N° d’ordre : 3125 -...

Université Bordeaux 1

Les Sciences et les Technologies au service de l’Homme et de l’environnement

N° d’ordre : 4821

THÈSE

PRÉSENTÉE A

L’UNIVERSITÉ BORDEAUX 1

ÉCOLE DOCTORALE DES SCIENCES ET ENVIRONNEMENT

Par Guillaume, BERNARD

POUR OBTENIR LE GRADE DE

DOCTEUR

SPÉCIALITÉ : Biogéochimie et écosystèmes

Mesures expérimentales et modélisation du remaniement

sédimentaire dans le bassin d’Arcachon

Directeurs de thèse : Antoine GREMARE et Pascal LECROART

Soutenue le : 10 Juillet 2013

Devant la commission d’examen formée de :

M. ANSCHUTZ, Pierre Professeur, Université Bordeaux 1 Président du jury

M. ORVAIN, Francis Maître de conférences HDR, Université de Caen Rapporteur

M. RIERA, Pascal Maître de conférences HDR, Université P.et M. Curie Rapporteur

M. KENNEDY Robert Principal researcher, NUI Galway Examinateur

M. MERMILLOD-BLONDIN Florian Chargé de recherches, CNRS Examinateur

M. MEYSMAN Filip J. R. Directeur de recherches, NIOZ Yerseke Examinateur

M. GREMARE, Antoine Professeur, Université Bordeaux 1 Directeur de thèse

M. LECROART, Pascal Professeur, Université Bordeaux 1 Directeur de thèse

M. DUCHENE, Jean-Claude Chargé de recherches HDR, CNRS Encadrant scientifique

Résumé Le remaniement sédimentaire, défini comme l’ensemble des mouvements de particules

sédimentaires induits par les organismes benthiques, est l’une des deux composantes du

phénomène de bioturbation. Il constitue un processus clé du fonctionnement des écosystèmes

côtiers. Ce manuscrit présente une étude intégrée de ce depuis l’échelle de la simple particule

sédimentaire jusqu’à celle de la communauté benthique in toto.

Le développement d’une nouvelle approche expérimentale basée sur l’acquisition à haute

fréquence et l’analyse de séries temporelles d’images de mouvements de luminophores le

long de la paroi d’aquariums plats a permis de mesurer directement les mouvements

élémentaires de particules de sédiment effectués par le bivalve A. alba. Cette approche a ainsi

conduit à la première détermination expérimentale d’ « empreintes » du remaniement

sédimentaire d’un invertébré marin, d’après le formalisme du modèle CTRW (Continuous

Time Random Walk).

Dans un second temps, le déploiement de cette nouvelle approche a permis d’évaluer, de

manière dynamique (i.e. pendant des expériences de 48h) et sur l’ensemble de la partie de la

colonne sédimentaire affectée par ce bivalve, le contrôle exercé par la température et par la

disponibilité de matière organique fraîche sur les caractéristiques du processus de

remaniement sédimentaire effectué par A. alba.

Enfin, l’intensité du remaniement sédimentaire effectué par l’ensemble de la communauté

benthique a été mesurée in-situ dans le Bassin d’Arcachon, à la fois dans un herbier à Zostera

noltii et dans une zone de vase nue d’où celui-ci a disparu. Ceci a permis de déterminer les

effets limitant de la présence d’herbier et de certaines espèces benthiques clés, sur le

remaniement sédimentaire.

Abstract Sediment particle mixing, defined as the movements of sediment particles induced by benthic

fauna, is one of the two components of bioturbation by benthic organisms. It is a key process

of the ecological functioning in coastal areas. This manuscript presents an integrated study of

sediment particle mixing process from the single sediment particle to the whole benthic

community.

The development of a new experimental approach, coupling high frequency acquisition of

time series images of luminophores motions along thin aquaria glass walls, allowed for the

direct measurement of elementary particle motions induced by the bivalve Abra alba. This

constitutes the first experimental assessment of sediment particle mixing “fingerprints” in a

marine invertebrate, according to the CTRW (Continuous Time Random Walk) model

formulation.

The deployment of this new approach also allowed for the determination of the control of

water temperature and of fresh organic matter availability on sediment particle mixing

induced by Abra alba. Moreover, the temporal (i.e., during 48h experiments) and spatial (i.e.,

over the whole section of the sediment column affected) dynamics of these effects were

considered.

At last, sediment particle mixing intensities induced by the whole benthic community were

assessed in-situ in Arcachon Bay, within both a Zostera noltii meadow and a bare sediment

mudflat where phanerogams were previously present. These results highlighted the restrictive

effect of phanerogams themselves and of a restricted number of key benthic species, on

sediment particle mixing.

Financements

Ce travail de thèse a pu être mené grâce à un contrat doctoral alloué par le Ministère

de l’Enseignement Supérieur et de la Recherche. Il a également été financé à travers les

programmes « BIOMIN » (LEFE-CYBER et EC2CO-PNEC), « IZOFLUX » (ANR blanc),

« Diagnostic de la Qualité des Milieux Littoraux » et « OSQUAR » (Conseil Régional

Aquitaine).

Remerciements

Je tiens tout d’abord à remercier infiniment mes deux directeurs de thèse,

Antoine Grémare et Pascal Lecroart.

Antoine, pour reprendre une syntaxe que vous affectionnez, MERCI : (1) de m’avoir

mis ce sujet entre les mains, d’avoir cru en moi (je me souviens du «Je ne suis pas mère

Theresa» lors de nos premières discussions à propos du sujet), (2) de m’avoir toujours

soutenu et épaulé et, (3) surtout de m’avoir énormément appris, et ce malgré votre emploi du

temps « ministériel ». Je pense, du moins j’espère, avoir grandi scientifiquement à vos côtés.

Pascal, même si nous ne nous sommes pas vu beaucoup durant ces presque 4 ans, vous

avez toujours répondu présent lorsque j’en avais besoin. Vous avez également été d’une

rapidité et d’une efficacité redoutables lors des phases de correction. Pour tout cela et pour

votre soutien, Merci !

Je tiens également à remercier sincèrement Jean-Claude Duchêne, sans qui ce travail

n’aurait jamais été possible. Votre expérience et votre science du codage ont été d’une valeur

inestimable (il est en effet facile de dessiner des algorithmes sur du papier, les traduire en

codes…beaucoup moins).

Je souhaite exprimer ma reconnaissance envers les rapporteurs de thèse ainsi que les

membres du jury, merci d’avoir accepté d’évaluer mon travail.

Je voudrais remercier Xavier de Montaudouin, directeur de l’équipe ECOBIOC au

sein de laquelle j’ai effectué cette thèse, de m’avoir accueilli, mais également de ses sages

conseils et de son soutien. Merci aussi à Fréderic Garabetian, directeur de la station marine

d’Arcachon.

Pour m’avoir initié à l’étude du monde fascinant du benthos pendant mon stage de

Master 2, je tiens à exprimer toute ma gratitude envers Guy Bachelet. C’est en partie grâce à

vous si j’en suis là.

Parce que j’ai eu la chance de travailler au sein du groupe des « bioturbateurs », merci

également à Olivier Maire, de qui j’ai en quelque sorte pris la suite, et à Aurélie Ciutat. Vos

conseils furent souvent avisés et votre soutien plus qu’appréciable. Alicia Romero-Ramirez,

aka la SPI-girl, merci pour ton amitié, ton soutien et ta fonction de traductrice officielle de

mes mails en anglais. Bon courage également à Cecile Massé pour ta fin de thèse (et merci

pour tes fameux gâteaux réconfortant !). Cecile et Aurélie, mes colocs de bureau, vous avez

supporté mes percussions sur bureau intempestives et mon sens du rangement « personnel »

avec un calme olympien, pour ça, merci et bravo ! Merci aux biogéochimistes, Pierre

Anschutz, Bruno Deflandre et Marie-Lise Delgard (Ma thèse sister, merci particulièrement

pour ton sens logistique sans lequel j’aurais été un peu perdu pendant nos manips in-situ).

Merci au Capitaine du PLANULA Francis Prince (la vase peut vraiment salir un bateau !) et

à Laurent. Merci également au « Labo 5 team » dont les membres éminants m’ont beaucoup

aidé: Benoît Gouilleux (l’homme bassin d’Arcachon, il ne fait qu’un avec son

environnement!), Nicolas Lavesque (Merci pour tout Nico, professionnellement et

humainement!!! C’est assez sobre !?), Hugues Blanchet (pareil que pour le grand landais!

« Marine invaders » deviendra un jeu légendaire…lorsque nous l’aurons créé) et Paolo

Bonifacio, qui survit tant bien que mal sous nos latitudes « polaires » et à qui je souhaite bon

courage pour la suite.

Enumérer tous les gens qui m’ont aidé et/ou soutenu pendant ces 3 ans et demi au labo

s’avère une tâche ardue, je vais remercier pêle-mêle, les stagiaires que j’ai pu encadrer:

Matthieu, Alice et Matthieu, les collègues et amis, thésards ou « piliers du labo »: François

« Fanfan » (Charentais maritimien, pour les bad, tennis, apéro, rigolades…), Sabrina

«bibich» et ton tact légendaire (et qui forme avec Laurence la Chiva de la chimie, faut pas

déconner!), Cerise (une répartie affutée et bien belge !!), Deb (pour la découverte du dixit),

Loïc (alias petit barbu) et Fabien. Merci également à Nicolas S., Val, Aurélie, Sandrine,

MC, Cathy, Michel, Michel, Florence, Nathalie, Flora, Patrice (des blagues et du volley),

Gaëlle, Alexia (souvient toi du zero de master1), Damien, Cindy, Henry, Pascal, Christian,

Céline, Wioleta… Je dois en oublier, ce qui ne veut pas dire que je ne pense pas aussi à eux.

Merci également à tous les amis qui m’entourent, que cela dure depuis la maternelle ou depuis

quelques mois.

Puisque cela m’est permis, j’aimerais remercier mes parents et ma famille qui m’ont

toujours soutenu, eh oui j’ai enfin fini mes études !! Pour finir, Un immense merci à Malka,

pour sa patience (parfois toute relative) pendant ces plus de 3 ans (en fait depuis beaucoup

plus), pour son soutien total mais aussi parfois ses « éléctrochocs » salvateurs, enfin bref,

pour tout ce qu’elle m’apporte!!

Table des matières

Résumé - Abstract ............................................................................................... 3

Financements ....................................................................................................... 5

Remerciements ..................................................................................................... 7

Table des matières ............................................................................................. 10

Chapitre 1 : Introduction générale .................................................................. 17

I. Les écosystèmes côtiers........................................................................................... 18

II. Le remaniement sédimentaire : une des composantes de la bioturbation ......... 21

2.1. Les organismes benthiques et le concept de bioturbation ......................................... 21

2.2. Le remaniement sédimentaire : définition et importance dans le fonctionnement des

écosystèmes ................................................................................................................ 22

2.3. Etat de l’art de la caractérisation et de la quantification du remaniement sédimentaire

.................................................................................................................................... 24

2.3.1. Les groupes fonctionnels du remaniement sédimentaire ..............................................25

2.3.2. Méthodes de quantification du remaniement sédimentaire ...........................................26

2.3.2.1. Méthodes directes .....................................................................................................27

2.3.2.2. Mesure de profils verticaux de traceurs sédimentaire et modélisation ....................29

2.3.2.2.1. Mesure des profils verticaux ..........................................................................29

2.3.2.2.2. Modélisation des profils .................................................................................31

III. Premier objectif de la thèse .................................................................................... 36

IV. Le bassin d’Arcachon, une lagune littorale affectée par le déclin mondial des

herbiers de phanérogames marines ...................................................................... 38

4.1. Présentation générale ................................................................................................ 38

4.2. Les communautés benthiques ................................................................................... 39

4.3. Les herbiers ............................................................................................................... 41

4.3.1. Répartition des herbiers dans le bassin d’Arcachon......................................................42

4.3.2. Rôle écologique des herbiers de phanérogames marines ..............................................42

4.3.3. Déclin et disparition des herbiers de phanérogames marines .......................................43

4.3.4. Etat de l’art de l’étude des interactions entre les herbiers marins et le remaniement

sédimentaire induit par la faune benthique ...................................................................45

V. Second objectif de la thèse ..................................................................................... 47

Chapitre 2 : Mesures expérimentales d’empreintes de remaniement

sédimentaire du bivalve Abra alba ................................................................... 48

Experimental assessment of particle mixing fingerprints in the deposit-

feeding bivalve Abra alba (Wood) .................................................................... 49

Abstract ............................................................................................................................... 50

I. Introduction ............................................................................................................ 51

II. Materials and methods ........................................................................................... 53

2.1. Bivalve collection and maintenance ......................................................................... 53

2.2. Experimental set-up .................................................................................................. 54

2.3. Image processing ...................................................................................................... 55

2.4. Data processing ......................................................................................................... 58

2.4.1. Frequency distributions ................................................................................................58

2.4.2. Overall mean values .....................................................................................................59

2.4.3. 2D Spatial analysis .......................................................................................................59

2.4.4. Vertical profiles ............................................................................................................60

III. Results ...................................................................................................................... 60

3.1. Classification of luminophore movements ............................................................... 60

3.2. Frequency distributions ............................................................................................. 62

3.3. Overall mean values .................................................................................................. 64

3.4. 2D Spatial analysis .................................................................................................... 65

3.5. Vertical profiles ........................................................................................................ 70

IV. Discussion ................................................................................................................ 73

4.1. Validation of the approach ........................................................................................ 73

4.2. Limitations of the approach ...................................................................................... 75

4.3. Particle mixing fingerprints in Abra alba .................................................................. 78

4.3.1. Vertical and horizontal components of particle mixing ................................................78

4.3.2. Spatial heterogeneity of particle mixing fingerprints ...................................................79

4.4. Consequences for the use of CTRW models in Abra alba ....................................... 80

4.4.1. Suitability of the distributions classically used in CTRW to describe particle mixing in

A. alba ...........................................................................................................................80

4.4.2. Taking into account spatial heterogeneity when modelling particle mixing in A. alba

................................................................................................................................................81

4.4.3. Consequences on the use of CTRW models .................................................................82

4.5. Temporal dynamics and possible coupling with other innovative approaches .......... 82

Acknowledgments ............................................................................................................... 83

References ............................................................................................................................ 83

Transition ............................................................................................................................ 88

Chapitre 3 : Mesures expérimentales de l’effet de la température et de la

disponibilité en matière organique sur le remaniement sédimentaire du

bivalve Abra alba : Utilisation d’une nouvelle technique d’analyse d’images

............................................................................................................................. 89

Experimental assessment of the effects of temperature and food availability

on particle mixing by the bivalve Abra alba using new image analysis

techniques. .......................................................................................................... 90

Abstract ............................................................................................................................... 91

I. Introduction ............................................................................................................ 92

II. Materials and methods ........................................................................................... 94

2.1. Bivalve collection and maintenance ......................................................................................94

2.2. Experimental set-up ...............................................................................................................95

2.3. Image processing ....................................................................................................................96

2.4. Data processing ......................................................................................................................96

2.4.1. Vertical profiles ............................................................................................................97

2.4.2. Overall mean values .....................................................................................................97

2.5. Data analysis ..........................................................................................................................98

2.5.1. Vertical profiles ........................................................................................................... 98

2.5.2. Overall mean values .................................................................................................... 98

III. Results .............................................................................................................. 99

3.1. Vertical profiles .....................................................................................................................99

3.1.1. Main effects ..................................................................................................................99

3.1.2. Normalized numbers of jumps ...................................................................................101

3.1.3. Inversed waiting times ................................................................................................103

3.1.4. Jump characteristics....................................................................................................104

3.1.5. Db ................................................................................................................................108

3.2. Overall mean values ..............................................................................................................110

3.2.1. Main effects ................................................................................................................110

3.2.2. Normalized numbers of jumps ...................................................................................110

3.2.3. Inversed waiting times ................................................................................................113

3.2.4. Jump characteristics....................................................................................................114

3.2.5. Db ................................................................................................................................116

IV. Discussion....................................................................................................... 117

4.1. Methodological considerations when assessing environmental effects on particle mixing

process using direct measurements of particle mixing fingerprints .......................................117

4.1.1. Descriptor of jump frequency .....................................................................................118

4.1.2. Experimental design ...................................................................................................118

4.1.3. Comparison of vertical profiles vs. overall mean values ............................................119

4.2. Seasonal changes in particle mixing fingerprints..................................................................121

4.3. Effect of food availability on particle mixing fingerprints ....................................................123

4.3.1. Effect on jump frequency ...........................................................................................123

4.3.2. Effect on jump lentghs and .....................................................................................125

4.3.3. Effect on Db ................................................................................................................126

V. Conclusions and perspectives .............................................................................. 126

References .......................................................................................................................... 127

Supporting informations .................................................................................................. 133

Transition .......................................................................................................................... 141

Chapitre 4 : Comparaison du remaniement sédimentaire dans un herbier à

Zostera noltii et dans un sédiment nu : Effet de la dynamique des

phanérogames et communautés benthiques endogées ................................. 142

Comparing sediment particle mixing a Zostera noltii meadow and a bare

sediment mudflat : Effects of seagrass dynamics and benthic infauna

composition ...................................................................................................... 143

Abstract ............................................................................................................................. 144

I. Introduction .......................................................................................................... 145

II. Material and methods........................................................................................... 148

2.1. Study area..............................................................................................................................148

2.2. Field sampling and experiments ...........................................................................................149

2.2.1. General strategy ..........................................................................................................149

2.2.2. Sediment particle mixing experiments .......................................................................149

2.2.2.1. Image analysis and vertical luminophore profiles computation .............................150

2.2.2.2. Modelling of sediment particle mixing intensity .....................................................150

2.2.2.3. Data processing ......................................................................................................151

2.2.3. Water and sediment characteristics ............................................................................151

2.2.4. Zostera noltii population characteristics.....................................................................151

2.2.5. Infauna ........................................................................................................................152

2.3. Statistical analysis .................................................................................................................152

2.3.1. Univariate analyses .....................................................................................................152

2.3.2. Infauna community structure ......................................................................................153

2.3.3. Linking DbN and synthetic descriptors ........................................................................153

2.3.4. Linking DbN and species distributions patterns ...........................................................153

III. Results .................................................................................................................... 154 3.1. Db

N ........................................................................................................................................154

3.2. Water and sediment characteristics .......................................................................................157

3.3. Zostera population characteristics ........................................................................................159

3.4. Benthic infauna characteristics .............................................................................................159

3.4.1. Univariate parameters .................................................................................................159

3.4.2. Community structure (multivariate) ..........................................................................160

3.5. Linking DbN and infauna synthetic descriptors .....................................................................168

3.6. Linking DbN and infauna species distributions patterns ........................................................170

IV. Discussion .............................................................................................................. 172

4.1. Sediment particle mixing intensity (DbN) .............................................................................172

4.2. Overall comparation of Zostera meadow and Bare sediment ...............................................174

4.2.1. Sediment particle mixing (DbN) .................................................................................174

4.2.2. Infauna ........................................................................................................................174

4.3. Spatio-temporal changes within Zostera meadow and Bare sediment during the period under

study .......................................................................................................................................177

4.3.1. Zostera meadow .........................................................................................................177

4.3.2. Bare sediment .............................................................................................................179

4.4. Control of sediment particle mixing intensity (DbN) by infauna composition ......................179

V. Conclusions ............................................................................................................ 182

Acknowledgments ............................................................................................................. 183

References .......................................................................................................................... 184

Chapitre 5 : Synthèse générale et perspectives ............................................ 192

I. La nouvelle approche expérimentale : intérêts, pertinence et perspectives .... 194

1.1. Intérêts ..................................................................................................................................194

1.2. Pertinence..............................................................................................................................194

1.3. Perspectives ..........................................................................................................................195

II. L’apport à la modélisation du remaniement sédimentaire ............................... 197

III. Le contrôle du remaniement sédimentaire : méthodologie, résultats et

perspectives ........................................................................................................... 199

3.1. Méthologie ............................................................................................................................199

3.2. Résultats ................................................................................................................................201

3.3. Perspectives ..........................................................................................................................201

IV. Mesures in-situ de l’intensité de remaniement sédimentaire, relation avec la

composition des communautés benthiques ......................................................... 204

4.1. La dispersion des données comme proxy de l’hétérogénéité spatiale : Détection d’effets de la

régression de l’herbier sur la distribution des organismes benthiques endogés et du processus

de remaniement sédimentaire ................................................................................................205

4.2. Identification d’espèces « clés » dans le contrôle du remaniement sédimentaire, mise en

évidence d’effets de stabilisation du sédiment .......................................................................207

4.3. Restriction de l’intensité du remaniement sédimentaire dans l’herbier abritant une plus

grande densité d’organisme que la vase nue ..........................................................................208

V. Bilan ....................................................................................................................... 210

Références biliographiques ............................................................................. 211

CHAPITRE 1 : Introduction générale

[17]

Chapitre 1:

Introduction générale


[18]

I. Les écosystèmes côtiers

Les écosystèmes côtiers et estuariens constituent l’interface entre les océans et les

continents. Cette position particulière a permis à ces écosystèmes de se développer sous de

multiples formes telles que les marais salés, mangroves, herbiers marins, récifs coralliens ou

plages sableuses dunaires (Barbier et al. 2011). Du fait de leur forte productivité et donc de la

disponibilité des ressources qu’ils génèrent, les écosystèmes côtiers ont de tous temps abrité

des activités humaines (pêcheries, centres portuaires, tourisme, industrie, production

d’énergie). En 2003, environ 3 milliard de personnes, soit 50% de la population mondiale, se

concentraient à moins de 200 km des côtes (Creel 2003, population reference bureau). Cette

utilisation des zones côtières par l’Homme a, depuis la préhistoire, conduit à la modification

de ces écosystèmes. Cette tendance s’est accélérée durant les derniers 150-300 ans avec le

développement de l’économie coloniale puis mondialisée (Figure 1.1), allant de pair avec le

progrès technique, l’industrialisation, l’augmentation des échanges internationaux ainsi

qu’une rapide croissance de la population (Lotze et al. 2006 ; Figure 1.1). De fait, durant

cette période, les pressions anthropiques croissantes sur ces écosystèmes (exploitation

excessive des ressources halieutiques, aménagement pour la navigation, urbanisation,

tourisme, pollutions physico-chimiques, introduction d’espèces, réchauffement global,

acidification des océans, eutrophisation), ont amené à ce constat : à ce jour, 50-67% des

marais salés, 35% des mangroves, 30% des récifs coralliens et 29-65% des herbiers marins

sont sévèrement dégradés, ou bien même ont tout simplement disparu (Lotze et al. 2006 ;

Barbier et al. 2011). De plus, on estime à 33% la réduction du nombre de pêcheries

considérées comme viables (Worm et al. 2006).


[19]

Figure 1.1 : Evolution historique de 12 écosystèmes côtiers et estuariens d’Amérique du

nord, d’Europe et d’Australie. Abondance relative moyenne de 30 à 60 espèces selon les

écosystèmes au cours du temps (A) et des périodes culturelles humaines (B) et évolution

conjointe de la population humaine (C et D). La datation des périodes culturelles varie selon

les écosystèmes étudiés. Ces périodes correspondent à l’émergence et au développement de

technologies et d’économies comparables : Pre, avant les premiers peuplements humains ;

HG, chasseur-cueilleurs (pas d’économie, petites populations, nomadisme, exploitation des

ressources à l’échelle individuelle pour la subsistance) ; Agr, période agricole (pas

d’économie, petites populations, sédentarisation, exploitation des ressources à l’échelle de

petits peuplements, cultures vivrières et artisanat) ; Est, période d’établissement de

l’économie et des échanges commerciaux (colonisation, développement des échanges entre

empires et zones colonisés), Dev, période de l’économie coloniale (accroissement de

l’économie et des échanges commerciaux, industrialisation, progrès technologiques,

exploitation intensive des grands mammifères, pêche sélective et côtière); Glo1, période de

développement de l’économie mondialisée 1900-1950 (accroissement de l’économie et des

échanges , forte accroissement de la population, forte exploitation des ressources, pêche non

sélective majoritairement côtière) ; Glo2 deuxième période de l’économie mondialisée, 1950-

2000 (pêche intensive, industrielle et non sélective, engins de pêches destructifs,

augmentation de la pollution, efforts de conservation) . Modifiée d’après Lotze et al. (2006).

Cette altération mondiale et drastique des écosystèmes côtiers ainsi que de la

biodiversité qu’ils abritent est d’autant plus préoccupante qu’elle affecte leurs fonctions


[20]

primordiales de : (1) réserve de ressources halieutiques, (2) réserve de nutriments et transfert

de ceux-ci aux écosystèmes adjacents, (3) lutte contre l’érosion, (4) détoxification et contrôle

de la pollution. Cependant, les capacités de résilience des écosystèmes côtiers, c’est-à-dire

leur faculté à recouvrer leur état initial, apparaissent d’un niveau élevé (Worm et al. 2006).

Ceci est particulièrement visible à travers les divers succès locaux, grâce à des politiques

adaptées, de tentatives de restauration qui expliquent le substantiel recouvrement d’environ

2% de certaines espèces dites « parapluies » (Roberge et Per Angelstam, 2004) via la

protection de leur habitat au sens large (Lotze et al. 2006). La protection de ces espèces,

souvent de grande importance économique et/ou patrimoniale, du fait de la grande étendue de

leur territoire propre, implique en effet la protection globale de l’ensemble de l’écosystème

qu’elles occupent. Ceci met en évidence le besoin urgent de mise en place de politique

globale de protection et pour ce faire d’une compréhension profonde des mécanismes

impliqués dans le fonctionnement de ces écosystèmes.

D’un point de vue fonctionnel, les écosystèmes côtiers sont caractérisés par : (1) la

grande proximité entre les sous-composantes pélagiques et benthiques, (2) le fait que ces deux

sous-composantes hébergent des processus de production primaire, et (3) le fort

hydrodynamisme et la forte diversité biologiques qui tendent à favoriser les échanges entre

composantes pélagiques et benthiques (notion de couplage pelagos-benthos) et à modifier les

modalités mêmes de ce couplage par rapport à ce qui est observé en milieu hauturier. La

vision traditionnelle qui consiste à considérer que la composante benthique est alimentée par

un flux vertical (ou advectif) descendant est ainsi insuffisante à décrire la complexité des

interactions pelagos-benthos intervenant en milieu côtier où il est par exemple reconnu que

les flux de sels nutritifs régénérés issus de la minéralisation en milieu benthique sont

susceptibles d’influer très significativement, voire même de contrôler la production primaire

pélagique (Cloern, 1982 ; Chauvaud et al. 2000 ; Grall et Chauvaud, 2002). Ces mêmes flux

de nutriments, associés à des temps de résidence plus longs dans le compartiment

sédimentaire doivent par exemple ainsi être pris en compte dans la cinétique de remédiation

du processus d’eutrophisation (Officer et al. 1982 ; Laruelle et al. 2009). On constate ainsi

que les concentrations en phosphore inorganique dans les eaux de la Baltique continuent à

augmenter malgré les mesures de réduction des apports continentaux (Pitkänen et al. 2001 ;

Conley et al. 2009), et ceci du fait, lors d’événements anoxiques, de la production de

phosphates issus de la dégradation de la matière organique stockée dans les sédiments. A

l’inverse, lors de périodes d’oxygénation des eaux, le phosphore tend à être fixé


[21]

préférentiellement dans les sédiments, mécanisme accentué via la bioturbation par les

organismes benthiques (Karlson et al. 2007). La bioturbation au sens large, et tout

particulièrement l’une de ses composantes que constitue le remaniement sédimentaire induit

autant de mécanismes par lesquels les organismes benthiques influent sur le fonctionnement

des écosystèmes côtiers. Les définitions et/ou les méthodes d’étude de ce dernier processus

qui fait l’objet de ce travail de thèse sont détaillées ci-dessous.

II. Le remaniement sédimentaire : une des

composantes de la bioturbation

2.1. Les organismes benthiques et le concept de bioturbation

Du fait de leur mobilité réduite, de leur durée de vie relativement longue et de leur

exposition à la fois aux perturbations venant de la colonne d’eau et des sédiments, les

organismes benthiques intègrent particulièrement bien les changements environnementaux

intervenant au sein des écosystèmes marins (Pearson et Rosenberg, 1978). De plus, ils

participent aux échanges entre la colonne d’eau et le compartiment sédimentaire (couplage

benthos-pelagos en milieu côtier surtout) à travers leurs activités de bioturbation. La

composition faunistique des communautés benthiques change d’une manière relativement

prévisible en fonction des perturbations (physiques ou liées aux apports de matière organique)

environnementales (Pearson et Rosenberg, 1978 ; Figure 1.2A). Parallèlement, des

changements de la profondeur : (1) d’oxydation du sédiment et (2) des traces d’activité

biologique (Rosenberg, 2001, Figure 1.2B) sont observés et attribués à la bioturbation

effectuée par ces organismes.


[22]

Figure 1.2 : Courbes généralisées d’abondance, biomasse et richesse spécifique de

l’endofaune le long d’un gradient de perturbation (d’après Pearson et Rosenberg, 1978) (A).

Images de profils sédimentaires correspondant aux stades de succession écologique de

l’endofaune benthique et représentation schématique du modèle de succession (B) (Modifié

d’après Nilsson et Rosenberg, 2000 et Rosenberg, 2001).

La bioturbation est ainsi définie comme l’ensemble des phénomènes par lesquels les

organismes (animaux ou végétaux) modifient la structure physico-chimique des sols et/ou des

sédiments qu’ils occupent (Richter, 1952 ; Rhoads, 1974 ; Meysman et al. 2006 ; Kristensen

et al. 2012). Ce terme général englobe les processus de bioirrigation (échange d’eau et de

solutés entre le sédiment et la colonne d’eau) ainsi que de remaniement sédimentaire

(mélange des particules de sédiment induit par les organismes) (Kristensen et al. 2012).

2.2. Le remaniement sédimentaire : définition et importance

dans le fonctionnement des écosystèmes


[23]

Dans les sédiments meubles, qui constituent 70% des fonds marins (Trush et Dayton,

2002), le remaniement sédimentaire est majoritairement induit par les organismes benthiques

via leurs activités de nutrition, de défécation, de locomotion, de fouissage, ou bien encore

d’édification de structures biogéniques comme des terriers ou des tubes (Meysman et al.

2006 ; figure 3). Ces activités induisent une hétérogénéité spatiale à la surface et en

profondeur dans le sédiment ainsi que des transferts particulaires bi-directionnels entre le

sédiment et la colonne d’eau, qui permettent : (1) la multiplication de niches écologiques qui

engendrent une augmentation de la biodiversité, et (2) une modification de l’hydrodynamisme

à la surface ainsi que de la texture du sédiment qui influencent les phénomènes de remise en

suspension et/ou de sédimentation préférentielle (Figure 1.3).

Figure 1.3 : Représentation schématique des mécanismes écologiques associés au processus

de remaniement sédimentaire induit par l’endofaune. Les flèches bleues indiquent les

directions des transferts des particules sédimentaires et de la matière organique particulaire

(MOP).

Ces mécanismes exercent une importance déterminante dans le contrôle : (1) de la

reminéralisation de la matière organique (Kristensen, 2000 ; Caradec et al. 2004 ; Braekman

et al. 2010) et des échanges de nutriments qui en résultent entre la colonne d’eau et le

sédiment, (2) des phénomènes d’érosion/accrétion, (3) du devenir des polluants chimiques et


[24]

des kystes dormants de certaines espèces phytoplanctoniques (Persson et Rosenberg 2003),

ainsi que (4) du devenir des graines et donc du recrutement de certaines phanérogames

marines (Cabaço et al. 2008 ; Delefosse et Kristensen, 2012 ; Balckburn et Orth, 2013). Le

remaniement sédimentaire induit par les organismes benthiques joue ainsi un rôle primordial

dans le fonctionnement global des écosystèmes côtiers peu profonds comme les lagunes

littorales où le couplage benthos-pelagos influence particulièrement une forte productivité

biologique. Ceci implique que, dans un contexte général d’altération de ces zones fortement

impactées par les activités anthropiques, le processus de remaniement sédimentaire se trouve

affecté, en lien avec des changements dans la structure des communautés benthiques (Solan et

al. 2004a). Si les effets d’évènements de type anoxie et/ou perturbation organique sont bien

connus et peuvent conduire à une baisse voire à une disparition complète des organismes

benthiques et donc du remaniement sédimentaire (Pearson and Rosenberg, 1979 ; Rosenberg,

2001 ; Solan et al. 2004a), ceux de changements plus complexes, constitués par exemple par

la disparition d’habitats clé de l’écosystème peuvent varier et indifféremment conduire, soit à

une augmentation soit à une réduction de l’intensité du remaniement sédimentaire selon les

changements de structure des communautés benthiques (Solan et al. 2004a).

2.3. Etat de l’art de la caractérisation et de la quantification

du remaniement sédimentaire

C’est Darwin (1881) qui le premier montra que l’activité des invertébrés endogés, en

l’occurrence celle des lombrics, jouait un rôle extrêmement important à grande échelle en

influençant notamment la formation de terre végétale. Un des premiers exemples d’étude du

remaniement sédimentaire induit par un invertébré marin peut être attribué à Davison (1891)

qui estima qu’une population d’arénicoles (Arenicola marina) occupant une surface d’une

acre (4046 m²) pouvait excaver jusqu’à 3000 tonnes de sable par an. Depuis, notamment à

partir du début des années 1970, l’intérêt scientifique est allé croissant quant au rôle de la

bioturbation dans le fonctionnement des écosystèmes marins. Il a ainsi conduit à une

caractérisation du remaniement sédimentaire induit par différents groupes fonctionnels

d’organismes, et, au développement de différentes méthodes de quantification de ce même

remaniement sédimentaire.


[25]

2.3.1. Les groupes fonctionnels du remaniement sédimentaire

Les organismes benthiques qui remanient les particules de sédiment peuvent être

classés en plusieurs groupes fonctionnels en fonction de leurs actions caractéristiques sur ces

mêmes particules. Les actions en question découlent de la position des particules dans la

colonne sédimentaire, du mode de nutrition des organismes considérés ou bien d’autres

aspects de leur mode de vie comme la construction de structures biogéniques ou bien la

mobilité. On recense ainsi classiquement quatre groupes principaux de remaniement

sédimentaire (François et al. 1997 ; Solan et Wigham 2005 ; Kristensen et al. 2012 ; Figure

1.4):

Figure 1.4: Représentation schématique des mouvements de particules induits par les 4

groupes fonctionnels principaux de remaniement sédimentaire : les biodiffuseurs (A), les

convoyeurs vers le haut (B), les convoyeurs vers le bas (C) et les régénérateurs (D). Modifié

d’après François et al. (1997) et Kristensen et al. (2012).


[26]

(1) Les biodiffuseurs, dont les activités conduisent à des mouvements de particules

pouvant être assimilés à de la diffusion moléculaire. Ces mouvements

interviennent continuellement, dans des directions aléatoires et sur de très courtes

distances (mouvements locaux) (Figure 1.4A). Dans ce groupe, on distingue : les

biodiffuseurs épigés qui vivent au-dessus ou à la surface du sédiment et remanient

uniquement l’interface eau-sédiment, les biodiffuseurs superficiels endogés qui

vivent et remanient les particules dans la partie supérieure de la colonne

sédimentaire (jusqu’à environ 5 cm) (Maire et al. 2006), et les biodiffuseurs à

galeries (Francois et al. 2002). Ces derniers induisent un remaniement

sédimentaire de type biodiffusif principalement via leurs activités de fouissage lors

de la construction de galeries ou de terriers dans les 10-30 premiers centimètres de

la colonne sédimentaire (Duport et al. 2006).

(2) Les convoyeurs vers le haut, qui sont orientés verticalement tête en bas et se

nourrissent en profondeur dans le sédiment (Volkenborn et al. 2007) (Figure

1.4B). Les particules sont alors transportées verticalement de manière

unidirectionnnelle donc de manière non-locale via le tube digestif. Lorsque ces

organismes vivent en contact avec l’interface eau-sédiment, un tel transport

conduit à l’éjection des particules directement à la surface du sédiment (Cadée,

1976).

(3) Les convoyeurs vers le bas, qui vivent, à l’inverse, orientés verticalement tête en

haut et se nourrissent à la surface (Figure 1.4C). Ce mode de vie induit un

transport non-local des particules à travers le tube digestif de la surface vers des

strates plus profondes de la colonne sédimentaire (Shull, 2001 ; Shull et Yasuda,

2001).

(4) Les régénérateurs, qui excavent le sédiment lors de la création et/ou la

maintenance de terriers temporaires qui s’effondrent ensuite sous l’action des

courants (Gardner et al. 1987 ; Figure 1.4D).

2.3.2. Méthodes de quantification du remaniement sédimentaire

Comme détaillées par Maire et al. (2008), ces méthodes peuvent être « directes »,

c’est-à-dire consister en une évaluation de la quantité de sédiment (volume ou masse)

transportée à l’interface eau-sédiment par les organismes benthiques pendant un laps de temps


[27]

donné, ou bien être basées sur le suivi du devenir de traceurs dans la colonne sédimentaire

puis de l’ajustement des profils verticaux obtenus à l’aide d’un modèle mathématique.

2.3.2.1. Méthodes directes

Maire et al. (2008) ont recensé cinq principales méthodes dites « directes » afin de

quantifier le remaniement sédimentaire :

(1) La collecte et/ou le moulage du sédiment excavé par les organismes benthiques

pendant un intervalle de temps donné puis la mesure de leur masse ou de leur

volume (Davison, 1891 ; Cadée, 1976).

(2) Le piégeage, à l’aide d’un dispositif disposé à la sortie d’un terrier, de l’ensemble

du sédiment excavé par les organismes benthiques pendant un intervalle de temps

donné puis la mesure de leur masse ou de leur volume (Rhoads, 1963, 1967 ;

Berkenbush et Rowden, 1999).

(3) La détection (approximative) du changement de niveau de l’interface eau-

sédiment sous l’action des organismes benthiques (Rhoads, 1967 ; Suchanek et al.

1986).

(4) L’analyse microtopographique de la surface du sédiment afin de détecter les

changements de niveau de l’interface eau-sédiment, mais cette fois à une échelle

spatiale nettement plus précise permettant de relier directement les changements de

niveau à l’activité des organismes (Maire et al. 2007b).

(5) L’analyse d’image. Cette méthode permet l’enregistrement d’images successives

de la surface du sédiment afin de quantifier le remaniement sédimentaire (la

surface de sédiment affectée) induit par des organismes présents se déplaçant et/ou

se nourrissant à la surface (Duchêne et Rosenberg, 2001 ; Maire et al. 2007a ;

Maire et al. 2007c) ou en sub-surface (Hollerz et Duchêne 2001 ; Lohrer et al.

2005).

Les trois premières de ces méthodes, également les plus anciennes, sont caractérisées

par une faible résolution spatio-temporelle. De plus, elles ne permettent de quantifier que la

composante du processus de remaniement sédimentaire intervenant à l’interface sédiment-

eau. Ces méthodes sont de plus (en partie) destructives puisqu’elles induisent le prélèvement

du sédiment excavé, empêchant ainsi leur répétition dans le temps. Ces caractéristiques

expliquent le relatif manque de précision des résultats issus du déploiement de ces techniques


[28]

et leur abandon progressif par la communauté scientifique (Maire et al. 2008). A l’inverse,

l’analyse microtopographique et l’analyse d’image, de par leurs meilleures résolutions

spatiales et/ou temporelles et leur caractère non-destructif, permettent des mesures plus

précises et dynamiques de l’activité et du remaniement sédimentaire. Elle reste néanmoins

cantonnée à la caractérisation de l’activité proche de la surface du sédiment. Ainsi, en utilisant

une technique d’analyse de séquences d’images successives de la surface du sédiment basée

sur les changements de niveau de gris des pixels constituant ces images, Gémare et al. (2004)

ont quantifié l’activité de surface, liée à la nutrition, induite par les mouvements de coquille et

des siphons de deux espèces proches au sein du genre Abra, et ce de manière dynamique. Ces

auteurs ont ainsi pu démontrer, grâce à la haute résolution temporelle apportée par cette

méthode, l’existence de réponses fonctionnelles différentes à un enrichissement en matière

organique fraiche du sédiment, chez A. ovata et A. nitida. L’utilisation de l’analyse d’image,

couplée cette fois à une analyse en microtopographie conduite à l’aide d’un laser motorisé, a

permis à Maire et al. (2007a) d’évaluer à une échelle de résolution spatiale très fine (i.e. de

l’ordre de 50 µm), le volume de sédiment excavé à la surface du sédiment par le bivalve Abra

ovata (segmentum) et de démontrer ainsi la forte corrélation existant (Figure 1.5A) entre

l’activité de prospection des siphons, obtenue par analyse d’image à haute fréquence de la

surface du sédiment, et le volume de sédiment affecté, obtenu simultanément par analyse

microtopographique (Figure 1.5B).

Figure 1.5 : Relation, en fonction du temps, entre l’activité siphonale (points noirs), liée à la

nutrition, et le volume de sédiment remanié (points blancs), chez 2 individus de l’espèce Abra

ovata (A). Evolution temporelle du volume de sédiment remanié par A. ovata obtenue grâce à

la technique de microtopographie (laser motorisé) : Exemple d’images montrant les résultats

de 4 scans successifs (B). EM : dôme d’éjection, PA : surface affectée par les siphons.

Modifié d’après Maire et al. (2007a).


[29]

L’analyse de séquences d’images à haute résolution spatio-temporelle a également été

utilisée pour déterminer les trajectoires de particules transportées par les tentacules de

l’annélide polychète Eupolymnia nebulosa à la surface du sédiment (Maire et al. 2007c ;

Figure 1.6).

Figure 1.6 : Exemple d’analyse d’image permettant de détecter l’activité à la surface du

sédiment du corps (en bleu) et des tentacules (en rouge) de l’annélide polychète Eupolymnia

nebulosa ainsi que le mouvement des particules de sédiment (en jaune) déplacées le long des

tentacules ciliés. D’après Maire et al. (2007c).

Pour arriver à une telle caractérisation, ces auteurs ont adapté des algorithmes

d’analyse d’image issus de logiciels initialement développés pour analyser les mouvements de

larves d’invertébrés dans la colonne d’eau (Duchêne et Nozais 1994; Duchêne et Queiroga

2001).

2.3.2.2. Mesure de profils verticaux de traceurs sédimentaire et modélisation

2.3.2.2.1. Mesure des profils verticaux

L’utilisation de traceurs sédimentaire permet d’accéder à la composante verticale du

remaniement sédimentaire ainsi qu’au devenir des particules initialement positionnées à la

surface du sédiment. La quantification de l’intensité du remaniement sédimentaire repose

alors sur la mesure de profils verticaux de concentration de traceurs. Ces profils sont ensuite

ajustés à l’aide d’un modèle mathématique sensé décrire les mouvements des particules de


[30]

traceur ayant entraîné la modification de leur distribution dans le sédiment, permettant ainsi

de dériver un indice reflétant l’intensité du remaniement sédimentaire. Ces traceurs peuvent

être naturellement présents dans le sédiment, ou bien artificiellement introduits.

Parmi les traceurs naturellement présents dans le sédiment, on trouve des

radionucléides (Aller, 1982 ; Lecroart et al. 2007a, 2007b, 2010 ) issus des retombées

atmosphériques (210

Pb, 234

Th, 228

Th, 32

Si, 14

C, 7Be), des grains de pollen (Davis, 1974)

provenant également de l’atmosphère, ou la chlorophylle a (Sun et al. 1991 ; Gérino et al.

1998 ; Josefson et al. 2002) qui peut être produite dans la colonne d’eau ou à la surface du

sédiment. Dès lors, la présence de ces traceurs à une certaine profondeur dans le sédiment et à

un temps donné va dépendre du remaniement sédimentaire, ainsi que de leur propre cinétique

de dégradation. Les traceurs ayant une cinétique de dégradation rapide comme la chlorophylle

a vont ainsi permettre de mesurer un remaniement sédimentaire ayant eu lieu pendant un

court (à l’échelle du mois) intervalle de temps depuis leur dépôt (Josefson, 2002) tandis que

les traceurs ayant une cinétique de réaction plus longue vont servir à mesurer ce processus à

des échelles de temps allant de la saison (Lecroart et al. 2005, 2007b ; Wheatcroft, 2006)

jusqu’à plusieurs dizaines d’années (Schmidt et al. 2007).

Les traceurs artificiellement introduits permettent une mesure du remaniement

sédimentaire à une échelle de temps allant de quelques heures à environ 1 mois après leur

dépôt à la surface du sédiment. Ces traceurs peuvent être des sables minéraux (D’Andrea et

al. 2004), des particules de sédiment ou de la matière organique marquées radioactivement

(Blair et al. 1996), des particules de sédiment enrichies en métaux nobles (Wheatcroft et al.

1994) ou des billes de verre (Shull et Yasuda, 2001). La littérature fait majoritairement état de

particules de plastique ou de sédiment recouvertes de peinture fluorescente aux rayonnements

ultra-violet (UV) respectivement dénommées microtaggants (Wheatcroft, 1991) et

luminophores (Mahaut et Graf, 1987).

La mesure des profils verticaux de concentration des traceurs dans la colonne

sédimentaire repose la plupart du temps sur la découpe de carottes sédimentaires en fines

tranches d’une épaisseur donnée après prélèvement lorsqu’il s’agit de traceurs naturellement

présents dans le sédiment, ou après incubation in-situ ou ex-situ (en conditions contrôlées en

mesocosme) pendant un temps donné lorsqu’il s’agit de traceurs artificiellement introduits.

Les profils ainsi mesurés sont par conséquent limités à une seule dimension (verticale) et à un


[31]

seul temps expérimental, ce qui empêche toute vision dynamique et en deux dimension du

remaniement sédimentaire (Figure 1.7).

Figure 1.7 : Principe général de quantification du remaniement sédimentaire par la méthode

de mesure de profils verticaux de traceurs sédimentaire. (1) : dépôt des traceurs qui sont

mélangés dans la colonne sédimentaire par les organismes. (2) : Mesure du profil vertical de

concentration de ces traceurs dans la colonne sédimentaire. (3) Ajustement de ce profil avec

un modèle mathématique pour dériver un indice de l’intensité du remaniement sédimentaire

au cours du laps de temps séparant le dépôt des traceurs à l’interface eau-sédiment et la

mesure du profil. Modifié d’après Maire et al. (2008).

Récemment, et afin de dépasser ces limitations, des protocoles couplant l’utilisation :

(1) d’aquariums plats remplis de sédiment (Gilbert et al. 2003 ; Maire et al. 2006, 2007a,b) ou

d’un profileur sédimentaire (SPI) (Solan et al. 2004) , (2) de luminophores, ainsi que (3) de

techniques d’acquisition de séquences d’images sous lumière UV, et (4) de techniques

d’analyse d’images ont permis d’obtenir une vision 2D dynamique du remaniement

sédimentaire induit par une seule espèce en laboratoire (Gilbert et al. 2003 ; Maire et al. 2006,

2007a,b ; Figures 1.8A&B) ou bien par une communauté benthique in-situ (Solan et al.

