address: [email protected] (M. Fettweis).

variability in the supply of fine-grained sediment, the remote or local availability of fine sediments, advective processes, erosion, deposition, and human activities (Velegrakis et al., 1997; Bass et al., 2002; Schoellhamer, 2002; Chang et al., 2007). The deepening of channels and construction of ports increases deposition of fine-grained sediments and has as consequence an increase of maintenance dredging and a local and instantaneous increase of SPM concentration in and around the disposal site (Truitt, 1988; Collins, 1990; Van den Eynde and Fettweis, 2006; Wu et al., 2006) followed by segregation between fine- and coarser-grained sediments on the disposal site (Du Four and Van Lancker, 2008). 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 as also wave heights. Waves have an important impact on cohesive sediment transport processes on continental shelves (e.g. Green et al., 1995; Cacchione et al., 1999; Traykovski et al., 2007;

0278-4343/$ - see front matter & 2010 Published by Elsevier Ltd. doi:10.1016/j.csr.2010.05.001

Please cite this article as: Fettweis, M., et al., Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea). Continental Shelf Research (2010), doi:10.1016/j.csr.2010.05.001

2

M. Fettweis et al. / Continental Shelf Research ] (]]]]) ]]]–]]]

Shi et al., 2008). 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, 2000). Fluidization or liquefaction of mud layers occurs, if the stress caused by wave pressure exceeds the yield stress of the sediments (Silva Jacinto and Le Hir, 2001). 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 investigation of storm influences on SPM concentration. The measuring locations are situated in a very energetic area that is among 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, 1990; Lafite et al., 2000). Fettweis et al. (2007) 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 (2007) 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 of sediments.

prominent sea bed feature in the area (Le Bot et al., 2005). 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 ( 73000 yr BP), modern mud (o100 yr BP) and freshly deposited mud (Fettweis et al., 2009). Generally, the freshly deposited mud occurs as thin fluffy layers or locally as gradually soft consolidated thicker packages ( 70.2–1 m). These thicker soft mud layers also consist 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 mediumconsolidated 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., 2005). 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 1997–2006), see Lauwaert et al. (2008). 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 consists of 70–85% of mud; from the harbours this amounts to more than 95%.

3. Materials and methods 3.1. Instrumentation and measuring sites

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 0 and 20 m below MLLWS (mean lowest low water spring), except in the mouth of the Westerschelde estuary where depths can reach more than 20 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 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 transport. The maximum current velocities are more than 1 m s  1. The winds are dominantly from the southwest and the highest waves occur during north-western winds (Fettweis and Van den Eynde, 2003). SPM forms a turbidity maximum between Oostende and the mouth of the Westerschelde. SPM concentration in the coastal zone varies between minimum 20–70 mg l  1 and maximum 100– 1000 mg l  1 during calm meteorological conditions at 3 m above the sea bed; lower values ( o10 mg l  1) occur in the offshore area (Fettweis et al., 2007). 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

Data collection was conducted between April 2005 and February 2007 using a tripod at two sites (Fig. 1). Both sites are characterised by the occurrence of near-bed Holocene mediumconsolidated mud, albeit covered with an ephemeral slightly muddy fine sand layer with a median grain size of about 170 mm. 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 and two SonTek Hydra systems for data storage and batteries (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 2005 and May 2006 at site MOW1. A Niskin bottle was closed every 20 min, thus resulting in about 40 samples per tidal cycle. Three sub samples were filtered on board of the vessel from each water sample, using preweighted filters (Whatman GF/C). After filtration, the filters were rinsed once with Milli-Q water ( 750 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 240) was assumed.

Please cite this article as: Fettweis, M., et al., Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea). Continental Shelf Research (2010), doi:10.1016/j.csr.2010.05.001

M. Fettweis et al. / Continental Shelf Research ] (]]]]) ]]]–]]]

3

Fig. 1. (A) Yearly and depth-averaged SPM concentration (mg l  1) in the southern North Sea, derived from 370 SeaWiFS images (1997–2002), see Fettweis et al. (2007a). (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 (1N) and longitude (1E).

Table 1 Oceanographic instruments mounted on the tripod and distance in m above bed (mab). The saturation concentration of the OBSs was 3.3 g l  1. Survey no. 2005/29 2006/23 2006/27

0.3 0.2 0.2

ADV

ADP

OBS1

OBS2

2.2 2.3 2.3

0.3 0.2 0.2

1.9 2.2 2.2

0.8 0.8 0.8

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 report on the possibility to estimate shear stress from the second moment (turbulence) statistics (Pope et al., 2006; Verney et al., 2007; Andersen et al., 2007). 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 t ¼C TKE, where C¼19 was adopted as proposed by Stapleton and

Huntley (1995) and Thompson et al. (2003). This linear relationship will however fail in the presence of waves. The inertialdissipation method uses the spectrum of velocity components and allows to apply a correction for the advection by waves (Trowbridge and Elgar, 2001; Sherwood et al., 2006). In this case the vertical component was used as it is least contaminated by waves. ADV Ocean data were discarded when the signal to noise ratio dropped below 15 dB and the correlation coefficient was lower than 0.8. The data were then transformed into a power density spectrum using a fast Fourier transform with a Hanning window and the spectral density, Eww, was normalized such that the integral over the spectrum yielded the variance over the burst. Afterwards it was transformed to Ewwo5/3(2p)  1, with o the radial frequency, as the turbulent dissipation, e, scales with Ewwo5/3(2p)  1 in the inertial subrange, defined as the range between 1 and 2.5 Hz (Trowbridge and Elgar, 2001). By calculating the mean over this range an estimate of Ewwo5/3(2p)  1 was obtained and e calculated using 

e ¼ Eww o5=3



12 ajU j2=3 55

3=2

Please cite this article as: Fettweis, M., et al., Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea). Continental Shelf Research (2010), doi:10.1016/j.csr.2010.05.001

4

M. Fettweis et al. / Continental Shelf Research ] (]]]]) ]]]–]]]

Table 2 Tripod deployments at MOW1 and Blankenberge. Survey no.

Location

Start

End date (dd/mm/yyyy hh:mm)

Duration (days)

2005/29 2006/23 2006/27

MOW1 Blankenberge Blankenberge

22/11/2005 08:11 08/11/2006 14:29 18/12/2006 10:44

05/12/2005 09:03 27/11/2006 09:04 07/02/2007 13:17

13.04 18.83 50.11

where a is the empirical Kolmogorov constant and U is the burst averaged velocity. To correct for the presence of waves the model uses a function I to correct e, see Trowbridge and Elgar (2001) and Sherwood et al. (2006). The shear stress at elevation z can then be obtained using t ¼ r(ekz)2/3, where k is the Von Ka´rma´n constant and r the water density. 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). Box cores were subsampled using cylindrical perspex tubes (diameter 13.5 cm) allowing to retrieve relatively undisturbed mud samples of the first 40 cm (on average) of the sea bed. Using a gamma-ray densitometer, bulk density profiles of these samples were determined in a non-destructive way. The measurements were performed at the University of Stuttgart using the SETEG-flume (Kern et al., 1999; Witt and Westrich, 2003): 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 so-called SEDCIA-system, enabling to measure the erosion rate. This system consists of a camera which observes the time-dependent 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). OPTOS-BCS covers an area between 511N and 51.921N in latitude and between 2.081E and 4.21E in longitude. The horizontal resolution is 0.240 (longitude) and 0.140 (latitude), corresponding both to a grid size of about 250 m. Boundary conditions of water elevation and depth-averaged 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 www.mumm.ac.be). Pison and Ozer (2003) have described the validation of the current velocities of OPTOS-BCS using ADCP measurements. The bottom shear stress for currents alone is calculated using the calculated current velocity in the lowest layer of the model and using a bottom roughness of 0.84 cm. The bottom 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 2005 are presented in Fig. 2. Spring tide occurred around December 1 (day 10), 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  1 (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 quarter-diurnal (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 (days 3–4). The SPM concentration measured by OBS1 (0.3 m above bottom, mab) increases to 1–3 g l  1, whereas at 2 mab (OBS2) the SPM concentrations remain less than 0.5 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, 2003). After the WNW storm the SPM concentration decreased in magnitude, however both OBSs 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 (0.5–1 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 OBSs, 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, 2006 and February 7, 2007. The data from 7 to 18 November 2006 and 24 December 2006–8 January 2007 are presented in Figs. 3 and 4. Remark that no bottom shear stresses are reported for the second period (Fig. 4) as the quality of the ADV data was insufficient. During both periods different storms have occurred. On 12–13 November, a NW storm generated significant wave heights of about 2.7 m. The secondhalf of December was characterized by low wind speeds from mainly W-SW. On December 31, wave heights of nearly 3 m were

Please cite this article as: Fettweis, M., et al., Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea). Continental Shelf Research (2010), doi:10.1016/j.csr.2010.05.001

5

Depth below surface (m)

16

4 depth

wave height

3

11

2 1

6

0

2

4

6

8

10

12

0 14

0

2

4

6

8

10

12

14

Wave height (m)

M. Fettweis et al. / Continental Shelf Research ] (]]]]) ]]]–]]]

Velocity (m/s)

1 0.5

Shear stress (Pa)

30

SPM conc (mg/l)

0

3000

measurements

model results

20 10 0 0

2

4

6

8

10

12

14

4

6 8 Time (days)

10

12

14

SPM1

2000

SPM2

1000 0

0

2

Depth below surface (m)

14

4 depth

wave height

3

9

2 1

4

0 0

2

4

6

8

10

12

14

0

2

4

6

8

10

12

14

0

2

4

6

8

10

12

14

Wave height (m)

Fig. 2. MOW1 site, tripod measurements of 22 November–5 December 2005 (survey 2005/29). 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 from the hydrodynamic model and SPM concentration at 0.3 mab (SPM1) and 1.9 mab (SPM2).

0.5

Shear stress (Pa)

0 40

SPM conc (mg/l)

Velocity (m/s)

1

4000

20 0

SPM1

SPM2

2000 0

0

2

4

6 8 Time (days)

10

12

14

Fig. 3. Blankenberge site, tripod measurements of 7–20 November 2006 (survey 2006/23). 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 0.2 mab (SPM1) and 2.2 mab (SPM2).

Please cite this article as: Fettweis, M., et al., Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea). Continental Shelf Research (2010), doi:10.1016/j.csr.2010.05.001

M. Fettweis et al. / Continental Shelf Research ] (]]]]) ]]]–]]]

Depth below surface (m)

14

Velocity (m/s)

wave height

3

9

2 1

4

0

5

10

0 15

0

5

10

15

10

15

1 0.5 0

SPM conc (mg/l)

4 depth

Wave height (m)

6

4000 SPM1

SPM2

2000 0

0

5 Time (days)

Fig. 4. Blankenberge site, tripod measurements of 27 December 2006–10 January 2007 (survey 2006/27). From up to down: depth below water surface (m) and significant wave heights; ADV current velocity (m s  1) and mean particle size (mm); shear stress (Pa) derived from the ADV; and SPM concentration at 0.2 mab (SPM1) and 2.2 mab (SPM2).

registered (Fig. 4). The beginning of January 2007 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 2 days after by OBS2 (Fig. 3). The OBS1 data are characterised by very high minima in SPM concentrations ( 40.8 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 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. Considerable variations in SPM concentrations exist during a tidal cycle: maximum concentrations were sometimes up to 50 times higher than the minimum concentrations. 3. The very high SPM concentrations measured near the bed are related to storm periods; our data suggest that HCMS occur near the bottom in the coastal turbidity maximum of the Belgian–Dutch nearshore zone. 4. 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, r, and the critical shear stress for erosion, tce, 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 tce can be observed in muddy sediments. The cores containing Holocene mud are both characterised by a 5–10 cm 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 tce ranging from 0.5 to 2.5 Pa. The Holocene mud layers are very strong, with a tce 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 tce 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 tce (0.5–1 Pa) and r (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 tce and r exist. The presence of shells has a negative effect on the stability of muddy sediments. Sandy fractions always exhibit low to medium stability (maximum tce:  4 Pa).

4.3. Hydrodynamic model results The numerical model has been used to simulate the bottom shear stress during autumn 2005 (22 November–5 December). In Fig. 2 the model results are compared to the shear stresses

Please cite this article as: Fettweis, M., et al., Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea). Continental Shelf Research (2010), doi:10.1016/j.csr.2010.05.001

M. Fettweis et al. / Continental Shelf Research ] (]]]]) ]]]–]]]

τce (Pa) 0

0

2

4

6

8

7

τce (Pa) 10

12

14

0

0

2

4

6

8

τce (Pa) 10

12

14

0

10

10

0

2

4

6

8

10

12

14

10 Tau-crit

20

30

depth below bed (cm)

depth below bed (cm)

depth below bed [cm]

Bulk density

20

30

Tau-crit bulk density

50 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 bulk density [g cm-3]

30

40

40

40

20

tau-crit bulk density

50 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 bulk density [g cm-3]

50 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 bulk density [g cm-3]

Fig. 5. Depth profile of critical erosion shear stress (tce) and bulk density of box core: (A) 7 cm of soft mud above 40 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 40 cm of alternating medium-consolidated mud and thin muddy sand layers with shells (Holocene). (C) 40 cm of freshly deposited mud from the navigation channel (Pas van het Zand) on top of mediumconsolidated mud (Holocene). The sandy and shelly layers have a lower tce and a higher bulk density.

derived from ADV measurements. 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 12 Pa. Without meteorological effects the maxima, occurring during spring tide in the channel towards the Westerschelde (Scheur), reaches about 2.5 Pa. The difference between model results and shear stress results from ADV can be ascribed to inaccuracies in calculating near-bed shear stress from ADV data and to the fact that the bottom roughness in model is not well known and remains a calibration parameter.

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, 2000). 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, 2006). 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., 2001), as no high-load rivers exist in the area (Fettweis et al., 2007). This SPM flux amounts to about 34  106 t year  1, from which about 50% is transported along the continental side (i.e. France–Belgian–Dutch nearshore area), see Fettweis et al. (2007). 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 range 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 1997–2004 in the turbidity maximum area has been calculated as about 1  106 t. Variations of the order of 30% occur between spring and neap tides, as also of the order of 40–60% in-between seasons. Based on the tripod measurements, the SPM mass during storm conditions has been estimated as 3–5  106 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 2007, SW winds prevailed resulting in advection of high salinity and low turbid water from the English

Please cite this article as: Fettweis, M., et al., Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea). Continental Shelf Research (2010), doi:10.1016/j.csr.2010.05.001

8

M. Fettweis et al. / Continental Shelf Research ] (]]]]) ]]]–]]]

Fig. 6. Hydrodynamic model results of maximum bottom shear stress (Pa) during a spring tide without wind (A) and during the November 2005 storm (B). Coordinates are in latitude (1N) and longitude (1E).

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. Our data show that high SPM concentrations are often more closely related to advection (Velegrakis et al., 1997; Blewett and Huntley, 1998) rather than instantaneous bed shear stress (Stanev et al., 2009). This confirms the idea that the Belgian coastal area can be seen as a congestion in the residual SPM transport of the southern North Sea rather than an important source of sediments (Fettweis and Van den Eynde, 2003).

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 41500 kg m  3). The area where these mud deposits occur in the first meter of the sea bed is  744 km2 (Fettweis et al., 2009). 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 Figs. 2–4 and 6). The sediments of the mud fields can thus not be eroded under calm meteorological conditions. Near bed shear stresses derived from the ADV data amount up to 40 Pa during the storm of November 2005 at MOW1 (Fig. 2) and of December 2006 (Fig. 3). 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

Please cite this article as: Fettweis, M., et al., Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea). Continental Shelf Research (2010), doi:10.1016/j.csr.2010.05.001

M. Fettweis et al. / Continental Shelf Research ] (]]]]) ]]]–]]]

Tidal range (m)

5

SPM1 (0.2 mab) SPM2 (2 mab)

4

3

2 0

500

1000

1500

2000

SPM concentration (mg/l) Fig. 7. Tidally-averaged SPM concentration as a function of tidal range at the Blankenberge site (November 2006–February 2007).

34

Salinity

32

31

0

200

400

600

800

1000

SPM concentration (mg/l) 35

Salinity

critical erosion shear stress of 10 Pa and an erosion rate of 0.5 g m  2 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 0.6  106 t day  1 could have been eroded by horizontal tidally induced shear stress 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, 2000; Silva Jacinto and Le Hir, 2001). 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 (2001) 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., 2009). 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

33

30

9

Other important erosion mechanisms are due to the mutual interaction of cohesive and non-cohesive sediments (Le Hir et al., 2007): 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. (2004) 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.

34

5.4. Freshly deposited mud as SPM source

33

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 tce of o1 up to  4 Pa, will be fully resuspended during 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

32 31 30 29

0

200

400 600 SPM concentration (mg/l)

800

1000

Fig. 8. SPM concentration from OBS2 (2 mab) as a function of salinity. The figures show the effect of alongshore advection of high saline and low turbid water during January 2007 at the Blankenberge site: (A) Negative correlation between salinity and SPM concentration during prevailing SW winds (5–7 January). (B) Two successive periods are shown; the first is characterised by high salinity and low SPM concentration (16–22 January) and the second by low salinity and high SPM concentration (22–31 January). The sudden decrease in salinity was induced by a change in wind direction from SW to SE.

Please cite this article as: Fettweis, M., et al., Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea). Continental Shelf Research (2010), doi:10.1016/j.csr.2010.05.001

10

M. Fettweis et al. / Continental Shelf Research ] (]]]]) ]]]–]]]

channels, where mud is dredged, equals to 715 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 30 km2. The bulk density of these sediments amounts to 1200–1400 kg m  3. If we assume a thickness of 20 cm very soft mud in these areas, then the total mass of mud available for resuspension equals 1.8  3.6  106 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 2005 and winter 2007, 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 fine-grained 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.

Acknowledgements This study was partly funded by the by the Belgian Science Policy within the framework of the QUEST4D project (SD/NS/06A) 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, 2007 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). We also are grateful to G. Voulgaris, for helping us out with the inertial-dissipation method. References Andersen, T.J., Fredsoe, J., Pejrup, M., 2007. In-situ estimation of erosion and deposition thresholds by acoustic Doppler velocimeter (ADV). Estuarine, Coastal and Shelf Science 75, 327–336. doi:10.1016/j.ecss.2007.04.039. Ariathurai, C.R., 1974. A Finite Element Model for Sediment Transport in Estuaries. University of California, Davis.

Bass, S.J., Aldridge, J.N., McCave, I.N., Vincent, C.E., 2002. Phase relationships between fine sediment suspensions and tidal currents in coastal seas. Journal of Geophysical Research 107 (C10), 1–14. Blewett, J., Huntley, D.A., 1998. Measurement of suspended sediment transport processes in shallow water off the Holderness Coast, UK. Marine Pollution Bulletin 37, 134–143. Cacchione, D.A., Wiberg, P.L., Lynch, J.F., Irish, J.D., Traykovski, P., 1999. Estimates of suspended-sediment flux and bedform activity on the inner portion of the Eel continental shelf. Marine Geology 154, 83–97. Collins, M., 1990. The behaviour of cohesive and non-cohesive sediments. In: Alzieu, C., Galenne, B. (Eds.), Proceedings of the International Seminar on the Environmental Aspects of Dredging Activities. Port Autonome de Nantes Saint-Nazaire, Nantes, pp. 15–32. ¨ A., 2007. Distinction between sortable silts Chang, T.S., Flemming, B., Bartholoma, and aggregated particles in muddy intertidal sediments of the East Frisian Wadden Sea, southern North Sea. Sedimentary Geology 202, 453–463. doi:10.1016/j.sedgeo.2007.03.009. de Wit, P.J., Kranenburg, C., 1997. The wave-induced liquefaction of cohesive sediment beds. Estuarine, Coastal and Shelf Science 45, 261–271. Du Four, I., Van Lancker, V., 2008. Changes of sedimentological patterns and morphological features due to the disposal of dredge spoil and the regeneration after cessation of the disposal activities. Marine Geology 25, 15–29. doi:10.1016/j.margeo.2008.04.011. Eisma, D., 1981. Supply and deposition of suspended matter in the North Sea. In: Nio, D.D., Shuttenhelm, R.T.E., van Weering, T.C.E. (Eds.), Holocene Marine Sedimentation in the North Sea Basin, vol. 5. IAS Special Publication, Blackwell Scientific, pp. 415–428. Fettweis, M., Van den Eynde, D., 2003. The mud deposits and the high turbidity in the Belgian–Dutch coastal zone. Southern bight of the North Sea. Continental Shelf Research 23, 669–691. doi:10.1016/S0278-4343(03)00027-X. Fettweis, M., Nechad, B., Van den Eynde, D., 2007. 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, 1568–1583. doi:10.1016/j.csr.2007.01.017. Fettweis, M., Houziaux, J.-S., Du Four, I., Van Lancker, V., Baeteman, C., Mathys, M., Van den Eynde, D., Francken, F., Wartel, S., 2009. Long-term influence of maritime access works on the distribution of cohesive sediments: analysis of historical and recent data from the Belgian nearshore area (southern North Sea). Geo-Marine Letters 29 (5), 321–330. doi:10.1007/s00367-009-0161-7. Gerritsen, H., Boon, J.G., van der Kaaij, T., Vos, R.J., 2001. Integrated modelling of suspended matter in the North Sea. Estuarine, Coastal and Shelf Science 53, 581–594. doi:10.1006/ecss.2000.0633. Green, M.O., Vincent, C.E., McCave, I.N., Dickson, R.R., Rees, J.M., Pearson, N.D., 1995. Storm sediment transport: observations from the British North Sea shelf. Continental Shelf Research 15, 889–912. ¨ Kern, U., Haag, I., Schurlein, V., Holzwarth, M., Westrich, B., 1999. Ein ¨ ¨ ¨ von Stromungskanal zur Ermittlung der tiefenabhangigen Erosionsstablitat ¨ Gewassersedimenten. Wasserwirtschaft DWA 89 (2), 72–77. Koutitas, C.G., 1988. Mathematical Models in Coastal Engineering. Pentech Press, London, UK 156 pp. Lafite, R., Shimwell, S., Grochowski, N., Dupont, J.-P., Nash, L., Salomon, J.-C., Cabioch, L., Collins, M., Gao, S., 2000. Suspended particulate matter fluxes through the Strait of Dover, English Channel: observations and modelling. Oceanologica Acta 23, 687–700. 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., 2008. Synthesis report on the effects of dredged material disposal on the marine environment. BMM, ILVO, AK & AMT rapport, BL/2008/01, 109 pp. Le Bot, S., Van Lancker, V., Deleu, S., de Batist, M., Henriet, J.-P., Haegeman, W., 2005. 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, 147–160. Le Hir, P., Monbet, Y., Orvain, F., 2007. Sediment erodability in sediment transport modelling: can we account for biota effects? Continental Shelf Research 27, 1116–1142. doi:10.1016/j.csr.2005.11.016. Li, Y., Mehta, A.J., 2000. Fluid mud in the wave-dominated environment revisited. In: McAnally, W.H., Mehta, A.J. (Eds.), Coastal and Estuarine Fine Sediment Dynamics, Proceedings in Marine Science, vol. 3, pp. 79–93. Luyten, P.J., Jones, J.E., Proctor, R., Tabor, A., Tett, P., Wild-Allen, K., 1999. COHERENS, a coupled hydrodynamical–ecological model for regional and shelf seas: user documentation. MUMM Report, Brussels, Belgium, 911 pp. /http:// www.mumm.ac.be/coherensS. Maa, P.-Y., Mehta, A.J., 1987. Mud erosion by waves: a laboratory study. Continental Shelf Research 7, 1269–1284. Pison, V., Ozer, J., 2003. Operational products and services for the Belgian coastal waters. In: Dahlin, H., Flemming, N.C., Nittis, K., Petersson, S.E. (Eds.), Building the European Capacity in Operational Modeling, Proceedings of the Third International Conference on EuroGOOS, Elsevier Oceanography Series, vol. 69, pp. 503–509. Pope, N.D., Widdows, J., Brinsley, M.D., 2006. Estimation of bed shear stress using the turbulent kinetic energy approach—a comparison of annular flume and field data. Continental Shelf Research 26, 959–970. doi:10.1016/j.csr.2006.02.010. Schoellhamer, D.A., 2002. Variability of suspended-sediment concentration at tidal to annual time scales in San Francisco Bay, USA. Continental Shelf Research 22, 1857–1866.

Please cite this article as: Fettweis, M., et al., Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea). Continental Shelf Research (2010), doi:10.1016/j.csr.2010.05.001

M. Fettweis et al. / Continental Shelf Research ] (]]]]) ]]]–]]]

Sherwood, C., Lacy, J., Voulgaris, G., 2006. Shear velocity estimates on the inner shelf off Grays Harbor, Washington, USA. Continental Shelf Research 26, 1995–2018. Shi, J.Z., Gu, W.-J., Wang, D.-Z., 2008. Wind wave-forced fine sediment erosion during slack water periods in Hangzhou Bay, China. Environmental Geology 55, 629–638. doi:10.1007/s00254-007-1013-2. Silva Jacinto, R., Le Hir, P., 2001. Response of stratified muddy beds to water waves. In: McAnally, W.H., Mehta, A.J. (Eds.), Coastal and Estuarine Fine Sediment Processes, Proceedings in Marine Science, vol. 3, pp. 95–108. ¨ Stanev, E.V., Dobrynin, M., Pleskachevsky, A., Grayek, S., Gunther, H., 2009. Bed shear stress in the southern North Sea as an important driver for suspended sediment dynamics. Ocean Dynamics 59, 183–194. doi:10.1007/ s10236-008-0159-0. Stapleton, R.L., Huntley, D.A., 1995. Seabed stress determination using the inertia dissipation method and the turbulent kinetic energy method. Earth Surface Processes and Landforms 20, 807–815. Streeter, V.L., 1996. Fluid Mechanics. McGraw-Hill Book Company, New York 705 pp. Thompson, C.E.L., Amos, C.L., Jones, T.E.R., Chaplin, J., 2003. The manifestation of fluid-transmitted bed shear stress in a smooth annular flume—a comparison of methods. Journal of Coastal Research 19, 1094–1103. Trowbridge, J., Elgar, S., 2001. Turbulence measurements in the surf zone. Journal of Physical Oceanography 31 (8), 2403–2417. Truitt, C.L., 1988. Dredged material behaviour during open water disposal. Journal of Coastal Research 4, 389–397. Traykovski, P., Wiberg, P.L., Geyer, W.R., 2007. 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, 375–399. doi:10.1016/j.csr.2005.07.008. Van Alphen, J.S.L.J., 1990. A mud balance for Belgian–Dutch coastal waters between 1969 and 1986. Netherlands Journal of Sea Research 25, 19–30.