2004 ; Figures 1.8C&D).

2.3.2.2.2. Modélisation des profils

Classiquement, les profils verticaux mesurés sont ajustés à un modèle mathématique

sensé décrire les mouvements des particules de traceur responsables de la modification de leur

distribution dans le sédiment.


[32]

Figure 1.8 : Visions 2D-dynamique du remaniement sédimentaire en laboratoire grâce à

l’utilisation d’un aquarium plat (A, B) et in-situ grâce au déploiement d’un profileur

sédimentaire (C,D): exemples de photographies sous lumière UV après dépôt de luminophore

à l’interface eau-sédiment au temps T0 (A,C) et après un temps d’incubation donné (B, D).

(photos A,B : Guillaume Bernard ; photos C,D : modifiées d’après Solan et al. 2004b).

Un indice de l’intensité du remaniement sédimentaire peut ensuite être calculé à partir

de l’ajustement du modèle au profil. Jusqu’à très récemment, le modèle le plus largement

employé était le modèle dit « biodiffusif », qui par analogie à la première loi de Fick,

considère que les mouvements de particules induits par les organismes benthiques sont tous

caractérisés par de très faibles amplitudes spatiales (mouvements locaux), des directions

totalement aléatoires et par le fait qu’ils sont infiniment fréquents (Guinasso and Schink,

1975; Boudreau, 1986a). Dans ce modèle, l’intensité du remaniement sédimentaire est


[33]

approchée via le coefficient de biodiffusion biologique Db par analogie avec le coefficient de

diffusion moléculaire. La concentration ( ) d’un traceur conservatif (e.g. luminophores) à

une profondeur x donnée, à un temps d’incubation t après l’introduction des traceurs à la

surface à t0, est alors obtenue par la solution suivante (Crank, 1975):

( )

√ (

) (1)

où N est le nombre de luminophores déposé à la surface en début d’expérience, et A la surface

de l’unité expérimentale (carotte ou aquarium).

Ces hypothèses s’avèrent parfois erronées ou irréalistes d’un point de vue biologique

(Meysman et al. 2003 ; Meysman et al. 2010), et ce malgré le fait que : (1) les profils

verticaux de traceurs sédimentaires s’ajustent souvent relativement bien avec un profil

exponentiel qui constitue la solution du modèle biodiffusif et (2) que des modèles plus

complexes (voir plus bas) peuvent également converger vers cette solution pour des durées

expérimentales assez longues (Meysman et al. 2010). Ceci constitue ce qui est communément

appelé le « paradoxe de la biodiffusion » (Meysman et al. 2003). Meysman et al. (2010), en

étudiant plus précisément les conditions d’applications du modèle biodiffusif, ont démontré

que la validité d’application de ce modèle était surtout conditionnée par la survenue, pendant

le laps de temps expérimental, d’un nombre suffisant de mouvements de particules. Les

limites du modèle biodiffusif expliquent le développement récent de modèles prenant mieux

en compte la diversité des actions caractéristiques des organismes sur les particules

sédimentaires, notamment quant à l’occurrence du transport non-local. Le but étant de dériver

des intensités de remaniement sédimentaire plus complexes, et découlant d’un temps

d’incubation plus court. A côté de modèles spécifiques, uniquement adaptés à la description

du remaniement sédimentaire induit par une espèce ou un groupe fonctionnel d’espèces

mélangeant le sédiment de la même façon (Boudreau, 1986b; Robbins, 1986; Boudreau and

Imboden, 1987 ; Soetaert et al., 1996; François et al., 2002), ces dernières années ont vu le

développement d’un modèle ubiquiste susceptible de s’adapter à tout type d’organisme et

d’échelle temporelle. Ce modèle, basé sur une description stochastique du remaniement

sédimentaire, est appelé Continuous Time Random Walk model (CTRW) (Meysman et al.

2008a,b, 2010). Il considère que les mouvements de particules ne sont uniformes ni dans le

temps, ni dans l’espace. Pendant un intervalle de temps donné, une particule peut soit, rester

immobile, soit au contraire « sauter » vers une nouvelle position. Les mouvements d’une

particule donnée sont donc caractérisés par une alternance de temps d’immobilité et de temps


[34]

de déplacements de direction et d’amplitude partiellement aléatoires (Figure 1.9A).

L’ensemble des mouvements de particules induits par une espèce ou une communauté donnée

peut alors être caractérisé par une « empreinte » (Figure 9B) constituée des distributions de

fréquence : (1) des temps pendant lesquels les particules restent immobiles entre deux

mouvements successifs (Wheatcroft, 1990), (2) des distances desquelles sont déplacées les

particules lors d’un mouvement élémentaire, et (3) des directions dans lesquelles ces mêmes

particules sont déplacées (Meysman et al. 2008a). Lorsque l’on ne considère qu’une seule

dimension (en général la verticale), seules les distributions des temps d’immobilité et de

l’amplitude des mouvements sur la dimension considérée conditionnent l’empreinte du

remaniement sédimentaire (Figure 1.9B).

Figure 1.9 : Principe de la description des mouvements d’une particule donnée, idéalisé

d’après le modèle CTRW (A). Exemple schématique du concept d’empreinte de remaniement

sédimentaire limité à une seule dimension à partir des distributions de fréquence des temps

d’immobilité et des distances de déplacement (B). Modifié d’après Meysman et al. (2008).

L’évolution d’un profil vertical de traceurs ( ) au cours du temps est alors

obtenue par l’équation générale (Meysman et al. 2008):


[35]

( ) ( ) [ ∫ ( )

] ∫ ∫

( )

( ) ( ) (2)

où ( ) est le profil vertical initial avec la profondeur x, la distribution des temps

d’immobilité des particules entre deux déplacements successifs et ( ) la distribution des

distances parcourues par les particules lors de déplacements élémentaires .

Le plus souvent, la distribution des temps d’immobilité est décrite par une loi de

Poisson, et celle des distances de déplacements élémentaires par une loi normale (Meysman et

al. 2008). ( ) et ( ) sont alors exprimés par :

( )

(

) (3)

( )

√ (

) (4)

où le temps d’immobilité moyen et la racine carrée de la variance (écart-type) des

distances de déplacement élémentaire.

Ces deux paramètres reflètent ainsi respectivement l’échelle de temps moyenne entre

deux déplacements et la distance caractéristique de ces déplacements de particule, induits par

un organisme ou une communauté donnée.

L’intensité de remaniement sédimentaire

exprimé en unité de surface par unité

de temps (le plus souvent en cm².an-1

) peut alors être déduite d’après la relation:

(5)

où DbNL

est le coefficient de biodiffusion « normal » par analogie avec le modèle

biodiffusif (Meysman et al 2010).

Il a été prouvé que ce modèle permettait une meilleure description du remaniement

sédimentaire comparé au modèle biodiffusif, notamment quant à la prise en compte des

mouvements non-locaux de particules et pour des durées expérimentales courtes (Maire et al.

2007b). Ce modèle a été utilisé pour quantifier l’intensité du remaniement sédimentaire induit

par les bivalves biodiffuseurs Abra ovata (Maire et al. 2007b) et A. alba (Braeckman et al.,

2010), l’annélide polychète biodiffuseur à galeries Nephtys sp. (Braeckman et al., 2010) et

l’amphipode biodiffuseur Corophium volutator (De Backer et al., 2011). Dans tous les cas,

les types de lois régissant les distributions constitutives de « l’empreinte » du remaniement


[36]

sédimentaire induite par les organismes étudiés ont été choisis a priori, et non à partir

d’observations quantitatives des mouvements de particules induits par ces mêmes organismes.

Ce point est d’autant plus important qu’il a été démontré que les types des lois de distribution

conditionnaient l’ajustement du modèle CTRW aux profils verticaux de traceurs (Meysman et

al. 2008, 2010). En ce sens, obtenir expérimentalement, à partir d’observations quantitatives,

des distributions des temps d’immobilité, de distances et de directions de déplacement de

particules constitue la condition sine qua non à la validation en bonne et due forme du modèle

CTRW. Cet objectif a d’ailleurs été identifié comme un des défis majeurs dans le domaine de

l’étude du processus de remaniement sédimentaire (Maire et al. 2007, 2008 ; Meysman et al

2008, 2010).

III. Premier objectif de la thèse

Dans ce contexte, le premier objectif de ce travail de doctorat consiste en une étude

mécanistique du remaniement sédimentaire. Il s’agit de mesurer expérimentalement une

empreinte de remaniement sédimentaire en utilisant le formalisme du modèle CTRW, non

plus à partir de lois de distribution choisies a priori, mais à partir d’observations quantitatives

des mouvements de particules sédimentaires induits par un organisme benthique cible.

Un tel objectif nécessite de suivre et de mesurer les mouvements élémentaires de

particules de sédiment. Pour ce faire, un nouveau protocole expérimental et analytique a été

développé. Ce protocole se base sur : (1) une optimisation de la résolution spatio-temporelle

de techniques existantes permettant d’accéder à une vision 2D et dynamique du remaniement

sédimentaire à l’échelle de la particule, à savoir l’utilisation conjointe de luminophores,

d’aquariums plats ainsi que d’un système d’acquisition d’images sous lumière ultraviolette

(Gilbert et al. 2003 ; Maire et al. 2006, 2007a, 2007b), (2) l’analyse des séquences d’images

ainsi obtenues grâce à un logiciel spécialement développé, issu de l’adaptation d’algorithmes

de « tracking » précédemment utilisés pour mesurer des déplacements de larves dans la

colonne d’eau (Duchêne et Nozais, 1994 ; Duchêne et Queiroga, 2001) ou de particules à la

surface du sédiment (Maire et al. 2007c, Figure 4) afin de mesurer les caractéristiques des

mouvements de luminophores.


[37]

Ce nouveau protocole a ensuite été utilisé pour mesurer l’empreinte du remaniement

sédimentaire induit par le mollusque bivalve Abra alba. Les connaissances quant à l’éthologie

et/ou l’action sur le sédiment des organismes au sein du genre Abra sont bien établies

(Hughes, 1973, 1975 ; Wikander, 1980, 1981 ; Duchêne et Rosenberg, 2001 ; Grémare et al.

2004 ; Maire et al. 2006, 2007a, 2007b ; Braeckman et al. 2010). Ces organismes sont des

déposivores de surface (Hughes, 1973) biodiffuseurs (Maire et al. 2006, 2007a,b), qui

déplacent les particules assez fréquemment et sur de courtes distances. Ces mouvements

interviennent de manière privilégiée dans des structures en forme de cônes inversés

constituées par le réseau de galeries siphonales (Wikander, 1980). Les remaniements

sédimentaires induits par A. alba et d’autres espèces appartenant au même genre ont de plus

déjà été mesurés via l’ajustement du modèle CTRW à des profils verticaux de luminophores

(Maire et al. 2007b ; Braeckman et al. 2010) obtenus après incubation dans des carottes

(Braekman et al. 2010) ou dans des aquariums plats (Maire et al. 2007b), et dans différentes

conditions de disponibilité de matière organique (Maire et al. 2007b) et de température (Maire

et al. 2007 ; Braeckman et al. 2010). Toutes ces indications désignent clairement Abra alba

comme un modèle biologique optimal dans le but : (1) de développer et tester le protocole

expérimental et analytique exposé ci-dessus (en comparant par exemple ces résultats avec les

données disponibles dans la littérature), et (2) d’évaluer l’effet de facteurs environnementaux,

connus comme affectant l’activité et le remaniement sédimentaire par le genre Abra, sur ces

« empreintes » mesurées expérimentalement.

La présentation des différents résultats correspondant à la poursuite de cette série

d’objectifs font l’objet des chapitres 2 et 3 de ce manuscrit de thèse. Ces deux chapitres

présentent respectivement: (1) le développement d’une nouvelle méthode expérimentale et

analytique permettant de mesurer directement les mouvements de particules et donc

l’empreinte du remaniement sédimentaire induit par le bivalve Abra alba, et (2) la

détermination de l’impact des principaux paramètres environnementaux que sont la

température et la disponibilité de la matière organique sur l’éthologie de A. alba ainsi que sur

l’empreinte du remaniement sédimentaire qui en résulte grâce à l’utilisation de notre nouvelle

approche expérimentale.

Dans un second temps, les travaux rapportés dans ce manuscrit de thèse se sont portés

sur la quantification du remaniement sédimentaire, mais cette fois dans une démarche d’étude


[38]

in-situ, et appliquée à une problématique écologique d’importance, tant à l’échelle locale du

bassin d’Arcachon qu’à une échelle plus globale incluant la quasi-totalité des écosystèmes

côtiers. Après une présentation du contexte, le second objectif de ce travail de thèse, faisant

l’objet du quatrième chapitre du manuscrit, sera explicité dans le point V de cette introduction

générale.

IV. Le bassin d’Arcachon, une lagune littorale

affectée par le déclin mondial des herbiers de

phanérogames marines

4.1. Présentation générale

Le bassin d’Arcachon est une lagune semi-fermée située au Sud du littoral atlantique

français (Figure 1.10). Il est soumis à un régime de marée meso à macro-tidal en fonction du

coefficient de marée. L’amplitude de la marée de type semi-diurne y est comprise entre 0,8 et

4,6 m (Plus et al. 2008). Cet espace de forme triangulaire d’une superficie de 174 km²

constitue la seule indentation le long du littoral sableux aquitain. Cette lagune est ouverte sur

l’océan par un système de passes au Sud-Ouest et reçoit des influences plus continentales,

majoritairement via le delta de la Leyre qui s’y déverse dans sa partie interne (à l’Est). Du fait

de cette double influence marine et continentale, trois masses d’eau distinctes présentant des

caractéristiques thermiques et halines propres peuvent y être distinguées: les eaux néritiques

externes, les eaux néritiques moyennes et les eaux néritiques internes (Boucher, 1968). La

zone intertidale représente une superficie de 110 km2, soit à peu près les deux tiers de la

superficie totale du bassin, le tiers restant est occupé par un réseau de chenaux dont la

profondeur n’excède pas 20 mètres (Figure 1.10). La majorité de la surface intertidale est

peuplée par des herbiers à Zostera noltii (70 km2) tandis que les bas niveaux des estrans sont

occupés par des concessions ostréicoles.


[39]

Figure 1.10 : Situation générale et bathymétrie du bassin d’Arcachon (Ganthy 2011)

4.2. Les communautés benthiques

L’influence relative et la conjonction de l’effet des facteurs que sont : (1) la

bathymétrie, (2) la salinité, (3) l’implantation d’herbiers à Zostères et, (4) dans une moindre

mesure l’existence de gisements ostréicoles (en culture ou en récifs « sauvages ») (Salvo

2010) tendent à structurer la répartition spatiale des différentes communautés d’organismes

benthiques (Blanchet et al. 2004, 2005). Au sein du domaine subtidal, les principaux facteurs

structurant les communautés benthiques sont l’influence océanique et les paramètres

sédimentaires induits par l’hydrodynamisme (Blanchet et al. 2005 ; Figure 1.11A), alors que

les facteurs prédominant dans le domaine intertidal ont plutôt été identifiés comme les

caractéristiques de masses d’eaux, la bathymétrie et la présence des herbiers à zostères naines

Zostera noltii (Blanchet et al. 2004 ; Figure 1.11B).


[40]

Figure 1.11: Répartition des différents peuplements benthiques identifiés par Blanchet (2004)

dans les domaines subtidal (A) et intertidal (B) du bassin d’Arcachon.


[41]

Dans le bassin d’Arcachon, la zone intertidale est ainsi particulièrement caractérisée

par un vaste herbier à Z.noltii, abritant une faune benthique qui lui est spécifiquement

associée (Blanchet et al. 2004).

4.3. Les herbiers

De manière générale, les herbiers de phanérogames marines peuvent être considérés

comme le pendant aquatique des prairies ou forêts terrestres (Duarte, 2002). Les

phanérogames marines constituent un groupe unique d’angiospermes, c’est à dire de plantes à

fleurs, qui se sont adaptées au milieu marin et notamment à la submersion. Environ 60

espèces de ces plantes, réparties en trois familles principales (Zosteraceae, Cymodoceaceae et

Posidoniaceae), ont été répertoriées (den Hartog and Kuo, 2006 ; Orth et al. 2006). Elles ont

colonisé la quasi-totalité des écosystèmes marins peu profonds des zones tempérées et

tropicales (Green et Short, 2003 ; Figure 1.12) où elles forment des herbiers plus ou moins

continus jouant un rôle primordial dans le fonctionnement de ces systèmes.

Figure 1.12 : Distribution mondiale des phanérogames marines en fonction des grandes

régions climatiques (d’après Green et Short, 2003 et Orth et al. 2006).

Ces plantes, qui peuvent culminer à plus d’un mètre au-dessus du fond (Koch et al.

2006), sont fixées dans le substrat par un système de racines et de rhizomes. Du fait du fort

niveau d’éclairement nécessaire à leur croissance/développement, les phanérogames marines


[42]

sont particulièrement sensibles aux changements environnementaux affectant la clarté de

l’eau (Orth et al. 2006).

4.3.1. Répartition des herbiers dans le bassin d’Arcachon

Deux espèces de phanérogames, appartenant au genre Zostera, se développent dans le

bassin d’Arcachon en fonction du niveau tidal, ce qui génère l’établissement de deux herbiers

distincts.

La zostère marine (Zostera marina) est présente dans une aire de répartition qui

s’étend du cercle polaire au sud de l’Espagne. Cette espèce colonise également les étangs

saumâtres et les lagunes du sud de la France. Dans le bassin d’Arcachon, elle occupe presque

exclusivement les chenaux, c’est-à-dire le domaine subtidal, et plus sporadiquement les

cuvettes perpétuellement immergées des estrans intertidaux.

La zostère naine (Zostera noltii) occupe généralement les zones intertidales ainsi que

quelques étangs littoraux depuis le sud de la Norvège jusqu’à la Mauritanie,. La plupart des

estrans intertidaux du bassin d’Arcachon sont colonisés par cette espèce, formant un herbier

communément considéré comme le plus vaste d’Europe (Auby et Labourg, 1996).

4.3.2. Rôle écologique des herbiers de phanérogames marines

Les herbiers marins constituent des écosystèmes hautement productifs, qui couvrent

seulement entre 0,1 et 0,2 % de la surface des océans, mais sont responsables de 15% du

stockage de carbone dans l’océan mondial (Duarte et Chiscano, 1999).

Les phanérogames marines sont considérées comme des organismes « ingénieurs de

l’écosystème autogènes » (autogenic ecosystem engineer) (Jones et al. 1994) qui, de par leur

implantation dans le sédiment via un dense réseau racinaire et le développement d’une

canopée complexe au-dessus de ce même sédiment, modifient leur environnement physique,

biogéochimique et biologique, facilitent l’implantation et l’accès aux ressources pour d’autres

espèces (Jones et al. 1997). Les herbiers marins augmentent localement la biodiversité

(Boström et Bonsdorf, 1997) en créant des habitats complexes, oxygénés, riches en nourriture

et fournissant un abri contre les prédateurs dans et au-dessus du sédiment qui abritent une


[43]

faune endo- et épigée abondante et diverse (Reise, 2002 ; Bouma et al. 2009). L’habitat

« herbier » sert notamment de zone de nurserie et/ou de nourricerie pour de nombreuses

espèces exploitées (Heck Jr et al. 2003 ; Figure 1.13A). La canopée réduit également les

contraintes hydrodynamiques qui s’exercent au niveau du sédiment, agissant comme un piège

à particules fines et riches en matière organique qui peuvent alors être stockées et/ou

recyclées dans le compartiment benthique de ces mêmes herbiers (Fonseca et Fisher, 1986 ;

Meadows et al., 2012). La réduction de l’hydrodynamisme tend également à limiter les effets

de l’érosion, stabilisant le substrat et réduisant les taux de remise en suspension des particules

fines susceptibles d’inhiber la production primaire (Figure 1.13A).

Dans le bassin d’Arcachon, le vaste herbier à Zostera noltii qui occupe les platiers

intertidaux, et dont la production annuelle était estimée en 1991 à 8880-12709 t C (Auby,

1991), contribue à lui seul à 20% de la production primaire totale du bassin. Cet herbier

constitue de plus un rempart contre l’érosion. Dans les zones colonisées par Z. noltii, le bilan

sédimentaire annuel est en effet positif (significatif d’un phénomène d’accrétion) alors qu’il

est négatif (érosion) dans les zones comparables (en termes de bathymétrie et

hydrodynamisme) non végétalisées (Ganthy et al. 2013). De plus, de Wit et al. (2001) ont

démontré l’effet tampon de ces herbiers sur la dynamique saisonnière des nutriments ainsi que

sur leur biodisponibilité (Welsh et al. 2000). Cet effet tampon se répercute tout naturellement

sur les communautés benthiques et particulièrement l’endofaune colonisant les herbiers à Z.

noltii dont les compositions faunistiques ne présentent que de faibles variations saisonnières

(Bachelet et al. 2000 ; Blanchet et al. 2004). Ces herbiers sont caractérisés par de fortes

abondances d’espèces opportunistes de petite taille dont la présence est la plupart du temps

reliée aux fortes concentrations en matière organique induites par : (1) le piégeage des

particules fines par la canopée et (2) la dégradation in-situ des feuilles et parties souterraines

de Z. noltii (Castel et al. 1989 ; Bachelet et al. 2000 ; Blanchet et al. 2004 ; Do et al. 2011 ;

Do et al. 2013).

4.3.3. Déclin et disparition des herbiers de phanérogames marines

A l’échelle du globe, Waycott et al. (2009) ont estimé à 29% le pourcentage de la

surface d’herbiers de phanérogames marines ayant disparu depuis la fin du XIXème

siècle. Ce

déclin, observé dans la grande majorité des écosystèmes côtiers, est clairement lié aux


[44]

perturbations provoquées par les activités anthropiques. Ces causes sont multiples (Duarte

2002) et varient selon les écosystèmes étudiés (Orth et al. 2006).

Figure 1.13: Représentation schématique des rôles principaux joués par les herbiers de

phanérogames dans le fonctionnement des écosystèmes côtiers tempérés (A), et des

mécanismes majeurs responsables du déclin de ces mêmes herbiers (B). Modifié d’après Orth

et al. (2006).

Dans les zones tempérées, les principaux mécanismes responsables du déclin des

herbiers, tels que compilés par Orth et al. (2006), sont : (1) l’eutrophisation, qui provoque une

réduction de la pénétration de la lumière et qui, associée au réchauffement climatique global

inhibe la croissance des phanérogames marines, (2) les maladies comme la « wasting disease

» qui affectent les herbiers à travers des épisodes de mortalité de masse, et dans une moindre

mesure (3) les interactions biologiques parmi lesquelles on peut citer la consommation de

phanérogames par les oiseaux herbivores et/ou des espèces de la macrofaune ou mégafaune,

ou bien encore (4) l’introduction d’espèces pouvant altérer mécaniquement ces plantes à

travers le phénomène de bioturbation (Figure 1.13B).

Dans le bassin d’Arcachon, les régressions des superficies occupées par Z. marina

(Figure 1.14A) entre 1989 et 2008 et Z. noltii (Figure 1.14B) entre 1989 et 2007 ont été

respectivement estimées à -74% et -33% par Plus et al. (2010).


[45]

Figure 1.14 : Evolution, dans le bassin d’Arcachon, des herbiers à Zostera marina entre 1989

et 2008 (A) et des herbiers à Zostera noltii entre 1989 et 2007 (B). Les zones figurées en

rouge, jaune et vert indiquent respectivement, un déclin, une progression et une stabilité de la

surface occupée par les phanérogames. Modifié d’après Plus et al. (2010).

Il est à noter que cette dynamique est dans un premier temps restée relativement lente

jusqu’en 2005 puis qu’elle s’est significativement accélérée (Plus et al. 2010), laissant

notamment des pans entiers de l’estran intertidal qui était auparavant recouvert par Z. noltii se

transformer en vastes vasières dénuées de végétation. Ce déclin au niveau local semble

pouvoir majoritairement être imputé à des températures des eaux anormalement hautes

enregistrées au milieu des années 2000, dont l’impact sur les zostères a pu être accentué par

l’effet d’une contamination par des herbicides (Auby et al. 2011). Depuis 2007, la tendance à

la disparition des herbiers dans le bassin semble toutefois légèrement s’infléchir (Auby et al.

2011).

4.3.4. Etat de l’art de l’étude des interactions entre les herbiers marins et

le remaniement sédimentaire induit par la faune benthique

La majorité des études traitant de l’interaction entre herbiers et bioturbation induite par

les organismes benthiques montrent des effets antagonistes liés à des actions biomécaniques

des phanérogames et des principales espèces bioturbatrices sur le sédiment. Suykerbuyk et al.

(2012) ont résumé ces interactions par le terme de « guerre biomécanique » (Biomechanical

warfare). L’établissement d’un dense réseau de racines/rhizomes au sein du sédiment

empêche ainsi le fouissage et donc l’installation des espèces bioturbatrices de grande taille

comme par exemple les crustacés décapodes thalassinidés (Berkenbusch et al. 2007a ;


[46]

Berkenbusch et al. 2007b ; Siebert and Branch 2007) et les annélides polychètes comme les

arénicoles (Arenicola marina) (Philippart 1994 ; Eklöf et al. 2011 ; Delefosse and Kristensen

2012 ; Suykerbuyk et al. 2012) ou Hediste disversicolor qui tendent ainsi à être exclues des

herbiers (Hughes et al., 2000 ; Berkenbusch and Rowden 2007 ; Berkenbusch et al., 2007 ;

Siebert and Branch 2005, 2007 ; Wesenbeeck et al 2007). En retour, le remaniement

sédimentaire très intense induit par ces mêmes espèces affecte la physiologie et/ou la

dynamique des phanérogames à travers : (1) un recouvrement des plantes elles-mêmes ou des

graines sous une couche de sédiment empêchant leurs développement et/ou germination

(Philippart 1994 ; Hughes et al 2000 ; Cabaço, 2008; Meadows et al. 2012 ; Suykerbuyrk

2012 ; Delefosse and Kristensen 2012), (2) une consommation des graines et/ou (3) une

altération des racines par friction avec les particules sédimentaires (Philippart 1994 ; Hughes

et al 2000 ; Cabaço, 2008). D’autres études ont au contraire montré que (1) l’enfouissement

ou le recouvrement des graines par certains organismes pouvait stimuler leur germination en

les déplaçant vers une profondeur optimale dans le sédiment (Delefosse and Kristensen, 2012

; Blackburn and Orth 2013), et que (2) à une échelle spatiale plus large, la bioturbation, en

interrompant les réseaux de racines/rhizomes, permettait de maintenir une certaine

hétérogénéité spatiale favorisant la reproduction sexuée des phanérogames au dépend de leur

seule reproduction asexuée (par fractionnement des rhizomes), et leur confère de plus grandes

capacités de résistance et de résilience aux événements climatiques (Townsend and Fonseca

1998 ; Meadows et al., 2012).

Cependant, à ce jour, très peu d’études (voire aucune) se sont focalisées sur la

quantification du remaniement sédimentaire induit par l’ensemble de la communauté

benthique des herbiers de phanérogames. Or, la composition de ces communautés diffère

généralement significativement de celle des communautés des habitats adjacents dépourvus

de végétation (Reise, 2002 ; Bouma et al. 2009 ; Boström et Bonsdorf, 1997 ; Fredriksen et

al. 2010). A l’intérieur des herbiers, cette composition est également affectée par la densité de

la végétation (Blanchet et al. 2004). En accord avec ces derniers points, le remaniement

sédimentaire, en tant que caractéristique fonctionnelle d’une communauté donnée, varie très

certainement entre les différents types d’habitat. Une quantification du remaniement

sédimentaire dans un herbier de phanérogame et une zone non-colonisée adjacente, ou selon

un gradient de densité de la végétation, permettrait ainsi d’évaluer : (1) l’importance de la

macrofaune benthique des herbiers sur les dynamiques sédimentaires qui y prennent place, et

(2) les conséquences du déclin des herbiers sur le processus de remaniement sédimentaire.


[47]

V. Second objectif de la thèse

Au regard de la sévère régression de la surface du plus grand herbier à zostère naine

(Zostera noltii) d’Europe observée depuis les années 1980 dans le bassin d’Arcachon (Plus et

al. 2010), le second objectif de cette thèse de doctorat consistait, quant à lui, en une

évaluation de l’impact de cette régression sur l’intensité de remaniement sédimentaire induit

par les communautés benthiques. La seconde partie de ce manuscrit expose donc les résultats

d’une étude comparative de la structure des communautés benthiques et de l’intensité du

remaniement sédimentaire qui en résulte dans un herbier à Z. noltii ainsi que dans une zone

d’où celui-ci a récemment disparu.

CHAPITRE 2 :

Mesures expérimentales d’empreintes de remaniement sédimentaire du bivalve Abra alba (Wood)

[48]

Chapitre 2:

Mesures expérimentales

d’empreintes de remaniement

sédimentaire du bivalve Abra alba

(Wood)

CHAPITRE 2 :


[49]

Experimental assessment of particle mixing fingerprints in

the deposit-feeding bivalve Abra alba (Wood)

Guillaume Bernard1,2

, Antoine Grémare1, Olivier Maire

1, Pascal Lecroart

1, Filip J. R.

Meysman3, Aurélie Ciutat

4, Bruno Deflandre

1, Jean Claude Duchêne

4

1 UNIV. BORDEAUX, EPOC, UMR 5805, F33400 Talence, France

2 Corresponding author. email: [email protected]

3The Royal Netherlands Institute of Sea Research (NIOZ) Korringaweg 7, 4401 NT Yerseke,

The Netherlands

4 CNRS, EPOC, UMR 5805, F33400 Talence, France

Keywords: Particle mixing, Abra alba, CTRW model, Image analysis, Bioturbation

Running title: Particle mixing fingerprints in Abra alba

Journal of Marine Research, 70, 689-718, 2012.

CHAPITRE 2 :


[50]

Abstract

Particle mixing induced by the deposit-feeding bivalve Abra alba was assessed using a

new experimental approach allowing for the tracking of individual particle displacements.

This approach combines the adaptation of existing image acquisition techniques with new

image analysis software that track the position of individual particles. This led to

measurements of particle mixing fingerprints, namely the frequency distributions of particle

waiting times, and of the characteristics (i.e. direction and length) of their jumps. The validity

of this new approach was assessed by comparing the so-measured frequency distributions of

jump characteristics with the current qualitative knowledge regarding particle mixing in the

genus Abra. Frequency distributions were complex due to the coexistence of several types of

particle displacements and cannot be fitted with the most commonly used procedures when

using Continuous Time Random Walk (CTRW) model. Our approach allowed for the spatial

analysis of particle mixing, which showed: (1) longer waiting times, (2) more frequent

vertical jumps, and (3) shorter jump lengths deep in the sediment column than close to the

sediment-water interface. This resulted in lower DbX and Db

Y (vertical and horizontal particle

mixing bioffusion coefficients) deep in the sediment column. Our results underline the needs

for: (1) preliminary checks of the adequacy of selected distributions to the

species/communities studied, and (2) an assessment of vertical changes in particle mixing

fingerprints when using CTRW.

CHAPITRE 2 :


[51]

I. Introduction

Marine benthic macrofauna strongly affect the fate of settled particulate organic matter

(POM) through bioturbation (Meysman et al., 2006), which encompasses two distinct

processes: (1) bioirrigation (i.e, the enhanced exchange of water and solutes across the

sediment-water interface due to burrow ventilation), and (2) particle mixing (i.e., particle

movements due the activity of benthic fauna) (Kristensen et al., 2012). Bioirrigation enhances

the oxygenation of the sediment and thereby promotes the degradation of POM (Aller and

Aller, 1998). Conversely, particle mixing stimulates the transfer of POM to deeper anoxic

layers where organic matter degradation processes are less efficient (Kristensen 2000).

Particle mixing results from burrowing, feeding, defecation and locomotion (Meysman et al.,

2006). It occurs globally across the ocean floor, and is of particular importance in areas where

physical disturbance is low (Lecroart et al., 2010). Benthic communities change along

disturbance gradients (Pearson and Rosenberg, 1978) together with rates of particle mixing.

These changes generate a complex interplay between benthic fauna and both the quantity and

quality of organic matter, which strongly affects the physical, chemical and geotechnical

properties of marine sediments (Rhoads, 1974; Aller, 1982; Rhoads and Boyer, 1982;

Meadows and Meadows, 1991; Gilbert et al., 1995; Rowden et al., 1998; Lohrer et al., 2004).

Unravelling these interactions requires an improvement of the methods currently used for

quantifying particle mixing (Maire et al. 2008).

Particle tracer methods are the most widely used and all rely on the same steps: (1) the

deposition of tracer particles at the sediment-water interface, (2) the determination of tracer

vertical profiles within the sediment column, and (3) the computation of particle mixing rates

by fitting mixing models to those profiles. The most widely implemented model is the

biodiffusion model, which assumes that Fick’s first law of diffusion is applicable to tracer

dispersion (Guinasso and Schink, 1975; Boudreau, 1986a; Wheatcroft et al., 1992; Gérino et

al., 1998). It is easy to implement and results in a single parameter that quantifies the rate of

particle mixing: the biodiffusion coefficient (Db). The biodiffusion model has often proved

suitable to fit tracer profiles (Lecroart et al., 2007, 2010), which constitutes a paradox since

its main assumptions (i.e., non-oriented, extremely frequent and extremely small particle

displacements) are most often not fulfilled (Meysman et al., 2003).

CHAPITRE 2 :


[52]

More sophisticated models have been developed, that claim to have a stronger

biological background (e.g. Boudreau, 1986b; Robbins, 1986; Boudreau and Imboden, 1987;

Soetaert et al., 1996; François et al., 2002). These so-called non-local models are more

difficult to handle from a mathematical standpoint. It is also much more difficult to acquire

appropriate data to evaluate them. This explains why they have only been implemented

occasionally (Rice, 1986; Shull, 2001; Solan et al., 2004; Delmotte et al., 2007). Meysman et

al. (2008a, 2008b, 2010) have proposed the CTRW model to describe particle mixing. In this

model, the effect of mixing is assessed by tracking the elementary motion of individual

particles. Particle displacement is described as a random process, and is governed by three

stochastic variables: (1) the jump direction, (2) the jump length, and (3) the waiting time

between two consecutive jumps of the same particle (Wheatcroft et al., 1990). Overall, the

joined frequency distributions of these variables form the “mixing fingerprint” of a benthic

community or a benthic organism (Meysman et al., 2008a). Meysman et al. (2008b) have

developed a one-dimensional CTRW model, in which particle displacement is governed by

two frequency distribution functions describing the waiting times (typically a Poisson

process) and the vertical components of jump lengths (typically a Gaussian distribution). This

approach was shown to have advantages over the simple biodiffusion model in describing

tracer profiles generated by the bivalve Abra ovata (Maire et al., 2007a). It has also been used

successfully with the bivalve A. alba, the polychaete Nephtys sp. (Braeckman et al., 2010)

and the amphipod Corophium volutator (De Backer et al., 2011). The application of the

CTRW model however remains limited due to its mathematical complexity. Moreover, even

if it constitutes a progress relative to the biodiffusive model, the a priori selection of simple

functions to describe the frequency distributions of waiting times and jump lengths remains

unverified (Meysman et al., 2010).

Tracer profiles are classically determined by slicing sediment cores and subsequently

quantifying the tracer within each sediment layer. Artificial tracers such as glass beads and

luminophores are quantified using image acquisition and analysis techniques (Maire et al.,

2008). The combination of these approaches have been successfully used to quantify particle

mixing both by benthic communities (Gilbert et al., 2003; Solan et al., 2004) and individual

species (Maire et al., 2006, 2007a, 2007b; Piot et al., 2008). The use of transparent aquaria

allows for a 2D dynamic view of particle mixing (Maire et al., 2006, 2007a, 2007b). Recent

technological advancements (increase the frequency of image capture) allow for a direct

assessment of waiting time frequency distributions. Similar possibilities exist for jump lengths

and directions. Maire et al. (2007c) have developed a specific algorithm to assess the

CHAPITRE 2 :


[53]

displacements of sediment particles along the tentacles of the deposit-feeding polychaete

Eupolymnia nebulosa that can indeed be transposed to the analysis of luminophore

trajectories.

A major challenge in particle mixing research is to get accurate estimates of mixing

fingerprints (Reed et al., 2006; Maire et al., 2007a; Meysman et al., 2008a, 2008b). This has

not been achieved yet due to technical limitations. The aim of the present study is to adapt

existing image analysis techniques to experimentally assess the frequency distributions of: (1)

waiting times, (2) jump directions, and (3) jump lengths to describe sediment particle mixing

by the bivalve A. alba.

II. Materials and methods

2.1. Bivalve collection and maintenance

The deposit-feeding bivalve Abra alba belongs to the super family Tellinoidea. It is a

dominant macrobenthic species in shallow subtidal areas along the European Atlantic coast

(Borja et al., 2004; Van Hoey et al., 2005). It is abundant and a dominant species in the

Arcachon Lagoon where abundances can reach up to 500 individuals.m-2

and shell lengths are

typically between 5 and 15 mm (Blanchet et al., 2005). Its body is usually buried a few

centimetres below the sediment surface (Braeckman et al., 2010). It reworks the upper layer

of the sediment when feeding. Foraging movements typically consist of circular motions of

the tip of the inhalant siphon at the sediment-water interface (Hughes, 1975).

During the present study, sediment samples were collected in June 2010 and May 2011

in the Courbey Channel (45°43’476 N, 1°37’758 W, 3-5 m depth, Arcachon Bay, France)

using a Van-Veen grab. Samples were passed through a 1 mm mesh, yielding ~500 clams (9-

12 mm shell length). Additional grabs were passed through a 1 mm sieve to remove

macrofauna. This sediment (47.7 % sand and 52.3 % fines; 1.40 % POC and 0.16 % PON)

was used both for maintenance and experimentation. Clams were kept in tanks (60 x 40 x

30cm) filled with field sediment and supplied with ambient running seawater prior

experimentation. They were fed once a week with crushed Tetramin® fish food (4.59

gPOC.week-1

).

CHAPITRE 2 :


[54]

Figure 2.1: Lateral (A) and frontal (B) views of the setup used during the 8 experiments.

2.2. Experimental set-up

The experimental set-up (Fig. 2.1) was modified from Maire et al. (2006). Thin aquaria

(L = 17 cm, W = 0.9 cm, H = 33 cm) were filled with 15 cm of field sediment and kept at

ambient seawater temperature for 3 days before each experiment. Three clams of known size

were then gently placed at the sediment surface, after which they typically buried within 30

seconds. After 24 h, 1.5 g of yellow luminophores (Ecotrace, Environmental Tracing®,

CHAPITRE 2 :


[55]

median diameter = 35 µm) were spread at the sediment surface. Aquaria were placed in front

of two UV lights ( = 365 nm, which allowed for the distinction between fluorescent

luminophores and the surrounding sediment) and of a µeye video captor (IDS®, definition of

2560 x 1920 pixels). The monitored field was 4.2 cm x 3.2 cm, which resulted in a resolution

of 16.5 µm.pixel-1

. The experiment began 24 h after luminophore introduction. This allowed

for: (1) the monitored field to be centred on an area reworked by a single bivalve, and (2) the

dispersion of luminophores. Each experiment lasted 48 h and image frequency acquisition

was 0.1 Hz. The series of images collected during each experiment were assembled in an AVI

video format. We will report on the results of 8 replicated experiments. Seven were carried

out between 22 June and 24 September 2010 with temperatures between 19.2 and 21.3°C and

one on 16 June 2011 at a temperature of 21.2°C.

2.3. Image processing

AVI films were processed using two specific algorithms that track the position of

individual luminophores within consecutive images. The goal was to categorize the motion of

individual particles following the CTRW formalism. Only the movements of isolated

luminophores make sense in such an analysis. Isolated luminophores were first binarised in

each individual image based on their red-green-blue levels, luminance and size. The (XY)

coordinates of their barycentre was taken as representative for their position. Waiting times

and jump characteristics algorithms were both implemented in a specific software

(ObviousAVIexplore), which was developed using Microsoft Visual® / C#.

The algorithm that describes a waiting event is presented in Figure 2.2a. A given

particle “waits” if it does not move outside of a sensitivity circle. This sensitivity circle

accounts both for changes in the apparent size of the luminophores due to fluctuations in light

intensity, and for small movements due to vibrations. Here, we used a radius of 66 µm for the

sensitivity circle (4 pixels). This value was based on the apparent size of luminophores

including halos (typically ~50 µm radius or 3 pixels), and the displacements of “fixed”

reference spots that were painted on the wall of the aquaria. The waiting event algorithm

works as follows: imagine that at time t, a luminophore has been waiting for m time intervals.

If at time t+1, the position of the luminophore is still located within the sensitivity circle, we

assume that there has been no jump between time t and t+1, and the waiting event continues.

The (XY) coordinates are updated, and the waiting time is incremented (i.e., set to m+1).

Conversely, if the position of the luminophore at time t+1 is located outside the sensitivity

CHAPITRE 2 :


[56]

circle, the luminophore has jumped between time t and t+1. The waiting time event has ended

and its waiting time is determined as being m time intervals. When a waiting time event has

ended, the following parameters are recorded: the index number of the starting image

(corresponding to the arrival of the luminophore at its initial position), the index number of

the final image (corresponding to the jump of the luminophore to its new position), the

waiting time, and the (XY) coordinates of the luminophore in the final image.

The algorithm that analyses the jumps is presented in Figure 2.2b. It accounts for the

fact that a particle jump does not necessarily occur within the monitored plane, which

complicates the tracking of particles. A given particle “jump” can be characterized if the

considered particle moves outside of a sensitivity circle, but stays within a broader search

circle. The delineation of a search circle is needed to ensure that the same particle is tracked

in consecutive images. Here, we used a value of 660 µm for the radius of the search circle.

This value was determined in calibration tests on selected video sections, where we optimized

the automatic detection of luminophore movements by visually following particle

movements.

The jump event algorithm works as follows: imagine that a particle has jumped between

t and t +1 (i.e. the particle is present with the sensitivity circle at time t, but no longer present

within the sensitivity circle at time t+1). There are multiple possibilities at time t+1. If there is

no particle present between the sensitivity and search circles, then it is assumed that the

particle has disappeared from the wall and moved into the interior of the aquarium and no

trajectory analysis can be performed. If one particle is present between the sensitivity and

search circles, this is regarded as the new position of the particle (note than when several

particles are present within the search circle, we select the particle with the position closest to

the original location). We then can draw a vector V1 that links the positions at time t and time

t+1. The trajectory analysis starts. If the same luminophore moves during more than two

consecutive images, this is considered as a single jump. V1 forms the first element of the

corresponding trajectory. The analysis then continues at time t+2 with three alternative

possibilities:

(1) There is a particle within the sensitivity circle centred on the luminophore’s position

at time t+1. This is most likely the original particle which has not moved. The jump has

ended, and the duration of the jump is one time interval. The overall jump vector is the vector

V1.

CHAPITRE 2 :


[57]

Figure 2.2: Principles used for the computation of waiting times (A). The two centred grey

circles are the luminophore and its halo positions. B is the barycentre of the luminophore. The

open circle with a dotted line is the sensitivity circle (see text for details). Principles used for

the computation of jump lengths and directions (B). The two centred grey circles are the

luminophore and its halo positions. B is the barycentre of the luminophore. The open circle

with a continuous line is the search circle. The open circle with a dotted line is the sensitivity

circle (see text for details).

CHAPITRE 2 :


[58]

(2) There is no particle within the search circle centred on the luminophore’s position at

time t+1. It is assumed that the particle has disappeared from the wall and moved into the

aquarium. The trajectory analysis is stopped. The duration of the jump is one time interval

and the overall jump vector is vector V1.

(3) There is one particle located between the sensitivity and search circles centred on

the luminophore’s position at time t+1 (here again, when several particles are present within

the search circle, we select the particle with the position closest to the location at time t+1).

The jump is continuing. The displacement is described by the vector V2, which links the

positions of the particle at times t+1 and t+2. The duration of the jump is set to 2 time

intervals. The overall jump vector between times t and t+2 is computed as the sum of vectors

V1 and V2. The trajectory analysis then continues at time t+3 with the three above mentioned

possibilities.

When a jump event has ended, the following parameters are recorded: the index

numbers of the starting and final images, the duration of the jump, the (XY) coordinates at the

start of the jump, the (XY) coordinates at the end of the jump, the length and the direction of

the jump vector.

Jumps can last for more than a single time interval and jump lengths can therefore be

longer than the search radius. This approach, however potentially undersamples large jumps,

but has the benefits that it: (1) avoids false positives (i.e., characterization of jumps that have

not truly occurred), and (2) provides an objective way to track particle displacements. It

should also be underlined that the total number of jumps, the numbers of waiting times and

analysed jumps are not necessarily equal. A jump is indeed recorded whenever a luminophore

disappears from its sensitivity circle, whereas it can be analysed only if it remains within its

search circle between two consecutive images. Furthermore, the determination of a waiting

time requires that both the arrival and the departure times of a luminophore at a given position

are known.

2.4. Data processing

2.4.1. Frequency distributions

CHAPITRE 2 :


[59]

Overall, we were able to measure 3,062,006 waiting times and to characterize

1,462,337 jumps. For each experiment, these data (e.g. 570,249 waiting times and 284,361

characterized jumps for Exp. 3) were used to compute the overall frequency distributions of:

(1) waiting times, (2) jump lengths, (3) jump directions, (4) the lengths of the vertical and

horizontal components of jump vectors, and (5) jump durations.

2.4.2. Overall mean values

The number of jumps detected in a given area is depending not only on particle

mixing intensity but also on the number of luminophores present. The normalized (i.e.,

divided by the number of luminophores counted in the area) number of jumps represents the

probability of jump of a luminophore within the considered area. Our results showed that: (1)

all the components of particle mixing fingerprints vary with depth within the sediment

column, and (2) the vertical profiles of the total and normalized numbers of jumps strongly

differ. The assessment of the overall values of a component of particle mixing fingerprints

directly derived from the sets of measured jumps and waiting times would thus be biased due

to a “luminophore density effect”. The computations of overall mean values of the waiting

time, jump length characteristics, and particle tracking biodiffusion coefficients were therefore

based on a bootstrap procedure involving 1000 re-samplings stratified relative to depth. The

vertical frequency distributions of the number of resampled events were kept strictly identical

to those of the normalized numbers of jumps. The number of jumps per re-sampling was set

to limit oversampling and therefore varied according to parameters and experiments.