11

Van den Eynde, D., Fettweis, M., 2006. Modelling of fine-grained sediment transport and dredging material dumpings at the Belgian continental shelf. Journal of Coastal Research SI39, 1564–1569. van Ledden, M., van Kesteren, W.G.M., Winterwerp, J., 2004. A conceptual framework for the erosion behaviour of sand–mud mixtures. Continental Shelf Research 24, 1–11. doi:10.1016/j.csr.2003.09.002. van Kessel, T., Kranenburg, C., 1998. Wave-induced liquefaction and flow of subaqueous mud layers. Coastal Engineering 34, 109–127. Verney, R., Deloffre, J., Brun-Cottan, J.C., Lafite, R., 2007. The effect of wave-induced turbulence on intertidal mudflats: impact of boat traffic and wind. Continental Shelf Research 27, 594–612. doi:10.1016/j.csr.2006.10.005. Verwaest, T., 2007. Using simple semi-empirical models for integrated assessment of scenarios for a navigation channel. The case of the port of Ostend, Belgium. Coastal Sediments 2007, New Orleans, USA. Velegrakis, A.F., Gao, S., Lafite, R., Dupont, J.P., Huault, M.F., Nash, L.A., Collins, M.B., 1997. Resuspension and advection processes affecting suspended particulate matter concentrations in the central English Channel. Journal of Sea Research 38, 17–34. Williamson, H., Torfs, H., 1996. Erosion of mud/sand mixtures. Coastal Engineering 29, 1–25. Winterwerp, J., 1999. On the dynamics of high-concentrated mud suspensions. Ph.D. Thesis, TU Delft, The Netherlands, 172 pp. Witt, O., Westrich, B., 2003. Quantification of erosion rates for undisturbed contaminated cohesive sediment cores by image analysis. Hydrobiologia 494, 271–276. Wright, L.D., Friedrichs, C.T., 2006. Gravity-driven sediment transport on continental shelves: a status report. Continental Shelf Research 26, 2092–2107. doi:10.1016/j.csr.2006.07.008. Wu, J., Liu, J.T., Shen, H., Zhang, S., 2006. Dispersion of disposed dredged slurry in the meso-tidal Changjiang (Yangtze River) Estuary. Estuarine, Coastal and Shelf Science 70, 663–672. doi:10.1016/j.ecss.2006.07.013.

Please cite this article as: Fettweis, M., et al., Storm influence on SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North Sea). Continental Shelf Research (2010), doi:10.1016/j.csr.2010.05.001

APPENDIX 6      Fettweis  M,  Baeye  M, Francken  F,  Lauwaert  B,  Van  den  Eynde  D,  Van  Lancker  V,  Martens C, Michielsen T. Monitoring the effects of disposal of fine sediments from  maintenance  dredging  on  SPM  concentration.  Marine  Pollution  Bulletin  (submitted) 

 

Monitoring the effects of disposal of fine sediments from maintenance  dredging on SPM concentration    Michael Fettweis1, Matthias Baeye2, Frederic Francken1, Brigitte Lauwaert1, Dries Van den Eynde1, Vera Van  Lancker1, Chantal Martens3, Tinne Michielsen3    1

Royal Belgian Institute for Natural Science (RBINS), Management Unit of the North Sea Mathematical Models 

(MUMM), Gulledelle 100, 1200 Brussels, Belgium  2

Ghent University, Renard Centre of Marine Geology, Krijgslaan 281 (S8), 9000 Gent, Belgium 

3

Ministry of Public Works and Mobility, Maritime Access Division, Tavernierkaai 3, 2000 Antwerp, Belgium 

  M. Fettweis (corresponding author) (e‐mail: [email protected], Tel.: +32‐2‐7732132, Fax: +32‐2‐ 7706972),  

 

Abstract  Harbour authorities worldwide are obliged to dredge their major shipping channels, and subsequently to dispose  the dredged spoil offshore. In May 2009, an experimental dredging study was carried out in the port of  Zeebrugge to investigate the efficiency of pumping fluid mud using a cutter dredger. Before, during and after the  experiment monitoring of SPM concentration using OBS and bed level variation using ADV altimetry was carried  out at a location about 5 km west of the disposal site. The data allowed to characterise the natural turbidity  regime and to assess the effects of the disposal operations on the sediment dynamics.   The median SPM concentration near the bed was more than 2 times higher during the field experiment. The time  lag between high wave heights and high SPM concentration suggested that the SPM was advected towards the  measuring location rather than eroded locally. The disposed material was mainly transported in the benthic layer  and resulted in a long‐term increase of SPM concentration near the bed and formation of fluid mud layers during  the neap tidal conditions. During the other periods fluffy layer were associated with quarter‐diurnal tidal  variations and with storm conditions.  

Keywords  dredging, disposal of dredged material, fluid mud, SPM, monitoring, outfall 

 

1

1 Introduction  Dredging and disposal of sediments is generally considered to be harmful to the environment and to aquatic life  (Lohrer and Wetz, 2003; Orpin et al., 2004; Ware et al., 2010). Numerous morphological, and sedimentological  effects (increased turbidity, enhanced sedimentation, changed distribution of sediments) result from dredging  and disposal works in nearshore areas (Nichols, 1988; Bray et al., 1996; O’Connor, 1999; Fredette and French,  2004; Van den Eynde, 2004; Wu et al., 2006; Bolam et al., 2006; Du Four and Van Lancker, 2008; OSPAR, 2008;  Okada et al., 2009; Siegel et al., 2009). The Water Framework Directive (2000/60/EC) and recently adopted EU  Marine Strategy Framework Directive (2008/56/EC) (see e.g. Borja, 2005; Devlin et al., 2007) requires the  monitoring of various environmental parameters for the establishment of present‐day water quality reference  conditions; specifies that there must be no temporal deterioration in chemical and biological status for many  water bodies; and identifies material in suspension as one of the main pollutants. The influence of large  engineering works, such as dredging and disposal, needs therefore to be understood to ensure cost‐effective  operations at sea, to better gauge the human footprint, and to develop environmental policies aiming at a more  sustainable management of the marine environment.   Many access channels and harbours suffer from sedimentation of fine sediments and formation of fluid  mud layers (Fettweis and Sas, 1999; Verlaan and Spanhoff, 2000; Winterwerp, 2005; PIANC, 2009; Van Maren et  al., 2009; De Nijs et al., 2009). Fluid mud is a high concentration aqueous suspension of fine‐grained sediment in  which settling is substantially hindered. Fluid mud consists of water, clay‐sized particles, and organic materials; it  displays a variety of rheological behaviours ranging from elastic to pseudo‐plastic (Mac Anally et al. 2007). Ship  manoeuvrability studies aided in the redefinition of the level of dredging required, which is often put at a density  of 1200 kg m‐3 (Delforterie et al., 2005; PIANC, 2008). Nevertheless, to keep navigation safe and economic,  extensive maintenance dredging is still needed. The efficiency of dredging operations in fluid mud is limited as  consolidation is hampered by ship movements, which generate significant flows and resuspension (PIANC, 2008)  or as accumulation may be so rapidly that it exceeds the capacity of available dredgers (Mac Anally et al., 2007).  Conventional dredging methods with trailer suction dredgers, and disposal of the dredged material at designated  locations, are inefficient for fluid mud and incur substantial costs.   In May 2009, an experimental study was carried out in the port of Zeebrugge to investigate whether  pumping using a cutter dredger could reduce the thickness of the mud layer with a density lower than 1200 kg m‐ 3

. An automatic method to intercept and pump away fluid mud using stationary pumping system was evaluated 

by Berlamont (1989). He concluded that mud from Zeebrugge could be pumped to a density of 1150 kg m‐3,  beyond which the radius of influence of the pump rapidly became very small. A similar approach was adopted  here, except that no stationary system was used, but a cutter dredger, which continuously dredged for periods of  a few days up to a week at a fixed location and a fixed depth corresponding to a density smaller than 1200 kg m‐3  before being moved to another location. The advantages of a stationary cutter dredger are the high pumping  capacities and the efficient removal of the dredged matter via floating pipelines. Although the evaluation of the  project was negative in terms of efficiently reducing the thickness of the fluid mud layer (Lauwaert et al., 2009),  insight was gained on the impact of an almost continuously disposal of fine‐grained dredged material on the  increase of suspended particulate matter (SPM) concentration close to a shoreline. The monitoring included in 

2

situ measurements carried out before, during and after the experiment. The specific aims of this study were to  characterise using statistical analysis the natural turbidity regime; to assess the effects of the disposal operations  on the sediment dynamics, especially on the formation of fluid mud layers; and to evaluate possible impacts of  disposal operations on the sediment composition.  

2. Methods  2.1 Study area  The study area is situated in the southern North Sea (Fig. 1). Water depths vary between 0–20 m below MLLWS  (mean lowest low water spring). Tides are semi‐diurnal with a mean tidal range at Zeebrugge of 4.3 m at spring  and 2.8 m at neap tide. The tidal current ellipses are elongated in the nearshore area and become gradually more  semi‐circular further offshore. Maximum current velocities are higher and minima lower in the nearshore area  than further offshore. The current velocities near Zeebrugge (nearshore) vary from 0.2–1.5 m s–1 during spring  tide and 0.2–0.6 m s–1 during neap tide; more offshore they range between 0.2‐0.6 m s‐1 during spring tide and  0.1‐0.3 m s‐1 during neap tide (see operational model results at www.mumm.ac.be). Flood currents are directed  towards the Northeast and ebb currents towards the Southwest. Winds blow predominantly from the southwest  and the highest waves occur during northwesterly winds. SPM forms a turbidity maximum between Oostende  and the mouth of the Westerschelde (Fig. 1). The strong tidal currents and the low fresh water discharge of the  Scheldt (yearly average is 100 m3 s–1) result in a well‐mixed water column with very limited salinity and  temperature stratification. Measurements indicate variations in SPM concentration in the nearshore area of 20– 70 mg l–1; reaching 100 to 3000 mg l–1 near the bed; lower values (<10 mg l–1) occur in the offshore (Fettweis et  al., 2010). The most important sources of SPM are the French rivers discharging into the English Channel, coastal  erosion of the Cretaceous cliffs at Cap Gris‐Nez and Cap Blanc‐Nez (France) and the erosion of nearshore  Holocene mud deposits (Fettweis et al., 2007).  On average 4.46×106 ton dry matter (tdm) is dredged annually in the port of Zeebrugge to maintain  navigation depth; this represents about 60% of the total amount of maintenance dredging in the Belgian  nearshore area (Lauwaert et al., 2009). The dredged matter consists of muddy sediments and is disposed on the  disposal sites S1 (47%), Zeebrugge Oost (44%) and S2 (9%). The sedimentation rate in the outer port of  Zeebrugge is on average about 1.7 tdm m‐2 per year. In 2007 and 2008, respectively, 0.7×106 tdm and 0.3×106  tdm of sediments were dredged in the Albert II dock (Fig. 1).  

2.2 Field experiment and monitoring  The experimental dredging study took place in the Albert II dock situated in the outer port of Zeebrugge between  5 May 12:00 and 2 June 2009 07:00. During the field experiment, the cutter suction dredger was operated at  different locations in the dock. The cutter head was positioned in mud with a density of 1200 kg m‐3. The dredged  matter was pumped using floating pipelines over the harbour breakwater into the sea at a location closer to the  shore and the port compared to the existing disposal sites (see Fig. 1). Pumping capacity and density of the  pumped matter was registered by the dredger.   The monitoring during the disposal experiment consist of near field and far field measurements. The near‐ field measurements were carried out inside and outside the dock. They consisted of daily bathymetrical and 

3

weekly mud density surveys in the dock and outside the port near the outfall. Bathymetrical surveys were  performed daily with 33/210 kHz echo sound measurements along fixed transects and of weekly with  multibeam. The far field monitoring was carried out at a fixed location near Blankenberge (51.33°N 3.11°E)  situated about 1 km offshore and 5 km west of the disposal site (Fig. 1) using a tripod which was developed for  collecting time‐series (up to 50 days) of SPM concentration and current velocity. The water depth at the site is  about 6 m MLLWS. A SonTek 5 MHz Acoustic Doppler Velocimeter (ADV) Ocean, a Sea‐Bird SBE37 CT system and  two OBS sensors (OBS1 at about 0.2 and OBS2 at about 2 m above bottom (hereafter referred to as mab) were  mounted on the frame. The ADV was attached at 0.35 mab such that the measuring volume was situated at 0.2  mab. The altimetry of the ADV was used to detect variation in bed level due to the occurrence of fluid mud  layers. Decreasing distance between probe and bed boundary correspond with the presence of high  concentration mud suspensions acting as an acoustic reflector. The tripod has been deployed for 240 days during  6 measuring periods before, during and after the experiment; the first two were situated in autumn ‐ winter  2006‐2007 (November 2006 – February 2007), the next three in winter and spring 2008 (January – June 2008)  and the last one during and after the field study in spring 2009 (May‐June 2009), see Table 1. 17% of the data  have been collected during or shortly after the field experiment. Respectively 15%, 38% and 47% of the data at  the measuring location are situated in autumn, winter and spring. As the SPM concentration is highest during  autumn and winter and lowest during spring and summer (Fettweis et al., 2007), the data are well distributed  over the high and low SPM concentration periods. Significant wave heights at a coastal station (Bol van Heist, see  Fig. 1) for the period 2006‐2009 have been used to characterise meteorological and sea state conditions. 

2.3 Statistical analysis  The SPM concentration variability at Blankenberge is the result of various processes related to tides, storms and  seasonal changes. In order to identify the effects from dredged material disposal, statistical methods are used.  SPM concentration is log‐normally distributed (Fettweis and Nechad, 2010). By using frequency distributions of  different data sets, we can calculate using standard statistic tests if two distributions are drawn from the same  distribution function. If the data collected during different sampling periods have similar log‐normal  distributions, geometric means and standard deviations, then we could conclude that ‐ within the range of  natural variability and measuring uncertainties ‐ similar sub‐samples from the whole population are obtained and  no changes have occurred due to external disturbances, such as dredging impacts. If disposal of dredged material  over the western breakwater of the port has an impact on SPM concentration then this should be detected in  variations of statistical parameters during the data collected during the field experiment.  

3. Results  3.1 Bathymetrical and density measurements  The dredging effort caused rapid (order of hours) formation of cone formed craters centred on the cutter head  location (Fig. 2), which disappeared again after relocation of the cutter. Influx of sediment related to shipping  activities and spring tide caused at some occasions the filling‐up of the crater during a short period. The dredging  caused a local deepening of the 1100 kg m‐3 density surface, however the influence remained local and did not  significantly changed the depth of the fluid mud density field in the dock. The pumping capacity was 3000 m3 h‐1,  resulting thus –using the average density of the pumped matter (including salt and sediment) of 1.055 t m‐3 –  in 

4

60×103 tdm of sediments that have been disposed during the duration of the experiment. As the density  recordings were inaccurate, this should be seen as an estimate.  

3.2 Significant wave height  The median significant wave height (Hs) during the period 2006‐2009 is 0.55 m. Lower values occur in spring  (0.48 m) and summer (0.53 m) and higher ones in autumn (0.62 m) and winter (0.60 m). The median Hs during  the tripod measurements were 0.54 m, with 0.50 m; 0.61 m and 0.53 m during respectively spring, autumn and  winter. The median Hs varied between 0.46 and 0.80; and the maximum Hs between 1.7 and 3.0 m (Table 1)  during the different deployments. Remark that during the field study (period 6) the median Hs is higher than  during the same season in 2008 and correspond with an autumn ‐ winter situation. The maximum Hs is however  lower (1.9 m), indicating that the storm intensity was limited.  

3.3 Time series  Time series of SPM concentration at 0.2 and 2 mab (m above bed) together with water depth, the significant  wave height and the ADV altimetry for the measuring periods 1, 5 and 6 are shown in Fig. 3‐5. Period 1  (November – December 2006) is characterized by the occurrence of different storms. On 12‐13 November, a NW  storm generated significant wave heights of about 2.8 m (day 317‐318). The highest SPM concentrations were  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 (>0.8 g l‐1). The OBS2 has measured an increase in  SPM concentration only during a short period after the storm. This indicates that vertical mixing was limited.  Similar data have been collected during the storm of December 31 (period 2), see Fettweis et al. (2010). ADV  altimetry shows a vertical rise of the acoustic reflective boundary after the storm (day 317 to 321) and indicating  the deposition of mud and the formation of a fluid mud layers. Its appearance coincides with low wave activity  and decelerating currents associated with neap tide when current velocities are not strong enough to resuspend  the mud layer on the bed. During higher wave activity or accelerating currents (around day 321), the fluid mud  layer disappears. The altimetry signal shows then a bed boundary fluctuating with the quarter diurnal tidal  currents; the change in altimeter height on day 336 is probably due to tripod instability. The decrease in  altimeter height on day 344 is due to mud accumulation. During the deposition events, the sea floor as detected  by the ADV altimetry raised about 10 cm, due to formation of fluid mud. Period 5 (April – June 2008) was  characterised by low meteorological disturbances. SPM concentration follows tidal and neap‐spring tidal signal  with higher SPM concentration around days 108‐114, 124‐130 and 142‐144 (Fig. 4). A clear shift between the  OBS1 (0.2 mab) and the OBS2 (2 mab) signal is observed from day 132 (May 2008). The highest SPM  concentrations occur at 0.2 mab during neap tide, whereas at 2 mab the highest values are around spring tide.  This indicates probably an increased deposition of mud in fluffy layers during neap tide. The acoustic bed  boundary remains at the same distance after stabilization of the tripod at the beginning of the deployment.  Deposition and consecutive resuspension occurs as temporal events coinciding with the ebb‐flood tidal signal  during neap tides and the availability of SPM. During the deposition events, the sea floor as detected by the ADV  altimetry raised on average by 10 cm, due to formation of fluffy layers. From day 140 on SPM concentration  decreased, resulting in no increase of the acoustic bed boundary. The field experiment (i.e. disposal operations)  took place between 5 May and 2 June 2009; the measurements continued until 15 June almost 14 days after the 

5

end of the disposal operations. May 2009 was marked by alternating W‐SW and E‐NE and relatively high wave  conditions (Table 1, Fig. 5). SPM concentration at 0.2 mab was strikingly high, with tide‐averaged values ranging  from 0.3‐1.6 g l‐1, when compared with the corresponding spring period of May 2008. These high values  remained until one week after the end of the experiment before decreasing to tide averaged values lower than  0.5 g l‐1. The high SPM concentrations in May 2009 are partially due to the occurrence of periods with high  waves. SPM concentration at 2 mab differs from the near‐bed one, and reveals a dynamic controlled by tidal and  neap‐spring tidal variation, whereas near the bed high concentrated benthic suspension or fluid mud layers have  dominated the sediment dynamics. The ADV altimetry reveals also the occurrence of an increase in acoustic bed  boundary of 8‐10 cm during neap tide(for day 134‐139 and 153‐159). For both mud depositions, favoured hydro‐ meteor conditions prevail (i.e. low wave activity and decelerating currents). The increase in bed level remains  during several days and hardly any quarter‐diurnal variation occur. This possibly indicates that fluid mud layers  were fomed with a higher erosion resistance rather than fluffy layers. After cessation of the disposal operations,  the SPM concentrations at 0.2 mab remained still very high during 1 week and disapeeard together with the fluid  mud layer. The major conclusion from the data can be summarised as follows:  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.

Considerable variations in SPM concentrations exist during a tidal cycle: maximum concentrations were  sometimes up to 50 times higher than the minimum concentrations. Maxima in SPM concentration occur  during ebb and flood.  

3.

The ADV altimetry altimetry data show quarter‐diurnal variations in bed level when SPM concentration is  sufficiently high; this is explained as formation and resuspension of fluffy layers during slack waters.  

4.

The very high SPM concentrations measured near the bed during winter and autumn are related to storms  and suggest that high concentrated benthic suspension layers have been formed, which may stay for a few  days.  

5.

During the field experiment (May – June 2009) high concentrated benthic suspension layers mainly  influenced the near‐bed dynamics, whereas at 2 mab a spring‐neap tidal signal was identified. The ADV  altimetry data indicate the formation and resuspension of fluid mud layers associated with neap‐spring  cycle, rather than fluffy layers. 

3.4 Statistics of SPM concentration  For each of the 6 measuring period probability distributions were constructed together with fitted lognormal  distributions (Fig. 6). The probability of the Χ² test was computed, hypothesising that SPM concentration data fit  a lognormal distribution. The geometric mean (x*), median (D50) and multiplicative standard deviation (s*) of  these distributions, together with the X² test results is shown in Table 2. The results show that the mean SPM  concentration during autumn and winter (periods 1, 2, 3, 4) is generally higher than during spring (period 5). The  mean and median SPM concentration at 0.2 mab during the field experiment (period 6a) is significantly higher  than during any of the other periods, whereas at 2 mab the same order of magnitude is observed than during a  winter situation (periods 1, 2 and 3). During the field experiment (5 May – 2 June) the mean increased to 612 mg  l‐1 (0.2 mab), i.e. more than twice the mean value before and after the experiment; but remained nearly similar 

6

at 2 mab (150 mg l‐1 vs. 128 mg l‐1).   Excluding samples from the measurements where the bottom wave orbital velocity Uw is lower (higher)  than a certain value allows calculating the median and multiplicative standard deviation of a population  representing good (stormy) weather conditions (Table 3‐4). The bottom wave orbital velocity has been calculated  based on the significant wave height (Hs), the water depth and the JONSWAP spectrum of waves (Soulsby, 1997).  An Uw of 0.03 m s‐1 (0.3 m s‐1) corresponds to a significant wave height of about 0.5 m (1.5 m) in a water depth of  about 8 m. The results show that the mean SPM concentration at 0.2 mab is generally lower during low wave  activity, except for period 2 and 3, whereas at 2 mab no clear relation can be observed. Before and after the field  experiment, lower wave influence is not significantly changing the mean SPM concentration at 2 mab. The low  mean SPM concentration during measuring period 3 is the result of calm weather (Hs = 0.46 m). The correlation  between median SPM concentration and SPM concentration during higher wave action (Uw > 0.3) is only obvious  for periods 1 and 6 (Table 4). For the other periods, the mean has similar values (period 4 and 5) or is even lower  than the mean for all data (Table 2).  The cumulative frequency distributions of SPM concentration are shown in Fig. 7. The probability to have  a SPM concentration at 0.2 mab higher than the median SPM concentration during the field experiments is on  average 0.21 (periods 1‐5, 6b), with 0.06 (period 4) and 0.30 (period 1) being the two extreme probabilities. At 2  mab the probabilities are on average higher (0.43: periods 1‐5, 6b) and the extreme values are closer together  (period 5: 0.32 – 0.52: period 6b).  

4. Discussions   In this study, the results based on time‐series measurements at a fixed location before, during and after an  experimental disposal of dredged matter, indicated a significant higher SPM concentration during the disposal.  Below we argue that the increase is not due to natural variability. The probability of having a SPM concentration  higher than the median SPM concentration at 0.2 mab during the field experiment is low.  

4.1 Wave influence  The Blankenberge site is situated in shallow waters (6 m MLLWS) where wave effects are important. The  correlation between median SPM concentration and SPM concentration during higher wave action (Uw > 0.3) is,  however, only obvious for periods 1 and 6a (Table 4). For the other periods, the median has a similar value  (period 4) or an even lower value than the median for all wave conditions (Table 2). This is in contrast with  observations made at MOW1 (Fig. 1) situated about 7 km offshore and at a water depth of about 10 m MLLWS,  where the median SPM concentration was clearly correlated with significant wave height (Fettweis and Nechad,  2010). The different dynamics between both stations cannot be explained by a different wave climate, as the  median Hs at both stations during measurements were very similar. It points possibly to the occurrence of a time  lag between waves and SPM concentrations at Blankenberge and that the suspended matter is rather advected  from elsewhere than locally eroded. The fact that advection is the dominant source of SPM and that during all  wave conditions the median SPM concentration during the field experiment (period 6a) was higher than during  the other periods, strengthen the argument that the high SPM concentration during this period is caused  significantly by the disposal of dredged material.   The data of May 2009 suggest that the high SPM concentrations are partially due to higher wave activity. 

7

Increase in SPM concentrations remains, however, limited to the near bed, suggesting that vertical mixing due to  waves was low. Fettweis et al. (2010) report that the increase in SPM concentration as observed during other  periods, was due to erosion of Holocene and recent muddy sediments during storms with Hs > 2 m. It is therefore  not very likely that the May 2009 storms (Hs < 1.8 m) have eroded sufficient sediments to explain the increase in  SPM concentrations.  

4.2 Ebb‐flood dynamics  During a tidal cycle, several peaks in SPM concentration are observed; generally, two peaks occur during ebb and  one during flood. The first ebb peak is generally lower and occurs when the increasing current velocity has  reached a critical value for resuspending the fluffy layer. The second one occurs at the end of ebb and is a  consequence of settling. This is confirmed by the fact that the SPM concentration peak at 0.2 mab is generally  observed after the peak at 2 mab. Maxima in SPM concentration during flood occur generally after slack water  and point thus to resuspension; the SPM concentration at 2 mab occurs after the peak at 0.2 mab. The mean of  the SPM concentration maxima during a tide is at least 1.7 times higher during the field experiment than during  the other periods (0.2 mab: 2670 mg l‐1 vs. 1566 mg l‐1; 2 mab: 941 mg l‐1 vs. 552 mg l‐1), whereas the mean of the  minima is similar (0.2 mab: 109 mg l‐1 vs. 99 mg l‐1; 2 mab: 35 mg l‐1 vs. 40 mg l‐1). These processes of  resuspension and rapid deposition have also been identified in the ADV altimetry data. The OBS measurements  indicate that in general the SPM concentration is higher during ebb at 0.2 mab, whereas at 2 mab it is generally  higher during flood. This is more pronounced during measuring period 6a, where the highest peaks at 0.2 mab  occur more frequently during ebb than flood. The SPM during the disposal experiment is thus concentrated in  the near bed layer rather than being well mixed in the water column, as is also observed by others (e.g. Wu et al.,  2006; Siegel et al., 2009), and indicates that the high SPM concentration near the bed is associated with  transport of fine sediments during ebb from the disposal site towards the measurement location; the  measurement location is situated in ebb direction of the disposal site. The SPM concentration and altimetry data  both suggest that a lutocline or benthic plume was formed during the field experiment and that the fate of the  fluid mud layer was controlled by the difference in bottom stress during neap and spring tidal periods.  