2.4.3. 2D Spatial analysis

The monitored sediment area was divided into cells of 660x660 µm (i.e. 40x40

pixels), thus resulting in a two-dimensional grid of 64 by 48 cells. For each cell, we

computed: (1) the total number of jumps, (2) the normalised number of jumps, (3) the mean

waiting time (Tc), (4) the mean length of the vertical and horizontal components of the jump

vectors, (5) the variances (X2 and Y

2) of the vertical and horizontal components of jump

vectors, and (6) the vertical/horizontal particle-tracking biodiffusion coefficients (DbX

and

DbY). Db

X was computed in each cell after Meysman et al. (2008a):

DbX

= x2/(2Tc)

CHAPITRE 2 :


[60]

A similar formula applied to the computation of DbY.

2.4.4. Vertical profiles

The sediment-water interface was determined based on the location (i.e., maximum X

coordinates within each cell column) of the cells where waiting times were recorded. It was

then flattened by vertically translating cell columns so that the sediment-water interface

corresponds to the first row of each column (Solan et al., 2004 ; Maire et al., 2006). The cell

X-position in the picture then corresponded directly to its depth within the sediment column.

We computed 1D vertical profiles by considering the subsets of individual jumps and/or

waiting times occurring within each 660 µm depth interval. This resulted in vertical profiles

of: (1) total numbers of jumps, (2) normalized numbers of jumps, (3) mean waiting times,

(4) proportions of perfectly vertical jumps (i.e., jumps with an origin and an endpoint within

the same pixel column), (5) means lengths of the vertical and horizontal components of

jump vectors, (6) X 2

and Y 2

, and (7) DbX and Db

Y. The effects of depth on all these

parameters were assessed using non-parametric Kruskal-Wallis ANOVAs. Differences

between vertical frequency distributions of the total number of jumps and of the normalized

number of jumps were assessed using a Kolgomorov-Smirnov test. For each experiment,

and based on their vertical profiles, the relationship between DbY and Db

X was assessed

using a simple linear correlation model.

All computations were carried out using specific routines in the open source R

programming framework v2.13.1. (http://www.R-project.org , 2011). Statistical analyses were

carried out using the SigmaStat® 11.0 software.

III. Results

3.1. Classification of luminophore movements

Through visual inspection of the 384 h video footage, we were able to distinguish

between 2 types of luminophore movements. The first one corresponded to an upward

displacement, immediately followed by a downward one. These vertical oscillations were

caused by the protrusion/retraction of the foot outside/inside the shell (Figs. 2.3a,b). They

http://www.r-project.org/

CHAPITRE 2 :


[61]

were also detected during inhalant siphon foraging and ejection of faeces or pseudofaeces via

the exhalant or the inhalant siphon, respectively (Figs. 2.3c,d). Since upward displacements

were immediately followed by downward ones, this typically resulted in overall jump vectors

with a length close to zero. The second type of particle movement was preferentially oriented

along the axis of siphonal galleries. Upward movements were caused by the extension of the

siphons, which induced luminophore jumps through friction (Fig. 2.3e). Similarly, downward

jumps were caused by siphon retraction (Fig. 2.3f). Corresponding jump lengths usually

differed from zero and their vertical components could be either negative (upward jumps) or

positive (downward jumps).

CHAPITRE 2 :


[62]

Figure 2.3: Description of the luminophore displacements induced by different types of

behaviours.

3.2. Frequency distributions

The frequency distribution of waiting times during Experiment 3 was dominated by

short values (i.e., 31.8% ≤ 1 minute, Fig. 2.4a). The right tail of this distribution was long and

the maximal waiting time was 47.8 h. For each experiment, these distributions were fitted

CHAPITRE 2 :


[63]

assuming a Poisson process (Table 2.1). r2 were between 0.119 (Exp. 7, p<0.0001) and 0.722

(Exp. 4, p<0.0001) and c between 0.32 (Exp. 7) and 1.16 h (Exp. 2).

The frequency distribution of jump directions during Exp. 3 (Fig. 2.4b) was polymodal.

Downward jumps were more frequent (56.5%) than upward ones (38.7%). Horizontal jumps

accounted for only 4.8%. Together with the vertical (both downward and upward), another

preferential jump direction was comprised between 180 and 225°, which corresponded to the

preferential orientation of the inhalant siphon within its gallery network during the monitored

period. The relative maxima observed at 45, 90, 135, 225, 270 and 315° corresponded to the

fact that they were, together with 0 and 180°, the only possible directions for luminophore

jumps of short (i.e., < 2 pixels) length.

Figure 2.4: Exp. 3. Frequency distributions of waiting times (A), jump directions (B), jump

lengths (C), and vertical components of lengths (D) (positive values are downward, negative

values are upward). The shaded area in B indicates a preferential jump orientation (see text

for details). Arrows indicate perfectly vertical downward (180°) and upward (0°) jumps. V

up: vertical upward, V down: vertical downward.

The frequency distribution of jump lengths during Exp. 3 was bimodal (Fig. 2.4c), with

modes corresponding to a jump length of 0.016 and 0.082 mm, respectively. The frequency

distribution of the vertical component of jump lengths during Exp. 3 was trimodal (Fig. 2.4d),

CHAPITRE 2 :


[64]

with a first mode (0.066 mm) corresponding to downward jumps, a second one (0 mm)

corresponding to a null vertical component of jump lengths (which can either correspond to

perfectly horizontal jumps or to jumps lasting for several time intervals with an overall

resulting vertical component of jump length of zero), and a third one (-0.066 mm)

corresponding to upward jumps.

Table 2.1: Fits of waiting times distributions assuming a Poisson process. r²: determination

coefficient, p: significance level, τc: fitted mean waiting (i.e., assuming a Poisson process for

waiting times).

Experiment Exp. 1 Exp. 2 Exp. 3 Exp. 4 Exp. 5 Exp. 6 Exp. 7 Exp. 8

r² 0.509 0.674 0.292 0.722 0.376 0.341 0.119 0.175

p <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001

τc (h) 0.73 1.16 0.92 0.65 1.11 0.93 0.32 1.39

The frequency distributions during the 7 other experiments were similar although the

relative magnitudes of modes slightly differed between experiments.

3.3. Overall mean values

Overall mean waiting times were between 0.86 (Exp. 4) and 2.15 h (Exp. 3) (Table 2.2),

which led to an overall mean of 1.75 h with a standard deviation of 0.41 h and a variation

coefficient of 23.5%.

X2 were between 0.009 (Exp. 4) and 0.033 mm² (Exp. 3) (Table 2.2), which led to an

overall mean of 0.024 mm² with a standard deviation of 0.007 mm² and a coefficient of

variation of 29.9%. Y2 were between 0.011 (Exp. 4) and 0.027 mm² (Exp. 6), which led to an

overall mean of 0.019 mm² with a standard deviation of 0.005 mm² and a coefficient of

variation of 26.8%).

DbX were between 0.486 (Exp. 1) and 0.707 cm

2.year

-1 (Exp. 8) (Table 2.2), which led

to an overall mean of 0.587 cm2.year

-1 with a standard deviation of 0.09 cm

2.year

-1 and a

variation coefficient of 15.0%. DbY were between 0.355 (Exp. 1) and 0.559 cm

2.year

-1 (Exp.

4), which led to an overall mean of 0.491 cm2.year

-1 with a standard deviation of 0.081

cm2.year

-1 and a variation coefficient of 16.5%.

CHAPITRE 2 :


[65]

3.4. 2D spatial analysis

The spatial distributions of luminophores at the beginning and end of Exp. 3 are shown

in Figures 2.5a and b, respectively. The initial image was taken 24 h after a layer of

luminophores was deposited at the sediment-water interface, and still showed large patches of

luminophores that have not been displaced. Most of these patches were dispersed during the

next 48 h, and the luminophores tended to be homogenously distributed at the end of the

experiment (Fig. 2.5b). Note however that the subsurface patches on the right side of image

remained almost unaffected, indicating an area that was not reworked by the siphon activity

of the clams. Also note that the sediment-water interface was not initially flat due to bivalve

activity during the 24 h preceding the experiment and then had a tendency to flatten out

during the course of the experiment.

The spatial distribution of the number of jumps during Exp. 3 showed that jumps: (1)

occurred over the whole Y axis of the monitored area, and (2) were scarcer at depth, where

they predominantly occurred in the area close to the bivalve (Fig. 2.6a). The spatial

distributions differed between the total and the normalized number of jumps although there

was also a clear tendency toward a decrease in the normalized number of jumps with

increasing depth within the sediment column (Fig. 2.6b, see also Fig. 2.9a).

The spatial distribution of waiting times during Exp. 3 is shown in Figure 2.6c. The

shortest waiting times were found at the sediment-water interface and in siphonal galleries,

while longer ones were found just below this interface in two distinct zones located outside

the galleries network on the left and on the right (Fig. 2.6c).

CHAPITRE 2 :


[66]

Tab

le 2

.2:

over

all

mea

n v

alues

of

wai

ting t

imes

, ju

mp c

har

acte

rist

ics,

Db

Y an

d D

bX d

uri

ng o

ur

8 e

xper

imen

ts (

see

tex

t fo

r det

ails

on t

he

com

pu

tati

on

pro

cedure

).

Dat

a fr

om

B

raec

km

ann

et

al.

(2

010)

are

pro

vid

ed

for

com

par

iso

n.

Tc:

w

aiti

ng

tim

e

X/Y

²:

var

ian

ce

of

ver

tica

l/hori

zonta

l co

mponen

ts o

f ju

mp l

ength

s, D

bX

/Y:

ver

tica

l/hori

zonta

l par

ticl

e-tr

ackin

g b

iodif

fusi

on

coef

fici

ent.

N

F:

No F

ood a

dded

.

Sp

ecie

s T

c (

h)

Y²

(mm

²)

Db

Y (cm

2.y

r-1

)

X²

(mm

²)

Db

X (cm

2.y

-1)

Ref

eren

ce

Ab

ra a

lba

(2

0.4

°C,

NF

, E

x 1

) 2

.12

0

.01

7

0.3

55

0.0

24

0.4

86

Pre

sent

stud

y

Ab

ra a

lba

(2

0.6

°C,

NF

, E

x 2

) 1

.78

0.0

18

0.4

47

0.0

24

0.5

88

Pre

sent

stud

y

Ab

ra a

lba

(2

1.3

°C,

NF

, E

x 3

) 2

.15

0

.02

6

0.5

38

0.0

32

0.6

68

Pre

sent

stud

y

Ab

ra a

lba

(2

0.7

°C,

NF

, E

x 4

) 0

.86

0

.01

1

0.5

59

0.0

09

0.4

90

Pre

sent

stud

y

Ab

ra a

lba

(2

0.4

°C,

NF

, E

x 5

) 1

.84

0.0

18

0.4

37

0.0

24

0.5

70

Pre

sent

stud

y

Ab

ra a

lba

(1

9.2

°C,

NF

, E

x 6

) 1

.99

0

.02

7

0.5

91

0.0

30

0.6

73

Pre

sent

stud

y

Ab

ra a

lba

(1

9.8

°C,

NF

, E

x 7

) 1

.68

0.0

17

0.4

46

0.0

20

0.5

11

Pre

sent

stud

y

Ab

ra a

lba

(2

1.2

°C,

NF

, E

x 8

) 1

.58

0.0

20

0.5

52

0.0

25

0.7

07

Pre

sent

stud

y

Ab

ra a

lba

(1

0°C

, N

F,

den

sity

:38

2.m

-2)

- -

- -

1.7

3-1

.82

Bra

eck

man

n e

t a

l. (

20

10

)

Ab

ra a

lba

(1

0°C

, N

F,

den

sity

:76

4.m

-2)

- -

- -

2.3

6-3

.63

Bra

eck

man

n e

t a

l. (

20

10

)

Ab

ra a

lba

(1

0°C

, N

F,

den

sity

: 1

27

3.m

-2)

- -

- -

4.4

7

Bra

eck

man

n e

t a

l. (

20

10

)

Ab

ra a

lba

(1

8°C

, N

F,

den

sity

:38

2.m

-2)

- -

- -

0.9

6

Bra

eck

man

n e

t a

l. (

20

10

)

Ab

ra a

lba

(1

8°C

, N

F,

den

sity

:76

4.m

-2)

- -

- -

1.9

8-3

.40

Bra

eck

man

n e

t a

l. (

20

10

)

Ab

ra a

lba

(1

8°C

, N

F,

den

sity

: 1

27

3.m

-2)

- -

- -

3.4

4-4

.22

Bra

eck

man

n e

t a

l. (

20

10

)

CHAPITRE 2 :


[67]

Figure 2.5: Exp. 3. Photographs showing the 2D spatial distributions of luminophores at the

beginning (A) and the end of the experiment (B). Arrows indicate remains of the layer of

luminophores initially deposited at the sediment-water interface (A) and an unaffected blob of

luminophores (B). Circles indicate the location of the shell and red lines show the location of

the two preferentially used galleries (B). See text for details.

The spatial distributions of: (1) the absolute values of vertical components of jump

vectors, (2) the absolute values of horizontal components of jump vectors, (3) X², and (4) Y²

are shown in Figures 2.7a, 2.7b, 2.7c and 2.7d, respectively. Absolute values of the vertical

components of jump vectors were maximal close to the sediment-water interface and then

progressively decreased with depth to reach a minimal value close to the location of the shell

(Fig. 2.7a). Absolute values of vertical components of jump vectors, together with X² and

Y² followed the same pattern (Figs. 2.7b, c and d, respectively). This general gradient was

altered by the presence of two preferential siphonal galleries characterized by larger vertical

and/or horizontal components of jump vectors and higher X² and Y² than surrounding

sediment. DbX and Db

Y (Figs 2.8a, b) were high close to the sediment-water interface and

within the preferential siphonal galleries. Conversely, they were low in the area surrounding

the shell and outside the network of siphonal galleries.

CHAPITRE 2 :


[68]

Figure 2.6: Exp. 3. 2D maps of the numbers of jumps (A), the normalized numbers of jumps

(B), and waiting times (C). White areas correspond to cells where no event was detected.

CHAPITRE 2 :


[69]

Figure 2.7: Exp. 3. 2D maps of: means of absolute vertical (A) and horizontal (B)

components of jump lengths; X2 (C) and Y

2 (D). Arrows indicate the localisation of the

most active siphonal galleries. White areas correspond to cells where no event was detected.

Figure 2.8: Exp. 3. 2D maps of DbX (A) and Db

Y (B). White areas correspond to cells where

no Db could be computed due to the lack of detected events.

CHAPITRE 2 :


[70]

3.5. Vertical profiles

Figure 2.9: Exp. 3. Vertical profiles of: the relative frequencies numbers and normalized

numbers of jumps (A), waiting times (B), and proportions of perfectly vertical jumps (C).

Horizontal bars are 95% confidence intervals.

CHAPITRE 2 :


[71]

Figure 2.10: Exp. 3. Vertical profiles of: means of absolute vertical (A) and horizontal (B)

components of jump lengths, X2 (C) and Y

2 (D). Horizontal bars are 95% confidence

intervals.

The vertical frequency distributions of the total number of jumps and of the normalized

number of jumps are shown in Figure 2.9a. These two distributions differed significantly

(Kolmogorov-Smirnov test, p<0.001). Total number of jumps showed a sharp maximum 5

mm deep in the sediment column, whereas normalized number of jumps tended to be higher

within the first 5 mm below the sediment-water interface. Both parameters then decreased to

reach almost null values 12 mm deep in the sediment.

During Exp. 3, all the parameters listed below were significantly affected by depth

within the sediment column (non-parametric Kruskal-Wallis ANOVAs, p<0.001 in all cases).

CHAPITRE 2 :


[72]

Waiting times first tended to increase with depth within the sediment column to reach a

maximum of ca. 3 h at a 2 mm depth (Fig. 2.9b). They then progressively decreased to ca. 2 h

10 mm deep in the sediment. Values below 10 mm were characterized by large confidence

intervals resulting from the low number of jumps deep in the sediment.

The proportion of perfectly vertical jumps tended to increase with depth within the

sediment column (Fig. 2.9c). It was about 3 % close to the sediment interface versus 12-14 %

deep in the sediment. Absolute values of both the vertical and the horizontal components of

jump vectors tended to decrease with depth (Figs. 2.10a,b). Vertical components decreased

from 0.260 close to the sediment-water interface to ca. 0.080 mm deep in the sediment (Fig.

2.10a), versus ca. 0.220 to 0.060 mm for horizontal ones (Fig. 2.10b). Corresponding

confidence intervals were small within the 2-10 mm depth range compared with those at the

immediate vicinity of the sediment water interface and deep in the sediment. X² and Y²

decreased from 0.031and 0.025 close to the water-sediment interface to 0.005 and 0.004 mm²

deep in the sediment, respectively (Figs. 2.10c, d).

Figure 2.11: Exp. 3. Vertical profiles of Db

X (A) and Db

Y (B).

DbX and Db

Y also decreased with depth (from 1.01 and 0.99 close to the sediment-water

interface to 0.02 and 0.08 cm2.year

-1 deep in the sediment, respectively) (Figs 2.11a, b). Db

X

and DbY correlated significantly during all experiments (Table 2.3).

CHAPITRE 2 :


[73]

IV. Discussion

4.1. Validation of the approach

To our knowledge, this is the first time that particle mixing fingerprints are

experimentally assessed in a marine benthic invertebrate. Our results can therefore only be

compared with: (1) qualitative knowledge regarding particle mixing induced by bivalves of

the genus Abra (Grémare et al., 2004; Maire et al., 2006, 2007a, 2007b), and (2) assessments

of c and DbX derived from the application of CTRW models to vertical luminophore profiles

recorded during experiments involving Abra alba (Braeckmann et al., 2010).

The frequency distribution of the vertical component of particle jump lengths in A. alba

is typically trimodal, with the first component oriented upward, the second oriented

downward and the third one with a jump length close to zero. This result is in good agreement

with the two kinds of luminophore movements identified by Maire et al. (2006) in both A.

ovata and A. nitida. These authors showed that these two bivalves induce particle jumps

through: (1) inhalant siphon displacements, and (2) shell motions due to the

extension/retraction of the foot outside/inside the shell. During the present study, we observed

an almost similar pattern in A. alba. The first kind of jumps is caused by the friction of the

luminophores onto the siphon. Corresponding jumps can be either upward in case of siphon

extension or downward in case of siphon retraction. Conversely, the jumps associated with

foot extension/retraction have an almost zero length. During the present study, luminophore

oscillations were also detected during maximal siphon extensions while foraging (inhalant

siphon) or producing faeces (exhalant siphon), which creates temporary tensions on the

sediment surrounding siphonal galleries. Also, slightly more complex, the pattern observed

for A. alba is therefore coherent with the observed trimodal frequency distributions of the

vertical component of jumps, which can be interpreted as resulting from the coexistence of:

(1) shell induced oscillations and movements induced by siphon tension on the sediment, and

(2) downward, and upward movements caused by siphons. This interpretation is further

confirmed by the decrease of the mean absolute value of the vertical component of jump

length with depth within the sediment column. This decrease indeed corresponds to an

increase in the frequency of jumps with a null vertical length component at depth (i.e., in

areas mostly affected by shell-associated jumps). Overall our results are thus consistent with

the current qualitative knowledge regarding particle mixing in the genus Abra, which supports

the validity of our assessments of jump characteristics.

CHAPITRE 2 :


[74]

Our assessment of particle mixing fingerprints can be compared with those derived

from the fit of CTRW models to vertical luminophore profiles during experiments involving

A. alba (Braeckman et al., 2010). These authors reported DbX (i.e., 1D vertical Db) between

0.96 and 4.47 cm2.y

-1 vs. only 0.355 and 0.707 cm

2.y

-1 during our own experiments. Both

ranges are low relative to literature data (Maire et al., 2006, 2007a,b), which supports the

apparently low intensity of particle mixing during our experiments (GB, personal

observation). Our DbX values are slightly lower than those of Braeckman et al. (2010). Direct

comparisons between the two studies are however complicated by several confounding

factors. An important difference between the two experimental setups is the timing of the

beginning of the experiment relative to luminophore introduction. Braeckman et al. (2010)

started their experiments immediately after luminophore addition versus 24 h later during the

present study. Maire et al. (2007) emphasized that, due to the filling of large burrows or/and

galleries, luminophores can be transferred to deep sediment layers immediately after their

deposition at the sediment-water interface. They suggested that this could partly contribute to

overestimates DbX during short-term experiments. Our own estimates of Db

X are not affected

by this bias, because: (1) they are based on experiments starting 24 h after luminophore

introduction, and (2) they are derived from the statistical analysis of individual jumps.

Consequently, the non-local (and non-directly associated with particle mixing) transfer of

luminophores to deep sediment layers during their introduction at the sediment-water

interface could explain some of the differences in the DbX recorded by Braeckman et al.

(2010) and during the present study.

Braeckman et al. (2010) reported no significant effect of temperature but a significant

effect of density. They controlled this last factor by manipulating the number of clams within

10 cm internal diameter cores whereas we monitored luminophore displacements over a much

smaller spatial scale typically corresponding to the portion of sediment reworked by a single

clam. It is therefore difficult to assess the exact clam density during our experiments and

discrepancies in the control of this factor may partly account for differences between the two

studies.

While modelling vertical sediment profiles with CTRW models, Braeckman et al.

(2010) used several assumptions regarding the frequency distributions of jumps lengths

(Gaussian distribution) and waiting times (Poisson process). Our own results suggest that

these two distributions are not fully adequate to describe particle mixing in A. alba (see

below). Our results also show that the mean values of jump lengths and waiting times clearly

decrease with depth within the sediment column. Meysman et al. (2008) showed that the

CHAPITRE 2 :


[75]

types of a priori selected frequency distributions have a significant impact on the modelling

of luminophore profiles and thus possibly on DbX values. Here again, this could account for

differences in DbX between our study and the one of Braeckman et al. (2010).

Another point is linked to the discrepancies between temporal scales classically

associated with DbX measurements through modelling of luminophore profiles (i.e., 14 d for

Braeckman et al. 2010) and with our approach (i.e., 10 s). Our results confirm the occurrence

of oscillating jumps with an overall length close to 0 as previously shown in other species of

the genus Abra (Grémare et al., 2004; Maire et al., 2006). These jumps do not significantly

affect luminophore profiles concentrations over temporal scales longer than their duration.

They therefore do not significantly affect the assessment of DbX through the modelling of

these profiles (De Baker et al., 2011). Conversely, taking into account oscillating jumps do

affect our own estimates of DbX, which could contribute to differences between our results

and those of Braeckman et al. (2010).

Overall, the good compatibility of the frequency distributions of the vertical component

of particle jumps recorded during the present study with the current knowledge regarding the

ethology of the genus Abra supports the validity of our approach. The comparison of our DbX

with those derived by Braeckman et al. (2010) is less conclusive due to the existence of

several confounding factors, which may account for differences between the two studies.

4.2. limitations of the approach

Our approach is based on the use of two different algorithms, both relying on specific

assumptions, which each induce limitations.

The first assumption, common to both algorithms, is linked with the use of thin aquaria

coupled with image analysis techniques. It assumes that the information captured within the

monitored plane (2D) is representative of processes occurring in the 3D sediment column.

This assumption is most questionable for the assessment of jump characteristics. “Wall

effects” may modify the trajectory of some jumps by constraining them to take place within

the monitored plane. One can therefore not exclude a bias linked to the subsampling (i.e.,

within the monitored plane) of the jumps occurring in 3D within the sediment column. We

however believe that this bias is weak for particle mixing induced by the genus Abra because,

despite of much higher DbX, Maire et al. (2006) have shown that luminophore profiles after 48

CHAPITRE 2 :


[76]

h of experiment do not differ along the walls and within the whole sediment columns of thin

aquaria containing A. ovata.

Another bias regarding waiting time measurements is the possibility of replacement of a

luminophore by another luminophore, which would not be considered as a jump and can lead

to an overestimation of waiting times. This probability can be approximated by the ratio of the

surface occupied by luminophores to the total surface of sediment. During our experiments,

this ratio was always less than 1%. Future experimentation should nevertheless consider the

best threshold in luminophore concentrations: (1) to allow for the assessment of a high-

enough number of jumps, and (2) not to bias the assessment of waiting times. One possibility

consists in using multi-color tracers, which would allow to keep a similar overall luminophore

concentration and to reduce the probability of replacement of a luminophore by another

luminophore of the same color.

Image frequency acquisition and overall experiment duration are key parameters when

assessing waiting times. it is not possible to detect waiting times shorter than the time interval

between the acquisitions of two consecutive images (i.e., 10 s). This may result in an over-

estimation of mean waiting times. There are however good rationale to believe that this bias is

negligible. In the genus Abra, particle mixing is mostly resulting from siphonal activity

(Maire et al., 2007b) and Grémare et al. (2004) have shown that a frequency acquisition of

0.05 Hz is sufficient to fully (i.e., >98%) describe this activity in A. ovata. Based on our own

observations (OM and AG), particle mixing is clearly higher in A. ovata than in A. alba. We

therefore believe that the image frequency acquisition used during the present study was

appropriate to fully describe particle mixing in A. alba.

It is also not possible to measure a waiting time longer than overall experiment duration.

The lack of recording of a waiting time can therefore be indicative of: (1) the total absence of

jumps (i.e., the total absence of particle mixing), and/or (2) the impossibility of sampling

waiting times longer than experiment duration. Even for shorter waiting times, there is a bias

towards the sampling of short waiting times using our approach. The probability that both the

start and the end of a waiting period occur during a temporal window of a given duration

indeed decreases with increasing waiting times. Although the number of long waiting times is

clearly low in A. alba, this may lead to an underestimation of waiting times. During

preliminary trials, we assessed the effect of experiment duration on the measurement of

waiting times by comparing mean waiting times obtained by resampling the same experiment

during overall durations comprised between 1 and 48 h. Our results showed that mean waiting

times: (1) tended to increase with experiment duration, (2) usually did not reach an asymptote

CHAPITRE 2 :


[77]

within 48 h, and (3) were affected by temporal changes in bivalve’s activity. Our waiting time

measurements were thus indeed affected by experiment duration. It should be stressed that

such a dependency is also affecting other approaches currently used to assess particle mixing

(Meysman et al., 2010). Anyhow, experiment duration should carefully be adjusted when

using our approach and a special care must be taken to carry out long-term experiments when:

(1) particle mixing is low, and/or (2) not constant relative to time.

The assessment of jump characteristics requires that a luminophore remains visible

within the monitored plane during the entirety of its jump. This is clearly not the case in a

large variety of sediment remixing modes, including non-local movements associated with

feeding and/or particle transfers within animals’ guts or even biogenic structures such as

burrows and tubes. The approach described in the present study is not adequate for those

types of particle mixing, which include both downward and upward conveyors (François et

al., 1997, 2002).

Even when the luminophore remains visible during the entirety of its jump, our

approach requires that its consecutive positions remain close enough for the reconstruction of

its trajectory. The algorithm used for the assessment of jump characteristics is indeed using

two circles (i.e., the sensitivity and the search circles) centred on the initial position of the

luminophore. The reconstruction of luminophore movements requires that the second of two

consecutive positions of the luminophore is located outside the sensitivity and inside the

search circle. For quick non local displacements, this can be achieved by increasing the size of

the search circle and/or image frequency acquisition. Increase in the size of the search circle is

limited because it would result in too many luminophores to be included (i.e., too many

possible ending points of elementary displacements), which would complicate trajectory

assessment. Increasing image frequency acquisition would result in much larger image files,

not compatible with our current possibilities of image processing. In this last case, it is

however possible to subsample these files and to derive the distributions regarding the

characteristics of jump length from the analysis of these subsamples. There is indeed no a

priori reason to suspect that the so-determined distributions would not be representative of

those which could have been derived from the analysis of the whole file.

Overall, the approach presented during the present study is suitable to assess particle

mixing in animals moving sediment particles frequently and over relatively short distances (or

over larger distances but slowly). It therefore appears mainly appropriate for the biodiffusors

and the gallery-diffusors groups, and is possibly transposable to regenerators (François et al.,

1997, 2002). Because of such restrictions in its applicability, this approach does not seem

CHAPITRE 2 :


[78]

appropriate to assess particle mixing by benthic communities as a whole. The direct

extrapolation of our results to field populations may itself prove difficult due to density

(Braeckman et al. 2010), individual size (Wikander 1980) and temperature (Maire et al.

2007c) in the genus Abra. Conversely, our approach is highly suitable to assess the effects of

these factors and of their interactions during laboratory experiments, which would allow for

the indirect assessment of sediment mixing by field populations.

4.3. Particle mixing fingerprints in Abra alba

4.3.1. Vertical and horizontal components of particle mixing

The positive correlations between DbX and Db

Y suggest that the vertical and horizontal

components of particle mixing are both cued by the same process, namely: clam activity as

already shown by Maire et al. (2007b) for A. ovata. Corresponding slopes were between 0.45

and 1.05 with a mean of 0.69, which suggests that the horizontal and the vertical components

of particle mixing induced by A. alba are of the same order of magnitude. This observation is

consistent with the rough estimates derived by Wheatcroft et al. (1990), which led these

authors to state that the horizontal components of particle mixing induced by benthic

invertebrates exceed vertical ones. Based on the relationships between DbX and Db

Y and the

direction of particle displacements proposed by Wheatcroft et al. (1990), this also suggests

that, in A. alba the mean angle of particle jump direction with the vertical should be close to

40° (45° during Exp. 3). Interestingly, during this last experiment, we recorded a preferential

range of jump direction between 180 and 270°, centred around 225°, which corresponds to

downward jumps oriented 45° from the vertical. This pattern mostly resulted from the

orientation of the siphonal galleries preferentially used during the experiment. Over short time

scales, changes in the ratio between horizontal and vertical components of sediment

reworking thus appears to be mainly controlled by the orientation of preferential siphonal

galleries. Our own observations show that the position of the siphon of A. alba changes while

exploring different sectors of the sediment. Therefore, over longer time scales, changes in this

ratio would probably be mainly controlled by the general geometry of the siphonal galleries

network. In this last case, one would expect this ratio to become less variable between

individuals/experiments, which could be tested by carrying out longer-term experiments.

CHAPITRE 2 :


[79]

4.3.2. Spatial heterogeneity of particle mixing fingerprints

Particle mixing induced by A. alba was spatially heterogeneous. The analysis of the

spatial distribution of waiting times during Exp. 3 allows for the distinction between 4 areas:

(1) The sediment-water interface, where particle mixing is high and mostly consists in

sediment movements induced by the distal part and the tip of the inhalant siphon. This area is

characterized by the shortest values of waiting times and the highest values of both the

horizontal and vertical components of jump vectors. It is also characterized by the highest

normalized number of jumps. Its overall thickness is ca. 2.5mm.

(2) The network of siphonal galleries, which has a conical shape and an overall volume

of ca. 2700 mm3 resulting in a surface of ca. 200 mm

2 in the monitored plane. This area is

characterized by intermediate and highly variable waiting times and jump lengths resulting

from the preferential use of two siphonal galleries during the period under study. Overall,

particle mixing is highly heterogeneous within this area.

(3) Two subsurface areas located outside the network of siphonal galleries, which are

characterized by very long waiting times. These two areas are ca. 12mm2 and are located on

both sides of the monitored field. Based on experimental data by Maire et al. (2007a),

Meysman et al. (2008a) stated that the vertical luminophore profiles generated by A. ovata

consisted in two separate zones: an upper “blocky zone” where particles movements are few

due to restricted access by the inhalant siphon, and a lower “smooth zone” where particles

movements due to siphon activity take place. According to Meysman et al. (2008b) these

“blocky zones” are due to lateral heterogeneity in particle mixing. During the present study,

we have worked at a small spatial scale (i.e., a few cm) typically associated with a single

bivalve. Our results nevertheless show the occurrence of subsurface areas characterized by

low particle mixing activity, which could be considered as the extension of interface “blocky

zones” (Meysman et al., 2008a).

(4) The area immediately surrounding the shell, which is characterized by low

horizontal components of jump vectors and low variability in both the horizontal and vertical

components of jump vectors. This corresponds to the strong dominance of almost vertical

jumps resulting from the protrusion/retraction of the foot inside/outside the shell.

The relative importance of these four areas and associated particle mixing processes

varies with depth within the sediment column, which induces vertical changes in particle

mixing fingerprints. The vertical profile of mean waiting times is characterized by a

CHAPITRE 2 :


[80]

subsurface maximum in relation with: (1) high particle mixing activity at the sediment-water

interface due to intense foraging by the inhalant siphon (see point 1 above), and (2) the

occurrence of the two subsurface areas located outside the siphonal gallery network and

characterized by almost null particle mixing activity (see point 3 above). Conversely, the

vertical profiles of mean absolute values of jump vectors and of X2 and Y

2 are all

characterized by maximal values at the sediment-water interface and then a continuous

decline with depth. This corresponds to the occurrence of: (1) long jumps associated with

siphon movements close to the sediment-water interface, and (2) a decrease in the variety of

jumps with depth in the sediment column. The vertical profiles of DbX and Db

Y are also

characterized by maximal values (ca. 10 times higher than 2 cm deep in the sediment) at the

sediment-water interface and then by a continuous decrease wit depth within the sediment

column. They are therefore much more similar to the vertical profiles of X2 and Y

2 than to

the vertical profile of waiting times. This suggests that vertical changes in DbX and Db

Y are

more cued by changes in jump characteristics than waiting time.

4.4. Consequences for the use of CTRW models in Abra alba

4.4.1. Suitability of the distributions classically used in CTRW to describe

particle mixing in A. alba

CTRW models are currently used to assess DbX (through mean waiting times and

2)

based on their fits to vertical luminophore profiles. This approach requires that the frequency

distributions can be described by simple functions and characterized by a limited number of

parameters. The most often used functions for describing the frequency distributions of

waiting times and jump lengths are the Poisson process and the Gaussian distribution (Maire

et al., 2007a; Braeckman et al., 2010; De Backer et al., 2011). The question of the choice of

these functions has been recently discussed by Meysman et al. (2008b). This selection is

made a priori without quantitative information regarding their pertinence to particle mixing

behaviour. To our knowledge, the present study is the first one, which allows for the direct

assessment of those distributions.

Our results support the “long right tail shape” of waiting times distributions, proposed

by Meysman et al. (2008a, 2008b, 2010). However, they also show that the fit of the Poisson

process to the frequency distributions of waiting times is quite approximate (Table 2.1). In

spite of the undersampling of long waiting times (see above), the frequency distributions of

CHAPITRE 2 :


[81]

waiting times were indeed less dominated by very short waiting times than those derived

using the Poisson process. Moreover, τc derived from fitting according to a Poisson process

were always smaller than directly calculated Tc. Our results therefore suggest that the Poisson

process is not fully suitable to describe the frequency distribution of waiting times in A. alba.

The vertical jump length frequency distributions recorded during the present study

exhibited three modes, each associated with specific shell or siphon movements (see above).

This is rather different from the Gaussian distribution generally assumed in CTRW (Meysman

et al. 2008a, 2008b, 2010) and proposed by Meysman et al. (2008b) for a “burrowing”

bivalve. These authors also proposed a more complex (i.e., with two modes) frequency

distribution for a “deposit feeding worm”. By simple combination of these two distributions,

one could thus expect that a benthic community composed of these two types of organisms

would induce a trimodal frequency distribution of vertical components of jump vectors. Our

results show that such a complex distribution can also result from different particle

movements induced by a single species. In any case, the Gaussian distribution is not fully

suitable to describe the frequency distribution of the vertical component of jump vectors in A.

alba.

Overall, it appears that the particle mixing fingerprints experimentally measured during

the present study do not follow the CTRW model most often used distributions. Further

applications of the CTRW model should therefore include preliminary checks of the adequacy

of these distributions to the species/communities studied.

4.4.2. Taking into account spatial heterogeneity when modelling particle

mixing in A. alba

The 1D-CTRW mixing model assumes that the frequency distributions of waiting times

and jump characteristics do not vary spatially (Meysman et al., 2008a). Our results show that

this is not the case in A. alba. Lateral heterogeneity in particle mixing fingerprints mainly

results from the occurrence of areas characterized by restricted access to the inhalant and

exhalant siphons. According to our observations, the preferential localisations of the siphons

are changing with time. Our results thus support the hypothesis of Meysman et al. (2008b),

who stated that, due to bivalve’s relocation, lateral heterogeneity in particle mixing should

become negligible with increasing experiment duration. One way of handling lateral

variability would thus be to use experimentally determined particle mixing fingerprints

CHAPITRE 2 :


[82]

provided that they are derived from long-enough experiments to result in horizontally

homogeneous particle mixing.

As far as vertical heterogeneity is concerned, our results show: (1) a clear increase of

waiting times with depth within the sediment column, (2) more frequent vertical jumps deep

in the sediment column, and (3) shorter jump lengths deep in the sediment column. These

patterns resulted in lower DbX as depth within the sediment column increases. This is linked to

the morphology and the ethology of A. alba and there is therefore no reason to believe that

vertical heterogeneity would diminish with experiment duration. Taking into account vertical

heterogeneity in CTRW models would therefore require to introduce a discretisation relative

to depth and to use different particle mixing fingerprints for each of the so-defined depth

intervals.

4.4.3. Consequences on the use of CTRW models

An important point consists in assessing to what extent the weaknesses of the

frequency distributions classically used in CTRW models indeed affect the assessment of

particle mixing fingerprints. In A. alba, this could be achieved by: (1) running a “classical”

CTRW model on our experimental data and comparing its outputs with our own overall mean

DbX, and (2) comparing the relative ability of a “classical” and an “experimentally based”

CTRW model to describe our experimental data. Depending on the outcome, one could

conclude on the validity of future use of classical CTRW models to assess particle mixing in

A. alba. In case of similar outputs, this conclusion will not necessarily hold for other

organisms, especially those belonging to other particle mixing groups. In case of different

outputs, one will have to unravel the effects linked to the selection of distributions and spatial

heterogeneity. In any case, this will make the use of CTRW to infer particle mixing hazardous

due to: (1) the necessary a priori knowledge regarding the shape of (simple enough)

distributions, and/or (2) the level of complexity associated with the description of spatial

(mostly vertical) heterogeneity.

4.5. Temporal dynamics and possible coupling with other

innovative approaches

CHAPITRE 2 :


[83]

Although not fully developed during the present study, our approach allows for the

assessment of particle mixing dynamics in 2D, which opens the field for a coupling with other

emerging techniques for the study of biogeochemical processes taking place at the sediment-

water interface. This could for example include the coupling with other non-invasive 2-D

imaging systems such as planar optodes (Volkenborn et al. 2012) or Diffusion Gradient Thin

gels (Teal et al. 2012) to better assess how the interactions between ethology, particle mixing,

bioirigation, oxygen and metals spatio-temporal distributions are controlled by environmental

factors such as temperature and POM availability.

Acknowledgments

This work is part of the PhD thesis of G Bernard (University Bordeaux 1). He was

supported by a doctoral grant from the French “Ministère de l’Enseignement Supérieur et de

la Recherche”. This work was funded through the BIOMIN (LEFE-CYBER and EC2CO-

PNEC), the “Diagnostic de la Qualité des Milieux Littoraux” and the « OSQUAR » (Conseil

Régional Aquitaine) programs.

References

Aller, R. C. 1982. The effects of macrobenthos on chemical properties of marine sediment

and overlying water, in Animal-Sediment Relations, P. L. McCall and P. J. S. Tevesz, ed.,

Plenum, New York, 53-102.

Aller R. C. and J. Y. Aller. 1998. The effect of biogenic irrigation intensity and solute

exchange on diagenetic reaction rates in marine sediments. J. Mar. Res., 52, 905-936.

Backer De, A., F. Coillie Van, F. Montserrat, P. Provoost, C. Colen Van, M. Vincx and S.

Degraer. 2011. Bioturbation effects of Corophium volutator: Importance of density and

behavioural activity. Est. Coast. Shelf. Sci., 91, 306-313.

Blanchet, H., X. de Montaudouin, P. Chardy and G. Bachelet. 2005. Structuring factors and

recent changes in subtidal macrozoobenthic communities of a coastal lagoon, Arcachon

Bay (France). Est. Coast. Shelf. Sci., 64, 561-576.

CHAPITRE 2 :


[84]

Borja, A., F. Aguirrezabalaga, J. Martínez, J. C. Sola, L. García-Arberas, J. M. Gorostiaga.

2004. Benthic communities, biogeography and resources management, in Oceanography

and marine environment of the Basque Country, A. Borja, M.Collins M, ed., Elsevier

Oceanogr. Ser., 70, 455-492.

Boudreau, B. P. 1986a. Mathematics of tracer mixing in sediments: I. Spatially-dependent,

diffusive mixing. Am. J. Sci., 286, 161-198.

Boudreau, B. P. 1986b. Mathematics of tracer mixing in sediment: II. Non-local mixing and

biological conveyor-belt phenomena. Am. J. Sci., 286, 199-238.

Boudreau, B. P. 1997. A one-dimensional model for bed-boundary layer particle exchange. J.

Mar. Syst., 11, 279-303.

Boudreau, B. P. and D. M. Imboden. 1987. Mathematics of tracer mixing in sediments. III.

The theory of nonlocal mixing within sediments. Am. J. Sci., 287, 693-719.

Braeckman, U., P. Provoost, B. Gribsholt, D. Van Gansbeke, J. J. Middelburg, K. Soetaert,

M. Vincx and J. Vanaverbeke. 2010. Role of macrofauna functional traits and density in

biogeochemical fluxes and bioturbation. Mar. Ecol. Prog. Ser., 399, 173-186.

Choi, J., F. François-Carcaillet and B. P. Boudreau. 2002. Lattice-automaton bioturbation

simulator (LABS): implementation for small deposit feeders. Comput. Geosci., 28, 213-

222.

Delmotte, S., F. J. R. Meysman, A. Ciutat, A. Boudou, S. Sauvage and M. Gérino. 2007.

Cadmium Transport in Sediments by Tubificid Bioturbation: an Assessment of Model

Complexity. Geochim. Cosmochim. Acta., 71, 844-862.

François, F., J. C. Poggiale, J. P. Durbec and G. Stora. 1997. A new approach for the

modelling of sediment reworking induced by a macrobenthic community. Acta. Biotheor.,

45, 295-319.

François, F., M. Gérino, G. Stora, J. P. Durbec and J. C. Poggiale. 2002. Functional approach

to sediment reworking by gallery-forming macrobenthic organisms: modeling and

application with the polychaete Nereis diversicolor. Mar. Ecol. Prog. Ser., 229, 127-136.

Gérino, M., R. C. Aller, C. Lee, J. K. Cochran, J. Y. Aller, M. A. Green and D. Hirschberg.

1998. Comparison of different tracers and methods used to quantify bioturbation during a

spring bloom: 234-Thorium, luminophores and chlorophyll a. Estuar. Coast. Shelf. Sci.,

46, 531-547.

Gilbert, F., P. Bonin and G. Stora. 1995. Effect of bioturbation on denitrification in a marine

sediment from the West Mediterranean littoral. Hydrobiologia, 304, 49-58.

Gilbert, F., S. Hulth, N. Stroemberg, K. Ringdahl and J. C. Poggiale. 2003. 2-D optical

quantification of particle reworking activities in marine surface sediments. J. Exp. Mar.

Biol. Ecol., 285/286, 251–263.

CHAPITRE 2 :


[85]

Grémare, A., J. C. Duchêne, R. Rosenberg, E. David and M. Desmalades. 2004. Feeding

behaviour and functional response of Abra ovata and A. nitida compared by image

analysis. Mar. Ecol. Prog. Ser., 267, 195-208.

Guinasso, N. L. and D. R. Schink. 1975. Quantitative estimates of biological mixing in

abyssal sediments. J. Geophys. Res., 80, 3032-3043.

Hughes, T. G. 1975. The sorting of food particles by Abra sp. (Bivalvia: Tellinacea). J. Exp.

Mar. Biol. Ecol., 20, 137-15.

Kristensen, E. 2000. Organic matter diagenesis at the oxic/anoxic interface in coastal marine

sediments, with emphasis on the role of burrowing animals. Hydrobiologia, 426, 1-24.

Kristensen, E., G. Penha-Lopes, M. Delefosse, T. Valdemarsen, C. O. Quintana and G. Banta.

2012. What is bioturbation? The need for a precise definition for fauna in aquatic sciences.

Mar. Ecol. Prog. Ser., 446, 285-302.

Lecroart, P., S. Schmidt and J. M. Jouanneau. 2007. Numerical estimation of the error of the

biodiffusion coefficient in coastal sediments. Est. Coast. Shelf. Sci., 72, 543-552.

Lecroart, P., O. Maire, S. Schmidt, A. Grémare, P. Anschutz and F. J. R. Meysman. 2010.

Bioturbation and short lived radioisotopes. Geochim. et Cosmochim. Acta., 74, 6049-6063.

Lohrer, A. M., S. F. Thrush, M. M. Gibbs. 2004. Bioturbators enhance ecosystem function

through complex biogeochemical interactions. Nature, 431, 1092-1095.

Maire, O., J. C. Duchêne, R. Rosenberg, J. Braga de Mendonça Jr. and A. Grémare. 2006.

Effects of food availability on sediment reworking in Abra ovata and Abra nitida. Mar.

Ecol. Prog. Ser., 319, 135-153.

Maire, O., J. C. Duchêne, A. Grémare, V. S. Malyuga and F. J. R. Meysman. 2007a. A

Comparison of Sediment Reworking Rates by The Surface Deposit-Feeding Bivalve Abra

Ovata During Summertime and Wintertime, with a Comparison Between Two Models of

Sediment Reworking. J. Exp. Mar. Biol. Ecol., 343, 21-36.

Maire, O., J. C. Duchêne, L. Bigot and A. Grémare. 2007b. Linking feeding activity and

sediment reworking in the deposit-feeding bivalve Abra ovata with image analysis, laser

telemetry, and luminophores tracers. Mar. Ecol. Prog. Ser., 351, 139-150.

Maire, O., J. C. Duchêne, J. M. Amouroux and A. Grémare. 2007c. Activity patterns in the

terebellid polychaete Eupolymnia nebulosa assessed using a new image analysis system.

Mar. Biol., 151, 737-749.

Maire, O., P. Lecroart, F. J. R. Meysman, R. Rosenberg, J. C. Duchêne and A. Grémare,

2008. Methods of sediment reworking assessment in bioturbation research: a review.

Aquat. Biol., 2, 219-238.

Meadows, P. S. and A. Meadows. 1991. The geotechnical and geochemical implications of

bioturbation in marine sedimentary ecosystems. Symp. Zool. Soc. Lond., 63,157-181.

CHAPITRE 2 :


[86]

Meysman, F. J. R., B. P. Boudreau and J. J. Middelburg. 2003. Relations between local,

nonlocal, discrete and continuous models of bioturbation. J. Mar. Res. 61, 391-410.

Meysman, F. J. R., J. J. Middelburg and C. H. R. Heip. 2006. Bioturbation: a fresh look at

Darwin's last idea. Trends. Ecol. Evol. 12-21, 688-695.

Meysman, F. J. R., V. S. Malyuga, B. P. Boudreau and J. J. Middelburg. 2008a. A generalized

stochastic approach to particle dispersal in soils and sediments. Geochim. Cosmochim.