4.3 Impact of disposal  The natural variability of SPM concentration in the area is very high, which is indicated by the high multiplicative  standard deviations of the probability distributions (Table 2). Orpin et al; (2004) argue that the natural variability  of the system could be used to define the initial limits of acceptable turbidity levels. Such an approach assumes  that a short‐term increase (several hours) that falls within the range of natural variability will not have any  significant ecological effect. Orpin et al. (2004) developed this strategy for coral communities, which are much  more sensitive to turbidity than the Macoma balthica community found in the high turbidity area of the study  site (Degraer et al., 2008). Changes in species density or faunal community may be attributable to changes in  sediment composition and increased SPM concentration. Nevertheless, applying the same trigger to indicate  acceptable upper limits of SPM concentration in the water column (2 mab) indicates that the increase is within  natural variability of the system. However, at 0.2 mab we found that the cumulative frequency of SPM  concentration at 0.2 mab during the field experiment is not included within one standard deviation of the curve  for all the data not collected during the field experiment (Fig. 7), which would indicate a significant change in 

8

turbidity and possibly bed sediment composition over a large area. The increase of SPM concentration in the  near bed layer together with deposition of mud might thus negatively affect the macrobenthos of the area. Van  Hoey et al. (2010) report that on the disposal site Zeebrugge Oost (Fig 1), situated west of the port, lower macro‐ benthos and epibenthos densities were found than elsewhere in the area. The impact of disposal of dredged  matter is, however, very site specific (Ware et al., 2010); it is therefore important to aim at understanding the  site‐specific environmental impact when such disposal projects are designed.  

5. Conclusions  Harbour authorities worldwide are obliged to dredge their major shipping channels, and subsequently to dispose  the dredged spoil offshore. This study provided a statistical analysis to evaluate the effects of disposal operations  during a one‐month field experiment in a highly turbid area. Monitoring of SPM concentration was carried out  before, during and after the experiment. The data have been treated statistically; the major conclusions are:  1.

The natural variability of SPM concentration is very high.  

2.

The SPM concentration near the bed (0.2 mab) was exceptionally high (median was more than 2 times  higher) during the field experiment. Wave influence was not responsible for the high SPM concentrations.  

3.

The disposal site is situated in ebb‐direction of the measuring location. During the experiment, a generally  higher SPM concentration near the bed during ebb and at 2 mab during flood was observed, suggesting that  the disposed material was mainly transported in the benthic layer. The time lag between high wave heights  and high SPM concentration suggests further that the SPM has been advected towards the measuring  location rather than eroded locally. 

4.

The disposal results in a long‐term increase of SPM concentration near the bed at the measuring location.  This together with ADV altimetry suggest that fluid mud layers have been formed during whole the disposal  experiment rather than being limited to neap tidal or storm conditions as observed during the other  periods.  

Acknowledgements  This study was funded by the Maritime Access Division of the Ministry of the Flemish Community in the  framework of the MOMO project and by the Belgian Science Policy within the framework of the QUEST4D  (SD/NS/06A) 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, Zeehond and DN23 for their  skilful mooring and recuperation of the tripod. The measurements would not have been possible without the  technical assistance of A. Pollentier, J.‐P. De Blauwe and J. Backers (measuring service of MUMM, Oostende). 

References  Berlamont, J. 1989. Pumping fluid mud: Theoretical and experimental considerations. J. Coast. Res. 5, 195‐205.  Borja, A. 2005. The new European Marine Strategy Directive: Difficulties, opportunities, and challenges. Mar.  Poll. Bull. 52, 239‐242. doi:10.1016/j.marpolbul.2005.12.007.  Bolam, S.G., H.L. Rees, P. Somerfield, R. Smith, K.R. Clarke, R.M. Warwick, M. Atkins, and E. Garnacho. 2006.  Ecological consequences of dredged material disposal in the marine environment: A holistic assessment of  activities around the England and Wales coastline. Mar. Poll. Bull. 52, 415‐426. 

9

doi:10.1016/j.marpolbul.2005.09.028  Bray, R.N., Bates, A.D., Land, J.M. 1996. Dredging and the Environment. In: Dredging: A Handbook for Engineers,  Elsevier, 371‐387.   Degraer, S., Verfaillie, E., Willems, W., Adriaens, E., Vincx, M., Van Lancker, V. 2008. Habitat suitability modelling  as a mapping tool for macrobenthic communities: An example from the Belgian part of the North Sea. Cont.  Shelf Res. 28, 369‐379. doi: 10.1016/j.csr.2007.09.001  Delefortrie, G., Vantorre, M., Eloot, K. 2005. Modelling navigation in muddy areas through captive model tests. J.  Mar. Sci. Technol. 10, 188‐202. doi: 10.1007/s00773‐005‐0210‐5  De Nijs, M.A.J., J. Winterwerp, Pietrzak, J.D. 2009. On harbour siltation in fresh‐salt water mixing region. Cont.  Shelf Res. 29, 175‐193. doi:10.1016/j.csr.2008.01.019  Devlin, M., Bets, M., Haynes, D. 2007. Implementation of the Water Framework Directive in European marine  waters. Mar. Poll. Bull. 55, 1‐2. doi:10.1016/j.marpolbul.2006.09.020   Du Four, I., Van Lancker. V. 2008. Changes of sedimentological patterns and morphological features due to the  disposal of dredge spoil and the regeneration after cessation of the disposal activities. Mar. Geol. 25, 15‐29.  doi:10.1016/j.margeo.2008.04.011 Fettweis, M., Sas, M. 1999. On the sedimentation of mud in access channels to the harbour of Antwerp. PIANC  Bull. 101, 53‐59.   Fettweis, M., Van den Eynde, D. 2003. The mud deposits and the high turbidity maximum in the Belgian‐Dutch  coastal zone, southern Bight of the North Sea. Cont. Shelf Res. 23, 669‐691. doi:10.1016/S0278‐ 4343(03)00027‐X  Fettweis, M., Nechad, B., Van den Eynde, D. 2007. 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, 1568‐1583. doi:10.1016/j.csr.2007.01.017  Fettweis, M., Francken, F., Van den Eynde, D., Verwaest, T., Janssens, J., Van Lancker, V. 2010. Storm influence on  SPM concentrations in a coastal turbidity maximum area with high anthropogenic impact (southern North  Sea). Cont. Shelf. Res. doi:10.1016/j.csr.2010.05.001  Fettweis, M., Nechad, B. 2010. Evaluation of in situ and remote sensing sampling methods for SPM  concentrations, Belgian continental shelf (southern North Sea). Ocean Dyn. (accepted)  Fredette, T.J., French, G.T. 2004. Understanding the physical and environmental consequences of dredged  material disposal: history in New England and current perspectives. Mar. Poll. Bull. 49, 93‐102.  doi:10.1016/j.marpolbul.2004.01.014  Lauwaert, B., Bekaert, K., Berteloot, M., De Backer, A., Derweduwen, J., Dujardin, A., Fettweis, M., Hillewaert, H.,  Hoffman, S., Hostens, K., Ides, S., Janssens, J., Martens, C., Michielsen, T., Parmentier, K., Van Hoey, G.,  Verwaest, T. 2009. Synthesis report on the effects of dredged material disposal on the marine environment  (licensing period 2008‐2009). MUMM, ILVO, CD, aMT, WL report BL/2009/01, 73pp.  http://www.mumm.ac.be/Downloads/News/synthesis_report_PW_2009.pdf Lohrer, A.M., Wetz, J.J. 2003. Dredging‐induced nutrient release from sediments to the water column in a  southeastern saltmarsh tidal creek. Mar. Poll. Bull. 46, 1156‐1163. doi:10.1016/S0025‐326X(03)00167‐X 

10

McAnally, W.H., Friedrichs, C., Hamilton, D., Hayter, E.J., Shrestha, P., Rodriguez, H., Sheremet, A., Teeter, A.  2007. Management of fluid mud in estuaries, bays, and lakes. Present state of understanding on character  and behavior. J. Hydraul. Engrg. 133, 9‐22. doi:10.1061/(ASCE)0733‐9429(2007)133:1(9)  Nichols, M.M. 1988. Consequences of dredging. In: Kjerfve, B. (Ed.), Hydrodynamics of Estuaries. CRC Press,  Florida, 89–99.  O'Connor, T.P.1999. A wider look at the risk of ocean disposal of dredged matter. Mar. Poll. Bull. 38, 760‐761.  Okada, T., Larcombe, P., Mason, C. 2009. Estimating the spatial distribution of dredged material disposed of at  sea using particle‐size distributions and metal concentrations. Mar. Poll. Bull. 58, 1164‐1177.  doi:10.1016/j.marpolbul.2009.03.023  Orpin, A.R., Ridd, P.V., Thomas, S., Anthony, K.R.N., Marshall, P., Oliver, J. 2004. Natural turbidity variability and  weather forecasts in risk management of anthropogenic sediment discharge near sensitive environments.  Mar. Poll. Bull. 49, 602‐612. doi:10.1016/j.marpolbul.2004.03.020  OSPAR. 2008. Assessment of the environmental impact of dredging for navigational purposes, OSPAR  Commission, Publication nr 366/2008, 17pp.   PIANC. 2008. Minimising harbour siltation, Report No 102, 75pp.  Siegel, H., Gerth, M. Heene, T., Ohde, T., Rüss, D., Kraft, H. 2009. Hydrography, currents and distribution of  suspendedmatter during a dumping experiment in the western Baltic Sea at a site near Warnemünde. J. Mar.  Syst. 75, 397‐408. doi:10.1016/j.jmarsys.2008.04.005  Soulsby, R. 1997. Dynamics of marine sands. Thomas Telford Publications, London. 249pp.  Van den Eynde, D. 2004. Interpretation of tracer experiments with fine‐grained dredging material at the Belgian  Continental Shelf by the use of numerical models. J. Mar. Syst. 48, 171‐189.  doi:10.1016/j.jmarsys.2003.03.003  Van Hoey, G., Hostens, K., Parmentier, K., Robbens, J., Bekaert, K., De Backer, A., Derweduwen, J., Devriese, L.,  Hillewaert, H., Hoffman, S., Pecceu, E., Vandendriessche, S., Wittoeck, J., 2009. Biological and chemical effects  of the disposal of dredged material in the Belgian Part of the North Sea (period 2007‐2008). ILVO‐report,  Oostende (Belgium), pp. 97.   Van Maren, D.S., Winterwerp, J.C., Sas, M., Vanlede, J. 2009. The effect of dock length on harbour siltation. Cont.  Shelf Res. 29, 1410‐1425.  doi:10.1016/j.csr.2009.03.003  Verlaan, P.A.J., Spanhoff, R. 2000. Massive sedimentation events at the mouth of the Rotterdam waterway. J.  Coast. Res. 16, 458‐469.  Ware, S., Bolam, S.G., Rees, H.L. 2010. Impact and recovery associated with the deposition of capital dredging at  UK disposal sites: Lessons for future licensing and monitoring. Mar. Poll. Bull. 60, 79‐90.  doi:10.1016/j.marpolbul.2009.08.031  Winterwerp, J.C., 2005. Reducing harbour siltation I: methodology. J. Waterw., Port, Coast. Ocean Engrg. 131,  258–266. doi:10.1061/(ASCE)0733‐950X(2005)131:6(258)  Wu, J., Liu, J.T., Shen, H., Zhang, S. 2006. Dispersion of disposed dredged slurry in the meso‐tidal Changjiang  (Yangtze River) Estuary. Estuar. Coast. Shelf Sci. 70, 663‐672. doi:10.1016/j.ecss.2006.07.013 

11

Table 1: Tripod deployments at Blankenberge and the median and maximum significant wave height (Hs) during  the measurement period. Period 6a corresponds with the field experiment.   

Start (dd/mm/yyyy 

End ((dd/mm/yyyy 

Duration 

Median (max) Hs 

hh:mm) 

hh:mm) 

(days) 

(m) 

1 

08/11/2006 14:30 

15/12/2006 08:30 

36.7 

0.83 (2.76) 

2 

18/12/2006 10:47 

07/02/2007 13:17 

50.1 

0.79 (2.96) 

3 

28/01/2008 15:38 

24/02/2008 13:18 

26.9 

0.44 (2.82) 

4 

06/03/2008 09:09 

08/04/2008 15:29 

33.7 

0.76 (3.03) 

5 

15/04/2008 08:58 

05/06/2008 07:48 

51.0 

0.46 (1.69) 

6 

04/05/2009 09:59 

15/06/2009 11:49 

41.9 

0.57 (1.89) 

6a 

05/05/2009 12:00 

02/06/2009 07:00 

27.8 

0.55 (1.89) 

6b 

09/06/2009 00:00 

15/06/2009 11:49 

7.5 

0.42 (1.12) 

   

12

Table 2: Median (D50) and geometric mean SPM concentration (x*) in mg l‐1 during the 6 deployments (table 1)  together with the Χ² test probability (p) compared with a lognormal distribution and the multiplicative standard  deviation (s*). 1‐5, 6b corresponds with all the data before and after the field experiment (6a).   

0.2 mab 

2 mab 

data 

1 

2 

3 

4 

5 

6a 

6b 

1‐5,6b

1 

2 

3 

4 

5 

6a 

6b 

1‐5,6b

D50 

341 

288 

199 

321 

280 

672 

345 

281 

137

143

116

150 

106 

150

158

131 

x* 

340 

308 

183 

290 

258 

612 

319 

279 

144

149

105

150 

102 

150

135

128 

s* 

2.9 

3.0 

2.4 

3.0 

2.7 

2.6 

2.2 

2.9 

2.3 

2.3 

2.5 

2.5 

2.4 

2.3 

2.3 

2.4 

p 

1.00  0.57  0.77  0.96  0.82  0.99  0.37 

0.93 

0.93

0.94

0.59

0.99  0.94  0.99

0.12

0.99 

  Table 3: Median (D50) and geometric mean SPM concentration (x*) in mg l‐1 during the 6 deployments (see table  1) and wave orbital velocities < 0.03 m/s. Also shown is the Χ² test probability (p) of the distributions compared  with a lognormal one and the multiplicative standard deviation (s*). 1‐5, 6b corresponds with all the data before  and after the field experiment (6a).   