Acta, 72, 3460-3478.

Meysman, F. J. R., V. S. Malyuga, B. P. Boudreau and J. J. Middelburg. 2008b. Quantifying

particle dispersal in aquatic sediments at short time scales: model selection. Aquat. Biol. 2,

239-254.

Meysman, F. J. R., B. P. Boudreau and J. J. Middelburg. 2010. When and why does

bioturbation lead to diffusive mixing? J. Mar. Res. 68, 881–920.

Pearson, T., and R. Rosenberg. 1978. Macrobenthic succession in relation to organic

enrichment and pollution of the marine environment. Oceanogr. Mar. Biol. Annu. Rev., 16,

229-331.

Piot, A., A. Rochon, G. Stora and G. Desrosiers. 2008. Experimental study on the influence of

bioturbation performed by Nephtys caeca (Fabricius) and Nereis virens (Sars) annelidae on

the distribution of dinoflagellate cysts in the sediment. J. Exp. Mar. Biol. Ecol., 359, 92-

101.

Reed, D. C., K. Huang, B. P. Boudreau and F. J. R. Meysman. 2006. Steady-state tracer

dynamics in a lattice-automaton model of bioturbation. Geochim. Cosmochim. Acta, 70,

5855-5867.

Rhoads, D. C. 1974. Organism-sediment relations on the muddy seafloor. Oceanogr. Mar.

Biol. Ann. Rev., 12, 263-300.

Rhoads, D. C. and L. F. Boyer. 1982. The effects of marine benthos on physical properties of

sediments. A successional perspective, in Animal–sediment relations—the biogenic

alteration of sediments. Topics in geobiology, Vol 2, P. L. McCall, M. J. S. Tevesz, ed.,


Rice, D. L. 1986. Early diagenesis in bioadvective sediments: relationships between the

diagenesis of beryllium-7, sediment reworking rates, and the abundance of conveyorbelt

deposit-feeders. J. Mar. Res., 44, 149-184.

Robbins, J. A. 1986. A model for particle-selective transport of tracers in sediments with

conveyor-belt deposit feeders. J. Geophys. Res. 91, 8542-8558.

Rosenberg, R. 1993. Suspension feeding in Abra alba (Mollusca). Sarsia, 78, 119-121

Rowden, A. A., C. F. Jago and S. E. Jones. 1998. Influence of benthic macrofauna on the

geotechnical and geophysical properties of surficial sediment, North Sea. Cont. Shelf. Res.,

18, 1347-1363.

CHAPITRE 2 :


[87]

Shull, D. H. 2001. Transition-Matrix Model of Bioturbation and Radionuclide Diagenesis.

Limnol. Oceanogr., 46, 905-916.

Shull, D. H. and M. Yasuda .2001. Size-selective downward particle transport by cirratulid

polychaetes. J. Mar. Res., 59(3), 453-473.

Soetaert K, P. M. J. Herman, J. J. Middelburg, C. Heip, H. S. De Stigter, T. C. E. Van

Weering, E. Epping and W. Helder. 1996. Modelling 210

Pb-derived mixing activity in

ocean margin sediments: diffusive versus non local mixing. J. Mar. Res., 54, 1207-1227.

Solan, M., B. D. Wigham, I. R. Hudson, R. Kennedy, C. H. Coulon, K. Norling, H. C. Nilsson

and R. Rosenberg. 2004. In Situ Quantification of Bioturbation Using Time-Lapse

Fluorescent Sediment Profile Imaging (F-Spi.), Luminophore Tracers and Model

Simulation. Mar. Ecol. Prog. Ser. 271, 1-12.

Teal, L. R., E. R. Parker and M. Solan. 2012. Simultaneous quantification of in-situ infaunal

activity and pore-water metal concentrations: establishment of benthic ecosystem process-

function relations. Biogeosciences Discuss., 9, 8541-8570.

Van Hoey, G., M. Vincx and S. Degraer. 2005. Small-scale geographical patterns within the

macrobenthic Abra alba community. Est. Coast. Shelf. Sci., 64, 751-763.

Volkenborn, N., L. Polerecky, D. S. Wethey, T. H. DeWitt and S. A. Woodin. 2012.

Hydraulic activities by gost shrimp Neotrypaea californiensis induce oxic-anoxic

oscillations in sediment. Mar. Ecol. Prog. Ser., 455, 141-156.

Wheatcroft, R. A., P. A. Jumars, C. R. Smith and A. R. M. Nowel. 1990. A mechanistic view

of the particulate biodiffusion coefficient: rest periods and transport directions. J. Mar.

Res., 48, 177-201.

Wheatcroft, R. A. 1992. Experimental tests for particle size-dependent bioturbation in the

deep ocean. Limnol. Oceanogr., 37, 90-104.

Wikander, P. B. 1980. Biometry and behaviour in Abra nitida (Müller) and A. longicallus

(Scacchi) (Bivalvia, Tellinacea). Sarsia, 65, 255-268.

CHAPITRE 2 :


[88]

Transition

Le déploiement de la nouvelle approche présentée dans ce chapitre, couplant

l’utilisation d’un système d’acquisition à haute fréquence d’images de la paroi d’un aquarium

plat sous lumière UV et la mise au point d’un logiciel spécifique, a permis la première mesure

expérimentale des mouvements élémentaires de particules engendrés par un invertébré

benthique. A partir des caractéristiques de l’ensemble de ces mouvements les empreintes de

remaniement sédimentaire chez le bivalve biodiffuseur Abra alba ont ainsi pu être mesurées.

Ces empreintes, constituées des distributions de fréquence des caractéristiques des

mouvements élémentaires de luminophores détectés (temps d’immobilité, longueurs et

directions des mouvements), se sont révélées en bonne concordance avec les connaissances

relatives à l’éthologie alimentaire du genre Abra. Un des autres principaux résultats de ce

travail réside dans la mesure d’empreintes de remaniement sédimentaire qui se sont révélés

particulièrement hétérogènes spatialement, aussi bien sur leur composante horizontale que

verticale.

Une deuxième étape, compliquée par cette hétérogénéité, consiste à tester l’aptitude de

cette approche à décrire l’effet de paramètres extrinsèques sur les empreintes de remaniement

sédimentaire. Le troisième chapitre de ce manuscrit s’attache par conséquent à décrire, via le

déploiement de cette nouvelle méthode, l’effet des paramètres environnementaux que sont la

température et la disponibilité de matière organique fraîche à l’interface eau-sédiment, sur les

empreintes de remaniement sédimentaire chez A. alba.

CHAPITRE 3 :

Mesures expérimentales de l’effet de la température et de la disponibilité en matière organique sur le

remaniement sédimentaire du bivalve Abra alba

[89]

Chapitre 3:

Mesures expérimentales de l’effet

de la température et de la

disponibilité en matière organique

sur le remaniement sédimentaire du

bivalve Abra alba : Utilisation d’une

nouvelle technique d’analyse

d’images

CHAPITRE 3 :



[90]

Experimental assessment of the effects of temperature and

food availability on particle mixing by the bivalve Abra

alba using new image analysis techniques.


, Jean-Claude Duchêne3,

Pascal Lecroart

1,

Olivier Maire

1,

Aurélie Ciutat3, Bruno Deflandre

1, Antoine Grémare

1




Keywords: Particle mixing, Abra alba, CTRW model, Image analysis, Bioturbation, feeding

behaviour, temperature, food availability

Running title: Temperature and food effects on particle mixing in Abra alba

CHAPITRE 3 :



[91]

Abstract

The effect of food addition on particle mixing in the deposit-feeding bivalve Abra alba

were assessed at two different seasons using an experimental approach allowing for the

tracking of individual particle displacements. This allowed for the computations of both

overall mean values and vertical profiles of: normalized numbers of jumps, inversed waiting

times, jump characteristics (mean and standard deviation of jump lengths), and particle

tracking biodiffusion coefficients (Db). Data originated from 32 experiments carried out under

4 combinations of 2 seasons (Se) and 2 food addition levels (Fo). For each of these treatments

parameters were computed for 5 experimental durations (Ed). The effects of Se, Fo and Ed

were assessed using PERMANOVAs carried out either on vertical profiles or on overall mean

values. The comparison of profiles resulted in a detection of a higher number of significant

effects than the comparison of overall mean values, thereby suggesting that it is more efficient

in detecting the effect of environmental factors on particle mixing by A. alba. Inversed

waiting times significantly decreased with Ed whereas the normalized number of jumps did

not thereby suggesting that it constitutes a better proxy of jump frequency when assessing

particle mixing based on the measure of individual particle displacements. Particle mixing

was low during autumn experiments and not affected by Fo, which was attributed to low

temperature. Conversely, particle mixing was high during summer and transitory inhibited by

Fo. This last result is coherent with the functional responses (both in terms of activity and

particle mixing) already measured for individual of the closely related clam A. ovata

originating from temperate populations. It also partly resulted from a transitory switch

between deposit- and suspension-feeding caused by the high concentration of suspended

particulate organic matter immediately following food addition.

CHAPITRE 3 :



[92]

I. Introduction

In aquatic environment, bioturbation is defined as “all transport processes carried out

by animals that directly or indirectly affect the sediment matrices” (Kristensen et al., 2012).

Such processes include both particle mixing and bioirrigation. Through bioturbation, benthic

fauna strongly affects the chemical, physical and geotechnical properties of marine sediments

(Gray, 1974; Rhoads, 1974; Aller, 1982; Rhoads and Boyer, 1982; Meadows and Meadows,

1991; Gilbert et al., 1995; Rowden et al., 1998; Lohrer et al., 2004). Particle mixing controls

the transfer of recently settled particles to deeper sediment layers and thereby affects the

remineralisation of particulate organic matter (Kristensen et al, 2000; Caradec et al., 2004). It

also affects various biological processes such as the burial of both dinoflagellate cysts

(Giangrande et al., 2002; Persson and Rosenberg, 2003; Piot et al. 2008) and phanerogam

seeds (Hughes et al. 2000; Cabaço et al., 2008; Delefosse et Kristensen, 2012 ; Balckburn et

Orth, 2013).

Particle mixing mainly results from locomotion, burrowing, defecation and feeding

activities of the benthic macrofauna (Meysman et al., 2006). The nature and intensity of all

these activities are depending on both intrinsic characteristics of benthic communities and on

their surrounding environment. The effect of disturbance (and especially organic matter

enrichment) on benthic community structure and functionalities (including bioturbation) is

well documented (Pearson and Rosenberg, 1979; Rosenberg, 2001). At the organism’s level,

key environmental factors such as organic matter availability and water temperature are well

known to tightly control the overall behaviour of benthic fauna; including burrowing and/or

feeding activities (Grémare et al., 2004; Stead and Thompson, 2006; Michaud et al. 2010)

thereby altering particle mixing modes and rates (Maire et al., 2006; 2007a; 2007b; Venturini

et al. 2011).

Particle mixing is classically quantified using particle tracers (Maire et al. 2008). As

opposed to natural ones (e.g. 7Be

210Pb,

234Th), which are naturally present in the sediment

column, artificial tracers, such as luminophores (i.e. sediment particles with a fluorescent

coating), are introduced at the sediment-water interface at the beginning of an experiment, and

their vertical distribution within the sediment column is then measured after an incubation

period of known duration. The so-obtained vertical tracer profile is then fitted using a

mathematical model. Several particle-mixing models are available. Due to its simplicity, the

biodiffusive model (Goldberg and Koide, 1962; Guinasso and Schink, 1975; Aller, 1982;

CHAPITRE 3 :



[93]

Boudreau, 1986a; Wheatcroft et al., 1992; Gerino et al., 1998; Lecroart et al., 2007, 2010) has

long been preferentially used despite the fact that its underlying hypotheses (i.e., highly

frequent and very short isotropic jumps) are often not fulfilled (Meysman et al., 2003). In this

model, particle mixing by benthic fauna is described by a single parameter: the biodiffusion

coefficient. Recent years have seen the emergence of the continuous random walk (CTRW)

model (Meysman et al, 2006, 2008a, b, 2010). The CTRW model implements a stochastic

description of particles mixing events. Particle displacement is then described as a random

process, and each individual particle displacement is governed by three stochastic variables:

(1) the jump direction, (2) the jump length, and (3) the waiting time between two consecutive

jumps of the same individual particle (Wheatcroft et al., 1990). Overall, the joined frequency

distributions of these random variables form the “mixing fingerprint” of a benthic community

or of a benthic organism (Meysman et al., 2008a). It is also possible to compute a particle-

tracking biodiffusion coefficient (Db) from those fingerprints (Wheatcroft et al., 1990;

Meysman et al., 2008a).

The CTRW model clearly constitutes an important step in better describing the

inherent complexity of a biologically mediated process such as particle mixing. It has already

been successfully used with the bivalves Abra ovata and A. nitida (Maire et al. 2007a), the

polychaete Nepthys sp. (Braeckman et al., 2010) and the amphipod Corophium volutator (De

Backer et al., 2011). In all these studies, mixing fingerprints were assessed: (1) assuming a

perfect spatial homogeneity of particle mixing, (2), based on a priori assumed frequency

distributions of waiting times and jump lengths and (3) through the fitting of a CTRW model

to vertical luminophore profiles after a known period of incubation.

These points are questionable (see for example Meysman et al., 2008a for a discussion

on the importance of the selection of a priori selected frequency distributions), explaining

why Bernard et al. (in revision) have recently developed an experimental approach allowing

for the direct and explicitly 2-D assessment of particle mixing fingerprints in the deposit

feeding bivalve Abra alba. These authors adapted existing high frequency images acquisition

and analysis techniques (Gilbert et al., 2003; Solan et al., 2004; Maire et al., 2006, 2007a,

2007b, 2007c; Piot et al., 2008) to track single luminophores motions along the wall of thin

aquaria and to directly derive the frequency distributions of waiting times, jump lengths and

directions. This allowed for the 2-D assessment of changes in particle mixing at a sub-

millimetre resolution, which revealed the highly spatially heterogeneous behaviour of the

particle mixing induced by A. alba under field-like conditions.

CHAPITRE 3 :



[94]

Particle mixing intensity in A. alba has previously been assessed trough the fit of the

CTRW model to experimentally derived luminophore profiles (Braeckman et al., 2010). These

authors reported a significant effect of water temperature but no significant effect of clam density

on particle mixing. The effects of temperature (Maire et al. 2006 and 2007a) and food

availability (Grémare et al. 2004, Maire et al. 2006) on feeding activity and particle mixing

have also been assessed in two closely related species: A. ovata and A. nitida. Corresponding

results have shown: (1) differences in the functional responses (both in terms of feeding

activity and particle mixing) of the two species to food addition, and (2) a significant effect of

temperature on particle mixing in both species. In spite of a strict similarity in experimental

procedures, there were important discrepancies between the studies carried out by Maire et al.

in 2006 and 2007a. During the first study, these authors indeed studied the two Abra species

and fitted a biodiffusive model to vertical luminophore profiles, whereas during the second

one they worked with only A. ovata and fitted a CTRW model to vertical luminophore

profiles. As underlined above, both approaches are no longer considered optimal in describing

particle mixing. The aim of the present study was therefore to use the new image analysis

techniques developed by Bernard et al. (in revision) to assess the effect of temperature and

food availability on particle mixing in A. alba.

II. Materials and Methods

2.1. Bivalve collection and maintenance

The deposit-feeding bivalve Abra alba (Tellinacea) is a dominant macrobenthic

species in shallow subtidal areas along the European Atlantic coast (Borja et al. 2004, Van

Hoey et al. 2005). The body of this clam is usually buried a few centimetres deep in the

sediment surface, while its siphons connect to the sediment-water interface. It reworks the

upper sediment layer by protruding its inhalant siphon and aspiring recently deposited organic

matter. Foraging movements consist of circular motions of the tip of the inhalant siphon at the

sediment-water interface (Hughes, 1975). During the present study, sediment samples were

collected in June 2010 and May 2011 in the Arcachon Bay (45°43’476 N, 1°37’758 W, 3-5 m

depth) using a Van-Veen grab. Sediment samples were sieved through a 1mm square mesh,

CHAPITRE 3 :



[95]

yielding ~500 clams (9-12 mm in total shell length). The sediment (47.7 % sand and 52.3 %

fines; 1.40 %DW POC and 0.16 %DW PON) was used both for maintenance or organisms

and experimentations. Clams were kept in tanks (60x40x30cm) filled with field sieved

sediment and supplied with running seawater prior experimentation. During that period of

time, they were fed once a week with crushed Tetramin® fish food.

2.2. Experimental set-up

The experimental set-up involved the use of thin aquaria, luminophores, UV lights,

and high frequency image acquisition (Maire et al. 2006 and 2007a, Bernard et al. 2012).

Thin aquaria (L=17 cm, W=0.9 cm, H=33 cm) were filled with 15 cm of field sieved

sediment. They were kept in a temperature-controlled climate room at ambient seawater

temperature for 3 days before each experiment. Three bivalves of known size were then

gently placed at the sediment surface, after which they typically buried within 30 seconds. If a

clam did not do so within a minute, it was replaced. After 24 hours, 1.5 gDW of yellow

luminophores (Geotrace Environmental Tracing®, median diameter=35 µm) were spread at

the sediment surface using a Pasteur pipette. Thin aquaria were placed 30 cm in front of two

UV lights (which allowed for the distinction between fluorescent luminophores and the

surrounding sediment particles) and of a µeye video captor (IDS®), which was positioned to

monitor luminophore movements. The µeye captor had a definition of 2560x1920 pixels,

while the monitored sediment area was 4.2 cm x 3.1 cm, which resulted in a resolution of 16.5

µm per pixel. Experiment began 24 hours after luminophore introduction. This allowed for:

(1) the monitored field to be centred on an area effectively reworked by a single clam, and (2)

the preliminary dispersion of luminophores. Each experiment lasted 48 hours and image

frequency acquisition was 0.1 Hz. The series of images collected during each experiment

were assembled in an AVI video format for further image analysis.

Here, we report on the results of 32 experiments during which both seawater

temperature (two Se conditions) and organic matter availability (two Food addition

conditions) were manipulated in a balanced experimental design leading to 8 replicated

experiments per combination of Se and Food addition conditions. The “Summer” condition

corresponded to experiments carried out at temperatures between 17.6°C and 22.3°C and the

“Autumn” condition to temperatures between 13.7°C and 16.6°C. The “Without food

addition” condition corresponded to experiments carried out in the absence of any organic

CHAPITRE 3 :



[96]

input. Conversely, the “With food addition” condition corresponded to the introduction of

94.5 mgDW of fresh Tetraselmis sp. detritus (Marine-Life®, 6.0 % DW PON, 45.4% DW

POC) one hour prior the beginning of each experiment. The amount of detritus added was set

to correspond to a nitrogen daily food ration to standing biomass ratio of 0.5, which has been

shown sufficient to support maximal daily weight specific growth rate in the opportunistic

polychaete Capitella capitata (Tenore and Chesney 1985).

2.3. Image processing

For each experiment, AVI films were processed using two specific algorithms based

on the analysis of the relative positions of luminophore barycentres within consecutive

images. These two algorithms are detailed in Bernard et al. (2012). Briefly, isolated

luminophores were first binarised for all images based on their red-green-blue levels,

luminance and size. The (XY) coordinates of their barycentre in the pixel grid were then

assessed for each individual image. The two algorithms respectively allow for the

measurements of: (1) luminophore waiting time between two consecutives jumps and (2)

luminophore jump characteristics (i.e., duration, direction and length). The first algorithm

uses a single sensitivity circle centred on the luminophore barycentre, which accounts for both

changes in the apparent size of luminophores due to fluctuations in light intensity, and to

small movements of the sensor and/or the aquarium induced by vibrations. The second

algorithm also uses a search circle that defines the maximum distance over which individual

particles can be tracked between two consecutive images. Based on preliminary trials

(Bernard et al., in revision), the radius of the sensitivity and the search circles were set to 66

and 660 µm, respectively. When a jump event has ended, the following parameters are

recorded in the jump results file: the index number of the starting image (corresponding to the

start of the jump), the index number of the final image (corresponding to the end of the

trajectory analysis), the duration of the jump, the (XY) coordinates of the particle at the start

of the jump, the (XY) coordinates at the end of the jump, the length and the direction of the

jump vector.

2.4. Data processing

Above-described image processing was used to assess both vertical depth profiles and

overall mean values of a set of parameters including: (1) the normalized numbers of jumps,

CHAPITRE 3 :



[97]

(2) inversed waiting times, (3) jump characteristics (mean and standard deviation of jump

lengths), and (4) Db.


During each experiment, the location of the water–sediment interface was assessed

based on its location within each 660 µm wide cell column defined (i.e., maximum X

coordinates within each cell column where waiting times were recorded). It was then

translated to the first row of each cell column. After this operation, the cell X-position in the

picture thus corresponded to its depth within the sediment column (Maire et al. 2006). We

then computed 1D vertical 1320 µm resolution profiles of: (1) normalized numbers of jumps,

(2) mean inversed waiting time (1/Tc), (3) means jump length, (4) standard deviation of jump

lengths (), and (5) Db. As pointed out by Bernard et al. (in revision), the detection of a jump

in a given area is depending not only on particle mixing intensity but also on the density of

luminophores in this area. The computation of the normalized (i.e., rescaled to the number of

luminophores counted in the area) number of jumps therefore provides a measure of the

probability of jump and therefore of particle mixing intensity. Inversed waiting time was used

instead of waiting time because it also measures the frequency of mixing events and co-varies

positively with the normalized number of jumps. Db was computed based on Meysman et al.

(2008a) using the following formula:

Db =

2/(4Tc)

For each experiment, all profiles were computed for 5 Experiment durations (i.e., 6,

12, 24, 36 and 48 hours).


Our results support previous observations (see Bernard et al. 2012) showing that: (1)

all the components of particle mixing fingerprints vary with depth, and (2) the vertical

profiles of the total and normalized numbers of jumps do differ. The computations of overall

mean values of inversed waiting time, jump length characteristics, and Db were therefore

based on a bootstrap procedure involving 1000 re-samplings that were stratified relative to

depth. During each re-sampling, the vertical frequency distribution of the number of

CHAPITRE 3 :



[98]

resampled events was kept strictly identical to the one of the normalized number of jumps.

The number of jumps per re-sampling was set to limit oversampling and therefore varied

according to parameters and experiments. As for vertical profiles, overall mean values were

computed for the 5 different durations listed above during each experiment.

All the above-described procedures were carried out using specific routines developed

with the R free software (v2.13.1., http://www.R-project.org , 2011).

2.5. Data analysis


The effects of tested factors (see below) on the location of the vertical profiles of all

the above-mentioned parameters were assessed using multivariate permutational ANOVAs

(PERMANOVA; Anderson 2001, McArdle & Anderson 2001) without preliminary data

transformation. We used the Euclidean distance to assess dissimilarities between profiles. Our

overall design consisted in 3 fixed factors, namely Season (Se, 2 levels), Food addition (Fo, 2

levels), and Experiment duration (Ed, 5 levels) together with a fourth random factor

“Replicates” (Rep), which was nested within Se and Fo. In case of significant interactions

between factors, pairwise tests were performed to characterize their modalities. The effects of

the tested factors on the dispersion (i.e., among experiment variability) of vertical profiles

were checked using the PERMDISP procedure (Anderson 2006; same distance and same

design as described above).


The effects of treatments on the overall mean values of all parameters (except for the

normalized number of jumps) were assessed through univariate PERMANOVAs using the

same distance and design as for vertical profile comparisons. Here again, pairwise tests were

performed in case of significant interactions between factors and differences in data

dispersion were assessed through PERMDISP (Anderson, 2006).

All the above-described procedures were performed using the PRIMER® v6 package

with the PERMANOVA+ add-on software (Clarke & Warwick 2001, Anderson et al. 2008).

http://www.r-project.org/

CHAPITRE 3 :



[99]

III. Results

3.1. Vertical profiles

3.1.1. Main effects

Table 3.1 shows the results of the PERMANOVA and the PERMDISP procedures

carried out on all vertical profiles.

Se significantly affected the location of the vertical profiles for all tested parameters.

Conversely, Fo only affected the location of the vertical profiles of jump characteristics. Ed

also significantly affected the location of the vertical profiles for all tested parameters but

normalized numbers of jumps. There were no interactions between the effects of Se and Fo

and between those of Se and Ed. Conversely, there was a significant interaction between the

effects of Fo and Ed on the location of the vertical profiles of: (1) normalized numbers of

jumps, (2) inversed waiting times, and (3) Db. Rep significantly affected the location of the

vertical profiles for all tested parameters. The interactions between the different levels of Se,

Fo and Ed only affected the location of the vertical profiles of the normalized numbers of

jumps.

CHAPITRE 3 :



[100]

Table 3.1 : Results from Permanova and Permdisp analyses for differences in vertical depth

profiles of normalized number of jumps (jumps.luminophore-1

), inversed waiting times (h-1

),

jump characteristics (means and standard deviations of jumps lengths in mm) and Db (cm2.yr

-

1), amongst Season (Se), food addition (Fo) and experimental duration (Ed), based on a

Euclidean resemblance matrix.

Factors

Normalized number of jumps

Inversed waiting times

Mean jump lengths

of jump lengths

Db

Se

df

MS

Pseudo-F

p(perm)

1

0.6

7.06*

0.0013

1

411.82

3.96

0.0054

1

0.49

5.03

0.0044

1

0.67

3.91*

0.0112

1

61.73

1.99

0.0276

Fo

df

MS

Pseudo-F

p(perm)

1

1.88E-2

0.22

0.7975

1

152.79

1.47

0.1673

1

0.26

2.68*

0.0451

1

0.55

3.2*

0.0214

1

42.66

1.38

0.1569

Ed

df

MS

Pseudo-F

p(perm)

4

1.2E-2

1.23

0.2751

4

325.52

7.8*

0.0001

4

3.8E-2

3.32

0.0003

4

5.78E-2

2.59

0.0003

4

74.96

4.68*

0.0001

SexFo

df

MS

Pseudo-F

p(perm)

1

2.06E-2

0.24*

0.7594

1

51.71

0.5

0.8437

1

0.16

1.63*

0.1663

1

0.23

1.31*

0.2405

1

27.19

0.88

0.557

SexEd

df

MS

Pseudo-F

p(perm)

4

7.71E-3

0.79*

0.6027

4

38.12

0.91*

0.5543

4

9.97E-3

0.87

0.6017

4

1.79E-2

0.8

0.6942

4

11.45

0.72*

0.9367

FoxEd

df

MS

Pseudo-F

p(perm)

4

3.73E-3

3.8

0.0006

4

84.76

2.03*

0.0063

4

9.83E-3

0.86*

0.6144

4

2.63E-2

1.18

0.2517

4

28.77

1.8*

0.0036

Rep

df

MS

Pseudo-F

p(perm)

28

8.54E-2

8.71*

0.0001

28

104.07

2.49

0.0001

28

9.7E-2

8.47*

0.0001

28

0.17

7.75*

0.0001

28

30.88

1.93

0.0001

SexFoxEd

df

MS

Pseudo-F

p(perm)

4

2.52E-2

2.57*

0.0144

4

22.54

0.54*

0.9814

4

9.92E-3

0.87*

0.5994

4

2.21E-2

0.99

0.4484

4

13.34

0.83*

0.7741

* : Permdisp, p<0.05

CHAPITRE 3 :



[101]

Se significantly affected the dispersion of the vertical profiles of normalized numbers

of jumps and of the standard deviations of jump lengths. Fo significantly affected the

dispersion of the vertical profiles of both jump characteristics. Ed significantly affected the

dispersion of the vertical profiles of inversed waiting times and Db. There were interactions

between the effects of Se and Fo on the dispersion of the vertical profiles of: (1) normalized

numbers of jumps, and (2) both jump characteristics. There were interactions between the

effects of Se and Ed on the dispersion of the vertical profiles of: (1) normalized numbers of

jumps, (2) inversed waiting times, and (3) Db. There were interactions between the effects of

Fo and Ed on the dispersion of the vertical profiles of: (1) inversed waiting times, (2) mean

jump lengths and (3) Db. Rep significantly affected the dispersion of the vertical profiles of:

(1) normalized numbers of jumps, and (2) both jump characteristics. The interactions between

the different levels of Se, Fo and Ed affected the dispersion of the vertical profiles of all tested

parameters but the standard deviation of jump lengths.

3.1.2. Normalized numbers of jumps

The vertical profiles of normalized number of jumps are shown in Figure 3.1 for all

combinations of the different levels of Se, Fo and Ed.

During the autumn experiments without food addition (Figure 3.1A), average

normalized numbers of jumps were low (less than 0.02 jumps.luminophore-1

recorded for the

total duration of the experiment) and variability was limited among experiments. Average

normalized numbers of jumps were maximal immediately below the sediment-water interface

and then decreased with depth in the sediment column. There were no effects of Ed on the

location and the dispersion of the vertical profiles of normalized numbers of jumps (Table

3.1).

Average normalized numbers of jumps were higher during the summer experiments without

food addition (Figure 3.1B) than during the autumn experiments without food addition with a

maximal value of 0.13 jumps.luminophore-1

(Table 3.1, Table S1).

CHAPITRE 3 :



[102]

A


(Number of jumps .luminophore-1)

0,00 0,02 0,04 0,06 0,08 0,10 0,12 0,14 0,16 0,18 0,20

De

pth

(m

m)

-30

-25

-20

-15

-10

-5

0

6H

12H

24H

36H

48H

B


(Number of jumps.luminophore-1)

0,00 0,02 0,04 0,06 0,08 0,10 0,12 0,14 0,16 0,18 0,20

De

pth

(m

m)

-30

-25

-20

-15

-10

-5

0

6H

12H

24H

36H

48H

C



0,00 0,02 0,04 0,06 0,08 0,10 0,12 0,14 0,16 0,18 0,20

Dep

th (

mm

)

-30

-25

-20

-15

-10

-5

0

6H

12H

24H

36H

48H

D



0,00 0,02 0,04 0,06 0,08 0,10 0,12 0,14 0,16 0,18 0,20

De

pth

(m

m)

-30

-25

-20

-15

-10

-5

0

6H

12H

24H

36H

48H

Figure 3.1: Vertical profiles of the normalized number of jumps.

Vertical profiles were recorded during the: autumn without food addition experiments (A),

summer without food addition experiments (B), autumn with food addition experiments (C)

and summer with food addition experiments (D). Colors correspond to different experiment

durations. Horizontal bars are standard errors and refer to between experiment variability.

Among experiments variability was also higher during the 6, 12 and 24h experiments

(Table S1). There were no significant effects of Ed on the vertical profile of normalized

numbers of jumps (Table S3) and the shapes of these profiles were similar to those recorded

during the autumn experiments without food addition. This trend was especially marked for

the 24, 36 and 48 h experiments.

The average normalized numbers of jumps in the upper part of the sediment column

seemed higher during the autumn experiments with food addition (Figure 3.1C) than during

the autumn experiments without food addition with a maximal value of 0.07

jumps.luminophore-1

. However, there were no significant effects of Fo on the location of

whole vertical profiles (Table 3.1, Table S2). There were no significant effects of Ed on the

location of vertical profiles as well (Table S3) and the shapes of these profiles were similar to

CHAPITRE 3 :



[103]

those recorded during the autumn experiments without food addition except for the 36h

experiments, which showed a marked subsurface (at ca. 7mm deep) maximum.

Ed significantly affected the location (but not the dispersion) of the vertical profiles of

normalized numbers of jumps during the summer experiments with food addition (Figure

3.1D, Table S3). Vertical profiles recorded during the 6 and 12 h experiments did not differ

between one another but did significantly differ from those recorded during the 24, 36 and

48h experiments. Corresponding maximal values were 0.05 and 0.14 jumps.luminophore-1

for

shorter and longer experiments, respectively. For the 6 and 12h experiments, there were no

significant differences in the location of the vertical profiles recorded during the summer and

the autumn experiments with food addition (Figures 3.1C and 3.1D, Table S1). Conversely,

there were significant differences in the location of the corresponding vertical profiles

recorded during the 24 and 48h experiments. In this last case, normalized numbers of jumps

were higher during summer (maximal value of 0.14 jumps.luminophore-1

) than during autumn

experiments (maximal value of 0.07 jumps.luminophore-1

). The vertical profiles recorded

during the 24, 36 and 48 h of summer experiments with food addition were all characterized

by a subsurface (at ca. 8 mm deep) maximum. This was rather different from the patterns

observed during the summer experiments without food addition. However it should be

underlined that, for all experiment durations, there were no statistically significant differences

in the location of the whole vertical profiles between the summer experiments with and

without food addition (Figures 3.1B and 3.1D, Table S2).

3.1.3. Inversed waiting times

The vertical distributions of the average inversed waiting times are shown in Figure

3.2 for all combinations of the different levels of Se, Fo and Ed. The effect of Se on the

location of vertical profiles was significant and similar to the one found for normalized

numbers of jumps with a trend toward higher values during summer than during autumn

experiments without food addition. The effect of Ed on the location of vertical profiles was

significant for all combinations of Se and Fo but the autumn experiments with food addition

(Table S3).

CHAPITRE 3 :



[104]

A

Mean Inversed Waiting time (h-1

)

0 2 4 6 8 10

Dep

th (

mm

)

-30

-25

-20

-15

-10

-5

0

6H

12H

24H

36H

48H

B


)

0 2 4 6 8 10

Dep

th (

mm

)

-30

-25

-20

-15

-10

-5

0

6H

12H

24H

36H

48H

C


)

0 2 4 6 8 10

Dep

th (

mm

)

-30

-25

-20

-15

-10

-5

0

6H

12H

24H

36H

48H

D

Mean Inversed Waiting time (h-1)

0 2 4 6 8 10

Dep

th (

mm

)

-30

-25

-20

-15

-10

-5

0

6H

12H

24H

36H

48H

Figure 3.2: Vertical profiles of inversed waiting times.

Vertical profiles were recorded during the: autumn without food addition experiments (A),

summer without food experiments (B), autumn with food addition experiments (C) and

summer with food addition experiments (D). Colors correspond to different experiment

durations. Horizontal bars are standard errors and refer to between experiment variability.

Whenever significant, the effect of Ed corresponded to a decrease in inversed waiting

times with experiment duration (Table S3), whereas normalized number of jumps tended to

increase with experiment duration during summer experiments with food addition (Figure

3.1D). Fo interacted significantly with Ed in affecting the location of vertical profiles (Table

3.1) by reducing: (1) the overall range of inversed waiting times (maximal values of inversed

waiting times of 3.5 vs. 5.6 h-1

during summer experiments with and without food addition,

respectively), and (2) the experiment duration required to reach a stable vertical profile (eg,

12 vs. 36 h during summer experiments with and without food addition, respectively).

3.1.4. Jump characteristics

CHAPITRE 3 :



[105]

The location of the vertical profiles of jump length and were both significantly

affected by Se, Fo and Ed without any significant interactions between any of them (Table

3.1).

Average jump lengths tended to decrease with depth (Figure 3.3) and were

significantly higher during summer than during autumn (Figure 3.3A). There was no effect of

Fo on mean jump length close (i.e., within the 7mm top sediment) to the sediment water-

interface. Average jump lengths were then higher without food addition from 7 to 19mm deep

in the sediment (Figure 3.3B). Average jump lengths also tended to increase with experiment

duration (Figure 3.3C, Table S4).

CHAPITRE 3 :



[106]

A

Mean jump lentghs (mm)

0,00 0,05 0,10 0,15 0,20 0,25

Dep

th (

mm

)

-30

-25

-20

-15

-10

-5

0

Autumn

Summer

B

Mean jump lengths (mm)

0,00 0,05 0,10 0,15 0,20 0,25

Dep

th (

mm

)

-30

-25

-20

-15

-10

-5

0

Food

No food

C

Mean jump lentghs (mm)

0,00 0,05 0,10 0,15 0,20 0,25

Dep

th (

mm

)

-30

-25

-20

-15

-10

-5

0

6 H

12 H

24H

36 H

48 H

Figure 3.3: Vertical profiles of mean jump lengths.

Vertical profiles are shown for pooled autumn and summer experiments (A), for pooled with

and without food addition experiment (B), and for all pooled experiments of the same

duration (C). Colors correspond to different experiment durations. Horizontal bars are

standard errors and refer to between experiment variability.

CHAPITRE 3 :



[107]

A

Standard deviation of jump lentghs (mm)

0,00 0,02 0,04 0,06 0,08 0,10 0,12 0,14

De

pth

(m

m)

-30

-25

-20

-15

-10

-5

0

Autumn

Summer

C


0,00 0,02 0,04 0,06 0,08 0,10 0,12 0,14

De

pth

(m

m)

-30

-25

-20

-15

-10

-5

0

6 H

12 H

24 H

36 H

48 H

B


0,00 0,02 0,04 0,06 0,08 0,10 0,12 0,14

De

pth

(m

m)

-30

-25

-20

-15

-10

-5

0

Food

No food

Figure 3.4: Vertical profiles of standard deviations of jump lengths.

Vertical profiles are shown for pooled autumn and summer experiments (A), for pooled with

and without food addition experiments (B), and for all pooled experiments of the same

duration (C). Colors correspond to different experiment durations. Horizontal bars are

standard errors and refer to between experiment variability.

CHAPITRE 3 :



[108]

followed the same general pattern with: (1) a diminution with depth in the sediment

column, (2) higher values without food addition, and (3) a general increase with experiment

duration. There was only a minor difference regarding the range of depths (i.e., from 3 to

15mm) where standard deviations were higher without than with food addition) (Figure 3.4).

3.1.5. Db

The locations of the vertical profiles of Db were significantly affected by Se and Ed. In

addition, there was a significant interaction between Ed and Fo (Table 3.1).

Overall, Db were higher during summer than during autumn experiments (Figure

3.5A). During both seasons, there were two peaks in vertical profiles: a first one close to the

sediment-water interface and a second one between 8 and 11mm deep in the sediment

column. During summer experiments, there was also a third peak located between 16 and

17mm deep in the sediment.

The vertical profiles of Db are shown in Figure 3.5B and 3.5C for all combinations of

Ed and Fo. The effect of Ed was significant in both food conditions (Table S5) with a trend

toward decreasing Db with increasing experiment duration. As observed for inversed waiting

times, this effect was less pronounced when food was added, due to: (1) a reduction in the

overall range of inversed waiting times (eg, maximal values of mean inversed waiting times

of 1.30 vs. 3.14 h-1

during experiments with and without food addition, respectively), and (2)

a reduction in the experiment duration required to reach a stable vertical profile (eg, 24 and

36h during experiments with and without food addition, respectively) (Table S5). Only

during the shortest experiment duration (i.e. 6h), vertical profiles exhibited significantly

higher Db without than with food addition (Figure 3.5B, 3.5C; Table S6).

CHAPITRE 3 :



[109]

A

Db (cm2.yr

-1)

0,0 0,2 0,4 0,6 0,8 1,0 1,2 1,4

Dep

th (

mm

)

-30

-25

-20

-15

-10

-5

0

Autumn

Summer

B

Db (cm2.y

r-1)

0 1 2 3 4 5

Dep

th (

mm

)

-30

-25

-20

-15

-10

-5

0

6H

12H

24H

36H

48H

C

Db (cm2.yr

-1)

0 1 2 3 4 5

Dep

th (

mm

)

-30

-25

-20

-15

-10

-5

0

6H

12H

24H

36H

48H

Figure 3.5: Vertical profiles of Db.

Vertical profiles are shown for pooled autumn and summer experiments (A), for pooled

experiments of the same duration without food addition (B), and for pooled experiments of

the same duration with food addition. Horizontal bars are standard errors and refer to between

experiment variability.

CHAPITRE 3 :



[110]

3.2. Overall mean values

3.2.1. Main effects

Table 3.2 shows the results of the PERMANOVA and PERMDISP procedures carried

out on: (1) the normalized number of jumps, and (2) the overall mean values resulting from

the depth-stratified bootstrap procedure described above and applied to inversed waiting

times, jump characteristics and Db.

Se significantly affected the mean normalized number of jumps and the mean of both

jump characteristics. There was no direct effect of Fo. Conversely Ed significantly affected

mean inversed waiting time and Db. There were no interactions between the effects of Se and

Fo and between those of Se and Ed. Conversely, there were significant interactions between

the effects of Fo and Ed on mean inversed waiting time and Db. Rep significantly affected the

mean values of all tested parameters except inversed waiting time. Interactions between Se,

Fo and Ed significantly affected mean normalized number of jumps and .

Se significantly affected the dispersion of: (1) the normalized number of jumps and (2)

both jump characteristics. Fo only affected the dispersion of jump lengths. Ed alone and in

interaction with: (1) Fo and (2) Se and Fo, significantly affected the dispersion of all

parameters but the mean normalized number of jumps. There were significant interactions

between the effects of Se and Fo on the dispersion of normalized number of jumps and jump

lengths. There were interactions between the effects of Se and Ed on the dispersion of all

tested parameters.

3.2.2. Normalized number of jumps

Changes in overall mean normalized number of jumps with Ed are shown in Figure

3.6 for all combinations of Se and Fo. During autumn, variability was limited among

experiments (Table 3.2). The mean normalized number of jumps was low (i.e., less than 0.04

jumps.luminophore-1

) and significantly higher during experiments with food addition. There

was no significant effect of Ed on the mean normalized number of jumps. During summer, the

mean normalized number of jumps was significantly higher (i.e., 0.07 jumps.luminophore-1

)

(Table 3.2). The mean normalized number of jumps tended to decrease with increasing

CHAPITRE 3 :



[111]

experiment duration during summer experiments without food addition. During summer

experiments with food addition, the mean normalized numbers of jumps were low during the

6 and 12h experiments and then significantly higher during the 24, 36 and 48h experiments

(Table S9).

The mean normalized numbers of jumps were significantly higher during the 48h

experiments with food addition than during the 48h experiments without food addition.

CHAPITRE 3 :



[112]

Table 3.2 : Results from Permanova and Permdisp analyses for differences in, normalized

number of jumps (jumps.luminophore-1

), and overall mean values of inversed waiting times

(h-1

), jump characteristics (means and standard deviations of jumps lengths in mm) and Db

(cm2.yr

-1), among Season (Se), food addition (Fo) and experimental duration (Ed), based on a


Factors


Inversed waiting times

Mean jump lengths

of jump lengths

Db

Se

df

MS

Pseudo-F

p(perm)

1

4.38E-2

7.41*

0.0072

1

0.36

0.4

0.5302

1

1.03E-2

2.32*

0.1313

1

9.65E-3

4.47*

0.036

1

0.36

2.18

0.1556

Fo

df

MS

Pseudo-F

p(perm)

1

1.08E-2

0.18

0.6759

1

1.32

1.45

0.2348

1

2.79E-3

0.63*

0.1313

1

1.95E-3

0.9

0.3645

1

0.16

0.98

0.3291

Ed

df

MS

Pseudo-F

p(perm)

4

1.22E-4

0.34

0.8492

4

31.44

53.107*

0.0001

4

1.57E-4

0.25*

0.9087

4

3.71E-4

1.13*

0.3539

4

3.48

64.93*

0.0001

SexFo

df

MS

Pseudo-F

p(perm)

1

1.82E-3

0.31

0.5891*

1

0.18

0.19

0.6617

1

2.99E-3

0.67*

0.4273

1

3.64E-3

1.79

0.2162

1

6.48E-2

0.39

0.5424

SexEd

df

MS

Pseudo-F

p(perm)

4

1.05E-4

0.29*

0.8877

4

0.27

0.45*

0.7669

4

3.61E-4

0.57*

0.6935

4

1.23E-4

0.37*

0.8364

4

5.92E-2

1.1*

0.361

FoxEd

df

MS

Pseudo-F

p(perm)

4

6.55E-4

1.8327

0.1265

4

2.3

3.89*

0.0062

4

2.32E-4

0.37*

0.8338

4

1.5E-4

0.46*

0.7701

4

0.15

2.84*

0.0293

Rep

df

MS

Pseudo-F

p(perm)

28

5.91E-3

16.53

0.0001

28

0.91

1.54

0.0637

28

4.44E-3

7.04*

0.0001

28

2.16E-2

6.59*

0.0001

28

0.17

3.08

0.0001

SexFoxEd

df

MS

Pseudo-F

p(perm)

4

9.16E-4 2.56

0.0378

4

9.51E-2

0.16*

0.9595

4

1.37E-3

2.17*

0.0736

4

8.94E-4

2.73*

0.0284

4

0.12

2.22*

0.0669


CHAPITRE 3 :



[113]

Experimental duration (hours)

0 10 20 30 40 50 60

No

rma

lize

d n

um

be

r o

f ju

mp

s (

jum

ps

.lu

min

op

ho

res

-1)

0,00

0,02

0,04

0,06

0,08

0,10

Autumn No food

Autumn food

Summer No food

Summer Food

Figure 3.6: Effect of experimental duration on mean normalized numbers of jumps.

Mean values are shown for summer and autumn experiments with and without food addition.

Vertical bars are standard errors.

3.2.3. Inversed waiting times

Changes in overall mean inversed waiting times with Ed are shown in Figure 3.7 for

all pooled experiments with and without food addition. Mean inversed waiting times

significantly decreased with experimental duration for both Fo conditions. This was also the

case for the variability among experiments. These trends were even slightly more pronounced

without food addition. Mean inversed waiting times were significantly higher during the 48h

experiments with food addition (0.74 h-1

) than during the 48h experiments without food

addition (0.59 h-1

) (Table S11).

CHAPITRE 3 :



[114]


0 10 20 30 40 50

Inve

rsed

wa

itin

g t

ime

(h

-1)

0

1

2

3

4

No food

Food

Figure 3.7: Effect of experimental duration on mean inversed waiting times.

Mean values are shown for pooled with and without food addition experiments. Vertical bars

are standard errors.

3.2.4. Jump characteristics

Changes in mean jump lengths with Experimental duration are shown in Figure 3.8

for all combinations of Se and Fo. Mean jumps lengths were not significantly affected by Ed

(Table 3.2), whereas among experiments variability tended to decrease with Ed. Among

experiments variability was also significantly higher: (1) during experiments with than during

experiments without food addition (Table 3.2), and (2) during experiments with food addition

conducted in summer than during those conducted in autumn. These trends were especially

pronounced for the 6, 12 and 24 hours long experiments.

CHAPITRE 3 :



[115]


0 10 20 30 40 50

Me

an

ju

mp

le

ntg

h (

mm

)

0,08

0,10

0,12

0,14

0,16

0,18

0,20

Autumn No food

Autumn Food

Summer No food

Summer Food

Figure 3.8: Effect of experimental duration on mean jump lengths.



Changes in mean with Ed are shown in Figure 3.9 for all combinations of Se and

Fo. These values were significantly higher in summer than in autumn (Table 3.2). During

autumn experiments and for the 6h experimental duration, the means of the standard deviation

of jump lengths tended to be higher (although non-significantly) without than with food

addition. They then tended to become lower for longer experimental durations. An almost

similar pattern was observed for the summer experiments, with higher values with food

addition for the 6 and 12h experiments and lower ones for longer durations.