0.2 mab 

2 mab 

data 

1 

2 

3 

4 

5 

6a 

6b 

1‐5,6b

1 

2 

3 

4 

5 

6a 

6b 

1‐5,6b

 

184 

310 

206 

217 

259 

559 

306 

250 

91 

209

137

126 

113 

141

130

134 

x* 

221 

341 

181 

203 

239 

470 

269 

237 

103

196

116

127 

103 

142

121

124 

s* 

2.7 

2.5 

2.3 

3.3 

3.1 

3.0 

2.2 

2.8 

2.5 

2.1 

2.5 

2.8 

2.6 

2.4 

2.3 

2.6 

p 

0.02  1.00  0.49  0.59  0.84  0.99  0.38 

0.95 

0.18

0.66

0.32

0.83  0.70  0.99

0.11

0.86 

  Table 4: Median (D50) and geometric mean SPM concentration (x*) in mg l‐1 during the 6 deployments (see table  1) and wave orbital velocities > 0.3 m s‐1. Also shown is the Χ² test probability (p) of the distributions compared  with a lognormal one and the multiplicative standard deviation (s*). 1‐5, 6b corresponds with all the data before  and after the field experiment (6a). For period 5 and 6b, not enough data correspond with these wave conditions  to give statistical meaningful values.   

0.2 mab 

2 mab 

data 

1 

2 

3 

4 

5 

6a 

6b 

1‐5,6b

1 

2 

3 

4 

5 

6a 

6b  1‐5,6b 

D50 

763 

197 

98 

303 

‐ 

595 

‐ 

244 

178

117

57 

167 

‐ 

177

‐ 

130 

x* 

609 

237 

115 

288 

‐ 

651 

‐ 

270 

197

114

61 

162 

‐ 

169

‐ 

129 

s* 

2.5 

2.6 

2.2 

2.5 

‐ 

2.1 

‐ 

2.8 

2.0 

1.8 

2.0 

2.1 

‐ 

2.2 

‐ 

2.1 

p 

1.00  0.09  0.07  0.98 

‐ 

0.99 

‐ 

0.52 

0.13

0.40

0.08

0.37 

‐ 

0.46

‐ 

0.91 

 

13

Fig. 1: (Above) Map of the southern North Sea showing the yearly averaged surface SPM‐concentration (mg l‐1) in  the southern North Sea, from MODIS images (2003‐2008); coordinates are in latitude (°N) and longitude (°E)..  (Below) Detail of the Zeebrugge area showing the measurement station at Blankenberge, the wave  measurement station at Bol van Heist, the location of the disposal site during the field experiment and the Albert  II dock. The background consist of the bathymetry and of the dredging and disposal intensity of 2008 (in ton dry  matter TDS).  

14

Fig. 2: Bathymetrical map of the 210 kHz echo soundings in the Albert II dock showing a relict (left) and an active  dredging crater.  

15

Fig. 3: Tripod measurements of 8 November ‐ 15 December 2006 (measuring period 1). From up to down: depth  below water surface (m) and significant wave heights; ADV altimetry; and SPM concentration at 0.2 mab (SPM1)  and 2.2 mab (SPM2).  

16

  Fig. 4: Tripod measurements of 15 April – 5 June 2008 (measuring period 5). From up to down: depth below  water surface (m) and significant wave heights; ADV altimetry; and SPM concentration at 0.2 mab (SPM1) and  2.2 mab (SPM2).  

17

  Fig. 5: Tripod measurements of 4 May – 15 June (measuring period 6). The field experiment lasted from 5 May  until 2 June. From up to down: depth below water surface (m) and significant wave heights; ADV altimetry; and  SPM concentration at 0.2 mab (SPM1) and 2.2 mab (SPM2).            

18

  Fig. 6: Probability density distribution of the SPM concentration data at 0.2 mab (left) and 2 mab (right) for  periods 1, 5, 6a (during field experiment) and all data except those during the field experiment (1‐5 and 6b) and  the corresponding log‐normal probability density functions and Χ² test probability p  (periods 2‐4 are not shown).  The data are binned in classes of 50 mg l‐1, the dashed lines correspond to the median x* times/over the  multiplicative standard deviation s*.   

19

 

  Fig. 7: Cumulative probability distribution of SPM concentration measured at 0.2 mab and 2 mab. Shown are the  distributions of the data not collected during the field experiment (1‐5, 6b) including standard deviation (thin  black lines) and during the field experiment (6a).    

20

APPENDIX 7 SPM concentratieprofielen uit 13-uursmetingen te MOW1 en Kwintebank

9 2001/06 A MOW1 Profile 2 (2.2h after start) 16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (1.2h after start) 16

10

8

6

10

8

6

4

4

2

2

0

0 0

100

200 300 SPM concentration (mg/l)

400

500

0

100

16

14

14

12

12

10

8

6

400

500

400

500

400

500

8

6

4

2

2

0 0

100

200 300 SPM concentration (mg/l)

400

500

0

100

Profile 5 (5.2h after start)

200 300 SPM concentration (mg/l) Profile 6 (6.2h after start)

16

16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

500

10

4

0

10

8

6

10

8

6

4

4

2

2

0

0 0

100

200 300 SPM concentration (mg/l)

400

500

0

100

Profile 7 (7.2h after start)

200 300 SPM concentration (mg/l) Profile 8 (8.2h after start)

16

16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

400

Profile 4 (4.2h after start)

16

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (3.2h after start)

200 300 SPM concentration (mg/l)

10

8

6

10

8

6

4

4

2

2

0

0 0

100

200 300 SPM concentration (mg/l)

400

500

0

100

200 300 SPM concentration (mg/l)

Profile 10 (10.2h after start) 16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (9.2h after start) 16

10

8

6

10

8

6

4

4

2

2

0

0 0

100

200 300 SPM concentration (mg/l)

400

500

0

100

400

500

400

500

Profile 12 (12.2h after start)

16

16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

Profile 11 (11.2h after start)

200 300 SPM concentration (mg/l)

10

8

6

10

8

6

4

4

2

2

0

0 0

100

200 300 SPM concentration (mg/l)

400

500

400

500

0

100

200 300 SPM concentration (mg/l)

Profile 13 (13.2h after start) 16

14

Depth above bottom (m)

12

10

8

6

4

2

0 0

100

200 300 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

21 2002/27 B MOW1 Profile 2 (1.5h after start) 16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (0.5h after start) 16

10

8

6

10

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

16

14

14

12

12

10

8

6

200

250

200

250

200

250

8

6

4

2

2

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

Profile 5 (4.5h after start)

100 150 SPM concentration (mg/l) Profile 6 (5.5h after start)

16

16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

250

10

4

0

10

8

6

10

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

Profile 7 (6.5h after start)

100 150 SPM concentration (mg/l) Profile 8 (7.5h after start)

16

16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

200

Profile 4 (3.5h after start)

16

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (2.5h after start)

100 150 SPM concentration (mg/l)

10

8

6

10

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

100 150 SPM concentration (mg/l)

Profile 10 (9.5h after start) 16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (8.5h after start) 16

10

8

6

10

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

200

250

200

250

Profile 12 (11.5h after start)

16

16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

Profile 11 (10.5h after start)

100 150 SPM concentration (mg/l)

10

8

6

10

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

200

250

0

50

100 150 SPM concentration (mg/l)

Profile 13 (12.9h after start) 16

14

Depth above bottom (m)

12

10

8

6

4

2

0 0

50

100 150 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

22 2003/04 A MOW1 Profile 2 (1.7h after start) 16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (0.7h after start) 16

10

8

6

10

8

6

4

4

2

2

0

0 0

150

300 450 SPM concentration (mg/l)

600

750

0

150

16

14

14

12

12

10

8

6

600

750

600

750

600

750

8

6

4

2

2

0 0

150

300 450 SPM concentration (mg/l)

600

750

0

150

Profile 5 (4.7h after start)

300 450 SPM concentration (mg/l) Profile 6 (5.7h after start)

16

16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

750

10

4

0

10

8

6

10

8

6

4

4

2

2

0

0 0

150

300 450 SPM concentration (mg/l)

600

750

0

150

Profile 7 (6.6h after start)

300 450 SPM concentration (mg/l) Profile 8 (7.6h after start)

16

16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

600

Profile 4 (3.6h after start)

16

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (2.6h after start)

300 450 SPM concentration (mg/l)

10

8

6

10

8

6

4

4

2

2

0

0 0

150

300 450 SPM concentration (mg/l)

600

750

0

150

300 450 SPM concentration (mg/l)

Profile 10 (10.6h after start) 16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (9.6h after start) 16

10

8

6

10

8

6

4

4

2

2

0

0 0

150

300 450 SPM concentration (mg/l)

600

750

0

150

300 450 SPM concentration (mg/l)

600

750

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

24 2003/15 Kwintebank Profile 2 (1.6h after start) 20

18

18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (0.6h after start) 20

12 10 8

12 10 8

6

6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

20

18

18

16

16

14

14

12 10 8

40

50

40

50

40

50

10 8 6

4

4

2

2 0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

Profile 5 (4.6h after start)

20 30 SPM concentration (mg/l) Profile 6 (5.6h after start)

20

20

18

18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

50

12

6

0

12 10 8

12 10 8

6

6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

Profile 7 (6.6h after start)

20 30 SPM concentration (mg/l) Profile 8 (7.6h after start)

20

20

18

18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

40

Profile 4 (3.6h after start)

20

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (2.6h after start)

20 30 SPM concentration (mg/l)

12 10 8

12 10 8

6

6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

20 30 SPM concentration (mg/l)

Profile 10 (9.6h after start) 20

18

18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (8.6h after start) 20

12 10 8

12 10 8

6

6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

40

50

40

50

Profile 12 (11.7h after start)

20

20

18

18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

Profile 11 (10.6h after start)

20 30 SPM concentration (mg/l)

12 10 8

12 10 8

6

6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

40

50

0

10

20 30 SPM concentration (mg/l)

Profile 13 (12.6h after start) 20 18 16

Depth above bottom (m)

14 12 10 8 6 4 2 0 0

10

20 30 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

25 2003/17 Kwintebank Profile 2 (1.5h after start) 18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (0.5h after start) 18

12 10 8 6

12 10 8 6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

18

16

16

14

14

12 10 8 6

40

50

40

50

40

50

10 8 6 4

2

2 0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

Profile 5 (4.5h after start)

20 30 SPM concentration (mg/l) Profile 6 (5.5h after start)

18

18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

50

12

4

0

12 10 8 6

12 10 8 6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

Profile 7 (6.5h after start)

20 30 SPM concentration (mg/l) Profile 8 (7.4h after start)

18

18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

40

Profile 4 (3.5h after start)

18

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (2.5h after start)

20 30 SPM concentration (mg/l)

12 10 8 6

12 10 8 6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

20 30 SPM concentration (mg/l)

Profile 10 (9.5h after start) 18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (8.4h after start) 18

12 10 8 6

12 10 8 6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

40

50

40

50

Profile 12 (11.5h after start)

18

18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

Profile 11 (10.5h after start)

20 30 SPM concentration (mg/l)

12 10 8 6

12 10 8 6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

40

50

0

10

20 30 SPM concentration (mg/l)

Profile 13 (12.4h after start) 18 16

Depth above bottom (m)

14 12 10 8 6 4 2 0 0

10

20 30 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

26 2003/22 MOW1 Profile 2 (2.0h after start) 16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (1.0h after start) 16

10

8

6

10

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

16

14

14

12

12

10

8

6

200

250

200

250

200

250

8

6

4

2

2

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

Profile 5 (5.0h after start)

100 150 SPM concentration (mg/l) Profile 6 (6.0h after start)

16

16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

250

10

4

0

10

8

6

10

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

Profile 7 (7.0h after start)

100 150 SPM concentration (mg/l) Profile 8 (8.0h after start)

16

16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

200

Profile 4 (4.0h after start)

16

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (3.0h after start)

100 150 SPM concentration (mg/l)

10

8

6

10

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

100 150 SPM concentration (mg/l)

Profile 10 (10.0h after start) 16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (9.0h after start) 16

10

8

6

10

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

200

250

200

250

Profile 12 (12.0h after start)

16

16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

Profile 11 (11.0h after start)

100 150 SPM concentration (mg/l)

10

8

6

10

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

200

250

0

50

100 150 SPM concentration (mg/l)

Profile 13 (13.0h after start) 16

14

Depth above bottom (m)

12

10

8

6

4

2

0 0

50

100 150 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

27 2003/25 Kwintebank Profile 2 (1.7h after start) 20

18

18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (0.7h after start) 20

12 10 8

12 10 8

6

6

4

4

2

2

0

0 0

20

40 60 SPM concentration (mg/l)

80

100

0

20

20

18

18

16

16

14

14

12 10 8

80

100

80

100

80

100

10 8 6

4

4

2

2 0 0

20

40 60 SPM concentration (mg/l)

80

100

0

20

Profile 5 (4.7h after start)

40 60 SPM concentration (mg/l) Profile 6 (5.7h after start)

20

20

18

18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

100

12

6

0

12 10 8

12 10 8

6

6

4

4

2

2

0

0 0

20

40 60 SPM concentration (mg/l)