Among experiments variability was significantly higher during autumn than during

summer experiments, especially with food addition. Among experiments variability was also

significantly higher with than without food addition, especially for the 6 and 12h experimental

durations in autumn and for the 12h experimental duration in summer (Table 3.2).

CHAPITRE 3 :



[116]


0 10 20 30 40 50

Sta

nd

ard

de

via

tio

n o

f ju

mp

le

ng

ths

(m

m)

0,10

0,11

0,12

0,13

0,14

0,15

Autumn No food

Autumn Food

Summer No food

Summer Food

Figure 3.9: Effect of experimental duration on mean .



3.2.5. Db

Changes in mean Db with Experimental duration are shown in Figure 3.10 for the two

conditions of Fo. Both mean Db values and their dispersion, significantly decreased with Ed

with and without food addition (Table 3.2). This trend was however more marked without

food addition. During experiments with food addition, there was no significant decrease in

mean Db values between 24 and 36h experimental durations (Table S12). Although not

significant, the mean Db tended to be higher without food addition for experimental durations

shorter than 24h, and lower without food addition for longer experimental durations.

CHAPITRE 3 :



[117]


0 10 20 30 40 50

Db

(cm

2.y

r-1)

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

No food

Food

Figure 3.10: Effect of experimental duration on mean Db.

Mean values are shown for pooled with and without food addition experiments. Vertical bars

are standard errors.

IV. Discussion

4.1. Methodological considerations when assessing

environmental effects on particle mixing process using direct

measurements of particle mixing fingerprints

Our results were obtained based on measurements of individual luminophore

displacements. This new experimental approach has been developed and successfully used to

describe particle mixing fingerprints in A. alba (Bernard et al., 2012) as indicated by the good

agreement between those mixing fingerprints and the current qualitative knowledge regarding

the feeding ethology of and the particle mixing induced by this clam. However, the present

study constitutes the first attempt to assess the effect of environmental factors on those

CHAPITRE 3 :



[118]

fingerprints using this approach. This requires several methodological considerations

regarding: (1) the appropriate descriptor of jump frequency, (2) possible experimental and

statistical designs, and (3) the possibility to measure and compare vertical profiles instead of

overall mean values of each considered parameter.

4.1.1. Descriptor of jump frequency

In the CTRW model, particle displacements are controlled by two main types of

parameters: the frequency of jumps and the characteristics (including their length) of

individual jumps (Meysman et al. 2008a,b and 2010). The frequency of jumps is usually

described through the frequency distributions of waiting times, which are defined as the time

intervals between two consecutive jumps of the same individual particle (Wheatcroft et al.,

1990; Meysman et al., 2008a,b). Our results clearly show that waiting times are affected by

experimental duration. Mean waiting times increase with experimental duration, which simply

results from the fact that it is not possible to measure a waiting time longer than the

experimental duration. The frequency of jumps can however also be approached through the

assessment of the probability of jumps of a luminophore during an elementary time interval.

During the present study, this was achieved through the computation of the normalized

number of jumps (i.e., the proportion of the total number of luminophores that have jumped

within an elementary time interval). This parameter is closely related to the “activity”

parameter proposed by Schiffers et al. (2011) with the difference that the normalized number

of jumps is based on the assessment of individual luminophore displacements and not on

changes in luminophore concentrations. During the present study, the normalized number of

jumps was integrated over time periods of increasing durations to become more comparable

with waiting times. Our results show that the normalized number of jumps was not affected

by experiment duration. They therefore suggest that it constitutes a better proxy of jump

frequency than waiting time when assessing particle mixing fingerprints based on the measure

of individual particle displacements.

4.1.2. Experimental design

CHAPITRE 3 :



[119]

Previous ethological and particle mixing studies carried out in the genus Abra have

shown a high degree of inter-individual variability (Grémare et al. 2004, Maire et al. 2006),

which certainly constitutes the main issue when assessing the effects of environmental factors

which could vary over short periods of time (e.g., Food addition). In order to overcome this

difficulty, some authors have used a before-after control design, which consists in

measurements of the same set of individuals before and after food addition (Duchêne and

Rosenberg 2001; Grémare et al. 2004). Together with the use of statistical tests for paired

samples, this clearly facilitated the assessment of the effect of environmental factors through a

better handling of inter-individual variability (Grémare et al. 2004). Unfortunately, this design

is not appropriate when assessing particle mixing through the coupling of the experimental

assessment of luminophore vertical profiles and modeling (Maire et al. 2007b). This is also

the case when measuring individual luminophore displacements such as during the present

study. The addition of phytoplankton detritus at the water-sediment interface during the

monitoring indeed induces: (1) changes in luminance and/or color composition of

luminophores located at the sediment-water interface, which may result in the detection of

false individual jumps, and (2) real but yet artefactual individual jumps directly caused by

changes in the hydrodynamics when adding food. In addition, and due to the dependency of

waiting times on Experimental duration, a sound assessment of the effect of Food addition on

Db would require long monitoring periods before and after food addition. The corresponding

time scale (i.e., order of day) is not compatible with a punctual addition due to the quick (i.e.,

order of hour) dilution of the food added as shown by Maire et al. (2006) for both A. ovata

and A. nitida. In this sense, it should be stressed that the new experimental approach used

during the present study is not compatible with a before-after control design and therefore

does not allow for an optimal handling of inter-individual variability.

4.1.3. Comparison of vertical profiles vs. overall mean values

The experimental approach used during the present study presents some advantages

relative to the classical assessment of particle mixing through the coupling of the

experimental assessment of luminophore vertical profiles and modeling. This last approach

indeed supposes spatially homogeneous particle mixing and produces: (1) single distributions

of waiting times and jump lengths and (2) a single value of Db for each experiment (i.e.,

combination of Season, Food addition and Experimental duration in the case of the present

CHAPITRE 3 :



[120]

study). Conversely, the use of our approach results in a 2D description of particle mixing

(Bernard et al., 2012), which presents a higher degree of finesse in describing particle mixing

and thus a higher potential in assessing the effect of environmental factors on particle mixing.

According to Bernard et al. (2012), the horizontal component of particle mixing is highly

variable among clams because mostly controlled by the geometry of the siphonal gallery

network and resulting from processes occurring over short time scales such as the preferential

use of a siphonal gallery during the duration of the experiment. There is thus no real sense in

considering explicitly this component when assessing the effects of environmental factors on

particle mixing. These authors also showed that, in A. alba, particle mixing fingerprints varied

with depth in the sediment column. This pattern is more constant between individuals because

corresponding to changes in different types of activities occurring at different depths due to

the morphology of the clams and of their positioning within the sediment column. The

comparison of vertical profiles of the values of the different parameters characterizing particle

mixing may therefore prove both more realistic and more discriminative (i.e., between the

different levels of tested environmental factors) than the crude comparison of their mean

values. During the present study, we compared the ability of these two approaches to show

significant effects of environmental parameters on particle mixing. Our results clearly show

that the comparison of profiles resulted in a detection of a higher number of significant effects

that the comparison of mean values. Overall, the comparisons of profiles resulted in the

detection of a significant effect for 8 factors or combinations of factors that were not detected

through the comparisons of mean values. Conversely, the comparisons of mean values

resulted in the detection of a significant effect for only one combination of factors that was

not detected through the comparisons of profiles (Tables 3.1 and 3.2). In this sense our

results clearly suggest that the comparison of profiles is much more efficient than the

comparison of mean values in detecting the effect of environmental factors on particle mixing

in A. alba. It should however be underlined that the comparison of mean values remains the

only one allowing for comparisons with literature data derived from a classical modeling

approach (Maire et al. 2007a; Breackman et al. 2010).

In conclusion, the use of a new experimental approach based on the assessment of

individual jumps has important methodological consequences when assessing the effect of

environmental factors on particle mixing. First, this procedure results in a clear dependency of

waiting times on experiment duration, and classical “before and after” experimental design

cannot be carried out to assess the effect of food availability. Our results nevertheless show

CHAPITRE 3 :



[121]

that these drawbacks can be overcome by: (1) using the normalized number of jump as a

proxy of the frequency of jumps, and (2) comparing vertical profiles instead of mean values

of considered parameters among treatments.

4.2. Seasonal changes in particle mixing fingerprints

The comparisons of vertical profiles clearly show that particle mixing fingerprints in

A. alba differ between the two tested seasons. Particle mixing was stronger in summer than in

autumn. All tested parameters were significantly affected by the factor Season (Table 3.1)

and the analysis of the corresponding profiles shows that these parameters were affected all

over the sediment column. Seasonal differences thus affected the three vertical functional

areas identified by Bernard et al. (2012), namely: the sediment-water interface, the network of

siphonal galleries and to a lower extent the shell area. Both proxies of jump frequency tended

to be higher in summer than in autumn all over the sediment column (Fig. 3.1 and 3.2). Jump

lengths were also higher and more variable in summer than in autumn (Fig. 3.3A and 3.4A).

Consequently Db were higher in summer than in autumn (Fig. 3.5A). In this last case,

differences were almost null close to the sediment-water interface and maximal deep in the

sediment column (i.e., within the shell area).

To our knowledge, only one study based on the CTRW model has been carried out to

assess the effect of season on particle mixing by A. alba (Braeckman et al. 2010). These

authors reported significant differences in the burying depths between summer and winter (the

clams being positioned deeper in the sediment column during winter) but no significant

differences in the proportions of luminophores found deeper than 0.5cm and in Db.

Conversely, we did not observe any clear differences in burying depth (Bernard personal

observation) but did observe significant differences in particle mixing intensity (including

Db). Two (nonexclusive) hypotheses may contribute to explain such a discrepancy.

Braeckman et al. (2010) used a classical CTRW approach and therefore derived Db from the

fitting of luminophore profiles using a CTRW model. Therefore, they only had access to

mean values of Db, whereas our approach was based on the assessment of individual jumps

and therefore allowed for the comparisons of both profiles and mean values (see above). It

should be underlined that using our own mean values, we did not detect any difference of the

Se factor between Db values computed in autumn and in summer whereas the comparison of

CHAPITRE 3 :



[122]

Db profiles allowed for the assessment of such significant differences. Discrepancies between

the two studies may thus first result from the use of different CTRW approaches. The

maximal duration of our experiments was 48h whereas those of Braeckman et al. (2010)

lasted for 14 days. We found that increasing experimental durations (i.e., from 6 to 48h)

resulted in a diminution of Db. A similar tendency has already been reported in A. nitida

(Maire et al. 2006). These last authors attributed this pattern to a depletion of luminophores at

the sediment-water interface during the curse of the experiments. They also showed that the

effects of some environmental factors could diminish with experimental duration because of

such a limitation. The long experimental duration used by Braeckman et al. (2010) may

therefore also contribute to the lack of seasonal effect on Db reported by these authors.

Maire et al. (2007a) looked at seasonal differences in particle mixing by the closely

related bivalve A. ovata using a classical CTRW modeling approach. Conversely to

Braeckman et al. (2010), these authors reported lower waiting times and more variable jump

lengths in summer than in autumn (mean values of 7.61 vs. 21.30h, and 0.21 vs. 0.07cm, for

waiting times and , respectively). Consequently, they reported higher Db in summer than in

autumn (mean values of 32.60 vs. 1.01 cm².yr-1

, respectively). Differences in the absolute

values of Db recorded during the studies by Maire et al. (2007a) and Bernard et al. (2012, who

used the same experimental approach as during the present study) are partly methodological

and have already been discussed in details (Bernard et al., 2012). From a qualitative

standpoint, these results are very similar to ours. Nevertheless, our results also show that jump

length and not only are higher in summer than in autumn. The main explanation proposed

by Maire et al. (2007a) to account for between seasons differences in particle mixing was an

effect of temperature (10°C vs 20°C in their study). In the genus Abra, particle mixing is

clearly resulting from clam’s activity (Maire et al. 2007b). Grémare et al. (2004) studied the

effect of temperature on the intensity of the siphonal activity in the bivalve A. nitida. They use

two independent indices for describing this activity, namely the percentage of time active

(ACT) and the mean activity per time active (MATA). In both cases they reported a

significant and positive effect of temperature. ACT is indicative of the frequency of feeding

and thus indirectly of the frequency of jumps induced by this activity. In this sense,

differences in both indices of frequency of jumps recorded during the present study could

result from the effect of temperature on clam’s activity. MATA is indicative of the surface

prospected by the inhalant siphon during feeding and is therefore indicative of the extension

of this siphon (i.e. the higher the MATA, the more extended the siphon). A positive effect of

CHAPITRE 3 :



[123]

temperature on MATA (as in Grémare et al. 2004) could therefore account for the occurrence

of longer and more variable jumps at the sediment-water interface during the present study.

The occurrence of longer and more variable jumps within the sediment column could result as

well from differences in the patterns of siphonal activity induced by temperature. Deposit-

feeding bivalves are known to destabilize the sediment between the mounds they create

(Levinton, 2001), i.e. in areas corresponding to the network of siphonal galleries. In this

sense, the higher siphonal activity found in summer could explain the occurrence of long and

variable jumps in breaking the elastic sediment matrix and therefore making particles easier to

displace over relatively longer distance. Such an impact of benthic organisms on visco-elastic

properties of sediment has already been demonstrated for burrowing worms (Dorgan et al.

2006).

4.3. Effect of food availability on particle mixing fingerprints

During the present study, particle mixing was overall negatively affected by Food

addition. However, the so-induced changes varied with: (1) the considered component of

particle mixing fingerprints, (2) Experimental duration and, (3) Season. These results can be

compared with those of Maire et al. (2006) who studied the effect of food addition on particle

mixing by the two closely related bivalves A. ovata and A. nitida.

4.3.1. Effect on jump frequency

The vertical profiles of the normalized number of jumps were significantly affected by

the interactions between Season, Experimental duration and Food addition. During autumn,

and for all experimental durations, vertical profiles were not significantly affected by Food

addition. During summer and without food addition, vertical profiles were not significantly

affected by Experimental duration. Conversely, with food addition, vertical profiles differed

significantly for short (i.e., 6 and 12h) and long (i.e., 12, 24, 36 and 48h) experiments. In this

last case, differences between profiles mostly resulted from the occurrence of lower

normalized numbers of jumps in the upper part of the sediment column when food was added.

A similar negative effect has been reported by Maire et al. (2006) in A. ovata. These

authors tested 3 levels of addition of phytoplanktonic detritus and reported a decrease in Db at

the highest concentration. Their results supported those of Grémare et al. (2004), who

CHAPITRE 3 :



[124]

previously reported a decrease in the the siphonal activity of the same species at high food

concentration. Both authors reported that the functional responses of the closely related

species A. nitida was completely different and characterized by a constant increase in siphonal

activity and particle mixing activities with food concentration. Their interpretation was that

the individuals of A. nitida originating from Swedish waters were more adapted to strong and

restricted in time phytoplanktonic blooms (Lindahl et al., 1998) than the individuals of A.

ovata originating from a Mediterranean lagoon. The individuals of A. alba used during the

present study were collected in the Arcachon Bay, where seasonal changes (i.e., in terms of

timing and intensity) in primary production are relatively close to those of the Mediterranean

Sea (Glé et al., 2008; Lazzari et al., 2012). Our hypothesis is thus that, during our

experiments, food addition has induced an inhibition of particle mixing in A. ovata as well. It

can be tested by carrying out food addition experiment at a lower concentration. The fact that

the inhibition recorded during the present study was restricted to short experimental durations

is probably linked to a progressive impoverishment of the sediment-water interface following

a punctual addition of phytoplankton detritus as already suggested by Maire et al. (2006).

Besides changes in the intensity of deposit-feeding, switch in feeding behavior may

however also contribute to account for the transitory inhibition of particle mixing. According

to Levinton (1991) “there appears to be no exclusively deposit-feeding Tellinacea”.

Moreover, several species of the genus Abra, including A. alba, have been shown to switch

between deposit and suspension feeding (Rosenberg, 1993). In this last case, the inhalant

siphon remains immobile in the water column for long periods of time (Grémare et al. 2004)

and then sediment particle movements resulting from siphonal become scarcer and also

shorter. Together with local hydrodynamics (Taghon et al., 1980; Levinton, 1991), which

were constant during our experiments, concentration of suspended particulate organic matter

is known to constitute one of the main controlling factors of the switch between deposit and

suspension feeding (Taghon and Greene, 1992 ; Riisgard and Kamerans, 2001). Grémare et

al. (2004) also observed suspension-feeding in A. nitida mostly immediately after food

addition. Our results suggest that this is also the case in A. alba since visual inspections of our

video records showed that suspension feeding did occur in the beginning (i.e. within the first

hour) of 10 out of our 16 with food addition experiments versus only 1 out of 16 of our

without food experiments (GB, personal observations). The transitory inhibition of particle

mixing by A. alba may thus partly result from a switch in feeding mode and not only from a

decline in deposit-feeding intensity as already suggested for A. ovata (Maire et al. 2006). This

inhibition was not observed in autumn, which could be related to the fact that, during this

CHAPITRE 3 :



[125]

particular season, feeding activity is severely limited by another environmental factor, namely

temperature (see above, Maire et al. 2007a).

4.3.2. Effect on jump lengths and

Vertical profiles of both jump characteristics (i.e., mean jump lengths and ) were

significantly affected by Food addition without any significant interaction with the two other

tested factors (Table 3.1). This effect consisted in shorter mean jump lengths and lower

within the depth range of the sediment column corresponding to the network of siphonal

galleries when food was added. Vertical profiles of these two parameters showed regular

decreases with depth when food was added. Conversely, without food addition, those vertical

profiles exhibited a subsurface peak located within the depth range corresponding to the

network of siphonal galleries. The difference in these two patterns suggests that foraging

behavior by the siphons below the sediment-water interface differed without and with food

addition. Here again, these results are indicative of an inhibition of particle mixing by food

addition. Two hypotheses may be raised to explain them. As explained above, food addition

may at certain concentration inhibit siphonal activity at the sediment-water interface and thus

particle mixing including jump characteristics in individuals of the genus Abra originating

from temperate populations (see above) as discussed above for jump frequency. However, this

hypothesis does not appear fully satisfactory because, the upper part of the profiles of both

jump lengths and were not affected by food addition. Maire et al. (2007b) reported that: (1)

siphonal activity of A. ovata tends to decrease with experiment duration, and (2) A. ovata

never explores the same subarea of the sediment surface before the total area delimited by the

extension of their inhalant siphon has been fully prospected. This suggests that the

impoverishment of the surface sediment in organic matter constitutes a key factor in

controlling the feeding activity in A. ovata. During our experiments, such an impoverishment

was probably faster during the without than the with food addition experiments due to: (1)

lower initial concentration during the without food addition experiments, and (2) the

transitory inhibition of deposit-feeding during the with food addition experiments (see above).

Amouroux et al. (1989) observed a regular current induced by the exhalent siphon of A. ovata

in the section of their burrow where faeces are stored and suggested that this was indicative of

a “gardening” behavior. During our experiments, longer and more variable jumps below the

sediment-water interface without food addition could therefore have been caused by

CHAPITRE 3 :



[126]

movements of the inhalant siphon foraging on gardened faeces in response to an

impoverishment of the sediment-water interface during the curse of our experiments.

4.3.3. Effect on Db

Together with , mean waiting times are one of the two components involved in the

computation of Db (Meysman et al., 2008a). During the present study, changes in vertical

profiles of Db were highly similar to those of inversed waiting times. They also showed a

transitory inhibition following food addition during summer experiments as already shown by

Maire et al. (2006) at high food concentration for individuals of A. ovata originating from a

temperate population. Overall Db thus appeared to be more controlled by Tc than by . Such a

dependency explains why Db were also significantly affected by Experimental duration,

which contributes to complicate the assessment of the effects of environmental parameters

during long term experiments by buffering Db values. This problem could be handled by

deriving estimates of Tc (and thus Db) from another proxy of jump frequency, which would

be independent of Ed. Our results suggest that this proxy could be the normalized number of

jumps.

V. Conclusions and perspectives

The use of a new approach allowing for the tracking of individual particles at a high

temporal resolution proved efficient in assessing the effects of Season and Food addition on

particle mixing in the deposit-feeding bivalve A. alba. Our main conclusions regarding these

effects are as follows:

(1) Sediment particle displacements were longer, more variable and more frequent

during summer. This restriction was attributed to the effect of temperature,

which has been shown to affect both siphonal activity and particle mixing in

closely related species. It precluded any significant effect of food addition

during autumn experiments.

CHAPITRE 3 :



[127]

(2) During summer, food addition induced a transitory inhibition of particle

mixing, which resulted from both: (1) a functional response of deposit-feeding

to food addition (as already described in a closely related species at high food

concentrations), and (2) a switch from deposit- to suspension-feeding

immediately following food addition.

From a methodological standpoint, our results show that the estimates of average

waiting time (generally used to assess the jump frequency of particle when describing

sediment particle mixing process with CTRW model) and thus Db are strongly Experimental

duration dependent. In this sense, the present study highlights the need for a better descriptor

of jump frequency at short time scale during experiments based on the tracking of individual

particles. We propose to use the normalized number of jumps for this purpose because: (1) it

was not significantly affected by experiment duration, and (2) it allowed for the detection of

the negative effect of food addition on the frequency of jump during summer experiments,

which was not the case of waiting time. A challenge for future research thus consists in

deriving a proxy of waiting times independent of experiment duration from this parameter. In

any case, future studies assessing on the effects of environmental variables on sediment

particle mixing process should carefully pay attention to observational time and space scales

to capture changes in behavior and their consequences, which can be restricted both in time

and to specific areas of the sediment column.

References




Amouroux, J. M., A. Grémare and J. Amouroux. 1989. Modelling of consumption and

assimilation in Abra alba (Mollusca, Bivalvia). Mar. Ecol. Prog. Ser., 51, 87-97.

Anderson M. J. 2001. A new method for non-parametric multivariate analysis of variance.

Austral Ecol., 26, 32-46.

CHAPITRE 3 :



[128]

Anderson, M. J. 2006. Distance-based tests for homogeneity of multivariate dispersions.

Biometrics, 62, 245-253.

Anderson, M. J., R. N. Gorley and K. R. Clarke. 2008. PERMANOVA+ for PRIMER: Guide

to Software and Statistical Methods. PRIMER-E: Plymouth, UK. 214 pp.







Oceanogr. Ser., 70, 455-492.






Cabaço, S., R. Santos and C. M. Duarte. 2008. The impact of sediment burial and erosion on

seagrasses: A review. Est. Coast. Shelf. Sci., 79, 354-366.

Caradec, S., V. Grossi, S. Hulth, G. Stora and F. Gilbert. 2004. Macrofaunal reworking

activities and hydrocarbon redistribution in an experimental sediment system. J. Sea. Res.,

52, 199-210.

Clarke, K. R. and R. M. Warwick. 2001. Change in marine communities: an approach to

statistical analysis and interpretation, 2nd edition. PRIMER-E: Plymouth.

Dorgan, K. M., P. A. Jumars, B. Johnson and B. P. Boudreau. 2006. Macrofaunal burrowing:

The medium is the message. Oceanogr. Mar. Biol., 44, 85-121.

Duchêne, J. C. and R. Rosenberg. 2001. Marine benthic faunal activity patterns on a sediment

surface assessed by video numerical tracking. Mar. Ecol. Prog. Ser., 223,113-119.




46, 531-547.

Giangrande, A., M. Montresor, A. Cavallo and M. Licciano. 2002. Influence of Naineris

laevigata (Polychaeta: Orbiniidae) on vertical grain size distribution, and dinoflagellate

resting stages in the sediment. J. Sea. Res., 47, 97-108.



CHAPITRE 3 :



[129]



Biol. Ecol., 285/286, 251–263.

Glé, C., Y. Del Amo, B. Sautour, P. Laborde and P. Chardy. 2008. Variability of nutrients and

phytoplankton primary production in a shallow macrotidal coastal ecosystem (Arcachon

Bay, France). Est. Coast. Shelf. Sci., 76, 642-656.

Goldberg, E. D. and M. Koide. 1962. Geochronological studies of deep sea sediments by the

ionium/thorium method. Geochim. Cosmochim. Acta., 26, 417-450.

Gray, J. S. 1974. Animal-sediment relationships. Oceanogr. Mar. Biol. Ann. Rev., 12, 223-

261.






Hughes, R. G., D. Lloyd, L. Ball and D. Emson. 2000. The effects of the polychaete Nereis

diversicolor on the distribution and transplanting success of Zostera noltii. Helgol. Mar.

Res., 54, 129-136.


Mar. Biol. Ecol., 20, 137-15.





Mar. Ecol. Prog. Ser., 446, 285-302.

Lazzari, P., C. Solidoro, V. Ibello, S. Salon, A. Teruzzi, K. Béranger, S. Colella and A. Crise.

2012. Seasonal and inter-annual variability of plankton chlorophyll and primary production

in the Mediterranean Sea: A modeling approach. Biogeosciences, 9, 217-233.





Levinton, J. S. 1991. Variable feeding behavior in three species of Macoma (Bivalvia:

Tellinacea) as a response to water flow and sediment transport. Mar. Biol., 110, 375-383.

Levinton, J. S. 2001. Marine biology: Function, biodiversity, ecology (second edition).

Oxford university press, New York. 515 pp.

CHAPITRE 3 :



[130]

Lindahl, O., A. Belgrano, L. Davidsson and B. Hernroth. 1998. Primary production, climatic

oscillations, and physic-chemical processes: The Gullmar fjord time-series data set (1985-

1996). ICES J. Mar. Sci., 55, 723-729.





Ecol. Prog. Ser., 319, 135-153.










Mar. Biol., 151, 737-749.

Maire, O., P. Lecroart, F. J. R. Meysman, R. Rosenberg, J. C. Duchene and A. Grémare,


Aquat. Biol., 2, 219-238.

McArdle, B. H. and M. J. Anderson. 2001. Fitting multivariate models to community data: A

comment on distance-based redundancy analysis. Ecology, 82(1), 290-297.



Meysman, F. J. R., B. P. Boudeau and J. J. Middelburg. 2003. Relations between local,








Acta, 72, 3460-3478.

CHAPITRE 3 :



[131]



239-254.

Michaud, E., R. C. Aller and G. Stora. 2010. Sedimentary organic matter distributions,

burrowing activity, and biochemical cycling: Natural patterns and experimental artifacts.

Est. Coast. Shelf. Sci., 90, 21-34.



229-331.

Persson, A. and R. Rosenberg. 2003. Impact of grazing and bioturbation of marine benthic

deposit feeders on dinoflagellate cysts. Harmful Algae, 2, 43-50.




101.


Biol. Ann. Rev., 12, 263-300.





Riisgård, H. U. and P. Kamermans. 2001. Switching between deposit and suspension feeding

in coastal zoobenthos. In: Ecological comparisons of sedimentary shores, Reise K. (Ed).

Ecological studies 151, Springer-Verlag, Heidelberg. pp. 73-101.

Rosenberg R. 2001. Marine benthic faunal successional stages and related sedimentary

activity. Sci. Mar., 65 (Suppl. 2), 107-119.



18, 1347-1363.

Schiffers, K., L. R. Teal, J. M. J. Travis and M. Solan. An open source simulation model for

soil and sediment bioturbation. Plos one, 6(12), e280228.


and R. Rosenberg. 2004. In Situ Quantification of Bioturbation Using Time-Lapse



Stead, R. A. and R. J. Thompson. 2006. The influence of an intermittent food supply on the

feeding behaviour of Yoldia hyperborea (Bivalvia: Nuculanidae). J. Exp. Mar. Biol. Ecol.,

332, 37-48.

CHAPITRE 3 :



[132]

Taghon, G. L. and R. R. Greene. 1992. Utilization of deposited and suspended particulate

matter by benthic “interface” feeders. Limnol. Oceanogr., 37(7), 1370-1391.

Taghon, G. L., A. R. M. Nowell and P. A. Jumars. 1980. Induction of suspension feeding in

spionid polychaetes by high particulate fluxes. Science, 210, 562-564.

Tebble, N. 1966. British bivalve seashells. A handbook for identification. Trustees of the

British Museum (Natural History), London, 212 pp.

Tenore, K. R. and E. J. Chesney.1985. The effects of interactions of rate of food supply and

population density on the bioenergetics of the opportunistic polychaete, Capitella capitata

(type 1). Limnol. Oceanogr., 30(6), 1188-1195.



Venturini, N., A. M. S. Pires-Vanin, M. Salhi, M. Bessonart and P. Muniz. 2011. Polychaete

response to fresh food supply at organically enriched coastal sites: Repercussion on

bioturbation potential and trophic structure. J. Marine Syst., 88, 526-541.





Res., 48, 177-201.

CHAPITRE 3 :



[133]

Supporting informations

Table S1: results of pairwise Permanova and Permdisp analyses for differences in vertical depth

profiles of normalized number of jumps and inversed waiting time among levels of Se within all

possible combinations of Fo and Ed, based on Euclidean resemblance matrix.


Inversed Waiting time

No food Food No food Food Autumn

vs Summer

Autumn vs Summer

Autumn vs Summer

Autumn vs Summer

6H t

p

1.70* 0.0322

1.49 0.054

1.04 0.3431

0.92 0.4612

12H t

p

1.89* 0.004

1.42 0.1146

1.50 0.0283

1.08 0.249

24H t

p

1.89* 0.0161

2.11* 0.0107

1.73 0.0327

1.14 0.1775

36H t

p

1.63 0.0578

1.17 0.219

1.09 0.2912

1.22 0.0712

48H t

p

1.67 0.0337

2.02* 0.0286

1.74* 0.0113

1.09 0.2676


Table S2 : results of pairwise Permanova and Permdisp analyses for differences in vertical depth

profiles of normalized number of jumps and inversed waiting time among levels of Fo within all

possible combinations of Se and Ed, based on Euclidean resemblance matrix.


Inversed Waiting time

Autumn Summer Autumn Summer No food

vs Food

No food vs Food

No food vs Food

No food vs Food

6H t p

0.60 0.8542

1.19 0.2631

0.99 0.4471

1.25 0.1494

12H t p

0.59 0.8302

0.92 0.4214

0.89 0.7192

1.12 0.2556

24H t p

1.05 0.3589

1.11 0.296

0.80 0.9989

1.09 0.2808

36H t p

1.12 0.2734

1.02 0.3527

1.04 0.3464

0.83 0.8778

48H t p

0.65 0.8131

1.11 0.2942

1.04* 0.3621

0.96 0.5507


CHAPITRE 3 :



[134]


profiles of normalized number of jumps and inversed waiting time among levels of Ed within all

possible combinations of Se and Fo, based on Euclidean resemblance matrix.

Normalized number of jumps Inversed Waiting time

Summer Autumn Summer Autumn

No food Food No food Food No food Food No food Food

6H vs 12H t p

1.04 0.3623

1.02 0.41

0.82 0.6688

0.69 0.625

1.15 0.2

1.86 0.0439

1.59 0.0421*

1.04 0.3579

6H vs 24H t p

1.26 0.2394

2.08 0.0099

0.78 0.6495

0.93 0.4319

1.79 0.0051*

1.98 0.0105*

2.17 0.0003*

1.10 0.2329

6H vs 36H t p

1.26 0.2456

2.27 0.0144

0.87 0.5262

1.25 0.174

2.06 0.0013*

1.66 0.0185

2.20 0.0004*

0.96 0.5198*

6H vs 48H t p

1.31 0.2387

1.94 0.048

0.65 0.86

1.19 0.2236

2.17 0.0005*

1.64 0.012

2.37 0.0004*

1.14 0.1626*

12H vs 24H t p

1.26 0.2569

2.14 0.0145

1.0513 0.323

0.95 0.4114

1.65 0.0144

1.38 0.1514

2.48 0.0001

1.23 0.2969

12H vs 36H t p

1.24 0.2565

2.56 0.0079

1.06 0.3206

1.22 0.1925

2.18 0.0002

1.0802 0.2886

2.50 0.0001

0.96 0.5602

12H vs 48H t p

1.32 0.2183

2.11 0.0395

0.72 0.7594

1.17 0.2381

2.13 0.0002*

1.11 0.2564

3.04 0.0001*

1.14 0.2056*

24H vs 36H t p

0.94 0.466

0.87 0.5118

1.01 0.3792

1.19 0.187

2.47 0.0008

0.9 0.6762

1.38 0.2439

0.73 0.7917

24H vs 48H t p

1.03 0.409

0.97 0.3995

0.76 0.6937

0.95 0.4249

2.46 0.001

0.93 0.6235

1.39 0.0025*

0.84 0.7052

36H vs 48H t p

0.88 0.5066

1.03 0.3411

1.13 0.2935

1.0894 0.3094

1.38 0.2125

0.79 0.5378

1.08 0.1797

1.4647 0.0577


CHAPITRE 3 :



[135]


profiles of mean jump lengths and standard deviation of jumps lentghs among levels of Ed, based on


mJL JL

6H vs 12H t p

1.63 0.0383

1.47 0.0636

6H vs 24H t p

1.69 0.0309

1.57 0.0447

6H vs 36H t p

1.91 0.0117

1.75 0.0156

6H vs 48H t p

2.34* 0.001

2.01 0.0053

12H vs 24H t p

1.42 0.087

1.35 0.1043

12H vs 36H t p

1.5683 0.0509

1.49 0.0564

12H vs 48H t p

1.88 0.0104

1.64 0.0258

24H vs 36H t p

1.22 0.1819

1.06 0.3157

24H vs 48H t p

1.49 0.0625

1.13 0.2555

36H vs 48H t p

1.5344 0.0498

1.44 0.0559


CHAPITRE 3 :



[136]


profiles of Db among levels of Ed within levels of Fo, based on Euclidean resemblance matrix.

Db

No food Food

6H vs 12H t p

1.43 0.0072

1.18 0.1682

6H vs 24H t p

1.86 0.0001

1.51 0.0177

6H vs 36H t p

2 0.0001

1.44 0.0256

6H vs 48H t p

2.09 0.0001

1.69 0.0017

12H vs 24H t p

1.7315 0.0002

1.6 0.014

12H vs 36H t p

2.09 0.0001

1.18 0.1972

12H vs 48H t p

2.09 0.0001

1.52 0.0294

24H vs 36H t p

2.7 0.0001

0.69 0.9678

24H vs 48H t p

2.55 0.0001

0.96 0.531

36H vs 48H t p

1.33 0.1905

1.44 0.0572



profiles of Db among levels of Fo within levels of Ed, based on Euclidean resemblance matrix.

Db

6H 12H 24H 36H 48H

Food vs No food t p

1.39 0.0246

1.05 0.3114*

1.03 0.3557

0.94 0.6294

1.01 0.4229


CHAPITRE 3 :



[137]

Table S7 : results of pairwise Permanova and Permdisp analyses for differences in overall mean

values of normalized number of jumps among levels of Se within all possible combinations of Fo and

Ed, based on Euclidean resemblance matrix.


No food Food

Autumn vs Summer

Autumn vs Summer

6H t p

2.07* 0.0329

0.82 0.4348

12H t p

2.44* 0.0016

0.67 0.5395

24H t p

2.31 0.0129

1.77 0.0859

36H t p

2.10 0.0399

1.29 0.2218

48H t p

2.03 0.0616

1.94 0.0455



values of normalized number of jumps among levels of Fo within all possible combinations of Se and

Ed, based on Euclidean resemblance matrix.


Autumn Summer

No food vs Food

No food vs Food

6H t p

1.00* 0.3329

0.73 0.4882

12H t p

1.47* 0.173

0.98 0.3768

24H t p

1.13 0.2778

0.35 0.7261

36H t p

1.59 0.1358

0.44 0.6789

48H t p

0.50 0.6388

0.78 0.4839


CHAPITRE 3 :



[138]


values of normalized number of jumps among levels of Ed within all possible combinations of Se and

Fo, based on Euclidean resemblance matrix.


Summer Autumn

No food Food No food Food

6H vs 12H t p

6.01E-2 0.9548

0.44 0.6829

0.72 0.4941

0.28 0.7855

6H vs 24H t p

0.60 0.5754

1.71 0.1302

0.24 0.8167

0.42 0.6781

6H vs 36H t p

0.66 0.5366

1.38 0.2053

0.28 0.7928

0.56 0.5926

6H vs 48H t p

0.96 0.3901

1.16 0.3002

0.31 0.7695

0.52 0.6154

12H vs 24H t p

0.89 0.412

3.10 0.0119

0.95 0.3728

1.07 0.31

12H vs 36H t p

0.87 0.4296

3.11 0.0143

0.94 0.3739

0.54 0.6062

12H vs 48H t p

1.26 0.2544

2.46 0.0295

0.67 0.5985

0.81 0.4385

24H vs 36H t p

0.72 0.502

0.63 0.5494

0.73 0.4946

0.89 0.3974

24H vs 48H t p

1.21 0.276

0.17 0.859

0.51 0.66

0.42 0.6894

36H vs 48H t p

1.05 0.3231

0.28 0.7711

0.24 0.82

1.66 0.1378


CHAPITRE 3 :



[139]


values of inversed waiting time among levels of Ed within all levels of Fo, based on Euclidean

resemblance matrix.

Inversed mean waiting time

No food Food

6H vs 12H t p

3.92 0.0011

3.21 0.0066

6H vs 24H t p

6.92* 0.0001

4.33* 0.0005

6H vs 36H t p

7.30* 0.0001

3.35* 0.0049

6H vs 48H t p

7.94* 0.0001

5.03* 0.0001

12H vs 24H t p

7.77 0.0001

4.44 0.0007

12H vs 36H t p

9.55* 0.0001

2.02 0.0643

12H vs 48H t p

8.43* 0.0001

4.74 0.0003

24H vs 36H t p

5.69 0.0001

0.43 0.7212

24H vs 48H t p

6.50* 0.0001

2.79* 0.0162

36H vs 48H t p

3.59 0.0017

2.13 0.0025



values of inversed waiting time among levels of Fo within all levels of Ed, based on Euclidean

resemblance matrix.

Inversed mean waiting time

6H 12H 24H 36H 48H

Food vs No food t p

1.96 0.0581

1.61 0.1259

0.49 0.6242

1.42 0.1292

2.0673 0.0439


CHAPITRE 3 :



[140]


values of Db among levels of Ed within all levels of Fo, based on Euclidean resemblance matrix.

Db

No food Food

6H vs 12H t p

4.62 0.0004

3.12 0.0083

6H vs 24H t p

8.03* 0.0001

4.49* 0.0001

6H vs 36H t p

8.58* 0.0001

3.92* 0.0014

6H vs 48H t p

8.45* 0.0001

5.25* 0.0001

12H vs 24H t p

8.03* 0.0001

4.58 0.0005

12H vs 36H t p

9.29* 0.0001

3.04 0.0083

12H vs 48H t p

8.00* 0.0001

5.16* 0.0002

24H vs 36H t p

6.03 0.0001

0.77 0.4593

24H vs 48H t p

6.32* 0.0001

2.75 0.0123

36H vs 48H t p

3.19 0.0038

2.30 0.0119


CHAPITRE 3 :



[141]

Transition

Le déploiement de la nouvelle approche, présentée dans le second chapitre de ce

manuscrit, a ainsi permis d’évaluer dans ce troisième chapitre, de manière dynamique et sur

l’ensemble de la partie de la colonne sédimentaire affectée, le contrôle exercé en premier lieu

par la température de l’eau puis par la disponibilité de matière organique fraîche à l’interface

eau-sédiment, sur les caractéristiques du processus de remaniement sédimentaire effectué par

A. alba.

Le second grand objectif de cette thèse de doctorat consistait, quant à lui, en une

évaluation de l’impact de la régression de l’herbier à Zostera noltii sur l’intensité de

remaniement sédimentaire induit par les communautés benthiques dans le bassin d’Arcachon.

Pour ce faire, il était initialement prévu de transposer in-situ, le protocole expérimental et

l’utilisation du nouveau logiciel développés en laboratoire. Cette transposition est rendue

possible par l’utilisation d’un profileur sédimentaire (Rhoads et Germano, 1982), en lieu et

place des aquariums plats, couplé à des éclairages UV (f-SPI) et à un système d’acquisition

d’images à haute fréquence à l’image de l’étude pionnière conduite par Solan et al. (2004b).

Dans les faits, cette démarche s’est trouvée contrariée par certaines des implications,

non pleinement anticipées, des choix effectués lors de la phase de mise au point de la nouvelle

approche méthodologique. Ces implications tiennent particulièrement : (1) à l’inaptitude des

algorithmes développés à mesurer les mouvements de particules non-locaux, impliquant un

passage via le tube digestif, mouvements pourtant prépondérants pour les organismes

appartenant aux groupes fonctionnels dits des « convoyeurs » (Kristensen et al. 2012), et (2) à

des limites technologiques liées à l’acquisition d’images, ne permettant pas pour l’heure de

suivre une portion (i.e., une surface) de sédiment qui soit représentative d’une communauté

benthique et ceci avec une résolution suffisamment importante pour qu’un luminophore soit

plus grand qu’un pixel élémentaire constitutif de cette même image.

Ceci explique que les expériences ex et in situ aient été conduites selon deux

approches méthodologiques différentes, même si leurs logiques reposaient toutes deux sur la

paramétrisation d’un même modèle (i.e., le Continuous Time Random Walk Model) de

remaniement sédimentaire (Meysman et al. 2008a, 2008b, 2010).

CHAPITRE 4 :

Comparaison du remaniement sédimentaire dans un herbier à Zostera noltii et dans un sédiment nu :

Effet de la dynamique des phanérogames et des communautés benthiques endogées

[142]

Chapitre 4:

Comparaison du remaniement

sédimentaire dans un herbier à

Zostera noltii et dans un sédiment

nu : Effet de la dynamique des

phanérogames et des communautés

benthiques endogées

CHAPITRE 4 :



[143]

Comparing sediment particle mixing in a Zostera noltii

meadow and a bare sediment mudflat: Effects of seagrass

dynamics and benthic infauna composition.


, Marie-Lise Delgard1, Olivier Maire

1, Aurélie Ciutat

2, Pascal

Lecroart1, Bruno Deflandre

1, Jean Claude Duchêne

2, Antoine Grémare

1




Keywords: seagrass, decline, Zostera noltii, bioturbation, infauna, spatial heterogeneity,

community structure, sediment particle mixing, Melinna palmata, structuring effect.

Running title: In-situ measurements of sediment particle mixing intensity within an intertidal

Zostera noltii meadow.

mailto:[email protected]

CHAPITRE 4 :



[144]

Abstract:

During a one-year survey carried out in the Arcachon Bay (SW France), we measured:

(1) sediment particle mixing intensity, (2) sediment and seagrass population characteristics,

and (3) macrozoobenthic community structures within a well-developed Zostera noltii

meadow and a bare sediment mudflat, where a Zostera noltii meadow has recently

disappeared. Sediment particle mixing intensities were measured through in-situ incubations

of sediment cores after a luminophore input and the subsequent fitting of luminophore vertical

depth-profiles using a Continuous Time Random Walk model. The so-measured sediment

particle mixing intensities (DbNL

) were between 2.99 ± 2.75 and 22.45 ± 43.73 cm².yr-1

within

the bare sediment mudflat and between 0.39 ± 0.30 cm².yr-1

and 18.07 ± 18.14 cm².yr-1

within

the Zostera noltii meadow. Spatial and temporal changes in both infauna community

structure and sediment particle mixing were lower in the Zostera noltii meadow than in the

bare mudflat, which tends to support the structuring and buffering effects of seagrass

meadows on biological sedimentary processes. The Zostera noltii meadow clearly declined

during the period under study as indicated by the significant decrease in its root biomass. This

decline was associated with: (1) an increase in both the mean value and the variability of

DbNL

, and (2) changes in infauna with an increase in the spatial variability of its composition

and a strong decrease of the dominant polychaete Melinna palmata. Within Zostera meadow,

a significant correlation was found between similarity matrices based on: (1) mean DbNL

and

(2) mean abundances of a set of three species, including M. palmata. This highlights the key

role of this species in controlling sediment particle mixing trough the creation of dense tube

mats, which contribute to sediment stabilization. Within the whole data-set, the similarity

matrix based mean DbNL

did not correlate with any of the similarity matrices based on the

abundance of all possible combination of infaunal species. Conversely, there was a positive

correlation between the similarity matrix based on DbNL

variability and the similarity matrix

based on the abundances of Abra segmentum, Glycera convoluta, Heteromastus filiformis and

Tubificoides benedii. This last result is in good agreement with: (1) the suspected role of these

species in controlling sediment particle mixing and (2) the higher sensibility to disturbance of

the variability than the mean values of ecological patterns.

CHAPITRE 4 :



[145]

I. Introduction

Seagrass meadows occupy about 0.1-0.2 % of the world ocean (Duarte 2002). They

are among the most productive ecosystems worldwide (Duarte and Chiscano, 1999).

Furthermore, they store a large fraction of this production, which makes them responsible for

about 15% of the carbon storage in the world Ocean (Duarte and Chiscano, 1999). Seagrasses

are known as “ecosystem engineers”, which (1) enhance biodiversity (Boström and

Bonsdorff, 1997), (2) provide oxygenated structured habitats and food for both epi- and

infauna (Reise, 2002 ; Bouma et al. 2009), and (3) reduce current velocity thereby enhancing

the entrapment of suspended organic matter (Fonseca and Fisher, 1986 ; Meadows et al.,

2012). Seagrass meadows are now declining worldwide due to anthropogenic disturbances

(Duarte, 2002). In temperate areas, major mechanisms responsible for seagrass loss include:

(1) eutrophication, which reduces light penetration and, combined with rising sea water

temperature and sea level, inhibits seagrass growth, (2) wasting diseases resulting in direct

seagrass mortality, and to a lower extent; (3) biological interactions such as herbivory by

birds or macro/megafauna and the introduction of species that can physically affect seagrasses

through bioturbation (Orth et al. 2006).

Bioturbation is defined as “all transport processes carried out by animals that directly

or indirectly affect the sediment matrices” (Kristensen et al., 2012). This process includes

both sediment particle mixing and bioirrigation. Through bioturbation, benthic fauna strongly

affect the chemical, physical and geotechnical properties of marine sediments (Gray, 1974;

Rhoads, 1974; Aller, 1982; Rhoads and Boyer, 1982; Meadows and Meadows, 1991; Gilbert

et al., 1995; Rowden et al., 1998; Lohrer et al., 2004). Sediment particle mixing mainly

results from locomotion, burrowing, defecation and feeding by benthic macrofauna

(Meysman et al., 2006). This process is of major importance in environments where physical

disturbance is low (Lecroart et al. 2007). It controls the fate of sedimented particles and

thereby affects the remineralisation (Kristensen et al, 2000; Caradec et al., 2004) and the

resuspension (Reise 2002) of particulate organic matter.

There are clear and strong interactions between seagrasses and benthic infauna

through: (1) the development of dense roots/rhizomes system by seagrasses, and (2) sediment

mixing by benthic infauna. Many authors have highlighted this complex relationship while

studying the effect of large bioturbators such as Arenicola marina (Philippart 1994 ; Eklöf et

CHAPITRE 4 :



[146]

al. 2011 ; Delefosse and Kristensen 2012 ; Suykerbuyk et al. 2012), Hediste (Nereis)

diversicolor (Hughes et al. 2000) or Thalassinid shrimps (Berkenbusch et al. 2007a ;

Berkenbusch et al. 2007b ; Siebert and Branch 2007) on seagrass growth (Philippart 1994),

seeds burial (Hughes et al. 2000 ; Delefosse and Kristensen 2012 ; Blackburn and Orth 2013),

transplanting success (Hughes et al. 2000 ; Siebert and Branch 2007 ; Wesenbeeck et al.

2007) or seagrass population characteristics and distribution (Philippart 1994 ; Hughes et al.

2000 ; Berkenbusch and Rowden 2007 ; Berkenbusch et al., 2007b ; Eklöf et al., 2011 ;

Suykerbuyk et al., 2012). Most often, they demonstrated a negative effect of bioturbation on

seagrass colonization (Philippart 1994 ; Hughes et al 2000 ; Cabaço, 2008; Meadows et al.

2012 ; Suykerbuyrk 2012) through increased: (1) physical alterations of shoots and rhizomes

(Philippart 1994 ; Hughes et al 2000 ; Cabaço, 2008), and (2) burial rates of seeds and

seedlings (Philippart 1994 ; Hughes et al 2000 ; Cabaço, 2008; Meadows et al. 2012 ;

Suykerbuyrk 2012 ; Delefosse and Kristensen 2012). They also showed a negative effect of

the establishment of dense roots/rhizomes networks, that tend to exclude burrowers (Reise

2002), on the density and burrowing speed of large bioturbators (Hughes et al., 2000 ;

Berkenbusch and Rowden 2007 ; Berkenbusch et al., 2007 ; Siebert and Branch 2005, 2007 ;

Wesenbeeck et al 2007). Conversely, some authors also pointed out that these local negative

interactions could result in an overall positive effect by maintaining spatial heterogeneity and

thereby enhancing seedling recruitment (Townsend and Fonseca 1998 ; Meadows et al.,

2012), or in burying seeds down to an optimal depth for their development (Delefosse and

Kristensen, 2012 ; Blackburn and Orth 2013). At the whole benthic infauna community level,

seagrass meadows clearly structure community pattern and stabilize sediment (Reise 2002).

To our knowledge, vertical sediment particle mixing intensity induced by the whole benthic

infauna community has never been quantitatively assessed within seagrass meadows.

However, studies assessing density and bioturbation of individual bioturbators such as

lugworms and mud shrimps inside and oustide seagrass meadows have shown that both

densities and bioturbation intensity tended to be higher in unvegetated areas than in seagrass

meadows (Berkenbusch and Rowden 2007 ; Berkenbusch et al., 2007 ; Siebert and Branch

2005, 2007 ; Wesenbeeck et al 2007). Overall, such feedback effects (“biomechanical

warfare”, Wesenbeeck et. al 2007) contribute to a dynamic equilibrium between seagrasses

and infauna (Huston 1979). This implies that seagrass meadows already affected by another

stressor (i.e. eutrophication) will be more sensitive and threatened by biomechanical

disturbance induced by sediment particle mixing (Orth et al., 2006 ; Berkenbusch and

Rowden 2007 ; Berkenbusch et al. 2007 ; Eklöf et al., 2011; Meadows et al., 2012). It also

CHAPITRE 4 :



[147]

suggests that, at a local scale, seagrass degradation or disappearance would lead to the

establishment of a more spatially heterogeneous infauna community (Hovel et al., 2002 ;

Boström et al. 2006 ; Bouma et al., 2009 ; Schückel et al., 2012), which could lead to a higher

spatial heterogeneity in sediment particle mixing intensity as well.

Arcachon Bay, a macrotidal lagoon of the south-western European coast, is colonized

by both Zostera noltii, which occupies intertidal areas, and Zostera marina, which inhabits

subtidal channels. In this bay, Zostera noltii meadows clearly prevent intertidal flats from

sediment erosion (Ganthy et al. 2013). It also strongly structures macrozoobenthic

communities (Castel et al. 1989 ; Bachelet et al. 2000 ; Blanchet et al. 2004; Do et al. 2011),

which present only limited seasonal changes within such meadows (Castel et al. 1989 ;

Bachelet et al. 2000). Blanchet et al. (2004) assessed that such a structuring effect particularly

occurred when shoot density were higher than 6000 shoots.m-2

and Bachelet et al. (2000)

highlighted the dominance within seagrass meadows of small annelids such as the oligochaete

Tubificoides benedii. This last author associated this trend with the enrichment of deep

sediment layers with decaying plant materials. The Zostera noltii meadow within the

Arcachon Bay used to cover the major (i.e., 70 out of 110 km²) part of intertidal flats and was

then considered as the largest European seagrass meadow (Auby and Labourg, 1996).

However, Plus et al. (2010) recently estimated that the surface occupied by Z. noltii has

diminished by a third between 1988 and 2008. This decline has been more pronounced since

2005 (Plus et al. 2010) resulting in the replacement of large zostera noltii meadows by bare

mudflats.

In this context, the present study focuses on sediment particle mixing as a functional

response of the whole infauna benthic community to both changes in its own structure and in

seagrass meadow characteristics. Its aim are therefore to: (1) quantitatively assess seasonal

changes in sediment vertical sediment particle mixing intensities induced by the whole

infauna community within both a Zostera noltii meadow and an adjacent bare sediment area

and (2) rely these intensities with infauna community composition and seagrass meadow

dynamics in order to (3) assess the impact of the decline of intertidal seagrass meadows on

biological sedimentary processes. This was achieved through the comparisons, over a one-

year survey, of: (1) sediment mixing intensity, (2) sediment and seagrass population

characteristics, and (3) macrozoobenthic community structures at two closely-related

intertidal stations: (1) a well-developed Zostera noltii meadow and (2) a bare sediment

CHAPITRE 4 :



[148]

mudflat where Zostera noltii meadow has recently (ca. 3-5 years before the beginning of the

present study) disappeared.

II. Material and methods

2.1. Study area

The present study was conducted at a site named “Germanan”, which is located in the center

of the Arcachon Bay (44°42'726 N, 1°07’940 W, Figure 4.1).

Figure 4.1: localization of the study site along the French Atlantic coast (A), and within

Arcachon Bay (B).

CHAPITRE 4 :



[149]

This site consisted in an intertidal flat. It was characterized by the presence of two

distinct habitats at the same hypsometric level (ca. 2 m deep at high tide), which were only

isolated by a small intertidal channel. The first habitat was located close to the navigation

channel and corresponded to an unvegetated mud flat (Bare sediment), whereas the second

one was colonized by a well-established Zostera noltii meadow (Zostera meadow).

2.2. Field sampling and experiments

2.2.1. General strategy

Experiments and samplings were both performed seasonally between October 2010

and October 2011 (5 campaigns during October 2010, February, April, July and October

2011) at both the Bare sediment and the Zostera meadow stations.

2.2.2. Sediment particle mixing experiments

Sediment particle mixing was assessed through in situ incubations of sediment cores

using luminophores as sediment particle tracers (Mahaut and Graf, 1987) together with

mathemathical model to fit luminophores vertical depth profiles (Maire et al., 2008).

One day before the beginning of each experiment (i.e., each combination of Stations x

Seasons), 6 cores (h = 30 cm, internal diameter = 9.6 cm) per station were carefully pushed

within field sediments where they were let for one full tidal cycle. At the beginning of the

experiment per se, 6.9 g of dried luminophores (green eco-trace®, environmental tracing

systems, D50 = 35 µm, density = 2.5 g.cm-2

) were suspended in seawater and gently spread at

the sediment surface of each core using a Pasteur pipette. For each combination of Stations x

Seasons, 3 cores were sampled after 7 and 14 days, respectively. Back to the laboratory, these

cores were frozen (-20°C) and sliced (0.5 cm thick sections down to 5 cm depth and 1 cm

thick sections down to 10 cm depth). After being freeze-dried, each slice was weighed and

homogenized. 2 g of sediment were carefully dispersed over a Petri dish and photographed

under UV light using a Nikon® D100 digital camera fitted with a yellow filter.

CHAPITRE 4 :



[150]

2.2.2.1. Image analysis and vertical luminophore profiles computation

Luminophore pixels were counted after a binarization step (based on RGB level) for

each image corresponding to one single slice using a custom home-made image analysis

software (Maire et al 2006). The relative concentrations of luminophores in each slice were

then used to compute corresponding vertical depth profiles.

2.2.2.2. Modelling of sediment particle mixing intensity

Sediment particle mixing intensity was assessed by fitting a Continuous Time Random

Walk model (CTRW, Meysman et al. 2008a) to luminophore vertical depth profiles. This

models described particle displacements as a sequence of random bioturbation events. This

model was chosen because it proved more efficient than the classical biodiffusion model in

describing sediment particle mixing process over short-time scale (Maire et al. 2008,

Meysman et al., 2010). CTRW model assumes that sediment particle mixing is controlled by

two probability distributions (Meysman et al., 2008a). The jump-length distribution defines

the elemental distance a particle travels within a given mixing event. The waiting-time

distribution describes the elemental time a particle waits in between two consecutive mixing

events. The jump-length and waiting-time distributions are most often assumed to follow a

Poisson process and a Gaussian distribution, respectively (Meysman et al., 2008b, 2010). For

each replicate of each combination of Seasons x Stations x Experiment durations, a single

normal biodiffusion coefficient ( in cm².yr

-1) value reflecting sediment particle mixing

intensity was obtained from fitted parameters according to Meysman et al. (2008b, 2010):

(1)

σ2 is the variance of the jump-lengths distribution and τc is average of the waiting time

distribution.

This operation was carried out using TURBO package (“functions for fitting

bioturbation models to tracer data”) within the open source R programming framework

(v2.13.1., http://www.R-project.org , 2011).

CHAPITRE 4 :



[151]

2.2.2.3. Data processing

For each Seasons x Stations combination, no significant differences in DbN were

detected between the two Experiment durations (univariate PERMANOVA, p<0.05).

Consequently, the 6 DbN values measured for each combination were considered as replicates

for further statistical analyses.

2.2.3. Water and sediment characteristics

Water temperature, sediment characteristics (granulometry, organic carbon and

nitrogen content), Zostera population characteristics (Shoot density, leaves and root

biomasses), macrofauna community structure (abundance and biomass at the species level),

and DbN were assessed for each Seasons x Stations combination. Four cores (Ø = 6 cm) were

sampled for further analyses of the top first centimeter of sediment. One core was used for

assessing sediment median grain size (D50) using a laser microgranulometer (MALVERN®

Master Sizer S). For carbon (POC) and nitrogen (PON) contents determination, the first top

centimeter of the sediment column were sliced, freeze-dried, homogenized and later

separately analyzed for organic carbon and nitrogen. Samples for carbon analysis were

decarbonated (HCl 0.3N). Both POC and PON were assessed using a CN auto-analyzer

(Thermo Flash® EA112).

2.2.4. Zostera noltii population characteristics

For each Seasons x Stations combination, 6 cores (internal diameter = 9.6 cm) were

sampled to assess Zostera noltii population characteristics. Sediment was sieved on 1 mm

square mesh to retain leaves and roots. Shoots were first counted and leaves and roots were

then carefully isolated and collected before being desiccated (60°C during 48h), weighed

(precision: 0.0001 g) and ashed (4h30 at 450 °C). Resulting ashes were weighed (precision:

CHAPITRE 4 :



[152]

0.0001 g) allowing for the computation of both leaf and roots biomass in g of Ash Free Dry

Weight (AFDW).

2.2.5. Infauna

For each Seasons x Stations combination, 5 replicates of sediment were sampled using

a 0.04 m² square corer (Castel et al., 1989) and gently sieved on a 1 mm square mesh.

Macrofauna was then fixed (4% buffered formaldehyde) and colored with Rose Bengale.

Back to the laboratory, each organism was identified at the species level, counted and its

biomass assessed at the species level as described above for Zostera. Infauna species (i.e.

macrofauna organisms living within the sediment column, Bouma et al., 2009) are mostly

responsible of vertical sediment particle mixing (Kristensen et al., 2012). They were thus

separated from other macrofauna according to Garcia (2010), and their species richness,

abundance (nb.indiv.m-2

) and biomass (g.AFDW.m-2) assessed.

2.3. Statistical analysis

2.3.1. Univariate analyses

Differences between Stations and Seasons in POC, PON, total macrofauna

abundances, biomasses and species richness, infauna abundance, biomass and species

richness and were assessed using univariate permutational ANOVAs (PERMANOVA;

Anderson 2001, McArdle & Anderson 2001) without preliminary data transformation.

Euclidean distance was used and the design consisted in 2 crossed factors, namely Seasons

(fixed, 5 levels) and Stations (fixed, 2 levels). Pairwise tests were also performed to highlight

differences among factor modalities. The effects of factors on spatial variability (i.e., among-

replicates variability) were tested using the PERMDISP procedure (Anderson 2006; same

distance and same design as described above).

CHAPITRE 4 :



[153]

2.3.2. Infauna community structure

Infauna community structure was investigated using non-metric multidimensional

scaling (n-MDS). This analysis was based on Bray-Curtis similarities calculated on un-

transformed abundance and biomass data. Differences in infauna assemblage compositions

between Seasons and Stations were tested using multivariate permutational ANOVAs

(PERMANOVA) with Bray-Curtis similarities and the same 2-ways crossed design described

in the Univariate analyses section. Corresponding pairwise tests were performed as well.

Differences in the dispersion of data between Seasons and Stations were also tested using the

PERMDISP procedure (using Bray-Curtis similarities). For both abundance and biomass data,

the species that contributed most to these differences were identified using the SIMPER

procedure (Clarke and Warwick, 2001).

2.3.3. Linking DbN and synthetic descriptors

Synthetic descriptors describing changes in abiotic and biotic parameters at the two

Stations and during all sampled Seasons were tentatively linked with both DbN and varDb

N

(variation coefficient of DbN) through a principal component analysis (PCA). This was

performed using water temperature, median sediment grain size, mean POC and PON, mean

Zostera shoot density, leaf and roots biomasses, and mean infauna species richness,

abundance and biomass as descriptive variables for ordination, and both DbN and varDb

N as

additional variables. Corresponding data were collected during February, April, July and

October 2011. All variables were normalized prior to the analysis.

2.3.4. Linking DbN and species distributions patterns

were linked to species distributions patterns using an inverse BIO-ENV procedure

(also called ENV-BIO, Clarke and Warwick 2001). The aim was to identify infauna species

potentially responsible for spatiotemporal changes (both in mean values and variability) in

. Variation coefficients were used as indicative of variability (i.e. spatial heterogeneity)

patterns of both and species abundance/biomass because they have proven more useful in

comparing variability among biological characteristics than standard deviations (Fraterrigo

and Rusak, 2008 ; Hewitt and Thrush 2009). ENV-BIO procedures were performed separately

CHAPITRE 4 :



[154]

on abundance and biomass data and carried out on three distinct datasets. The first one

corresponded to all Seasons x Stations combinations whereas the two others corresponded to

only Bare sediment and Zostera meadow stations, respectively. For each dataset, only the

species that represent at least 3% of the total abundance and 3% of the total biomass within at

least one replicate were selected. ENV-BIO analyses were run separately to identify (1)

species whose average abundance/biomass patterns correlated best (BEST procedure, Clarke

and Warwick 2001) with patterns, and (2) species whose variation coefficients of their

abundance/biomass patterns correlated best with var patterns. Correlations were assessed

using Spearman coefficients and corresponding significances were tested with permutation

tests involving 999 random permutations.

All the above-described statistical analyses were performed using the PRIMER® v6

package with the PERMANOVA+ add-on software (Clarke & Warwick 2001, Anderson et al.

2008).

III. Results

3.1. DbN

Means and standard deviations of DbN measured during the present study in both Bare

sediment and Zostera meadow are shown in Figure 4.2. Mean DbN was maximal during

October 2010 (22.45 ± 43.73 cm².yr-1

) and minimal during October 2011 (2.99 ± 2.75 cm².yr-

1) in Bare sediment, whereas it was maximal during October 2011 (18.07 ± 18.14 cm².yr

-1)

and mimimal during February 2011 (0.39 ± 0.30 cm².yr-1

) in Zostera meadow. Overall, DbN

were characterized by a high spatial variability as indicated by high standard deviations. Due

to this variability, we detected no global effect of Seasons and Stations factors or any

significant interaction between these two factors (Table 4.1A), although DbN apparently

tended to be higher and more spatially variable in Bare sediment, except during October 2011

(Figure 4.2).

Within Zostera meadow, DbN recorded in February 2011 were significantly lower than

those recorded in October 2010, April 2011 and July 2011 (Table 4.1B). Difference was

CHAPITRE 4 :



[155]

almost significant between DbN

recorded in February and October 2011. Corresponding DbN

were also significantly more variable during October 2011 than during the 4 other sampled

seasons. Within Bare sediment, DbN were significantly less spatially variable during April and

October 2011 than during the other 3 sampled seasons. DbN were significantly higher within

Bare sediment than within Zostera meadow during February 2011.

Sampling season

Oct.-10 Feb.-11 Apr.-11 Jul.-11 Oct.-11

DbN

(cm

².yr-1

)

0

10

20

30

40

5060

Bare sediment

Zostera meadow

Figure 4.2: Mean (±sd) particle mixing intensities DbNL

recorded within both bare sediment

(open bars) and Zostera meadow (black bars) during the 5 sampling seasons.

They were also significantly more spatially variable within Bare sediment than within

Zostera meadow during October 2010, February and July 2011 (Table 4.1B). Conversely,

DbN were significantly more spatially variable within Zostera meadow than within Bare

sediment during October 2011.

CHAPITRE 4 :



[156]

Tab

le 4

.1:

PE

RM

AN

OV

As.

A

: ef

fect

s of

the

mai

n f

acto

rs (

df:

deg

ree

of

free

dom

) on D

bN

L v

alues

. B

: p

-val

ues

obta

ined

by p

airw

ise

com

par

iso

ns

bet

wee

n a

ll p

oss

ible

Sea

sons

x Sta

tions

com

bin

atio

ns.

V

alues

in b

old

indic

ate

com

bin

atio

ns

that

sig

nif

ican

tly d

iffe

rs (

p<

0.0

5)

and

D i

ndic

ates

com

bin

atio

ns

that

pre

sen

t si

gnif

ican

tly d

iffe

rent

dis

per

sions

(PE

RM

DIS

P a

nal

ysi

s, p

<0.0

5).

A

D

bN

d

f 4

Sea

son

s P

seud

o-F

0

.65

46

P

(p

erm

) 0

.68

52

d

f 1

Sta

tio

ns

Pse

ud

o-F

1

.00

77

P

(p

erm

) 0

.36

24

d

f 4

Se

x S

t P

seud

o-F

1

.60

29

P

(p

erm

) 0

.16

7

Oct

ob

er

20

10

Feb

ruar

y 2

01

1

Ap

ril 2

01

1

July

20

11

Oct

ob

er

20

11

B

B

are

sed

imen

t

Zost

era

mea

do

w

Ba

re

sed

imen

t

Zost

era

mea

do

w

Ba

re

sed

imen

t

Zost

era

mea

do

w

Ba

re

sed

imen

t

Zost

era

mea

do

w

Ba

re

sed

imen

t

Zost

era

mea

do

w

Oct

ob

er

20

10

B

are

sed

imen

t

Zost

era

mea

do

w

0.8

10

3D

Feb

ruar

y 2

01

1

Ba

re s

edim

ent

0.8

88

9

0.3

56

6D

Zost

era

mea

do

w

0.0

02

5D

0.0

02

5D

0.0

46

2D

Ap

ril 2

01

1

Ba

re s

edim

ent

0.3

28

4D

0.2

44

0

0.1

55

2D

0.0

63

0D

Zost

era

mea

do

w

0.3

86

4D

0.2

00

2

0.1

79

0D

0.0

11

5D

0.9

51

0

July

20

11

B

are

sed

imen

t 0

.45

69

1

0

.48

04

0

.70

09

D

0.9

77

8

0.9

95

0D

Zost

era

mea

do

w

0.2

54

2D

0.1

04

5

0.1

79

6D

0.0

03

6D

0.8

76

4

0.6

46

5

1D

Oct

ob

er

20

11

B

are

sed

imen

t 0

.23

83

D

0.0

81

2

0.1

62

5D

0.0

35

3D

0.8

13

7

0.5

94

2D

0.9

86

0D

0.9

55

2

Zost

era

mea

do

w

0.9

03

3

0.1

78

0D

0.9

19

5

0.0

52

9D

0.1

37

2D

0.1

40

3D

0.4

63

7

0.1

38

2D

0.1

32

7D

CHAPITRE 4 :



[157]

3.2. Water and sediment characteristics

Water temperature and main sediment characteristics (D50, POC and PON) recorded

during the present study are listed in Table 4.2.

Water temperature was 14.5, 6, 16.5 and 21.5 °C during October (2010 and 2011),

February, April and July 2011, respectively. Within Bare sediment, D50 were between 26.2

µm during April 2011 and 37.8 µm during February 2011, versus 25.0 µm during October

2011 and 61.4 µm during February 2011 within Zostera meadow. D50 tended to be higher

within Zostera meadow than within Bare sediment from October 2010 to April 2011.

Conversely, they seemed lower within Zostera meadow than within Bare sediment during

October 2011 (Table 4.2).

The main effects of Seasons and Stations factors on surface sediment POC and PON

are shown in Table 4.3. Both POC and PON were minimal during April 2011 within Bare

sediment (1.64 ± 0.31 %POC and 0.17 ± 0.03 %PON) as well as within Zostera meadow (1.03

± 0.04 %POC and 0.14 ± 0.02 %PON). Maximal values of both POC and PON contents were

recorded during October 2011 within Bare sediment (2.68 ± 0.35 %POC and 0.29 ± 0.04

%PON), and during July 2011 within Zostera meadow (2.86 ± 0.47 %POC and 0.31 ± 0.06

%PON). Both POC and PON varied significantly between Seasons but not between Stations.

They were also significantly affected by the interactions between these two factors. POC and

PON were significantly higher within Bare sediment than within Zostera meadow only during

October 2011 (Table 4.3).

CHAPITRE 4 :



[158]

Tab

le 4

.2:

Wat

er t

emp

erat

ure

, m

edia

n s

edim

ent

gra

in s

ize

(D5

0)

and m

eans

and s

tand

ard d

evia

tions

of

sedim

ent

org

anic

conte

nts

(%

PO

C a

nd

%P

ON

), Z

ost

era

popula

tion c

har

acte

rist

ics

(shoot

den

sity

, le

af a

nd r

oot

bio

mas

ses)

and I

nfa

una

char

acte

rist

ics

(spec

ies

rich

nes

s, a

bundan

ce

and b

iom

ass)

mea

sure

d w

ithin

Zost

era m

eadow

and b

are

sedim

ent

stat

ions

and d

uri

ng t

he

5 s

ampli

ng s

easo

ns.

Val

ues

in b

old

in

dic

ate

signif

ican

t dif

fere

nce

s am

ong s

tati

ons

for

the

giv

en s

ampli

ng s

easo

n (

univ

aria

te P

ER

MA

NO

VA

pai

rwis

e co

mpar

ison,

p<

0.0

5).

Wit

hin

a

giv

en s

tati

on,

val

ues

lin

ked

by t

he

sam

e le

tter

do n

ot

signif

ican

tly d

iffe

r am

ong t

he

consi

der

ed s

easo

ns

(univ

aria

te P

ER

MA

NO

VA

pai

rwis

e

com

par

ison,

p<

0.0

5).

O

cto

be

r 2

01

0

Feb

ruar

y 2

01

1

Ap

ril 2

01

1

July

20

11

Oct

ob

er

20

11

Ba

re

sed

imen

t Zo

ster

a

mea

do

w

Ba

re

sed

imen

t Zo

ster

a

mea

do

w

Ba

re

sed

imen

t Zo

ster

a

mea

do

w

Ba

re

sed

imen

t Zo

ster

a

mea

do

w

Ba

re

sed

imen

t Zo

ster

a

mea

do

w

Wat

er

tem

per

atu

re (

°C)

14

.5

14

.5

6

6

16

.5

16

.5

21

.5

21

.5

14

.5

14

.5

D50

(µ

m)

27

.2

43

.9

37

.8

61

.4

26

.2

46

.5

32

.5

32

.2

29

.8

25

.0

%

PO

C

(m

ean

±

sd)

- -

2.2

3 ±

0.6

1ac

2.3

2 ±

0.4

7a

1.6

4 ±

0.3

1a

1.0

3 ±

0.0

4b

2.6

3 ±

0.1

9b

2.8

6 ±

0.4

7ac

2.6

8 ±

0.3

5b

c

2.0

8 ±

0.0

7c

%

PO

N

(mea

n

±

sd)

- -

0.1

9 ±

0.0

2a

0.1

7 ±

0.0

2a

0.1

7 ±

0.0

3a

0.1

4 ±

0.0

2b

0.2

5 ±

0.0

1b

0.3

1 ±

0.0

6ab

c

0.2

9 ±

0.0

4b

0.2

3 ±

0.0

2c

Sho

ot

den

sity

(sho

ot

.m-2

) 0

1

0 3

39

±

98

7a

0

12

56

0 ±

26

44

a 0

1

4 3

62

±

39

02

a 0

1

0 0

29

±

59

29

ab

0

6 0

52

±

13

17

b

Lea

f b

iom

ass

(g.A

FD

W.m

-2)

0

36

.18

±

7.8

4a

0

9.8

4 ±

1.8

9b

0

22

.14

±

6.2

9c

0

40

.27

±

4.9

2a

0

38

.10

±

7.2

8a

Ro

ot

bio

mas

s

(g.A

FD

W.m

-2)

0

12

3.1

0 ±

10

.09

a 0

9

5.6

9 ±

18

.93

b

0

72

.25

±

18

.57

bd

0

47

.58

±

8.0

4 c

0

54

.55

±

18

.38

cd

Infa

una

Sp

ecie

s

rich

ness

(n

b

of

spec

ies)

1

2 ±

3ac

1

2 ±

1a

12

± 2

ac

9 ±

3ab

1

0 ±

1a

9 ±

3ab

5

± 2

b

8 ±

2b

14

± 2

c 1

1 ±

2ab

Infa

una

abund

ance

(nb

. In

div

.m-2

) 5

40

5 ±

18

07

a

7 9

60

±

14

55

a

2 4

90

±

43

8b

5 4

40

±

14

63

bc

2 8

95

±

12

98

ab

8 1

75

±

28

03

ab

49

4 ±

35

0c

6 9

62

±

75

5ab

2 5

95

±

68

1b

3 2

85

±

18

14

c

Infa

una

bio

mas

s

(g.A

FD

W.m

-2)

4.1

1 ±

3.5

0ab

16

.82

±

11

.62

6

.12

± 3

.14

a 1

5.5

2 ±

1

5.5

7

8.5

4 ±

6.5

2a

15

.47

±

4.0

7

1.9

8 ±

1.2

4b

11

.83

±

5.8

1

5.4

7 ±

2.5

7a

14

.27

±

9.3

1

CHAPITRE 4 :



[159]

3.3. Zostera population characteristics

Means and standard deviations of Zostera shoot densities, leaf biomasses and root

biomasses recorded during the present study are shown in Table 4.2. All three parameters

varied significantly with Seasons (Table 4.3). There was a significant decrease in both shoot

density (10 339 ± 987 shoots.m-2

in October 2010 versus 6052 ± 1317 shoots.m-2

in October

2011, Table 4.2) and root biomass (123.10 ± 10.09 gAFDW.m-2

in October 2010 versus

54.55 ± 18.38 g.AFDW.m-2

in October 2011, Table 4.2). Conversely, there was a

significantly higher leaf biomass during October 2010, October 2011 and July 2011 (36.18 ±

7.84, 38.10 ± 7.28 and 40.27 ± 4.92 gAFDW.m-2

, respectively, Table 4.2) than during April

(22.14 ± 6.29 gAFDW.m-2

) and February (9.84 ± 1.89 gAFDW.m

-2).

Table 4.3: PERMANOVA analysis: main effects of Seasons, Stations and Seasons x Stations

(Se x St) factors on POC, PON, Zostera shoot density, leaf and root biomasses, and infauna

species richness, abundance and biomass. Values in bold indicate significant (p<0.05) effects.

POC PON

Shoot

density

Leaf

biomass

Root

biomass

Infauna

Species

Richness

Infauna

Abundance Infauna

Biomass

df 3 3 4 4 4 4 4 4

Seasons Pseudo-F 12.552 11.119 4.879 27.542 17.231 9.7705 9.9054 0.5129 P (perm) 0.0005 0.0005 0.0051 0.0001 0.0001 0.0001 0.0001 0.7476

df 1 1 - - - 1 1 1 Stations Pseudo-F 2.6675 1.2405 - - - 1.9927 67.665 17.786 P (perm) 0.1221 0.2835 - - - 0.1706 0.0001 0.0002

df 3 3 - - - 4 4 4 Se x St Pseudo-F 5.1367 3.5223 - - - 2.3002 5.3252 0.1797 P (perm) 0.0138 0.0410 - - - 0.0744 0.0020 0.9528

3.4. Benthic infauna characteristics

3.4.1. Univariate parameters

Means and standard deviations of species richness, abundance and biomass of benthic

infauna recorded during the present study are shown in Table 4.2. The main effects of the

Seasons, Stations and Seasons x Stations factors on these 3 parameters are shown in Table

CHAPITRE 4 :



[160]

4.3. Species richness significantly varied with Seasons but not with Stations and not with the

interaction between these two factors. Species richness was minimal during July 2011 within

both Bare sediment and Zostera meadow (5 ± 2 and 8 ± 2 species, respectively). Species

richness was maximal during October 2011 within Bare sediment (14 ± 2 species) and during

October 2010 within Zostera meadow (12 ± 1 species).

Infauna abundance varied significantly with Seasons, Stations and Seasons x Stations

(Table 4.3). Within Bare sediment, infauna abundance was significantly lower during July

2011 (494 ± 350 indiv.m-2

) than during the four other sampling seasons (Table 4.2).

Conversely, it was maximal during October 2010 (5 405 ± 1807 indiv.m-2

). This last value

was significantly higher than the one recorded during October 2011 (2595 ± 681 indiv.m-2

)

(Table 4.2). A similar trend toward a significantly lower infauna abundance during October

2011 than during October 2010 was found within Zostera meadow (3 285 ± 1814 indiv.m-2

in

October 2011 versus 7 960 ± 1455 indiv.m-2

in October 2010, Table 4.2), where maximal

abundance was recorded during April 2011 (8175 ± 2803 indiv.m-2

). Infauna abundances

were always significantly higher in Zostera meadow than in Bare sediment, except during

October 2011, where it was only slightly non-significantly higher (3285 ± 1814 indiv.m-2

within Zostera meadow versus 2595 ± 681 indiv.m-2

in Bare sediment, Table 4.2).

Infauna biomass significantly varied with Stations but not with Seasons and not with

the interaction between these two factors (Table 4.3). Lowest biomasses were recorded

during July 2011 within both Bare sediment (1.98 ± 1.24 g.AFDW.m-2

) and Zostera meadow

(11.83 ± 5.81 g.AFDW.m-2

). Highest biomasses were recorded during April 2011 (8.54 ± 6.52

g.AFDW.m-2

) and October 2010 (16.82 ± 11.62 g.AFDW.m-2

) within Bare sediment and

Zostera meadow, respectively. Infauna biomasses were significantly higher within Zostera

meadow than within Bare sediment during October 2010, July and October 2011. Conversely,

there was no significant difference in infauna biomass between stations during both February

and April 2011.

3.4.2. Community structure (multivariate)

The results of the nMDS suggested that the compositions of benthic infauna differed

between Bare sediment and Zostera meadow. (Figure 4.3).

CHAPITRE 4 :



[161]

Figure 4.3: nMDS ordination of infauna assemblage compositions. Data are based on non-

transformed abundances.

This was confirmed by PERMANOVA and PERMISP results, which showed that both

the mean composition and the variability of benthic infauna composition were significantly

affected by Stations (Table 4.4), Benthic infauna compositions within Bare sediment were

more variable than within Zostera meadow. Both Seasons and Seasons x Stations significantly

affected the variability of benthic infauna composition (Table 4.4).

Table 4.4: PERMANOVA analysis: main effects of factors on the infauna abundance and

biomass assemblages. Values in bold indicates whether effects are significant (p < 0.05). D

indicates if factor are significantly affects the multivariate dispersion of data.

Infauna

Abundance Infauna

Biomass

df 4 4

Seasons Pseudo-F 8.6796 2.4229 P (perm) 0.0001

D 0.0002

df 1 1 Stations Pseudo-F 38.721 15.681 P (perm) 0.0001

D 0.0001

df 4 4 Se x St Pseudo-F 38.721 2.4175 P (perm) 0.0001

D 0.0001

D

CHAPITRE 4 :



[162]

Overall, there were clearly stronger seasonal changes in benthic infauna composition

within Bare sediment than within Zostera meadow (Figure 4.3, Table 4.6A). An exception to

this general pattern was October 2011, with: (1) different benthic infauna composition within

Bare sediment and Zostera meadow, and also (2) a more variable benthic infauna composition

within Zostera meadow than in October 2010 (Figure 4.3, Table 4.6A). Infauna biomasses

showed the same general pattern, but with less marked differences in mean compositions and

especially in variability (Table 4.4, Table 4.6B).

Table 4.5: Within-station average similarity (Bray-Curtis) percentages of infauna

assemblages obtained for each sampling period by SIMPER analysis based on non-

transformed infauna species abundances and biomasses.

SIMPER analysis carried out on abundance data showed that Zostera meadow assemblages

had an overall similarity of 61.89 % mostly due to the polychaetes Melinna palmata and

heteromastus filiformis, whereas Bare sediment assemblages had an overall similarity of

43.98 % mostly due to the polychaetes Melinna palmata, Heteromastus filiformis,

Aphelochaeta marioni and Nephtys hombergii, to the bivalve Abra segmentum and to the

oligochaete Tubificoides benedii. Within-stations similarities changed with Seasons from

51.1% in October 2011 to 87.7% in July 2011 within Zostera meadow, and from 50.7% in

July 2011 to 70.8% in February 2011 within Bare sediment (Table 4.5). They were always

higher within Zostera meadow than within Bare sediment, except in October 2011 (Table

4.5). Within-stations similarities were lower when computed based on biomass (Table 4.5).

They were higher within Bare sediment than within Zostera meadow in October 2010 and

2011. SIMPER analysis also showed that both between-stations and seasonal differences in

infauna compositions were mostly driven by differences in the abundances of the five above-

mentioned species (Table 4.7). The two dominant polychaetes Heteromastus filiformis and

Melinna palmata contributed to all between-stations dissimilarities because of their higher

abundances within Zostera meadow than within Bare sediment.

October 2010

February 2011

April 2011

July 2011

October 2011

Bare

sediment Zostera meadow

Bare sediment

Zostera meadow

Bare sediment

Zostera meadow

Bare sediment

Zostera meadow

Bare sediment

Zostera meadow

Abundance 60.7 79.6 70.8 75.2 63.5 73.1 50.7 87.7 65.1 51.1

Biomass 51.3 37.7 33.6 45.5 56.3 74.3 22.1 67.9 45.2 35.4

CHAPITRE 4 :



[163]

Tab

le 4

.6:

Pai

rwis

e co

mpar

isons:

Aver

age

dis

sim

ilar

ity p

erce

nta

ges

am

ong i

nfa

una

abundan

ce (

A)

and b

iom

ass

(B)

asse

mbla

ges

giv

en b

y S

IMP

ER

anal

yse

s. V

alues

in b

old

indic

ate

asse

mbla

ges

that

sig

nif

ican

tly d

iffe

r (P

ER

MA

NO

VA

pai

rwis

e te

sts,

p<

0.0

5)

and D

indic

ates

ass

embla

ges

that

pre

sent

signif

ican

tly d

iffe

rent

mu

ltiv

aria

te d

isper

sions

(PE

RM

DIS

P a

nal

ysi

s, p

<0.0

5).

A

O

cto

be

r 2

01

0

Feb

ruar

y 2

01

1

Ap

ril 2

01

1

July

20

11

Oct

ob

er

20

11

Ba

re

sed

imen

t

Zost

era

mea

do

w

Ba

re

sed

imen

t

Zost

era

mea

do

w

Ba

re

sed

imen

t

Zost

era

mea

do

w

Ba

re

sed

imen

t

Zost

era

mea

do

w

Ba

re

sed

imen

t

Zost

era

mea

do

w

Oct

ob

er

20

10

B

are

sed

imen

t

Zost

era

mea

do

w

54

.4

Feb

ruar

y 2

01

1

Ba

re s

edim

ent

49

.1

60

.8

Zost

era

mea

do

w

51

.8

32

.1

53

.0

Ap

ril 2

01

1

Ba

re s

edim

ent

61

.1

68

.5

44

.2

59

.0

Zost

era

mea

do

w

62

.5

24

.1

67

.5

35

.6

70

.4

July

20

11

B

are

sed

imen

t 8

9.8

9

3.1

D

77

.0D

91

.7D

76

.8

94

.5

Zost

era

mea

do

w

61

.0D

20

.3

65

.6D

33

.6D

66

.6D

20

.9D

91

.9D

Oct

ob

er

20

11

B

are

sed

imen

t 5

3.5

6

5.3

3

7.2

D

62

.2

51

.1

72

.2D

74

.6

68

.7D

Zost

era

mea

do

w

62

.8

57

.6D

50

.9

47

.2D

58

.8

62

.9D

81

.0

62

.5D

55

.7

CHAPITRE 4 :



[164]

B

O

cto

be

r 2

01

0

Feb

ruar

y 2

01

1

Ap

ril 2

01

1

July

20

11

Oct

ob

er

20

11

Ba

re

sed

imen

t

Zost

era

mea

do

w

Ba

re

sed

imen

t

Zost

era

mea

do

w

Ba

re

sed

imen

t

Zost

era

mea

do

w

Ba

re

sed

imen

t

Zost

era

mea

do

w

Ba

re

sed

imen

t

Zost

era

mea

do

w

Oct

ob

er

20

10

B

are

sed

imen

t

Zost

era

mea

do

w

75

.9

Feb

ruar

y 2

01

1

Ba

re s

edim

ent

59

.1

71

.7

Zost

era

mea

do

w

71

.0

60

.4

67

.9

Ap

ril 2

01

1

Ba

re s

edim

ent

52

.1

75

.3

60

.9D

61

.9

Zost

era

mea

do

w

80

.4

54

.5

77

.0D

58

.0

71

.7

July

20

11

B

are

sed

imen

t 8

2.8

8

8.9

8

3.2

9

0.7

8

6.2

D

95

.3D

Zost

era

mea

do

w

76

.3

51

.8

72

.1D

49

.7

66

.5

34

.3

91

.6D

Oct

ob

er

20

11

B

are

sed

imen

t 5

7.6

7

2.9

6

0.3

6

8.1

5

2.9

7

4.8

D

80

.7D

70

.5D

Zost

era

mea

do

w

78

.7

64

.8

68

.7

62

.4

77

.0D

65

.8D

89

.4D

61

.3D

76

.1D

CHAPITRE 4 :



[165]

These two species contributed as well to seasonal differences in both Bare sediment

and Zostera meadow with lower abundances in July 2011 within Bare sediment and a strong

decrease of M. palmata in October 2011 within Zostera meadow (Table 4.7).

Biomass-based SIMPER analyses also pointed out the importance of Melinna palmata

and to a lower extent of Heteromastus filiformis and Aphelochaeta marioni in driving both

between-stations and within-stations seasonal dissimilarities in infauna composition.It amlso

highlighted revealed the importance of the bivalve Ruditapes phillipinarum, which

contributed to : (1) inter-stations dissimilarities during all sampled seasons (Table 4.8) with

higher values within Zostera meadow than within Bare sediment, and (2) all seasonal

dissimilarities within Zostera meadow. The biomasses of another bivalve Cerastoderma edule

contributed to all seasonal dissimilarities within both Bare sediment and Zostera meadow and

were also higher within Zostera meadow than within Bare sediment, except in February 2011

(Table 4.8).

CHAPITRE 4 :



[166]

Tab

le 4

.7:

Abundan

ce m

eans

and s

tandar

d d

evia

tion o

f in

fauna

spec

ies

that

rep

rese

nt

at l

east

3%

of

the

tota

l ab

undan

ce i

n a

t le

ast

one

repli

cate

, to

get

her

wit

h r

esult

s of

SIM

PE

R a

nal

ysi

s. S

pec

ies

nam

es

in b

old

ind

icat

e th

ose

that

co

ntr

ibute

d t

o 9

0%

of

dis

sim

ilar

ity b

etw

een o

ver

all

ba

re s

edim

ent

and

Zo

ster

a m

ead

ow

ass

em

bla

ges

.

Let

ters

in s

easo

ns

SIM

PE

R c

olu

mn i

nd

icat

e sp

ecie

s, w

ith

in b

are

sed

imen

t (s

mal

l le

tter

s) a

nd

wit

hin

Zo

ster

a m

ead

ow

(en

cap

sula

ted

let

ters

), t

hat

co

ntr

ibute

d t

o 9

0%

of

dis

sim

ilar

ity

bet

wee

n O

cto

ber

20

10

and

Feb

ruar

y 2

01

1 (

a/A

), O

cto

ber

20

10

and

Ap

ril

20

11

(b

/B),

Oct

ob

er 2

010

and

July

20

11

(c/

C),

Oct

ob

er 2

010

and

Oct

ob

er 2

01

1 (

d/D

), F

ebru

ary a

nd

Ap

ril

20

11

(e/E

), F

ebru

ary a

nd

July

20

11

(f/

F),

Feb

ruar

y a

nd

Oct

ob

er 2

01

1 (

g/G

), A

pri

l an

d J

uly

20

11

(h/H

), A

pri

l an

d O

cto

ber

201

1 (

i/I)

, an

d b

etw

een J

uly

and

Oct

ob

er 2

011

(j/

J).

Val

ues

in b

old

ind

icat

ed s

pec

ies

that

co

ntr

ibu

ted

to

90

% o

f d

issi

mil

arit

y b

etw

een b

are

sed

imen

t an

d Z

ost

era

mea

do

w s

tati

ons

duri

ng t

he g

iven s

easo

n.

Oct

ob

er

20

10

Feb

ruar

y 2

01

1

Ap

ril 2

01

1

July

20

11

Oct

ob

er

20

11

Sea

son

s SI

MP

ER

Ba

re

sed

imen

t Zo

ster

a

mea

do

w

Ba

re

sed

imen

t Zo

ster

a

mea

do

w

Ba

re

sed

imen

t Zo

ster

a

mea

do

w

Ba

re

sed

imen

t Zo

ster

a

mea

do

w

Ba

re

sed

imen

t Zo

ster

a

mea

do

w

Ab

ra s

egm

entu

m

b-C

-d-e

-f-F

-h-g

-G-H

-i-

j-J

30

± 4

1

15

± 1

4

65

± 1

4

40

± 3

8

77

5 ±

31

8

15

± 1

4

12

5 ±

12

2

18

1 ±

90

1

25

± 1

15

1

20

± 5

4

Am

pel

isca

bre

vico

rnis

5 ±

11

0

0

0

1

0 ±

13

0

0

0

2

0 ±

32

0

Ap

hel

och

aet

a m

ari

on

i a-

A-b

-B-c

-C-d

-D-e

-f-

g-

G-

h-i

-I-j

-J

23

10

±

18

27

4

60

± 3

73

3

10

± 1

67

1

0 ±

14

3

0 ±

67

1

0 ±

22

5

6 ±

11

3

19

± 2

4

50

5 ±

25

5

12

5 ±

88

Cer

ast

od

erm

a e

du

le

1

5 ±

14

6

0 ±

42

1

5 ±

22

4

0 ±

42

3

0 ±

21

1

0 ±

14

1

9 ±

24

1

3 ±

14

1

5 ±

22

4

0 ±

38

Cly

men

ura

cly

pea

ta

e-h

-i

15

± 2

2

40

± 5

2

5 ±

11

0

7

0 ±

84

2

0 ±

21

0

0

2

5 ±

43

4

0 ±

76

Ga

lato

wen

ia o

cula

ta

0

0

0

0

0

0

0

0

0

1

5 ±

33

Gly

cera

co

nvo

luta

a-

c-d

-f-J

1

85

± 7

6

65

± 2

9

55

± 2

7

80

± 5

7

45

± 2

7

50

± 4

0

6 ±

13

1

3 ±

14

3

5 ±

29

7

0 ±

27

Hed

iste

div

ersi

colo

r g

0

5 ±

11

3

5 ±

29

0

2

0 ±

27

0

0

6

± 1

3

0

0

Het

ero

ma

stu

s fi

lifo

rmis

a-

A-b

-B-c

-C-d

-D-e

-E-f

-F-

g-G

-h-H

-i-I

-j-J

4

05

± 1

60

8

55

± 3

68

4

25

± 1

65

1

06

5 ±

4

55

3

70

± 1

77

6

80

± 7

05

0

2

56

± 1

49

1

90

± 1

17

8

30

± 4

09

Med

iom

ast

us

fra

gili

s A

-C-E

-G-H

0

3

5 ±

22

2

0 ±

27

1

70

± 1

32

5

± 1

1

35

± 5

5

19

± 2

4

15

6 ±

55

1

5 ±

22

0

Mel

inn

a p

alm

ata

a-

A-b

-B-c

-C-d

-D-e

-E-f

-F-

g-G

-h-H

-i-I

-j-J

20

00

±

88

4

61

15

±

16

61

1

17

0 ±

4

34

3

53

5 ±

9

28

1

16

5 ±

6

24

6

74

5 ±

2

01

3

15

6 ±

52

6

05

0 ±

8

05

1

01

5 ±

4

77

1

47

5 ±

1

68

8

Nem

erti

0

0

5 ±

11

3

5 ±

42

4

5 ±

45

2

0 ±

21

0

0

0

0

Nep

hty

s h

om

ber

gii

e-g-

i-j-

J 2

5 ±

18

3

0 ±

41

6

0 ±

38

0

2

0 ±

21

5

± 1

1

75

± 3

5

0

10

0 ±

50

1

0 ±

22

No

tom

ast

us

late

rice

us

g-J

40

± 8

9

0

0

20

± 4

5

0

5 ±

11

0

3

1 ±

63

4

0 ±

45

8

5 ±

78

Pse

ud

op

oly

do

ra p

ulc

hra

30

± 3

3

0

20

± 3

2

0

0

0

0

0

25

± 2

5

0

Pyg

osp

io e

leg

an

s a-

b-B

-e-H

-f-g

-h-i

-j

0

0

12

0 ±

82

0

1

80

± 2

01

1

10

± 2

32

1

9 ±

38

0

4

0 ±

55

1

5 ±

14

Ru

dit

ap

es p

hili

pp

ina

rum

0

15

± 2

2

15

± 2

2

35

± 6

5

0

55

± 9

7

0

6 ±

13

5

± 1

1

45

± 2

7

Stre

blo

spio

sh

rub

solii

a-

b-c

-d-g

-i-j

1

75

± 2

29

0

1

0 ±

14

0

1

0 ±

22

0

0

0

1

05

± 7

6

10

± 2

2

Tub

ific

oid

es b

ened

ii

a-

A-B

-C-d

-D-e

-E-f

-F-g

-G

-h-H

-i-I

-j-J

4

5 ±

62

1

60

± 1

04

1

00

± 9

2

35

5 ±

27

6

80

± 8

2

39

5 ±

35

5

19

± 3

8

20

6 ±

13

9

25

0 ±

97

3

80

± 1

59

CHAPITRE 4 :



[167]

Tab

le 4

.8:

Bio

mas

s m

eans

and

sta

ndar

d d

evia

tions

of

infa

una

spec

ies

repre

senti

ng a

t le

ast

3%

of

the

tota

l bio

mas

s in

at

leas

t on

e re

pli

cat

e,

toget

her

wit

h r

esult

s of

SIM

PE

R.