80

100

0

20

Profile 7 (6.7h after start)

40 60 SPM concentration (mg/l) Profile 8 (7.7h after start)

20

20

18

18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

80

Profile 4 (3.7h after start)

20

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (2.7h after start)

40 60 SPM concentration (mg/l)

12 10 8

12 10 8

6

6

4

4

2

2

0

0 0

20

40 60 SPM concentration (mg/l)

80

100

0

20

40 60 SPM concentration (mg/l)

Profile 10 (9.7h after start) 20

18

18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (8.7h after start) 20

12 10 8

12 10 8

6

6

4

4

2

2

0

0 0

20

40 60 SPM concentration (mg/l)

80

100

0

20

80

100

80

100

Profile 12 (11.7h after start)

20

20

18

18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

Profile 11 (10.7h after start)

40 60 SPM concentration (mg/l)

12 10 8

12 10 8

6

6

4

4

2

2

0

0 0

20

40 60 SPM concentration (mg/l)

80

100

80

100

0

20

40 60 SPM concentration (mg/l)

Profile 13 (12.7h after start) 20 18 16

Depth above bottom (m)

14 12 10 8 6 4 2 0 0

20

40 60 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

28 2004/04 Kwintebank Profile 2 (2.3h after start) 18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (1.3h after start) 18

12 10 8 6

12 10 8 6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

18

16

16

14

14

12 10 8 6

40

50

40

50

40

50

10 8 6 4

2

2 0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

Profile 5 (5.3h after start)

20 30 SPM concentration (mg/l) Profile 6 (6.3h after start)

18

18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

50

12

4

0

12 10 8 6

12 10 8 6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

Profile 7 (7.3h after start)

20 30 SPM concentration (mg/l) Profile 8 (8.3h after start)

18

18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

40

Profile 4 (4.3h after start)

18

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (3.3h after start)

20 30 SPM concentration (mg/l)

12 10 8 6

12 10 8 6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

20 30 SPM concentration (mg/l)

Profile 10 (10.3h after start) 18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (9.3h after start) 18

12 10 8 6

12 10 8 6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

40

50

40

50

Profile 12 (12.3h after start)

18

18

16

16

14

14

Depth above bottom (m)

Depth above bottom (m)

Profile 11 (11.3h after start)

20 30 SPM concentration (mg/l)

12 10 8 6

12 10 8 6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

20 30 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

29 2004/05 Kwintebank Profile 2 (2.0h after start) 22

20

20

18

18

16

16

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (1.0h after start) 22

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

25

50

75 SPM concentration (mg/l)

100

125

150

0

25

50

22

20

20

18

18

16

16

14 12 10 8

150

100

125

150

100

125

150

100

125

150

12 10 8 6

4

4

2

2 0 0

25

50

75 SPM concentration (mg/l)

100

125

150

0

25

50

Profile 5 (5.0h after start)

75 SPM concentration (mg/l) Profile 6 (6.0h after start)

22

22

20

20

18

18

16

16

Depth above bottom (m)

Depth above bottom (m)

125

14

6

0

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

25

50

75 SPM concentration (mg/l)

100

125

150

0

25

50

Profile 7 (7.0h after start)

75 SPM concentration (mg/l) Profile 8 (8.0h after start)

22

22

20

20

18

18

16

16

Depth above bottom (m)

Depth above bottom (m)

100

Profile 4 (4.0h after start)

22

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (3.0h after start)

75 SPM concentration (mg/l)

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

25

50

75 SPM concentration (mg/l)

100

125

150

0

25

50

75 SPM concentration (mg/l)

Profile 10 (11.0h after start) 22

20

20

18

18

16

16

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (9.0h after start) 22

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

25

50

75 SPM concentration (mg/l)

100

125

150

0

25

50

100

125

150

100

125

150

Profile 12 (12.0h after start)

22

22

20

20

18

18

16

16

Depth above bottom (m)

Depth above bottom (m)

Profile 11 (11.0h after start)

75 SPM concentration (mg/l)

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

25

50

75 SPM concentration (mg/l)

100

125

150

0

25

50

75 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

32 2004/24 MOW1 Profile 2 (1.7h after start)

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (0.7h after start)

10

8

6

10

8

6

4

4

2

2

0

0 0

250

500 SPM concentration (mg/l)

750

1000

0

250

14

12

12

10

8

6

750

1000

750

1000

750

1000

8

6

4

2

2

0 0

250

500 SPM concentration (mg/l)

750

1000

0

250

Profile 5 (4.7h after start)

500 SPM concentration (mg/l) Profile 6 (5.7h after start)

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

1000

10

4

0

10

8

6

10

8

6

4

4

2

2

0

0 0

250

500 SPM concentration (mg/l)

750

1000

0

250

Profile 7 (6.7h after start)

500 SPM concentration (mg/l) Profile 8 (7.7h after start)

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

750

Profile 4 (3.7h after start)

14

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (2.7h after start)

500 SPM concentration (mg/l)

10

8

6

10

8

6

4

4

2

2

0

0 0

250

500 SPM concentration (mg/l)

750

1000

0

250

500 SPM concentration (mg/l)

Profile 10 (9.7h after start)

14

12

12

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (8.7h after start)

14

10

8

6

10

8

6

4

4

2

2

0

0 0

250

500 SPM concentration (mg/l)

750

1000

0

250

14

12

12

10

8

6

1000

750

1000

750

1000

10

8

6

4

4

2

2

0

0 0

250

500 SPM concentration (mg/l)

750

1000

0

250

Profile 13 (12.7h after start)

500 SPM concentration (mg/l) Profile 14 (13.3h after start)

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

750

Profile 12 (11.7h after start)

14

Depth above bottom (m)

Depth above bottom (m)

Profile 11 (10.7h after start)

500 SPM concentration (mg/l)

10

8

6

10

8

6

4

4

2

2

0

0 0

250

500 SPM concentration (mg/l)

750

1000

0

250

500 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

33 2004/25 A MOW1 Profile 2 (1.8h after start)

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (0.8h after start)

8

6

4

2

8

6

4

2

0

0 0

50

100

150 SPM concentration (mg/l)

200

250

300

0

50

100

12

10

10

8

6

4

2

300

200

250

300

200

250

300

200

250

300

8

6

4

0 0

50

100

150 SPM concentration (mg/l)

200

250

300

0

50

100

Profile 5 (4.8h after start)

150 SPM concentration (mg/l) Profile 6 (5.8h after start)

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

250

2

0

8

6

4

2

8

6

4

2

0

0 0

50

100

150 SPM concentration (mg/l)

200

250

300

0

50

100

Profile 7 (6.9h after start)

150 SPM concentration (mg/l) Profile 8 (7.8h after start)

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

200

Profile 4 (3.8h after start)

12

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (2.8h after start)

150 SPM concentration (mg/l)

8

6

4

2

8

6

4

2

0

0 0

50

100

150 SPM concentration (mg/l)

200

250

300

0

50

100

150 SPM concentration (mg/l)

Profile 10 (9.8h after start)

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (8.8h after start)

12

8

6

4

2

8

6

4

2

0

0 0

50

100

150 SPM concentration (mg/l)

200

250

300

0

50

100

200

250

300

200

250

300

Profile 12 (11.8h after start)

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 11 (10.8h after start)

150 SPM concentration (mg/l)

8

6

4

2

8

6

4

2

0

0 0

50

100

150 SPM concentration (mg/l)

200

250

300

200

250

300

0

50

100

150 SPM concentration (mg/l)

Profile 13 (12.8h after start)

12

Depth above bottom (m)

10

8

6

4

2

0 0

50

100

150 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

35 2005/02 MOW1 Profile 2 (2.0h after start) 14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (1.1h after start) 14

8

6

8

6

4

4

2

2

0

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

0

100

200

14

12

12

10

10

8

6

600

400

500

600

400

500

600

400

500

600

6

4

2

2

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

0

100

200

Profile 5 (5.0h after start)

300 SPM concentration (mg/l) Profile 7 (7.0h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

500

8

4

0

8

6

8

6

4

4

2

2

0

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

0

100

200

Profile 8 (8.0h after start)

300 SPM concentration (mg/l) Profile 9 (9.0h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

400

Profile 4 (4.0h after start)

14

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (3.0h after start)

300 SPM concentration (mg/l)

8

6

8

6

4

4

2

2

0

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

0

100

200

300 SPM concentration (mg/l)

Profile 11 (12.1h after start) 14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 10 (11.1h after start) 14

8

6

8

6

4

4

2

2

0

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

400

500

600

0

100

200

300 SPM concentration (mg/l)

400

500

600

Profile 12 (13.1h after start) 14

12

Depth above bottom (m)

10

8

6

4

2

0 0

100

200

300 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

36 2005/07 A MOW1 Profile 2 (1.7h after start) 14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (0.7h after start) 14

8

6

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

14

12

12

10

10

8

6

200

250

200

250

200

250

6

4

2

2

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

Profile 5 (4.8h after start)

100 150 SPM concentration (mg/l) Profile 6 (5.7h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

250

8

4

0

8

6

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

Profile 7 (6.7h after start)

100 150 SPM concentration (mg/l) Profile 8 (7.7h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

200

Profile 4 (3.7h after start)

14

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (2.7h after start)

100 150 SPM concentration (mg/l)

8

6

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

100 150 SPM concentration (mg/l)

Profile 10 (9.7h after start) 14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (8.7h after start) 14

8

6

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

200

250

200

250

Profile 12 (11.7h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 11 (10.7h after start)

100 150 SPM concentration (mg/l)

8

6

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

200

250

0

50

100 150 SPM concentration (mg/l)

Profile 13 (13.0h after start) 14

12

Depth above bottom (m)

10

8

6

4

2

0 0

50

100 150 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

38 2005/15 A Kwintebank Profile 2 (1.8h after start) 22

20

20

18

18

16

16

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (0.7h after start) 22

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

22

20

20

18

18

16

16

14 12 10 8

40

50

40

50

40

50

12 10 8 6

4

4

2

2 0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

Profile 5 (4.7h after start)

20 30 SPM concentration (mg/l) Profile 6 (5.7h after start)

22

22

20

20

18

18

16

16

Depth above bottom (m)

Depth above bottom (m)

50

14

6

0

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

Profile 7 (6.8h after start)

20 30 SPM concentration (mg/l) Profile 8 (7.8h after start)

22

22

20

20

18

18

16

16

Depth above bottom (m)

Depth above bottom (m)

40

Profile 4 (3.7h after start)

22

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (2.8h after start)

20 30 SPM concentration (mg/l)

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

20 30 SPM concentration (mg/l)

Profile 10 (9.1h after start) 22

20

20

18

18

16

16

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (8.7h after start) 22

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

40

50

40

50

Profile 12 (11.8h after start)

Profile 11 (10.7h after start) 22

22

20

20

18

18

16

16 Depth above bottom (m)

Depth above bottom (m)

20 30 SPM concentration (mg/l)

14 12 10 8

14 12 10 8

6

6

4

4

2

2 0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

20 30 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

39 2005/15 B MOW1 Profile 2 (1.8h after start) 16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (0.8h after start) 16

10

8

6

10

8

6

4

4

2

2

0

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

0

100

200

16

14

14

12

12

10

8

6

600

400

500

600

400

500

600

400

500

600

8

6

4

2

2

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

0

100

200

Profile 5 (4.8h after start)

300 SPM concentration (mg/l) Profile 6 (5.8h after start)

16

16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

500

10

4

0

10

8

6

10

8

6

4

4

2

2

0

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

0

100

200

Profile 7 (6.8h after start)

300 SPM concentration (mg/l) Profile 8 (7.8h after start)

16

16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

400

Profile 4 (3.8h after start)

16

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (2.8h after start)

300 SPM concentration (mg/l)

10

8

6

10

8

6

4

4

2

2

0

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

0

100

200

300 SPM concentration (mg/l)

Profile 10 (9.8h after start) 16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (8.8h after start) 16

10

8

6

10

8

6

4

4

2

2

0

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

0

100

200

400

500

600

400

500

600

Profile 12 (11.8h after start)

16

16

14

14

12

12

Depth above bottom (m)

Depth above bottom (m)

Profile 11 (10.8h after start)

300 SPM concentration (mg/l)

10

8

6

10

8

6

4

4

2

2

0

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

400

500

600

0

100

200

300 SPM concentration (mg/l)

Profile 13 (12.9h after start) 16

14

Depth above bottom (m)

12

10

8

6

4

2

0 0

100

200

300 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

40 2005/29 MOW1 Profile 2 (1.2h after start) 14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (0.1h after start) 14

8

6

8

6

4

4

2

2

0

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

0

100

200

14

12

12

10

10

8

6

600

400

500

600

400

500

600

400

500

600

6

4

2

2

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

0

100

200

Profile 5 (4.1h after start)

300 SPM concentration (mg/l) Profile 6 (5.1h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

500

8

4

0

8

6

8

6

4

4

2

2

0

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

0

100

200

Profile 7 (6.1h after start)

300 SPM concentration (mg/l) Profile 8 (7.1h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

400

Profile 4 (3.1h after start)

14

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (2.0h after start)

300 SPM concentration (mg/l)

8

6

8

6

4

4

2

2

0

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

0

100

200

300 SPM concentration (mg/l)

Profile 10 (9.1h after start) 14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (8.1h after start) 14

8

6

8

6

4

4

2

2

0

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

0

100

200

14

12

12

10

10

8

6

500

600

400

500

600

400

500

600

8

6

4

4

2

2

0

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

0

100

200

Profile 13 (12.0h after start)