Sp

ecie

s n

ames

in

bo

ld i

nd

icat

e th

ose

th

at c

ontr

ibu

ted

to

90

% o

f d

issi

mil

arit

y b

etw

een

over

all

ba

re s

edim

ent

and

Zo

ster

a m

ead

ow

ass

emb

lages

.

Let

ters

in

sea

son

SIM

PE

R c

olu

mn

in

dic

ate

spec

ies,

wit

hin

ba

re s

edim

ent

(sm

all

lett

ers)

an

d w

ith

in Z

ost

era

mea

do

w (

enca

psu

late

d l

ette

rs),

th

at c

on

trib

ute

d t

o 9

0%

of

dis

sim

ilar

ity b

etw

een

Oct

ob

er 2

010

and

Feb

ruar

y 2

01

1 (

a/A

), O

cto

ber

20

10

and

Ap

ril

20

11 (

b/B

), O

ctob

er 2

01

0 a

nd J

uly

201

1 (

c/C

), O

cto

ber

201

0 a

nd O

cto

ber

201

1 (

d/D

), F

ebru

ary a

nd

Ap

ril

201

1 (

e/E

), F

ebru

ary

and

Ju

ly 2

01

1 (

f/F

), F

ebru

ary a

nd

Oct

ob

er 2

011

(g/G

), A

pri

l an

d J

uly

20

11

(h

/H),

Ap

ril

and

Oct

ob

er 2

011

(i/

I),

and

bet

wee

n J

uly

an

d O

cto

ber

201

1 (

j/J)

. V

alu

es i

n b

old

in

dic

ate

spec

ies

that

con

trib

ute

d t

o 9

0%

of

dis

sim

ilar

ity b

etw

een

ba

re s

edim

ent

and

Zo

ster

a m

ead

ow

sta

tion

s du

rin

g t

he

giv

en s

easo

n.

Oct

ob

er

20

10

Feb

ruar

y 2

01

1

Ap

ril 2

01

1

July

20

11

Oct

ob

er

20

11

Se

aso

ns

SIM

PER

B

are

se

dim

ent

Zost

era

m

ead

ow

B

are

se

dim

ent

Zost

era

m

ead

ow

B

are

se

dim

ent

Zost

era

m

ead

ow

B

are

se

dim

ent

Zost

era

m

ead

ow

B

are

se

dim

ent

Zost

era

m

ead

ow

Ab

ra s

egm

entu

m

b-D

-e-G

-h-i

-J

0.0

29

±

0.0

29

0

.23

6 ±

0

.25

2

0.0

57

±

0.0

31

0

.16

7 ±

0

.19

9

0.4

38

±

0.4

77

0

.21

8 ±

0

.30

9

0.0

59

±

0.0

57

0

.24

0 ±

0

.16

2

0.0

62

±

0.0

48

0

.58

7 ±

0

.52

7

Ap

hel

och

aet

a m

ari

on

i a-

A-b

-B-c

-C-d

-D

0.3

42

±

0.2

89

1

.22

8 ±

2

.21

9

0.0

26

±

0.0

22

0

.00

3 ±

0

.00

4

0.0

03

±

0.0

07

0

.00

2 ±

0

.00

3

0.0

05

±

0.0

10

0

.00

1 ±

0

.00

1

0.0

52

±

0.0

23

0

.01

8 ±

0

.01

5

Cer

ast

od

erm

a e

du

le

a-A

-b-B

-c-C

-d-D

-e-E

-f-

F-g-

G-h

-H-i

-I-j

-J

1.7

22

±

3.3

26

2

.41

5 ±

2

.04

8

1.4

46

±

2.1

92

0.9

91

±

1.4

22

0

.02

9 ±

0

.01

9

0.3

55

±

0.6

13

0

.82

9 ±

1

.16

7

1.0

24

±

1.1

95

0

.79

8 ±

1

.16

2

1.5

07

±

1.7

34

Cly

men

ura

cly

pea

ta

0

.00

8 ±

0

.01

4

0.0

38

±

0.0

42

0

.00

4 ±

0

.00

9

0

0.1

10

±

0.1

27

0

.05

4 ±

0

.07

5

0

0

0.0

36

±

0.0

56

7

0.0

18

±

0.0

28

Dio

pa

tra

bis

caye

nsi

s d

-g-H

-i-j

0

0

.36

1 ±

0

.68

5

0.0

68

±

0.1

52

0

0

0

.12

2 ±

0

.18

2

0

0.4

46

±

0.8

93

0

.38

8 ±

0

.76

9

0

Eun

ice

ha

rass

ii

0

0

0

0.0

46

±

0.1

02

0

0

0

0

0

0

Gly

cera

co

nvo

luta

A

-b-c

-C-d

-e-g

-h-i

-j

0.1

75

±

0.1

01

0

.47

5 ±

0

.50

5

0.0

79

±

0.1

04

0

.07

3 ±

0

.08

4

0.4

05

±

0.6

72

0

.18

3 ±

0

.11

2

0.0

03

±

0.0

06

0

.02

3 ±

0

.03

3

0.5

61

±

0.7

75

0

.11

4 ±

0

.09

1

Het

ero

ma

stu

s fi

lifo

rmis

a-

B-c

-C-e

-f-F

-g

0.1

18

±

0.0

67

0

.60

6

±0.4

21

0

.72

1 ±

1

.25

8 0

.43

4 ±

0

.20

4

0.1

12

±

0.0

55

0

.19

1 ±

0

.19

9

0

0.0

72

±

0.0

44

0

.08

8 ±

0

.07

6

0.4

93

±

0.3

32

Lori

pes

lact

eus

d-i

-j

0.0

21

±

0.0

47

0

0

0

0

0

0

0

.02

8 ±

0

.05

6

0.1

15

±

0.2

57

0

.24

4 ±

0

.54

6

Med

iom

ast

us

fra

gili

s

0

0.0

22

±

0.0

16

0

.00

8 ±

0

.01

4

0.0

81

±

0.0

75

0

.00

3 ±

0

.00

7

0.0

08

±

0.0

13

0

.04

5 ±

0

.08

8

0.0

34

±

0.0

18

0

.00

9 ±

0

.01

2

0

Mel

inn

a p

alm

ata

a-

A-b

-B-c

-C-d

-D-e

-E-

f-F-

g-G

-h-H

-i-I

-j-J

1

.32

4 ±

0

.20

1

9.9

27

±

12

.41

3

1.5

64

±

0.5

17

4.7

69

±

1.0

61

2

.07

2 ±

1

.02

9

11

.98

0 ±

3

.32

7

0.1

50

±

0.0

97

8

.14

4 ±

1

.61

4

1.8

11

±

0.7

43

3

.67

9 ±

4

.16

0

Mys

ta p

icta

b

0

.10

0 ±

0

.22

1

0.1

25

±

0.2

80

0

0

0

0

0

0

0

.00

7 ±

0

.01

5

0

Nep

hty

s h

om

ber

gii

a-b

-c-d

-g-h

-i-j

0

.06

5 ±

0

.05

2

0.0

52

±

0.0

82

0

.21

5 ±

0

.14

6

0

0.1

28

±

0.2

04

0

.07

6 ±

0

.16

9

0.2

38

±

0.2

22

0

0

.88

1 ±

0

.68

7

0.0

10

±

0.0

21

No

tom

ast

us

late

rice

us

0

.03

2 ±

0

.07

2

0

0

0.0

10

±

0.0

21

0

0

.01

8 ±

0

.03

9

0

0.0

12

±

0.0

24

0

.05

1 ±

0

.06

9

0.0

87

±

0.0

74

Ru

dit

ap

es d

escu

ssa

tus

0

0

0

0

.09

8 ±

0

.21

8

0

0

0

0

0

0

Ru

dit

ap

es p

hili

pp

ina

rum

a-

A-B

-C-d

-D-e

-E-f

-F-

g-G

-H-i

-I-j

-J

0

1.1

06

±

1.5

18

1

.77

3 ±

3

.05

8 8

.59

5 ±

1

4.7

28

0

0

.90

2 ±

1

.25

0

0

1.7

35

±

3.4

70

0

.25

8 ±

0

.57

6

7.2

70

±

8.2

04

Tub

ula

riu

s p

oly

mo

rph

us

0

.03

5 ±

0

.05

0

0.0

07

±

0.0

16

0

.00

5 ±

0

.01

0

0

0

0

0

0

0.0

06

±

0.0

12

0

.01

1 ±

0

.01

2

CHAPITRE 4 :



[168]

3.5. Linking DbN and infauna synthetic descriptors

The relations linking (and var Db

N) and synthetic descriptors of seagrass, sediment

and infauna were assessed through a PCA (Figure 4.4). Figure 4.4A shows the projections of

both descriptors and objects (i.e., combinations of Stations x Seasons) on the first plane of this

PCA. Components 1 and 2 accounted for respectively 48.5 and 22.9 % of the total variance.

Figure 4.4B shows the same projections on the plane defined by components 1 and 5 (which

accounted for only 3.4% of the total variance).

Component 1 discriminated Zostera meadow from Bare sediment. It correlated best

with shoot density (Spearman r = -0.960), root biomass (Spearman r = -0.950), infauna

biomass (Spearman r=-0.940) and infauna abundance (Spearman r = -0.870). Component 2

clearly discriminated Seasons and correlated best with water temperature (Spearman r = -

0.800) and leaf biomass (Spearman r = -0.700). and varDb

N were only poorly described

by the plan defined by the two first components of the PCA. Conversely, they were rather

well described by the plane defined by components 1 and 5 of the PCA. However, in this last

case, there was no clear relationship between these two variables and any of the considered

synthetic descriptors as well.

CHAPITRE 4 :



[169]

Figure 4.4: Description of Db

NL by PCA analysis, projection of the combinations of Seasons x

Stations and environmental parameters on the planes defined by principal components 1 and 2

(A), and principal components 1 and 5 (B). DbNL : particle mixing intensity ; varDbN : particle mixing

intensity variability ; Temp : Water temperature ; D50 : sediment median grain size ; COP : sediment %POC ;

NOP : sediment %PON ; Shoot dens : shoot density ; leaf B : leaf biomass ; root B : root biomass ; infauna S :

infauna species richness ; infauna N : infauna abundance ; infauna B : infauna biomass.

CHAPITRE 4 :



[170]

3.6. Linking DbN and infauna species distributions patterns

The results of the ENV-BIO procedure are shown in Table 4.9 for correlations

between (var

) and infauna species compositions within: (1) all the data set (Table

4.9A), (2) Bare sediment stations (Table 4.9B) and Zostera meadow stations (Table 4.9C).

When considering the whole data set, the ENV-BIO procedure highlighted a

significant correlation ( = 0.728 ; p = 0.046) between the similarity matrices based on:

(1)var , and (2) the abundances of five species: namely, Abra segmentum, Glycera

convoluta, Heteromastus filiformis, Tubificoides benedii and Ruditapes phillipinarum (Table

4.9A). Another significant correlation ( = 0.964 ; p = 0.024) was detected within Zostera

meadow between the similarity matrices based on: (1) ,and (2) the abundances of three

species, namely: Aphelochaeta marioni, Mediomastus fragilis and Melinna palmata.

Overall, infauna biomass correlated less with (and var

) than infauna abundance

as indicated by the lack of any significant correlation reported in the corresponding sections

of Table 4.9A-C.

CHAPITRE 4 :



[171]

Table 4.9: Best results of the ENV-BIO analysis within: all data set (a), bare sediment (b) and

Zostera meadow (c).

A

B

DbN vs Abundance Db

N variablilty vs

Abundance variability

DbN vs Biomass Db

N variability vs

Biomass variability

ρ (Spearman) 0.491 0.83 0.491 0.5

P level 0.854 0.329 0.984 0.835

species Abra segmentum ; Mediomastus fragilis ; Nephtys hombergii

Melinna palmata ; Tubificoides benedii

Cerastoderma edule ; Glycera convoluta ; Heteromastus filiformis ; Nephtys hombergii ; Pygospio elegans

Streblospio shrubsolii ; Tubificoides benedii

C

DbN vs Abundance Db

N variablilty vs


DbN vs Biomass Db

N variability vs

Biomass variability

ρ (Spearman) 0.964 0.842 0.879 0.503

P level 0.024 0.187 0.448 0.855

species Aphelochaeta marioni ; Mediomastus fragilis ; Melinna palmata

Abra segmentum ; Cerastoderma edule

Aphelochaeta marioni ; Mediomastus fragilis

Cerastoderma edule ; Tubificoides benedii

DbN vs Abundance Db

N variablilty vs


DbN vs Biomass Db

N variability vs

Biomass variability

ρ (Spearman) 0.368 0.728 0.425 0.686

P level 0.569 0.046 0.607 0.09

species Mediomastus fragilis

Abra segmentum ; Glycera convoluta ; Heteromastus filiformis ; Tubificoides benedii ; Ruditapes phillipinarum

Abra segmentum ; Diopatra biscayensis ; Mediomastus fragilis

Heteromastus filiformis ; Mediomastus fragilis ; Ruditapes phillipinarum ; Streblospio shrubsolii ; Tubificoides benedii

CHAPITRE 4 :



[172]

IV. Discussion

4.1. Sediment particle mixing intensity (DbN)

During the present study, we measured sediment particle mixing intensity through (1)

in situ incubations of sediment cores after a pulse of luminophores at the sediment-water

interface, and (2) subsequent fitting of a CTRW model to luminophore vertical profiles. The

so-obtained DbN values were between 0.39 ± 0.30 cm².yr

-1 and 22.45 ± 43.73 cm².yr

-1 (Figure

2), which is of the same order of magnitude as available literature data regarding sediment

particle mixing intensities in coastal environments as measured through a large variety of

methods (Clifton et al., 1995 ; Herman et al. 2001 ; Josefson et al. 2002 ; Gilbert at al. 2003 ;

Widdows et al. 2004 ; Wheatcroft 2006 ; Duport et al. 2007 ; Gérino et al. 2007 ; Teal et al.

2008 ; Leorri et al. 2009). Based on a large data set (n=87) of Db derived from different

methods, Teal et al. (2008) for example reported a global average value of 23.00 ± 50.17

cm².yr-1

within the Temperate Northern Atlantic realm but also underlined the relative lack of

data regarding intertidal areas. Nonetheless, in intertidal mud-flats, Db from 1.8 to 108 cm².yr-

1 (Clifton et al., 1995), of 5 cm².yr

-1 (Herman et al. 2001), and from 6 to 52 cm².yr

-1

(Widdows et al. 2004) were measured using 7Be. Db from 8 to 43 cm².yr

-1 (Clifton et al.,

1995), of 4 cm².yr-1

(Herman et al. 2001), and from 44 to 67 cm².yr-1

(Widdows et al. 2004)

were reported using chlorophyll a, labeled algal carbon and 210

Pb as tracers, respectively.

These data are largey consistent with ours regarding Bare sediment, which were between 2.99

± 2.75 and 22.45 ± 43.73 cm².yr-1

(Figure 4.2).

To our knowledge, the present study is the first one that quantitatively assessed in situ

DbN within a seagrass meadow. In a salt marsh habited by Spartina alterniflora, Leorri et al.

(2009) nevertheless reported average Db values between 0.47 and 1.29 cm².yr-1

, which can be

compared to the low DbN values (i.e., from 0.39 ± 0.30 to 7.06 ± 4.32 cm².yr

-1) measured

during the present study within Zostera meadow between October 2010 and July 2011

(Figure 4.2).

Literature values obtained using strictly similar tracer (luminophores) and model

(CTRW) methods as ours are much scarcer. Up to now, such values are only available for

single individual species such as the bivalves Abra ovata (segmentum) (Maire et al., 2007a)

CHAPITRE 4 :



[173]

and A. alba (Braeckman et al., 2010), the polychaete Nephtys sp. (Braeckman et al., 2010)

and the amphipod Corophium volutator (De Backer et al., 2011). Corresponding average DbN

are between 1.01 and 45.37 cm².yr-1

, 0.96 and 4.47 cm².yr-1

, 0.29 and 1.22 cm².yr-1

, 0.0007

and 0.02 cm².day-1

, for A. ovata, A. alba, Nephtys sp. and Corophium volutator, respectively.

Here again, these values are fully compatible with our own data.

Overall, our DbN

measurements were also characterized by high among-replicates

variability (Figure 4.2). This is consistent with previous studies, which have shown high

variability in vertical luminophore profiles within replicated sediment cores during in situ

sediment particle mixing experiments (Gilbert et al 2003 ; Wheatcroft et al. 2006 ; Duport et

al. 2007 ; Gérino et al. 2007). The High variability within treatment in DbN is indicative of

small-scale spatial heterogeneity. It clearly complicates the assessment of both Seasons and

Stations effects, mostly explaining why we failed in detecting any significant effect of these

two factors using PERMANOVAs (Table 4.1a). An important aim of the present study was to

relate spatio-temporal changes in with changes in the characteristics (including the

composition) of benthic infauna. Classically, this can also be achieved through the correlation

of changes in : (1) mean values of DbN and (2) species abundances/biomasses (Gérino et al.

2007). During the present study, this was also complicated by high within treatment

variability. We therefore decided to also correlate the variability of DbN and species

abundances/variabilities assuming that a significant correlation would also be indicative of a

control of sediment particle mixing by benthic infauna composition. More generally, within

treatment variability reflects the degree of small-scale spatial heterogeneity in sediment

particle mixing process. Fraterrigo and Rusak (2008) highlighted the extremely sensitiveness

of variability in ecological patterns and processes to disturbance. As far as benthic

communities are concerned, variability in macrobenthos abundance, diversity and

composition for example increase with increasing disturbance (Warwick and Clarke, 1993;

Hewitt and Thrush 2009). Our own results showed that variability in both and

macrobenthos characteristics both changed with the modalities of Seasons and Stations. It was

therefore important to assess the effects of these two factors not only on the magnitude but

also on the variability of: (1) , and (2) infauna characteristics.

This approach was carried out: (1) by comparing Zostera meadow and Bare sediment,

and (2) by looking at spatio-temporal changes within Zostera meadow and Bare sediment

during the period under study. It should be underlined that these two procedures corresponded

CHAPITRE 4 :



[174]

to different time scales. The first one indeed referred to the several-year period since the loss

of the Zostera meadow, whereas the second one referred to the one year period of the study,

which was characterized by a decline of the Zostera meadow.

4.2. Overall comparison of Zostera meadow and Bare sediment

4.2.1. Sediment particle mixing (DbN)

Our results did not show any significant effect of the presence of the Zostera meadow

on DbN. This was indicated by: (1) the failure of the PERMANOVA to detect any significant

effect of Stations on DbN, and (2) the lack of any significant correlation between the

parameters indicative of the presence of the Zostera meadow and DbN as indicated by the

PCA. As stated above, this mostly resulted from high among-replicates variability. The

analysis of raw data nevertheless allowed for the detection of some trends in DbN in relation

with the presence of Zostera meadow. Except during October 2011, DbN tended to be lower

and less variable within Zostera meadow than within Bare sediment (Figure 4.2, Table 4.1b).

It therefore suggests that sediment particle mixing was less intense and more spatially

homogeneous within Zostera meadow. This is consistent with the fact that seagrasses are

generally considered as sediment stabilizers (Orth, 1977; Townsend and Fonseca 1998; Reise

2002 ; Meadows et al. 2012) through the creation of dense roots/rhizomes networks (Reise

2002). Through this effect, the presence of seagrasses should indeed result in lower sediment

particle mixing due to (1) sediment compaction which has a negative effect on particle

movements induced by the burrowing and displacements of benthic infauna (Hughes et al.

2000 ; Berkenbush and Rowden 2007) and (2) the exclusion and/or inhibition of the activity

of large bioturbators such as lugworms and thalassinid shrimps (Hughes et al. 2000 ;

Berkenbush and Rowden 2007 ; Wesenbeeck et al. 2007).

4.2.2. Infauna

The lugworm Arenicola marina is considered as large and efficient bioturbator and

known to interact with seagrasses so that its abundance is often used to indirectly assess

sediment particle mixing intensity within seagrass meadows (Philippart 1994 ; Berkenbush

and Rowden, 2007 ; Berkenbush et al., 2007 ; Siebert and Branch, 2007 ; Wesenbeeck et al.,

CHAPITRE 4 :



[175]

2007 ; Eklöf et al. 2011; Suykerbuyk et al. 2012). This species was however not found during

the present study at either of the two studied stations, which could be explained by the fact

that Arcachon bay shelters large farmed and wild oysters (Crassostrea gigas) populations,

which induce the presence of numerous shell fragments at the surface and within the sediment

column (personal observations), which are known to prevent colonization and burrowing by

lugworms (Sukerbuyk et al., 2012).

Infauna compositions clearly differed between Zostera meadow and Bare sediment

(Figure 3). Similar differences have already been observed both in Arcachon Bay (Blanchet et

al. 2004 ; Do et al 2011 ; Do et al. 2013) as well as in other seagrass meadows (Boström and

Bonsdorff, 1997 ; Boström et al., 2006 ; Bouma et al.,2009 ; Fredriksen et al. 2010). They

result from several processes linked to the presence of seagrasses, which contribute to create a

specific habitat. This notably includes: (1) the creation of dense roots/rhizomes networks,

which constitute a protection for many prey species (Summerson and Peterson, 1984), and (2)

the accumulation of organic matter through both enhanced sedimentation (Fonseca and

Fisher, 1986 ; Meadows et al. 2012 ; Wilkie et al. 2012 ; Ganthy et al. 2013) and the decay of

plant materials (Castel et al. 1989 ; Rossi and Underwood, 2002).

During the present study, species richness never differed between Zostera meadow and

Bare sediment (Table 4.2) and the abundances of species specific to either environment

contributed only very little (i.e., 0.7%) to the overall dissimilarity in infauna composition

between the two environments. The composition of infauna in Zostera meadow was

characterized by high abundances of the deposit-feeding polychaetes Melinna palmata,

Heteromastus filiformis, Aphelochaeta marioni and of the oligochaete Tubificoides benedii

(Table 4.7) as already reported (Bachelet et al. 2000 ; Blanchet et al. 2004 ; Do et al. 2011,

2013). Interestingly, these species were also present but in lower abundance in Bare sediment

(Table 4.7). Differences between Zostera meadow and Bare sediment thus mostly resulted

from: (1) higher abundances of Melinna palmata, Heteromastus filiformis, tubificoides

benedii and Abra segmentum, and (2) lower abundances of Aphelochaeta marioni within

Zostera meadow (Table 4.7). This suggests that Bare sediment infauna community

corresponds to an impoverished Z. noltii meadow sub-community. A similar pattern (i.e.

higher infauna abundance within meadow but similar species richness and composition as

within Bare sediment) was observed by Fredriksen et al. (2010) when comparing infauna

communities within a Zostera marina meadow and a bare sediment along the Norwegian

coast. This was attributed to the fact that the corresponding stations were directly adjacent so

CHAPITRE 4 :



[176]

that the bare sediment could also benefit from particle trapping and primary production by the

Z. marina meadow. This could also be the case during the present study since the two

sampled stations were located less than 50 m apart. Along the same line, it should also be

underlined that the Bare sediment station sampled during the present study was previously

occupied by a Z. noltii meadow, which disappeared ca. 3-5 years before the beginning of the

present study (Plus et al. 2010).

For all given Seasons, but October 2011, the compositions of infauna were also much

more (spatially) homogeneous within Zostera meadow than within Bare sediment as indicated

by: (1) the higher among-replicates Bray-Curtis similarity within Zostera meadow than

within Bare sediment (Table 4.5), and (2) the corresponding dispersion of replicates in the

nMDS based on infauna compositions (Figure 4.3). A similar effect was suggested by

Blanchet et al. (2004). These authors showed that benthic fauna compositions at stations with

low above- and below-ground Zostera biomass could not be considered as similar due to high

variability. Such differences in variability within Zostera meadow and Bare sediment can be

related with differences in spatial homogeneity between the two habitats. They may refer: (1)

to food availability, and (2) the occurrence of a roots/rhizomes network within the Zostera

meadow. Within Zostera meadow, decaying leaves and enhanced sedimentation provide an

abundant and homogeneous food source for benthic infauna whose abundance and

composition are subsequently not restricted by food availability, resulting in a more spatially

homogeneous species distribution. Accordingly, infauna abundance and composition within

Bare sediment are mostly conditioned by competition for a spatially heterogeneous food

resource (for example more available in: (1) spots constituted of buried decaying plant

materials, and (2) little hollows), which is contributing to a more spatially heterogeneous

pattern of species distribution (Levinton and Kelaher, 2004). This hypothesis is supported by:

(1) the more variable sediment POC contents recorded within Bare sediment than within

Zostera meadow during the present study but July 2011 (Table 4.2), and (2) the higher

variability in abundance within Bare sediment of two opportunistic species: the oligochaete T.

benedii and the polychaete M. palmata (table 4.7), which are known to be associated with

high concentrations in sedimentary organics resulting from: (1) decaying plant materials

buried below the sediment (Rossi and Underwood, 2002), and/or (2) deposited at the sediment

surface due to the proximal vicinity of Zostera meadow.

Rossi and Underwood (2002) demonstrated that increased abundance of oligochaetes

induced by the burial of seagrass wracks also occurred for simulated wracks consisting of

CHAPITRE 4 :



[177]

plastic ribbons. This suggests that, besides organic enrichment, the occurrence of a complex

structure within the sediment column can also enhance the abundances of these organisms.

Within Zostera meadow, such a structure can be constituted by the dense roots/rhizomes

network, which provides shelter from predation for small species (Summerson and Peterson,

1984). Higher spatial heterogeneity within Bare sediment may thus also result from the

patchy distribution of buried Zostera wracks compared to the more homogeneous

roots/rhizomes network within the Zostera meadow. Furthermore, a similar effect could be

attributed to polychaete tube mats (Brenchley, 1982) such as those elaborated by the

gregarious tube builder M. palmata (Olabarria et al, 2010) whose abundances recorded during

the present study but October 2011, tended to be less spatially variable in Zostera meadow

than in Bare sediment. According to Brenchley (1982), such tube mats act on their

sedimentary and biological environments in a similar manner as Zostera roots/rhizomes

networks in enhancing sediment compaction and reducing the burrowing speed of large

bioturbators and predators. Brenchley (1982) also demonstrated that these effects were of

higher amplitude when both Zostera roots/rhizomes network and biogenic tube mats were

present.

4.3. Spatio-temporal changes within Zostera meadow and Bare

sediment during the period under study

4.3.1. Zostera meadow

All sampled Seasons, but October 2011, were characterized by: (1) relatively low DbN

associated with low var , and (2) almost similar and low among-replicates variability in the

compositions of benthic infauna. Such a co-variation between variability of sediment particle

mixing and infauna composition constitutes a first rationale supporting an effect of infauna

composition on sediment particle mixing. According to this hypothesis, low variability in

infauna composition should result in low variability in , thereby facilitating the assessment

of the effect of environmental factors on sediment particle mixing. There are two lines of

evidence, which suggests that this is indeed the case.

CHAPITRE 4 :



[178]

First, DbN measured within Zostera meadow during February 2011 were significantly

lower and less variable than during all other sampled seasons (Figure 4.2, Table 4.2), which

is in good agreement with the strong negative effect of low temperature on sediment particle

mixing (Grémare et al. 2004 ; Maire et al. 2006, 2007 ; Bernard et al. in prep). Along the

same line, the low DbN recorded in April 2011 could be related with a lower organic content

of the sediment at this particular season (Table 4.2), which here again is in good agreement

with the positive effect of food availability on sediment particle mixing (Wheatcroft, 2006,

Maire et al. 2006, 2007 ; Bernard et al. in prep).

Second, there was a clear decline of root biomass within the Zostera meadow (Table

4.2) between October 2010 and October 2011. The abundances of infauna during October

2011 did not significantly differ within Zostera meadow and Bare sediment (Table 4.2).

Infauna compositions (i.e., based on abundances and biomasses) within Zostera meadow were

also significantly more heterogeneous during October 2011 than October 2010 (Figure 4.3,

Table 4.6a,b). This mainly resulted from the decrease of the polychaete Melinna palmata

from 6115 ± 1661 ind. m-2

to 1475 ± 1688 ind.m-2

between October 2010 and 2011, and to a

lower extent from: (1) the decrease of the opportunistic polychaete Aphelochaeta marioni

from 460 ± 373 ind. m-2

to 125 ± 88 ind. m-2

and, (2) the increase of the oligochaete

Tubificoides benedii from 160 ± 104 to 380 ± 159 ind.m-2

between October 2010 and 2011.

Such enhanced abundances of Tubificoides benedii following a seagrass mortality event has

already been reported and attributed to organic enrichment through the decay of buried plant

material (Rossi and underwood, 2002), which is known to constitute a food source for

oligochaetes in general and T. benedii in particular (Rasmussen, 1973). Since the variability in

the composition of infauna and DbN were higher in Bare sediment than in Zostera meadow, a

sound hypothesis consists in attributing differences (both in absolute values and variability) in

infauna composition and in DbN recorded in October 2011 to a degradation of the Zostera

meadow and thus to a convergence with Bare sediment conditions. Blanchet et al. (2004)

reported a significant effect of Zostera noltii meadow on the composition of benthic infauna

in Arcachon Bay for shoot densities higher than ca. 6000 shoots.m-2

. This threshold value

corresponded to the shoot density recorded within the Zostera meadow during October 2011

(Table 4.2), which supports the occurrence of a positive an either direct or indirect (i.e.,

through changes in infauna composition) effect of a decline of the Zostera meadow on

sediment particle mixing. A contrario, the facts that: (1) both DbN and varDb

N were low during

October 2010, April and July 2011 and, (2) infauna composition (based on abundances) did

CHAPITRE 4 :



[179]

not significantly differ between these three seasons (Table 4.2) (Figure 4.3, table 4.6)

support the importance of this threshold and therefore is in accordance with the postulated

role of Zostera meadow in buffering seasonal environmental changes (de Wit et al. 2001 ;

Bachelet et al., 2000).

4.3.2. Bare sediment

The analysis of temporal changes in Bare sediment was more complicated due to

higher among-replicates variability both in DbN and infauna composition. It was therefore

much more difficult to relate changes in DbN with those of environmental parameters than

within Zostera meadow. Low seawater temperature in February 2011 was for example not

associated with low DbN

and varDbN. The apparent heterogeneity between Db

N recorded in

October 2010 and October 2011 was also difficult to relate with changes in any recorded

environmental parameter. Conversely the low DbN (with a low associated variability, Figure

4.2, Table 4.1b) recorded in April 2011 could be related with a lower organic content of the

sediment at this particular season as it was also the case for Zostera meadow (Table 4.2).

There were also significant seasonal changes in infauna composition (Table 4.6) and

conversely no significant change in the seasonal changes in the variability of infauna

composition (Table 4.6). These temporal patterns were therefore difficult to put in correlation

with changes in DbN and associated (i.e., between seasons) variability. The analysis of Bare

sediment data thus led to different conclusions regarding the possible effect of infauna on

sediment particle mixing. On the one hand the association between high among-replicates

variability in both DbN and infauna suggests a possible interaction. On the other hand, the lack

of clear correlation between temporal changes in: (1) DbN and infauna composition, and (2)

DbN and infauna composition variability does not support this hypothesis.

4.4. Control of sediment particle mixing intensity (DbN)

by

infauna composition

During the present study, sediment particle mixing was induced by infauna

communities, which were mainly dominated in terms of abundance by: (1) small annelids

CHAPITRE 4 :



[180]

(Melinna palmata, Heteromastus filiformis, Aphelochaeta marioni, tubificoides benedii and to

a lower extent Mediomastus fragilis), and (2) the small bivalve Abra segmentum. In terms of

biomasses, the bivalves Ruditapes phillipinarum and Cerastoderma edule were also

dominant. The large free-living polychaetes Glycera convoluta and Nephtys hombergii were

scarcer (Tables 4.7, 4.8).

During the present study, we looked at a possible effect of infauna composition by

using an ENV-BIO procedure carried out through 12 modalities (3 data sets, 2 modalities

corresponding to mean values and variation coefficients, and 2 modalities corresponding to

the use of abundances and biomasses as the basis for the computation of similarity between

infauna compositions).

Two significant results were obtained when using abundances and conversely no

significant result were obtained when using biomasses as a basis for the computation of

similarity in infauna compositions (Table 4.9). This was surprising since biological processes,

including sediment particle mixing, are more often cued by organisms’ biomasses rather than

by organisms’ abundances (Rice 1986; Wheatcroft 1990). A possible explanation is that,

during the present study, sediment particle mixing was mostly cued by small organisms (see

below), which did not exhibit major spatio-temporal changes in their individual biomass.

The only significant result when using mean values was obtained for the Zostera

meadow data set (Table 4.9), which is not surprising since this data set was also characterized

by the lowest among-replicates variability in both DbN and infauna composition. According to

the ENV-BIO analysis, the combined abundances of the three polychaetes Aphelochaeta

marioni, Mediomastus fragilis and Melinna palmata correlated best with spatio-temporal

changes in DbN. Interestingly, the analysis: (1) of temporal changes in the abundances of these

organisms, and (2) of their sediment particle mixing modes are coherent with their postulated

role in the control of sediment particle mixing.

The cirratulid A. marioni was more abundant when mean sediment particle mixing

intensity was high (i.e., during October 2010 and 2011) and was scarcer during the other

sampling seasons (Table 4.7), which is coherent with the fact that this species is a downward

conveyor (Bouchet et al., 2009 ; Garcia, 2010). Conversely, the capitellid M. fragilis was

absent in October 2011 when mean was the highest (Table 4.7, Figure 4.2), which is

coherent with the fact that this species is a head-down upward conveyor (Quintana et al.,

2007 ; Garcia, 2010). The ampharetid M. palmata was always the dominant species within

CHAPITRE 4 :



[181]

Zostera meadow during the present study. Its abundance dramatically decreased in October

2011 when was the highest, which here again was consistent with its sediment mixing

mode. This tube-building polychaete is gregarious (Oyenakan 1988) and constitutes dense

populations, particularly when sediment organic matter concentrations are high (Cacabelos et

al., 2011) such as within Zostera meadows (Blanchet 2004 ; Dauvin et al., 2007 ; Table 4.7).

M. palmata constructs mucus lined tube covered with sediment particles (Fauchald and

Jumars, 1979) forming dense tube mats that impact sediment structure (Cacabelos et al.,

2011) from the surface to a few centimeter deep, leading to a sediment compaction effect,

which is superimposed to the one induced by Zostera roots/rhizomes networks (Brenchley,

1982). It is consequently considered as a sediment stabilizer, which is coherent with the

correlation between its low abundance and the high DbN measured during October 2011.

Given its often (very) high abundances, M. Palmata is a key species in contributing to low

DbNL

within Zostera meadow.

As stated above, macrobenthic species abundance, species richness and species

composition increase in variability when communities are submitted to increasing levels of

disturbance (Warwick and Clarke, 1993; Hewitt and Thrush 2009). Variability in ecological

patterns and processes is also more sensitive to disturbance than their mean values (Fraterrigo

and Rusak, 2008).

This may explain that a significant correlation between DbN and species abundances

patterns was obtained for the whole data set when using variation coefficient as an index of

variability, whereas it was not the case when using mean values (Table 4.9). The combined

coefficient of variation of Abra segmentum, Glycera convoluta, Heteromastus filiformis,

Tubificoides benedii, Ruditapes phillipinarum correlated best with spatio-temporal changes in

varDbN. In most cases (i.e., 4 out of 5) changes in the abundances of these species together

with their sediment particle mixing modes were coherent with available literature regarding

their potential role in the control of sediment particle mixing. The small deposit-feeding

biodiffusor bivalve A. segmentum reworks sediment down to a few centimeters (Maire et al.

2006, 2007 ; Garcia 2010). The high variability in its abundances recorded within Bare

sediment during October 2010 and July 2011 could therefore partly explain the high

corresponding varDbN. G. convoluta is a gallery-diffusor (François et al, 1997). It locally

increases DbN by extending its semi-permanent burrow while prospecting the sediment

(Garcia, 2010). The capitellid polychaete H. filiformis and the oligochaete T. benedii are

CHAPITRE 4 :



[182]

both upward conveyors (Quintina et al., 2007 ; Garcia 2010), which therefore contribute to

decrease DbN when present.

The status of genus Ruditapes as a sediment reworker is more controversial. Indeed,

although François et al. (1999) reported that R. decussatus induced only low sediment particle

mixing, Sgro et al. (2005) found that the burrowing activity of R. phillipinarum significantly

reduces sediment stability. During the present study, R. philipinarum was the only of the 5

species selected through the ENV-BIO procedure whose variability in abundance correlated

negatively with variability in DbN. This mostly resulted from: (1) its high abundances in

Zostera meadow where varDbN is low, and (2) its absence from Bare sediment (where varDb

N

is high) during 3 out of 5 sampled seasons. Moreover, the forced exclusion of R. philipinarum

from the ENV-BIO procedure resulted in only a slight diminution of the correlation

coefficient between the infauna variability and the varDbN similarity matrices (=0.68,

p=0.01). There are thus good rationales in stating that the selection of R. philipinarum by the

ENV-BIO procedure is largely artefactual and does not result from a strong contribution of

this particular species to the control of sediment particle mixing.

V. Conclusions

The present study is, to our knowledge, the first one that assessed sediment particle

mixing intensity induced by the whole benthic infauna community within a seagrass meadow.

The relatively low amplitude of these presently measured intensities looked nonetheless in

good agreement with (1) intensities of a similar range available in literature obtained within

another intertidal environment colonized by root-vascular phanerogam and (2) well-known

insights in the sediment-stabilization role of seagrass meadows.

Sediment particle mixing process and infauna community structure were, within the

studied Zostera noltii meadow, less intense and heterogeneous and also less subject to

seasonal variations that within adjacent bare sediment mud-flat. These trends toward a higher

spatial and temporal stability of Z. noltii meadow tend to confirm the structuring and

buffering effects of seagrass meadow on biological sedimentary processes.

CHAPITRE 4 :



[183]

Z. noltii meadow clearly declined during the period under study, which was assessed

through the observation of root biomass and shoot density losses rather than through the

investigation of leaf biomass dynamic. Such a decline was associated, once the shoot-density

fall down to a threshold previously reported in literature, with (1) increasing of both

amplitude and variability of sediment particle mixing intensity and (2) changes in infaunal

composition.

The present study, in showing a significant correlation between mean DbN and mean

abundances of a set of three species including M. palmata, therefore underlined that within

Zostera meadow, sediment mixing process is regulated by the dense population of the

polychaete Melinna palmata that plays a synergic and probably additional role than Zostera

roots/rhizomes network itself. Consequently, when Zostera meadow and associated

M.palmata population decline, sediment particle mixing intensity becomes more intense and

particularly more heterogeneous due to the diverse mixing activities of Abra segmentum,

Glycera convoluta, Heteromastus filiformis and Tubificoides benedii, as assessed through the

significant correlation between the similarity matrix based on DbN

variability and the

similarity matrix based on the abundances of these 4 species. This is coherent with (1) the

suspected role of these species in controlling sediment particle mixing and (2) the higher

sensibility to disturbance of the variability than the mean values of ecological patterns.

Acknowledgments

This work is part of the PhD thesis of Guillaume Bernard (University Bordeaux 1).

Guillaume Bernard was supported by a doctoral grant from the French “Ministère de

l’Enseignement Supérieur et de la Recherche”. This work was funded through the IZOFLUX

(ANR blanche), BIOMIN (LEFE-CYBER and EC2CO-PNEC), the “Diagnostic de la Qualité

des Milieux Littoraux” and the « OSQUAR » (Conseil Régional Aquitaine) programs.

CHAPITRE 4 :



[184]

References







Auby, I. and P-J. Labourg. 1996. Seasonal dynamics of Zostera noltii Hornem. In the bay of

Arcachon (France). J. Sea Res., 35 (4), 269-277.

Bachelet, G., X. de Montaudouin, I. Auby and P-J. Labourg. 2000. Seasonal changes in

macrophyte and macrozoobenthos assemblages in three coastal lagoons under varying

degrees of eutrophication. ICES J. Mar. Sci., 57, 1495-1506.




Berkenbusch, K. and A. A. Rowden. 2007. An examination of the spatial and temporal

generality of the influence of ecosystem engineers on the composition of associated

assemblages. Aquat. Ecol., 41, 129-147.

Berkenbusch, K., A. A. Rowden and T. E. Myers. 2007. Interactions between seagrasses and

burrowing ghost shrimps and their influence on infaunal assemblages. J. Exp. Mar. Biol.

Ecol., 341, 70-84.

Blackburn, N. J. and R. J. Orth. 2013. Seed burial in eelgrass Zostera marina: The role of

infauna. Mar. Ecol. Prog. Ser., 474, 135-145.

Blanchet, H. 2004. Structure et fonctionnement des peuplements benthiques du bassin

d’Arcachon. Phd thesis, University Bordeaux 1, 224 pp.

Blanchet, H., X. de Montaudouin, A. Lucas and P. Chardy. 2004. Heterogeneity of

macrozoobenthic assemblages within a Zostera noltii seagrass bed : diversity, abundance,

biomass and structuring factors. Est. Coast. Shelf Sci., 61, 111-123.

Boström, C. and E. Bonsdorff. 1997. Community structure and spatial variation of benthic

invertebrates associated with Zostera marina (L.) beds in the northern Baltic Sea. J. Sea.

Res., 37, 153-166.

CHAPITRE 4 :



[185]

Boström, C., E. Bonsdorff, P. Kangas and A. Norkko. 2002. Long-term changes of a

brackish-water eelgrass (Zostera marina) community indicate effects of coastal

eutrophication. Est. Coast. Shelf Sci., 55, 795-804.

Boström, C., E. L. Jackson and C. A. Simenstad. 2006. Seagrass landscapes and their effects

on associated fauna: A review. Est. Coast. Shelf Sci., 68, 383-403.

Bouchet, V. M. P., P.-G. Sauriau, J.-P. Debenay, F. Mermillod-Blondin, S. Schmidt, J.-C.

Amiard and B. Dupas. 2009. Influence of the mode of macrofauna-mediated bioturbation

on the vertical distribution of living foraminifera: First insight from axial

tomodensitometry. J. Exp. Mar. Biol. Ecol., 371, 20-33.

Bouma T. J., V. Ortells and T. Ysebaert. 2009. Comparing biodiversity effects among

ecosystem engineers of contrasting strength: macrofauna diversity in Zostera noltii and

Spartina anglica vegetations. Helgol. Mar. Res., 63, 3-18.




Brenchley, G. A. 1982. Mechanisms of spatial competition in marine soft-bottom

communities. J. Exp. Mar. Biol. Ecol., 60, 17-33.


seagrasses: A review. Est. Coast. Shelf Sci., 79, 354-366.

Cacabelos, E., J. Moreira, A. Lourido and J. S. Troncoso. Ecological features of Terebellida

fauna (Annelida, Polychaeta) from Ensenada de San Simon (NW Spain). Anim. Biodivers.

Conserv., 34(1), 141-150.



52, 199-210.

Castel, J., P-J. Labourg, V. Escaravage, I. Auby and M. E. Garcia. 1989. Influence of seagrass

beds and Oyster parks on the abundance and biomass patterns of meio- and macrobenthos

in tidal flats. Est. Coast. Shelf Sci., 28, 71-85.



Clifton, R. J., P. G. Watson, J.T. Davey and P. E. Frickers. 1995. A study of processes

affecting the uptake of contaminants by intertidal sediments, using radioactive tracers:

7Be, 137Cs and unsupported 210Pb. Est. Coast. Shelf Sci., 41, 459-474.

Dauvin, J.-C., T. Ruellet, E. Thiebaut, F. Gentil, N. Desroy, A.-L. Janson, S. Duhamel, J.

Jourde and S. Simon. 2007. The presence of Melinna palmata (Annelida: Polychaeta) and

CHAPITRE 4 :



[186]

Ensis directus (Mollusca: Bivalvia) related to sedimentary changes in the bay of Seine

(English Channel, France). Cah. Biol. Mar., 48: 391-401.

Delefosse, M. and E. Kristensen. 2012. Burial of Zostera marina seeds in sediment inhabited

by three polychaetes: Laboratory and field studies. J. Sea. Res., 71, 41-49.

De Wit, R., L. J. Stal, B. A. Lomstein, R. A. Herbert, H. van Gemerden, P. Viaroli, V. U.

Cecherelli, F. Rodrìguez-Valera, M. Bartoli, G. Giordani, R. Azzoni, B. Schaub, D. T. Welsh,

A. Donnelly, A. Cifuentes, J. Antòn, K. Finster, L. B. Nielsen, A. G. Underlien Pedersen, A.

Turi Neubauer, M. A. Colangelo and S. K. Heijs 2001. ROBUST: The ROle of BUffering

capacities in Stabilising coastal lagoon ecosystems. Cont. Shelf Res., 21, 2021–2041.

Do, V. T., X. de Montaudouin, N. Lavesque, H. Blanchet and H. Guyard. 2011. Seagrass

colonization: Knock-on effects on zoobenthic community, populations and individual

health. Est. Coast. Shelf Sci., 85, 458-469.

Do, V. T., H. Blanchet, X. de Montaudouin and N. Lavesque. 2013. Limited consequences of

seagrass decline on benthic macrofauna and associated biotic indicators. Estuar. Coast.,

published online.

Duarte, C. M. 2002. The future of seagrass meadows. Environ. Conserv., 29 (2), 192-206.

Duarte, C. M. and C. L. Chiscano. 1999. Seagrass biomass and production : a reassessment.

Aquat. Bot., 65, 159-174.

Duport, E., F. Gilbert, J. –C. Poggiale, K. Dedieu, C. Rabouille and G. Stora. 2007. Benthic

macrofauna and sediment reworking quantification in contrasted environments in the Thau

Lagoon. Est. Coast. Shelf Sci., 72, 522-533.