300 SPM concentration (mg/l) Profile 14 (13.0h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

400

Profile 12 (11.1h after start)

14

Depth above bottom (m)

Depth above bottom (m)

Profile 11 (10.1h after start)

300 SPM concentration (mg/l)

8

6

8

6

4

4

2

2

0

0 0

100

200

300 SPM concentration (mg/l)

400

500

600

0

100

200

300 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

41 2006/06 MOW1 Profile 2 (2.7h after start) 14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (0.7h after start) 14

8

6

8

6

4

4

2

2

0

0 0

200

400 600 SPM concentration (mg/l)

800

1000

0

200

14

12

12

10

10

8

6

800

1000

800

1000

800

1000

6

4

2

2

0 0

200

400 600 SPM concentration (mg/l)

800

1000

0

200

Profile 5 (7.7h after start)

400 600 SPM concentration (mg/l) Profile 6 (8.1h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

1000

8

4

0

8

6

8

6

4

4

2

2

0

0 0

200

400 600 SPM concentration (mg/l)

800

1000

0

200

Profile 7 (9.7h after start)

400 600 SPM concentration (mg/l) Profile 8 (10.7h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

800

Profile 4 (4.0h after start)

14

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (3.7h after start)

400 600 SPM concentration (mg/l)

8

6

8

6

4

4

2

2

0

0 0

200

400 600 SPM concentration (mg/l)

800

1000

0

200

400 600 SPM concentration (mg/l)

Profile 9 (11.7h after start) 14

12

Depth above bottom (m)

10

8

6

4

2

0 0

200

400 600 SPM concentration (mg/l)

800

1000

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

42 2006/10 A MOW1 Profile 2 (1.7h after start) 14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (0.7h after start) 14

8

6

8

6

4

4

2

2

0

0 0

200

400 600 SPM concentration (mg/l)

800

1000

0

200

14

12

12

10

10

8

6

800

1000

800

1000

800

1000

6

4

2

2

0 0

200

400 600 SPM concentration (mg/l)

800

1000

0

200

Profile 5 (4.8h after start)

400 600 SPM concentration (mg/l) Profile 6 (5.8h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

1000

8

4

0

8

6

8

6

4

4

2

2

0

0 0

200

400 600 SPM concentration (mg/l)

800

1000

0

200

Profile 7 (6.8h after start)

400 600 SPM concentration (mg/l) Profile 8 (7.7h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

800

Profile 4 (3.7h after start)

14

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (2.8h after start)

400 600 SPM concentration (mg/l)

8

6

8

6

4

4

2

2

0

0 0

200

400 600 SPM concentration (mg/l)

800

1000

0

200

400 600 SPM concentration (mg/l)

Profile 10 (11.8h after start) 14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (10.7h after start) 14

8

6

8

6

4

4

2

2

0

0 0

200

400 600 SPM concentration (mg/l)

800

1000

0

200

800

1000

800

1000

Profile 12 (13.2h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 11 (12.7h after start)

400 600 SPM concentration (mg/l)

8

6

8

6

4

4

2

2

0

0 0

200

400 600 SPM concentration (mg/l)

800

1000

0

200

400 600 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

45 2007/11 A Kwintebank Profile 2 (1.0h after start) 22

20

20

18

18

16

16

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (0.1h after start) 22

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

10

20

30

40

50

0

10

22

22

20

20

18

18

16

16

14 12 10 8

40

50

40

50

40

50

40

50

12 10 8 6

4

4

2

2 0 0

10

20

30

40

50

0

10

Profile 5 (4.4h after start)

20

30

Profile 6 (5.4h after start)

22

22

20

20

18

18

16

16

Depth above bottom (m)

Depth above bottom (m)

30

14

6

0

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

10

20

30

40

50

0

10

Profile 7 (6.4h after start)

20

30

Profile 8 (7.5h after start)

22

22

20

20

18

18

16

16

Depth above bottom (m)

Depth above bottom (m)

20

Profile 4 (3.5h after start)

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (2.5h after start)

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

10

20

30

40

50

0

10

20

30

Profile 10 (9.5h after start) 22

20

20

18

18

16

16

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (8.4h after start) 22

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

10

20

30

40

50

0

10

20

30

40

50

40

50

Profile 12 (11.5h after start)

22

22

20

20

18

18

16

16

Depth above bottom (m)

Depth above bottom (m)

Profile 11 (10.5h after start)

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

10

20

30

40

50

40

50

0

10

20

30

Profile 13 (12.4h after start) 22 20 18

Depth above bottom (m)

16 14 12 10 8 6 4 2 0 0

10

20

30

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

46 2007/11 B MOW1 Profile 2 (2.5h after start) 14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (1.5h after start) 14

8

6

8

6

4

4

2

2

0

0 0

50

100

150

200

0

50

14

14

12

12

10

10

8

6

200

150

200

150

200

150

200

6

4

2

2

0 0

50

100

150

200

0

50

Profile 5 (5.4h after start)

100 Profile 6 (6.4h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

150

8

4

0

8

6

8

6

4

4

2

2

0

0 0

50

100

150

200

0

50

Profile 7 (7.4h after start)

100 Profile 8 (8.4h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

100 Profile 4 (4.4h after start)

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (3.4h after start)

8

6

8

6

4

4

2

2

0

0 0

50

100

150

200

0

50

100

Profile 10 (10.4h after start) 14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (9.4h after start) 14

8

6

8

6

4

4

2

2

0

0 0

50

100

150

200

0

50

100

150

200

150

200

Profile 12 (12.4h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 11 (11.2h after start)

8

6

8

6

4

4

2

2

0

0 0

50

100

150

200

150

200

0

50

100

Profile 13 (13.3h after start) 14

12

Depth above bottom (m)

10

8

6

4

2

0 0

50

100

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

47 2007/16 MOW1 Profile 2 (1.5h after start) 14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (0.6h after start) 14

8

6

8

6

4

4

2

2

0

0 0

50

100

150

0

50

14

14

12

12

10

10

8

6

100

150

100

150

100

150

6

4

2

2

0 0

50

100

150

0

50

Profile 5 (4.6h after start)

Profile 6 (5.5h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

150

8

4

0

8

6

8

6

4

4

2

2

0

0 0

50

100

150

0

50

Profile 7 (6.5h after start)

Profile 8 (7.5h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

100 Profile 4 (3.5h after start)

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (2.6h after start)

8

6

8

6

4

4

2

2

0

0 0

50

100

150

0

50

Profile 10 (9.5h after start) 14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (8.6h after start) 14

8

6

8

6

4

4

2

2

0

0 0

50

100

150

0

50

14

14

12

12

10

10 Depth above bottom (m)

Depth above bottom (m)

100

150

100

150

Profile 12 (11.5h after start)

Profile 11 (10.5h after start)

8

6

8

6

4

4

2

2

0

0 0

50

100

150

100

150

0

50

Profile 13 (12.5h after start) 14

12

Depth above bottom (m)

10

8

6

4

2

0 0

50

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

49 2007/25 B MOW1 Profile 2 (1.5h after start) 14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 1 (0.5h after start) 14

8

6

8

6

4

4

2

2

0

0 0

50

100

150

200 250 300 SPM concentration (mg/l)

350

400

450

500

0

50

100

150

14

12

12

10

10

8

6

450

500

350

400

450

500

350

400

450

500

350

400

450

500

6

4

2

2

0 0

50

100

150

200 250 300 SPM concentration (mg/l)

350

400

450

500

0

50

100

150

Profile 5 (4.5h after start)

200 250 300 SPM concentration (mg/l) Profile 6 (5.5h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

400

8

4

0

8

6

8

6

4

4

2

2

0

0 0

50

100

150

200 250 300 SPM concentration (mg/l)

350

400

450

500

0

50

100

150

Profile 7 (6.6h after start)

200 250 300 SPM concentration (mg/l) Profile 8 (7.5h after start)

14

14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

350

Profile 4 (3.5h after start)

14

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (2.5h after start)

200 250 300 SPM concentration (mg/l)

8

6

8

6

4

4

2

2

0

0 0

50

100

150

200 250 300 SPM concentration (mg/l)

350

400

450

500

0

50

100

150

200 250 300 SPM concentration (mg/l)

Profile 10 (9.5h after start) 14

12

12

10

10

Depth above bottom (m)

Depth above bottom (m)

Profile 9 (8.5h after start) 14

8

6

8

6

4

4

2

2

0

0 0

50

100

150

200 250 300 SPM concentration (mg/l)

350

400

450

500

350

400

450

500

0

50

100

150

200 250 300 SPM concentration (mg/l)

350

400

450

500

Profile 11 (10.6h after start) 14

12

Depth above bottom (m)

10

8

6

4

2

0 0

50

100

150

200 250 300 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

50 2008/02 A MOW1 Profile 1 (2.1h after start) 14

12

12

10

10 Depth above bottom (m)

Depth above bottom (m)

Profile 2 (3.1h after start) 14

8

6

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

14

12

12

10

10

8

6

200

250

200

250

200

250

6

4

2

2

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

Profile 5 (6.0h after start)

100 150 SPM concentration (mg/l) Profile 6 (7.0h after start)

14

14

12

12

10

10 Depth above bottom (m)

Depth above bottom (m)

250

8

4

0

8

6

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

Profile 7 (8.0h after start)

100 150 SPM concentration (mg/l) Profile 8 (10.0h after start)

14

14

12

12

10

10 Depth above bottom (m)

Depth above bottom (m)

200

Profile 4 (5.1h after start)

14

Depth above bottom (m)

Depth above bottom (m)

Profile 3 (4.1h after start)

100 150 SPM concentration (mg/l)

8

6

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

100 150 SPM concentration (mg/l)

Profile 10 (12.0h after start) 14

12

12

10

10 Depth above bottom (m)

Depth above bottom (m)

Profile 9 (11.1h after start) 14

8

6

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

200

250

200

250

Profile 12 (14.0h after start)

14

14

12

12

10

10 Depth above bottom (m)

Depth above bottom (m)

Profile 11 (13.0h after start)

100 150 SPM concentration (mg/l)

8

6

8

6

4

4

2

2

0

0 0

50

100 150 SPM concentration (mg/l)

200

250

0

50

100 150 SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and the log-fitted profiles. Left: distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

51 2008/02 B Kwintebank Profile 1 (1.2h after start)

Profile 2 (2.2h after start)

22

22 avg data (0.5m) log-profile

avg data (0.5m) log-profile

20

18

18

16

16 Depth above bottom (m)

Depth above bottom (m)

20

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

Profile 3 (3.2h after start)

avg data (0.5m) log-profile

20

18

18

16

16 Depth above bottom (m)

Depth above bottom (m)

50

22 avg data (0.5m) log-profile

20

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

Profile 5 (5.2h after start)

20 30 SPM concentration (mg/l)

40

50

Profile 6 (6.2h after start)

22

22 avg data (0.5m) log-profile

20

avg data (0.5m) log-profile

20

18

18

16

16 Depth above bottom (m)

Depth above bottom (m)

40

Profile 4 (4.2h after start)

22

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

Profile 7 (7.2h after start)

20 30 SPM concentration (mg/l)

40

50

Profile 8 (8.2h after start)

22

22 avg data (0.5m) log-profile

20

avg data (0.5m) log-profile

20

18

18

16

16 Depth above bottom (m)

Depth above bottom (m)

20 30 SPM concentration (mg/l)

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

20 30 SPM concentration (mg/l)

40

50

Profile 9 (9.2h after start)

Profile 10 (10.2h after start)

22

22 avg data (0.5m) log-profile

avg data (0.5m) log-profile

20

18

18

16

16 Depth above bottom (m)

Depth above bottom (m)

20

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

Profile 11 (11.2h after start)

40

50

Profile 12 (12.2h after start)

22

22 avg data (0.5m) log-profile

20

avg data (0.5m) log-profile

20

18

18

16

16 Depth above bottom (m)

Depth above bottom (m)

20 30 SPM concentration (mg/l)

14 12 10 8

14 12 10 8

6

6

4

4

2

2

0

0 0

10

20 30 SPM concentration (mg/l)

40

50

0

10

20 30 SPM concentration (mg/l)

40

50

Profile 13 (13.2h after start) 22 avg data (0.5m) log-profile

20 18

Depth above bottom (m)

16 14 12 10 8 6 4 2 0 0

10

20

30

40

50

SPM concentration (mg/l)

Above: Measurements averaged over 0.5 m depth cells and log-fitted profiles.

Left: Distribution from fitted profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

APPENDIX 8 Saliniteits- en temperatuursprofielen

9 2001/06 A MOW1

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

21 2002/27 B MOW1

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

22 2003/04 A MOW1

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

24 2003/15 Kwintebank

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

25 2003/17 Kwintebank

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

26 2003/22 MOW1

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

27 2003/25 Kwintebank

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

28 2004/04 Kwintebank

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

29 2004/05 Kwintebank

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

32 2004/24 MOW1

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

33 2004/25 A MOW1

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

35 2005/02 MOW1

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

36 2005/07 A MOW1

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

38 2005/15 A Kwintebank

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

39 2005/15 B MOW1

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

40 2005/29 MOW1

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

41 2006/06 MOW1

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

42 2006/10 A MOW1

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

45 2007/11 A Kwintebank

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

46 2007/11 B MOW1

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

47 2007/16 MOW1

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

49 2007/25 B MOW1

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

50 2008/02 A MOW1

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).

51 2008/02 B Kwintebank

Salinity and temperature profiles as a function of time (x-axis in hours) and depth (y-axis depth above bottom in m).