Eklöf, J. S., T. van der Heide, S. Donadi, E. M. van der Zee, R. O’Hara and B. K. Eriksson.

2011. Habitat-mediated facilitation and counteracting ecosystem engineering interactively

influence ecosystem responses to disturbance. Plos one, 6(8), e23229.

Fauchald, K and P. A. Jumars. 1979. The diet of worms: a study of polychaete feeding guilds.

Oceanogr. Mar. Ann. Rev., 17, 193-284.

Fonseca M. S. and J. S. Fisher. 1986. A comparison of canopy friction and sediment

movement between four species of seagrass with reference to their ecology and restoration.

Mar. Ecol. Prog. Ser., 29, 15-22.



45, 295-319.

Fraterrigo, J. M. and J. A Rusak. 2008. Disturbance-driven changes in the variability of

ecological patterns and processes. Ecol. Lett., 11, 756-770.

CHAPITRE 4 :



[187]

Fredriksen, S., A. de Backer, C. Boström and H. Christie. 2010. Infauna from Zostera marina

L. meadows in Norway. Differences in vegetated and unvegetated areas. Mar. Biol. Res.,

6, 189-200.

Ganthy, F., A. Sottolichio and R. Verney. 2013. Seasonal modification of tidal flat sediment

dynamics by seagrass meadows of Zostera noltii (Bassin d’Arcachon, France). J. Mar.

Syst., 109-110, S233-S240.

Garcia, C. 2010. Approche fonctionnelle des communautés benthiques du bassin oriental de la

Manche et du sud de la mer du Nord. Phd thesis, University Lille 1, 399 pp.

Gérino, M., M. Frignani, C. Mugnai, L. G. Bellucci, D. Prevedelli, A. Valentini, A. Castelli,

S. Delmotte and S. Sauvage. 2007. Bioturbation in the Venice Lagoon: Rates and

relationship to organisms. Acta Oecol., 32, 14-25.





Biol. Ecol., 285/286, 251–263.

Gilbert, F., S. Hulth, V. Grossi, J. –C. Poggiale, G. Desrosiers, R. Rosenberg, M. Gérino, F.

François-Carcaillet, E. Michaud and G. Stora. 2007. Sediment reworking by marine

benthic species from the Gullmar Fjord (Western Sweden): Importance of faunal

biovolume. J. Exp. Mar. Biol. Ecol., 348, 133-144.


261.




Herman, P. M. J., J. J. Middelburg and C. H. R. Heip. 2001. Benthic community and sediment

processes on an intertidal flat: results from the ECOFLAT project. Cont. Shelf Res., 21,

2055-2071.

Hewitt, J. E. and S. F. Trush. 2009. Do species’ abundances become more spatially variable

with stress? T. O. Ecol. J., 2, 37-46.

Hovel, K. A., M. S. Fonseca, D. L. Myer, W. J. Kenworthy and P. E. Whitfield. 2002. Effects

of seagrass landscape structure, structural complexity and hydrodynamic regime on

macrofaunal densities in North Carolina seagrass beds. Mar. Ecol. Prog. Ser., 243, 11-24.

Hughes, R. G., D. Loyd, L. Ball and D. Emson. 2000. The effects of the polychaete Nereis

diversicolor on the distribution and transplanting success of Zostera noltii. Helogol. Mar.

Res., 54, 129-136.

CHAPITRE 4 :



[188]

Huston, M. 1979. A general hypothesis of species diversity. Am. Nat., 113(1), 81-101. Mar.

Josefson, A. B., T. L. Forbes and R. Rosenberg. 2002. Fate of phytodetritus in marine

sediments: functional importance of macrofaunal community. Mar. Ecol. Prog. Ser., 230,

71-85.





Mar. Ecol. Prog. Ser., 446, 285-302.



Leorri, E., R. E. Martin and B. P. Horton. 2009. Field experiments on bioturbation in salt

marshes (Bombay hook national wildlife refuge, Smyrna, DE, USA): implication for sea-

level studies. J. Quarternary Sci., 24, 139-149.

Levinton, J. and B. Kelaher. 2004. Opposing organizing forces of deposit-feeding marine




Mahaut M.L. and Graf G. (1987) A luminophore tracer technique for bioturbation studies.

Oceanol. Acta, 10, 323-328



Ecol. Prog. Ser., 319, 135-153.

Maire, O., J. C. Duchêne, A. Grémare, V. S. Malyuga and F. J. R. Meysman. 2007. A




Maire, O., P. Lecroart, F. J. R. Meysman, R. Rosenberg, J. C. Duchêne and A. Grémare,


Aquat. Biol., 2, 219-238.





CHAPITRE 4 :



[189]

Meadows P. S., A. Meadows and J. M. H. Murray. 2012. Biological modifiers of marine

benthic seascapes: Their role as ecosystem engineers. Geomorphology, 157-158, 31-48.





Acta, 72, 3460-3478.



239-254.



Murphy, E. A. K. and K. M. Dorgan. 2011. Burrow extension with a proboscis: mechanics of

burrowing by the glycerid Hemipodus Simplex. J. Exp. Biol., 2014, 1017-1027.

Olabarria, C., M. Incera, J. Garrido and F. Rossi. 2010. The effect of wrack composition and

diversity on macrofaunal assemblages in intertidal sediments. J. Exp. Mar. Biol. Ecol., 396,

18-26.

Orth, R. J. 1977. The importance of sediment stability in seagrass communities. In: Coull, B.

C. (ed). Ecology of marine benthos. Univ. South Carolina Press, Columbia, 281-300.

Orth R. J., T. J. B. Carruthers, W. C. Dennison, C. M. Duarte, J. W. Fourqurean, K. L. Heck

jr., A. R. Hughes, G. A. Kendrick, W. J. Kenworthy, S. Olyarnik, F. T. Short, M. Waycott

and S. L. Williams. 2006. A global crisis for seagrass ecosystems. BioScience, 56(12),

987-996.

Oyenakan, J.A. 1988. Population dynamics and secondary production in Melinna palmata

(Polychaeta: Ampharetidae). Mar. Biol., 98, 247-251.

Philippart, C. J. M. 1994. Interactions between Arenicola marina and Zostera noltii on a tidal

flat in the Wadden Sea. Mar. Ecol. Prog. Ser., 111, 251-257.

Plus, M., S. Dalloyau, G. Trut, I. Auby, X. de Montaudouin, E. Emery, C. Noël and C. Viala.

Long-term evolution (1988-2008) of Zostera spp. Meadows in Arcachon Bay (Bay of

Biscay). Est. Coast. Shelf Sci., 87(2), 357-366.

Quintina C. O., M. Tang and E. Kristensen. 2007. Simultaneous study of particle reworking,

irrigation transport and reaction rates in sediment bioturbated by the polychaetes

Heteromastus and Marenzellaria. J. Exp. Mar. Biol. Ecol., 352, 392-406.

Rasmussen, E. 1973. Systematics and ecology of the Isefjord marine fauna (Denmark).

Ophelia, 11, 1-495.

CHAPITRE 4 :



[190]

Reise K. 2002. Sediment mediated species interactions in coastal waters. J. Sea. Res., 48, 127-

141.


Biol. Ann. Rev., 12, 263-300.





Rice, D. L., T. S. Bianchi and E. H. Roper. 1986. Experimental studies of sediment reworking

and growth of scoloplos spp. (Orbiniidae: Polychaeta). Mar. Ecol. Prog. Ser., 30, 9-19.

Rossi, F. and A. J. Underwood. 2002. Small-scale disturbance and increased nutrients as

influences on intertidal macrobenthic assemblages: experimental burial of wrack in

different intertidal environments. Mar. Ecol. Prog. Ser., 241, 29-39.



18, 1347-1363.

Schückel, U., M. Beck and I. Kröncke. 2012. Spatial variability in structural and functional

aspects of macrofauna communities and their environmental parameters in the Jade Bay

(Waden sea Lower Saxony, southern North Sea). Helgol. Mar. Res., published online.

Sgro, L., M. Mistri and J. Widdows. 2006. Impact of infaunal Manila clam, Ruditapes

philippinarum, on sediment stability. Hydrobiologia, 550, 175-182.

Siebert, T. and G. M. Branch. 2005. Interactions between Zostera capensis, Callianassa

Kraussi and Upogebia africana: deductions from field surveys in Langebaan lagoon, South

Africa. Afr. J. Mar. Sci., 27(2), 345-356.

Siebert, T. and G. M. Branch. 2007. Influences of biological interactions on community

structure within seagrass beds and sandprawn-dominated sandflats. J. Exp. Mar. Biol.

Ecol., 340, 11-24.

Summerson, H. C. and C. H. Peterson. 1984. Role of predation in organizing benthic

communities of a temperate-zone seagrass bed. Mar. Ecol. Prog. Ser., 15, 63-77.

Suykerbuyk, W., T. J. Bouma, T. van der Heide, C. Faust, L. L. Govers, W. B. J. T. Giesen,

D. J. de Jong and M. M. van Katwijk. 2012. Suppressing antagonistic bioengineering

feedbacks doubles restoration success. Ecol. Appl., 22(4), 1224-1231.

Teal, L. R., M. T. Bulling, E. R. Parker and M. Solan. 2008. Global patterns of bioturbation

intensity and mixed depth of marine soft sediments. Aquat. Biol., 2, 207-218.

Townsend, E. C. and M. S. Fonseca. 1998. Bioturbation as a potential mechanism influencing

spatial heterogeneity of North Carolina seagrass beds. Mar. Ecol. Prog. Ser., 169, 123-132.

CHAPITRE 4 :



[191]

Warwick, R.M. and K. R. Clarke. 1993. Increased variability as a symptom of stress in marine


Wesenbeeck van, B. K., J. van de Koppel, P. M. J. Herman, J. P. Bakker and T. J. Bouma.

2007. Biomechanical warfare in ecology; negative interactions between species by habitat

modification. Oikos, 116, 742-750.

Wheatcroft, R. A. 2006. Time-series measurements of macrobenthos abundance and sediment

bioturbation intensity on a flood-dominated shelf. Prog. Oceanogr., 71, 88-122.



Res., 48, 177-201.

Widdows, J., A. Blauw, C. H. R. Heip, P. M. J. Herman, C. H. Lucas, J. J. Middelburg, S.

Schmidt, M. D. Brinsley, F. Twisk and H. Verbeek. 2004. Role of physical and biological

processes in sediment dynamics of a tidal flat in Westerschele Estuary, SW Netherland.


Wilkie, L., M. T. O’Hare, I. Davidson, B. Dudley and D. M. Paterson. 2012. Particle trapping

and retention by Zostera noltii: A flume and field study. Aquat. Bot., 102, 15-22.

Wilson, W. H. 1991. Competition and predation in marine soft-sediment communities. Annu.

Rev. Ecol. Syst., 21, 221-241.

CHAPITRE 5: Synthèse générale et perspectives

[192]

Chapitre 5:

Synthèse générale et perspectives


[193]

La principale originalité du travail présenté dans ce manuscrit réside dans le fait qu’il

constitue une étude intégrée de la caractérisation et de la quantification du processus de

remaniement sédimentaire, depuis l’échelle de la simple particule sédimentaire jusqu’à celle

de la communauté benthique in toto. Pour ce faire, ce travail associe plusieurs composantes :

(1) la mise au point d’une approche expérimentale originale, (2) l’étude expérimentale ex situ

de l’effet de paramètres environnementaux, et (3) l’étude in situ de l’effet de la régression des

herbiers de zostères.

L’ensemble a conduit à la rédaction de trois articles consacrés respectivement à : (1) la

mise au point méthodologique conduite sur le modèle biologique Abra alba, (2) l’étude

expérimentale ex situ de l’effet de deux paramètres environnementaux sur les caractéristiques

du remaniement sédimentaire induit par cette espèce, et (3) la quantification in situ du

remaniement sédimentaire dans un contexte de régression des herbiers dans le bassin

d’Arcachon. Le premier est récemment paru dans Journal of Marine Research (cf. Annexe I).

Les deux autres seront rapidement soumis pour publication.

Ce dernier chapitre vise tout à la fois à rappeler les principaux résultats obtenus lors de

ce travail de thèse, et à présenter les perspectives qui en découlent. Il est structuré en quatre

parties consacrées respectivement à:

(1) la nouvelle approche expérimentale,

(2) l’apport de mes résultats à la modélisation du remaniement sédimentaire,

(3) l’étude expérimentale du contrôle du remaniement sédimentaire, et

(4) la mesure in situ du remaniement sédimentaire dans le bassin d’Arcachon

Chacune de ces parties contient tout à la fois un bref rappel des principaux résultats

obtenus ainsi que des éléments de perspectives la concernant. Afin de faciliter la lecture, ces

derniers apparaissent en italiques et sur fond grisé. Des visions d’ensemble des résultats

obtenus et de la vision prospective sont par ailleurs fournies dans une figure bilan à la fin de

ce chapitre de synthèse (Figure 5.2).


[194]

I. La nouvelle approche expérimentale : intérêts,

pertinence et perspectives

1.1. Intérêts

Les aspects théoriques des recherches concernant le remaniement sédimentaire ont

récemment progressé avec la généralisation de l’utilisation des modèles de type Continuous

Time Random Walk (CTRW) qui constituent un socle commun susceptible de décrire une

large variété de modes de remaniement (Meysman et al. 2008). L’un des points

d’achoppement de l’utilisation de ces modèles reste toutefois la sélection a priori de patrons

de distributions de fréquence des temps de repos et des distances parcourues lors de l’un de

leur déplacement. Ces deux éléments constituent les analogues « d’empreintes du

remaniement sédimentaire ». Meysman et al. (2008) a ainsi démontré que la sélection de

patrons de distribution différents pouvait affecter de manière significative les résultats issus

de ce type de modèles. Mesurer directement les caractéristiques des mouvements de particules

sédimentaires induits par un ou des organismes benthiques constitue par conséquent un défi

majeur (Maire et al. 2008 ; Meysman et al. 2010) que l’approche méthodologique originale

développée au cours de cette thèse et présentée dans le deuxième chapitre de ce manuscrit a

clairement relevé.

1.2. Pertinence

L’approche en question a été développée en utilisant le bivalve biodiffuseur (Maire et

al. 2006) Abra alba en tant que support biologique. Elle s’appuie sur des adaptations de

dispositifs expérimentaux et d’algorithmes d’analyse d’images préexistants (Duchêne et

Nozais, 1994 ; Duchêne et Queiroga, 2001 ; Maire et al. 2007c). Ces adaptations consistent

notamment en l’amélioration des résolutions spatiale et temporelle des séquences d’images

acquises, ainsi qu’en des développements logiciels majeurs conduits en collaboration avec

Jean Claude Duchêne. Elles ont permis la mesure des caractéristiques des mouvements de

particules induits par cette espèce. Les résultats obtenus sont pleinement cohérents avec l’état

des connaissances actuelles relatives à l’éthologie au sein du genre Abra. Plus précisément,


[195]

les caractéristiques des mouvements mesurés montrent une parfaite adéquation avec les deux

grands types d’actions sur le sédiment effectués par ces bivalves, à savoir : (1) un transport

relativement intense des particules au sein du réseau en forme de cône inversé formé par les

galeries siphonales, et (2) des mouvements plus erratiques et de très faible amplitude (car

consistant le plus souvent en des oscillations) induits par les enchainements d’extension et

rétraction du pied associés à des ouvertures et fermetures des valves ou bien par des tensions

temporaires sur le sédiment créées par l’élongation des siphons. Ce couplage direct entre

comportement et mouvements de particules suggère l’aptitude de cette nouvelle méthode à

décrire le remaniement sédimentaire induit par des organismes biodiffuseurs tels qu’A. alba.

L’utilisation de cette nouvelle approche a également conduit, pour la première fois, à accéder

à une vision dynamique et 2D du remaniement sédimentaire. Ce faisant, elle a permis de

mettre en évidence et de décrire l’hétérogénéité spatiale du remaniement sédimentaire à

l’échelle de la sphère d’influence de l’organisme sur le sédiment. Dans le cas précis de A.

alba, l’hétérogénéité spatiale du remaniement sédimentaire s’exprime non seulement en

termes de temps de repos et de longueurs de déplacements, mais également d’anisotropie de

ces mêmes déplacements. Les importances relatives des mouvements horizontaux et verticaux

se sont en effet révélées variables selon les portions de sédiment considérées.

Ces propriétés, liées à l’aspect dynamique de l’approche développée associée à la prise

en compte du caractère spatialement hétérogène du remaniement sédimentaire effectué par A.

alba, ont notamment permis de détecter l’effet transitoire, donc limité dans le temps, de

l’ajout de matière organique fraiche à l’interface eau-sédiment sur l’intensité du déplacement

des particules. Ceci est d’autant plus notable qu’il n’était pas possible, en raison de la

dépendance à la durée expérimentale, de détecter cet effet en utilisant le temps moyen

d’immobilité comme indice. Etant donné que l’une des caractéristiques théorique du modèle

CTRW, par opposition au modèle biodiffusif, résidait dans sa capacité à décrire les

mouvements de particules pour des durées expérimentales courtes, implémenter un proxy de

l’intensité de déplacement de particules indépendant de cette durée comme le Normalized

number of jumps peut constituer une voie intéressante pour l’amélioration de ce modèle.

1.3. Perspectives

Si les résultats issus de ce travail suggèrent clairement la capacité de l’approche

méthodologique développée à décrire le remaniement sédimentaire induit par le bivalve Abra


[196]

alba, il reste néanmoins encore plusieurs points à préciser quant à son utilisation future. Le

premier est lié à la paramétrisation des trois principales variables impliquées que sont : (1)

le diamètre du cercle de sensibilité (qui permet de déterminer l’occurrence d’un mouvement),

(2) le cercle de recherche (qui permet de caractériser ce même mouvement), et (3) de la

fréquence d’acquisition d’images. Durant la présente étude, le diamètre du cercle de

sensibilité a été fixé, de manière objective, à partir de l’analyse des oscillations de points

fixes positionnés dans le plan étudié. Cette démarche semble pleinement satisfaisante et n’a

pas lieu d’être remise en cause. Le diamètre du cercle de recherche a quant à lui été fixé par

la comparaison des mouvements caractérisés avec l’analyse visuelle des séries d’analyses

d’images utilisées. Cette approche a donné satisfaction mais elle n’en reste pas moins

qualitative. Il conviendrait dorénavant de la compléter par la réalisation d’une analyse de

sensibilité qui consisterait en l’analyse des mêmes séries d’images pour des cercles de

recherche différents. Le diamètre approprié du cercle de recherche devant être

nécessairement négativement corrélée avec la fréquence d’acquisition des images, il

conviendrait en fait que cette analyse prenne en compte l’interaction (i.e., les combinaisons)

entre ces deux paramètres.

Le second point a trait à la durée de l’expérimentation. Il concerne plus

spécifiquement la mesure des temps de repos. Les résultats montrent clairement que les

valeurs moyennes des temps de repos tendent à augmenter avec la durée expérimentale. Cette

dépendance se répercute directement sur la mesure des coefficients de diffusion biologique.

Elle résulte en partie du fait que les temps de repos longs sont rares et qu’ils sont par

conséquent mieux échantillonnés lorsque la durée expérimentale augmente. Une part de ce

résultat reste néanmoins purement artefactuelle puisque, par définition, il est impossible

d’enregistrer un temps de repos dont la durée est supérieure à celle de la durée

expérimentale. En l’état, toute variation temporelle du temps de repos peut donc refléter une

modification réelle de l’intensité du remaniement sédimentaire et/ou un artefact résultant de

la modification de la durée expérimentale prise en compte. Une première manière d’aborder

ce problème consisterait à réaliser des expériences longues (i.e. d’une durée nettement

supérieure à celles réalisées lors de la présente étude) et à analyser expérimentalement la

relation liant durée expérimentale et temps de repos. Le but consisterait à déterminer une

durée expérimentale permettant d’atteindre une asymptote de la valeur moyenne des temps de

repos. Un tel travail devrait être effectué pour chaque modèle biologique et il n’est pas

évident que la durée expérimentale minimale ainsi déterminée s’avérerait automatiquement


[197]

compatible avec la stabilité de certaines conditions expérimentales (e.g. la quantité et la

qualité de matière organique disponible). Une seconde approche, de nature différente,

consiste à identifier un autre paramètre, indépendant de la durée expérimentale, pour décrire

la fréquence de mouvements des particules sédimentaires. Les résultats de ce travail

(Chapitre 3) suggèrent que ce paramètre pourrait être le nombre de mouvements normalisé

par le nombre de luminophores. Ce paramètre constitue en fait une probabilité instantanée de

mouvement qu’il devrait être possible de relier au temps de repos via la notion d’espérance

d’une variable aléatoire consistant en des multiples de l’intervalle de temps séparant

l’acquisition de deux images successives. Il deviendrait alors possible de dériver des temps de

repos et donc des coefficients de diffusion biologique a priori non affectés par la durée

expérimentale.

Les enregistrements (allant jusqu’à des durées de 6 jours) nécessaires à la réalisation

de l’ensemble de ces premières perspectives de travail a d’ores et déjà été acquis. La dernière

d’entre elles nécessitera cependant très probablement la mise en place d’une collaboration

avec des mathématiciens spécialistes dans le domaine des probabilités.

II. L’apport à la modélisation du remaniement

sédimentaire

Le concept d’ « empreintes de remaniement sédimentaire » fait directement référence

à la description stochastique de l’ensemble des mouvements de particules induit par un ou des

organismes benthiques telle qu’énoncée dans le modèle CTRW (Meysman et al. 2008).

L’obtention expérimentale de valeurs des caractéristiques de ces mouvements grâce au

déploiement de la nouvelle méthode présentée ci-dessus constitue par conséquent un cadre

adéquat pour discuter de la validité des hypothèses de ce modèle.

Les résultats obtenus confirment, de manière qualitative, certaines des hypothèses

constitutives du modèle CTRW. Les distributions des fréquences des temps de repos des

particules et de leurs longueurs de déplacements sont par exemple fortement biaisées vers les

fortes valeurs, avec de faibles occurrences des valeurs élevées de chacun de ces deux

paramètres. De plus, pour Abra alba, les coefficients de biodiffusion calculés (Particle


[198]

tracking biodiffusion coefficients) à partir des valeurs moyennes des temps de repos et de la

variance de la longueur des déplacement sont du même ordre de grandeur que ceux issus de

l’ajustement « classique » d’un modèle CTRW aux profils verticaux de concentrations de

luminophores mesurés après incubation de carottes de sédiment en présence de cette même

espèce (Braekman et al., 2010). Ceci tend à valider l’utilisation d’empreintes de remaniement

sédimentaire basées sur des distributions de fréquence biaisées vers les fortes valeurs qui

caractérisent les mouvements de particules sédimentaires. D’un point de vue plus quantitatif,

les distributions de fréquences mesurées ne se sont néanmoins avérées que faiblement

compatibles avec les lois de distribution théoriques le plus souvent utilisées dans le cadre du

modèle CTRW (i.e., la loi de Poisson pour les temps de repos et la loi normale pour les

longueurs de déplacement). Les caractéristiques des mouvements élémentaires mesurés lors

de ce travail attestent par ailleurs du caractère anisotrope du remaniement sédimentaire induit

par Abra alba. Les empreintes de remaniement sédimentaire mesurées expérimentalement se

sont enfin révélées très fortement variables spatialement, notamment dans leur dimension

verticale.

L’ensemble de ces points tend à invalider certaines des pratiques (e.g. l’utilisation de

distributions théoriques) ainsi que certaines des hypothèses (e.g. l’invariance spatiale)

constitutives du modèle CTRW. Il reste toutefois à déterminer si ces imperfections remettent

significativement en cause les résultats obtenus par l’utilisation « classique » du modèle

CTRW. Pour traiter de ce point, il conviendrait : (1) dans un premier de temps de démontrer

l’aptitude d’un modèle CTRW paramétré à l’aide des empreintes de remaniement

sédimentaire mesurées lors de la présente étude (incluant ou non leur discrétisation verticale)

à décrire efficacement les variations temporelles observées des profils verticaux de

concentrations de luminophores, puis (2) dans un deuxième temps de comparer les

caractéristiques des empreintes sédimentaires mesurées expérimentalement avec celles

dérivées de l’utilisation « classique » du modèle CTRW. Là encore, ces travaux pourraient

intégralement s’appuyer sur des données expérimentales déjà acquises.


[199]

III. Le contrôle du remaniement sédimentaire :

méthodologie, résultats et perspectives

3.1. Méthodologie

L’approche méthodologique développée consiste à suivre les mouvements de

luminophores dans le plan vertical constitué par la paroi d’un aquarium plat. Elle base sa

reconstitution des mouvements de particules sur la notion de proximité (i.e., entre une

particule disparue au temps t et une autre apparue au temps t+1) à l’intérieur d’un cercle de

recherche. Cette manière de procéder exclue de fait toute utilisation de cette approche pour la

caractérisation de modes de remaniement sédimentaire faisant intervenir essentiellement des

mouvements de type non locaux. Cette contrainte exclue par exemple les organismes

appartenant aux groupes fonctionnels des « convoyeurs » (vers le haut ou vers le bas) dont le

mode de remaniement sédimentaire repose essentiellement sur des transits de particules au

travers du tube digestif. Cette contrainte semble nettement moins prégnante (et donc la

nouvelle méthode mieux adaptée) dans les autres groupes fonctionnels de remaniement

sédimentaire, notamment les biodiffuseurs (dont les biodiffuseurs à galeries) et possiblement

les régénérateurs.

Analyser à partir de séquences d’image les mouvements de luminophores implique par

ailleurs que la taille de ces derniers soit : (1) supérieure à celle d’un pixel de l’image, ce qui

conditionne de fait la taille du champ capturé, et (2) voisine de la médiane granulométrique du

sédiment utilisé. A titre d’exemple, lors des expériences présentées dans les deux premiers

chapitres de ce manuscrit, avec des luminophores d’un diamètre moyen de 35µm et un

capteur optique dont la définition était proche de 5 millions de pixels (2560 x 1920 pixels), le

champ des images acquises mesurait ainsi seulement 4,2 x 3,1 cm. De telles contraintes,

purement techniques limitent de fait l’utilisation de la nouvelle approche à des organismes

pas/peu mobiles exerçant un remaniement sédimentaire locale.

Il serait très certainement intéressant d’appliquer la nouvelle approche expérimentale

à un organisme qui : (1) répondrait à ces deux contraintes, et (2) présenterait une grande

importance écologique, par exemple en contrôlant du fait de sa seule activité, l’intensité du

remaniement sédimentaire induit par l’ensemble de la communauté l’abritant. Une bonne

espèce candidate est l’annélide polychète Melinna palmata, comme détaillé dans la partie

IV.2 de cette synthèse.


[200]

Les travaux déjà conduits sur l’éthologie alimentaire et le remaniement sédimentaire

au sein du genre Abra ont démontré l’existence d’une forte variabilité individuelle (Maire et

al. 2006). Cette propriété complique l’analyse des effets des facteurs environnementaux. Il est

néanmoins largement possible de s’en affranchir en soumettant consécutivement un même lot

d’individus à plusieurs conditions environnementales (Grémare et al. 2004). L’approche

développée ici ne permet malheureusement pas de mettre en œuvre ce type de plan

expérimental. Les valeurs des temps de repos qu’elle permet d’obtenir sont en effet fonction

de la durée expérimentale (voir ci-dessus). En pratique, il s’avère par conséquent impossible

d’introduire de la matière organique particulaire en milieu d’expérience ce qui aurait pour

effet de perturber l’agencement spatial des luminophores, et/ou de prolonger la durée des

expériences au-delà de la durée temporelle suivant l’introduction de cette matière pendant

laquelle l’activité et le remaniement sédimentaire sont effectivement modifiés. En

contrepartie, la nouvelle approche permet d’accéder à une vision spatiale du remaniement

sédimentaire, ici souvent résumée sous la forme de profils verticaux. Cette structuration des

données m’a conduit à utiliser une approche statistique originale basée sur l’utilisation

d’analyses de variances multiples par permutations (PERMANOVAs, Anderson 2001,

McArdle & Anderson 2001) pour comparer, non plus des valeurs moyennes, mais des profils

verticaux des paramètres constitutifs des empreintes du remaniement sédimentaire. Cette

méthodologie a permis de mettre en évidence des effets significatifs de la température et de la

disponibilité de la matière organique sur les empreintes de remaniement sédimentaire alors

que ceci n’était pas le cas pour la comparaison des seules valeurs moyennes de ces mêmes

paramètres.

Certaines des options retenues lors de la mise en œuvre des PERMANOVAs

mériteraient néanmoins très certainement d’être explorées plus avant. Il s’agit notamment

de : (1) la métrique utilisée pour évaluer le niveau de similarité des profils verticaux

(distance euclidienne faisant intervenir les valeurs du paramètre analysé à chaque intervalle

de profondeur défini à partir de l’interface eau-sédiment), (2) de la valeur définissant chaque

intervalle de profondeur (1,32 mm dans le cas de la présente étude), et enfin (3) du seuil

d’activité (10 mouvements/48h) délimitant la gamme totale de profondeur prise en

considération. Ce travail pourrait notamment prendre la forme : (1) d’une analyse de

sensibilité, et (2) d’une interaction plus théorique avec des biostatisticiens et des spécialistes

de l’étude du remaniement sédimentaire. Ses résultats pourraient à terme être étendus à


[201]

d’autres domaines de l’écologie et de la biogéochimie des sédiments marins comme l’analyse

de la variabilité des profils verticaux d’éléments biogènes dont l’oxygène (voir plus bas).

3.2. Résultats

L’accès au lien direct entre éthologie et mouvements de particules mesurés à haute

résolution spatiale et temporelle, rendu possible par le déploiement de la nouvelle approche

méthodologique, a permis de caractériser l’effet de la température ainsi que de la matière

organique disponible à l’interface eau-sédiment sur les « empreintes du remaniement

sédimentaire » induit par Abra alba. Les résultats obtenus suggèrent que, chez cette espèce,

c’est la température qui contrôle en premier lieu le remaniement sédimentaire. L’activité

siphonale est en effet plus importante en conditions estivales ; ce qui entraine des

mouvements de particules plus fréquents, de plus grande amplitude et d’une plus grande

variabilité. Cet effet de la température se trouve indirectement confirmé par le fait, qu’en

conditions automnales, la forte restriction du remaniement sédimentaire par la faible

température empêche la détection d’un éventuel effet d’un ajout de matière organique

fraiche ; alors qu’à l’inverse, en conditions estivales, ce même ajout restreint transitoirement

l’amplitude et la fréquence des mouvements de particules. Cette inhibition pourrait résulter

d’un changement de stratégie alimentaire depuis un mode «déposivore de surface»

impliquant une activité siphonale importante, à un mode «suspensivore» associé à des siphons

immobiles et pointant dans la colonne d’eau lorsque la concentration en matière organique

particulaire en suspension est élevée. De manière plus générale, l’existence chez Abra alba

d’un continuum conditions environnementales-éthologie-remaniement sédimentaire pointe la

nécessité de mieux prendre en compte les variables environnementales dans le classement des

espèces responsables du remaniement sédimentaire en groupes fonctionnels (François et al.

1997 ; Solan et Wigham 2005 ; Kristensen et al. 2012), notamment lors d’études portant sur le

calcul d’indices synthétiques visant à décrire le remaniement sédimentaire à partir de données

faunistiques (Solan et al. 2004a).

3.3. Perspectives

L’approche méthodologique nouvelle développée dans le cadre de ce travail permet

d’envisager plusieurs perspectives d’application relatives à l’étude des mécanismes


[202]

contrôlant le remaniement sédimentaire ainsi qu’à l’interaction de ce processus avec la

minéralisation de la matière organique sédimentée.

Le premier de ces deux points concerne tout d’abord la prise en compte d’autres

facteurs abiotiques tels que les contaminants. En effet, des contaminants organiques tels que

le DDT, à des concentrations non-létales, peuvent inhiber l’intensité de remaniement

sédimentaire (Mulsow et al. 2002). L’utilisation de l’approche méthodologique développée ici

permettrait ainsi, de par ses aspects dynamique et en deux dimensions, de déterminer

directement l’effet de tels contaminants sur le comportement d’organismes benthiques cibles

et sur leur empreinte de remaniement sédimentaire.

Un deuxième aspect concerne l’étude des effets de paramètres biotiques et tout

particulièrement de la densité et de la biodiversité. Dans le premier cas, l’étude pourrait, au

moins dans un premier temps, concerner Abra alba. Une manipulation réaliste de la densité

nécessiterait alors d’utiliser des capteurs à plus haute résolution de manière à pouvoir

étudier un champ de plus grande dimension dans lequel la densité d’individu pourrait varier

de manière significative. L’intérêt d’une telle étude consisterait à tester la forme exacte de la

relation liant densité et intensité du remaniement sédimentaire, forme qui est a priori souvent

considérée comme linéaire sans réel support expérimental (Solan et al. 2004a).

Sur la base : (1) de l’analogie entre la paroi verticale d’un aquarium plat et la vitre

d’un profileur sédimentaire, et (2) de travaux préliminaires ayant utilisé un profileur

sédimentaire pour quantifier le remaniement sédimentaire (Solan et al. 2004b), il était

initialement prévu dans le cadre de ce travail de transposer la nouvelle approche

expérimentale à l’étude du remaniement sédimentaire induit par des communautés benthiques

in situ. Cette perspective est finalement apparue comme non réaliste du fait de certaines des

limitations exposées ci-dessus. Une étude des effets de la diversité de la macrofaune

benthique reste néanmoins possible. Elle suppose :

(1) de travailler sur des groupements d’espèces artificiellement reconstitués comme

c’est souvent le cas pour les études liant biodiversité et fonctionnement (Mermillod-

Blondin et al. 2005).

(2) de travailler au sein de groupements composés exclusivement voire quasi-

exclusivement d’organismes appartenant aux groupes fonctionnels des biodiffuseurs

(incluant les biodiffuseurs à galeries) et des régénérateurs.


[203]

(3) d’acquérir des images représentatives de la distribution spatiale des organismes qui

composent cette communauté ; ces images devant par ailleurs présenter une résolution

suffisamment importante (i.e., taille du luminophore > taille du pixel).

(4) de paramétrer les algorithmes d’analyse d’image, c’est-à-dire d’adapter la taille du

cercle de recherche aux espèces présentes. Cette étape peut être envisagée sous deux

angles différents. Le premier résiderait en une seule analyse avec des réglages

« moyens » aptes à caractériser le plus possible de mouvements de luminophores dans

l’ensemble du plan. Le second consisterait en plusieurs analyses avec des réglages

distincts selon la portion de sédiment. Cette dernière approche, bien que nettement plus

chronophage, conduirait probablement à des mesures nettement plus fines, puisque à un

type de mouvement prenant place dans une zone du sédiment serait associé un réglage

optimal du cercle de recherche ainsi que de la fréquence d’acquisition.

L’accession à une vision dynamique et en deux dimensions du remaniement sédimentaire

ouvre enfin la voie à l’amélioration de la compréhension des mécanismes biologiques

impliqués dans le contrôle des processus biogéochimiques prenant place à proximité de

l’interface eau-sédiment. Ces derniers, et notamment la distribution spatiale de l’efficience du

recyclage de la matière organique par les microorganismes semblent en effet étroitement liés

à l’hétérogénéité spatiale du remaniement sédimentaire et de la bioirrigation induite par les

organismes benthiques (Bertics et Ziebis 2010). Des travaux dans ce sens pourraient

notamment inclure le couplage : (1) de mesure d’activité (Maire et al. 2007b), (2) de la

nouvelle méthode de mesure des mouvements de particules, et (3) d’autres techniques

procurant une vision dynamique et 2D de processus biogéochimiques (e.g. optodes planaires,

Volkenborn et al. (2012) ; ou bien gels DGT, Diffusion Gradient Thin gels, Teal et al. (2012).

Une telle combinaison, déjà partiellement mise en œuvre par Teal et al. (2012), permettrait

par exemple de mettre en relation l’hétérogénéité spatiale de l’activité et du remaniement

sédimentaire par une sélection d’organismes endobenthiques avec les distributions spatiales

d’éléments chimiques impliqués dans les processus redox tels que l’oxygène, les nitrates ou

certains métaux.


[204]

Figure 5.1: Exemples d’approches 2-D pour l’analyse conjointe de la bioirrigation, du

remaniement sédimentaire et de mesure des flux de métaux au sein d’aquariums plats. A :

distribution de l’oxygène induite par l’annélide polychète Arenicola marina dans un sédiment

perméable obtenue grâce à l’utilisation d’optodes planaires (modifié d’après Volkenborn et al.

(2010)), approche dynamique. B : distribution spatiale de l’intensité du remaniement

sédimentaire vertical (Dbx) effectué par A. alba en 48h obtenue grâce à la mesure des

mouvements de luminophores à partir de vidéo sous UV (d’après Bernard et al. (2012)),

approche dynamique. C : distribution spatiale des flux de fer mesurés dans un gel DGT placé

dans le sédiment (modifié d’après Tankere-Muller et al. (2007)), approche intégrative.

IV. Mesures in-situ de l’intensité de remaniement

sédimentaire, relation avec la composition des

communautés benthiques

Le troisième chapitre de ce manuscrit présente les résultats d’une étude conduite au

cours d’un cycle saisonnier et visant à comparer le remaniement sédimentaire, les

compositions des communautés benthiques, les caractéristiques sédimentaires et les


[205]

caractéristiques de la population de phanérogames dans un herbier à Zostera noltii et dans un

sédiment vaseux non végétalisé.

En l’état actuel des connaissances, cette étude est la première à mesurer l’intensité du

remaniement sédimentaire induit par l’ensemble de la communauté benthique au sein d’un

herbier de phanérogames marines. En ce sens, elle se démarque des études antérieures qui se

sont essentiellement attachées à décrire les interactions biomécaniques intervenant entre le

remaniement sédimentaire induit par certaines espèces bioturbatrices de grande taille et le

développement du réseau de racines/rhizomes des phanérogames (Hughes et al., 2000 ;

Berkenbusch and Rowden 2007 ; Berkenbusch et al., 2007 ; Siebert and Branch 2005, 2007 ;

Wesenbeeck et al 2007).

Les résultats obtenus ont notamment permis : (1) de mettre en évidence l’effet

structurant sur les variations spatiale et saisonnière de l’herbier sur les communautés

benthiques endogées et l’intensité du remaniement sédimentaire associée, et (2) d’identifier

certaines espèces clés dans le contrôle du remaniement sédimentaire.

4.1. La dispersion des données comme proxy de

l’hétérogénéité spatiale : Détection d’effets de la régression de

l’herbier sur la distribution des organismes benthiques endogés

et du processus de remaniement sédimentaire

L’intensité du remaniement sédimentaire et la composition des communautés

benthiques endogées se sont avérées beaucoup plus variables dans les vases nues que dans

l’herbier à Z. noltii. Si ce fait complique fortement l’analyse statistique et notamment la

détection d’effet significatif des facteurs considérés sur les niveaux moyens de ces

paramètres, la prise en compte de la dispersion des données en tant que proxy de

l’hétérogénéité spatiale permet également la mise en évidence des effets d’éventuelles

perturbations sur la composition des communautés benthiques (Warwick and Clarke, 1993;

Hewitt and Thrush 2009) ainsi que sur l’intensité du remaniement sédimentaire. La variabilité

des patrons écologiques et des processus associés s’avère en effet extrêmement sensible aux

perturbations (Fraterrigo et Rusak, 2008). Les résultats obtenus confortent cet état de fait dans

la mesure où la disparition des phanérogames, qu’elle intervienne à l’échelle pluriannuelle

(telle que mise en évidence par la comparaison des stations d’herbier et de vases nues) ou bien


[206]

en l’espace de seulement quelques mois (telle que mise en évidence par le cycle saisonnier de

la station herbier) se traduit par des augmentations de l’hétérogénéité spatiale de la

composition des communautés benthiques endogés et de l’intensité du remaniement

sédimentaire. L’analyse des tendances des valeurs moyennes associées suggère dès lors une

plus faible intensité du remaniement sédimentaire dans l’herbier que dans les vases nues. Ceci

est très certainement lié à la présence dans l’herbier d’un dense réseau de racines et rhizomes

ainsi qu’aux fortes abondances d’organismes tubicoles comme M. palmata, limitant la

composante verticale descendante du remaniement sédimentaire par des processus de

compaction (Brenchley 1982 ; Reise 2002), ou bien convoyeurs vers le haut comme

l’oligochète Tubificoides benedii par transport actif des particules sédimentaires vers le haut

(Bianchi, 1988). Les fortes abondances de ces dernières espèces semblent résulter de fortes

teneurs en matière organique particulaire. Lors de la disparition de l’herbier, ou plus

précisément lorsque la densité des pieds de zostères décroit en deçà d’un seuil critique

d’environ 6000 pieds.m-2

(Blanchet et al. 2004), le réseau de racines et rhizomes souterrains

devient discontinu, ce qui a pour conséquence la création d’une hétérogénéité spatiale de la

distribution des débris végétaux et par effet « cascade » des hétérogénéités spatiales de la

composition des communautés benthiques endogées et du remaniement sédimentaire, ainsi

que de la valeur moyenne de ce même remaniement sédimentaire.

Pour estimer plus précisément le niveau de ce seuil critique de densité de l’herbier en

deçà duquel son rôle structurant est altéré, il conviendrait de réaliser de nouvelles mesures

de l’intensité de remaniement sédimentaire et de la composition des communautés benthiques.

De telles mesures devraient être conduites sur une large gamme de densité d’herbier et être

basées sur un plus grand nombre de réplicats. Une telle augmentation de l’effort

d’échantillonnage permettrait de détecter, non plus uniquement des effets statistiques liés à la

dispersion des données, mais également ceux associés aux valeurs moyennes des différents

paramètres. De plus, étant donné : (1) l’effet important semblant être joué par les parties

souterraines des phanérogames et, (2) le caractère d’ancien herbier de la station de vases

nues étudiées dans le cadre du présent travail ; des mesures de la biomasse de racines et de

rhizomes permettraient de valider l’hypothèse selon laquelle les patchs de végétaux en

décomposition dans le sédiment conditionnent la forte hétérogénéité spatiale de la répartition

des organismes benthiques et du remaniement sédimentaire après disparition de l’herbier.


[207]

4.2. Identification d’espèces « clés » dans le contrôle du

remaniement sédimentaire, mise en évidence d’effets de

stabilisation du sédiment.

Ce travail a également permis de mettre en évidence le rôle « clé » de certaines

espèces endogées dans le contrôle du remaniement sédimentaire (voir également plus haut,

partie 4.1.). Les fortes abondances de l’annélide polychète Melinna palmata sont ainsi

significativement corrélées aux faibles intensités de remaniement sédimentaire. Ce lien

suggère le rôle important d’inhibiteur du remaniement sédimentaire pour cette espèce.

Melinna palmata est une espèce tubicole appartenant à la famille des ampharetidae qui

s’établit en population dense, induisant par là même une forte compaction du sédiment

(Brenchley, 1982). Le fait que cette espèce se nourrisse en prélevant les particules à

l’interface eau-sédiment à l’aide de ses tentacules buccaux où elle produit également ses

pelotes fécales (Fauchald et Jumars, 1979) apparaît également parfaitement cohérent avec une

inhibition de la composante verticale du remaniement sédimentaire. Ce mode de remaniement

sédimentaire, essentiellement horizontal, ne coïncide par ailleurs avec aucune groupes

fonctionnels existants (François et al. 1997 ; Solan et Wigham 2005 ; Kristensen et al. 2012)

ce qui n’est pas étonnant dans la mesure où leur définition est essentiellement basée sur la

composante verticale de ce même remaniement.

En ce sens, obtenir expérimentalement, à partir d’études purement mécanistiques en

laboratoire, les caractéristiques des mouvements de particules induits par M. palmata,

identifiée ici comme dominante dans l’herbier, permettrait de caractériser plus précisément

l’effet négatif qu’exercerait cette espèce sur la composante verticale du remaniement

sédimentaire dans l’herbier. L’impact de M. palmata sur les particules de sédiment est

essentiellement en surface où son action délimite clairement des zones de prospection et

d’accumulation de particules fines. Cet organisme fait actuellement l’objet du travail de thèse

de Cécile Massé qui étudie notamment le contrôle exercé par l’activité de bioturbation de M.

palmata sur la structuration des communautés microbiennes associées au sédiment. Certains

des résultats présentés dans ce manuscrit ayant confirmé l’importance de la composante

horizontale du remaniement sédimentaire (Wheatcroft, 1991), et Maire et al (2007c) ayant

démontré la faisabilité d’un suivi d’activité et d’une première analyse sommaire du

mouvement des particules sédimentaires dans le plan horizontal constitué par l’interface eau-


[208]

sédiment, une perspective de travail intéressant consisterait à appliquer la nouvelle approche

méthodologique à l’étude des mouvements horizontaux de particules induits par M. palmata.

Un tel objectif pourrait être réalisé grâce à l’analyse de séquences d’images « vues de

dessus » (Maire et al. 2007c) sous lumière UV et en présence de luminophores dispersés à

l’interface eau-sédiment, et ce via l’utilisation du logiciel présenté dans la première partie de

ce manuscrit. La principale difficulté dans cette approche de mesure d’empreinte de

remaniement sédimentaire, résiderait cependant, comme expliqué précédemment, dans la

mesure du temps de transit des particules dans le tube digestif. Il conviendrait ainsi de

procéder par une introduction des luminophores au début de l’expérience afin de déterminer,

en plus des mouvements de luminophores le long des tentacules buccaux, les flux de ces

mêmes luminophores entrant ou sortant des tubes créés par cet organismes, et ce grâce à la

détection des directions des mouvements et de la localisation des tubes.

4.3. Restriction de l’intensité du remaniement sédimentaire

dans l’herbier abritant une plus grande densité d’organismes

que la vase nue

En dépit des différences de variabilité évoquées ci-dessus, les communautés

benthiques endogées associées aux stations herbier et vases nues restent étroitement

apparentées. Les mêmes espèces sont présentes au sein des deux stations mais avec de plus

faibles abondances au sein des vases nues qui semblent par conséquent constituer une forme

appauvrie de la communauté de l’herbier. La présence de ces espèces caractéristiques de

l’herbier au sein des vases nues peut s’expliquer : soit par la proximité immédiate de l’herbier

(Fredriksen et al. 2010), soit par la rémanence de débris végétaux dans ce qui constitue en

réalité un ancien herbier (Plus et al. 2010). Le point important, est que pour des compositions

spécifiques somme toute assez voisines, le remaniement sédimentaire tend à être plus

important pour la communauté présentant les plus faibles abondances. Ce patron qui semble

paradoxal traduit en fait la difficulté à dériver un potentiel de remaniement sédimentaire à

partir de la seule composition de la macrofaune benthique comme proposé par Solan et al

(2004a) via la notion de Bioturbation Potential Index (BPI).

Ces auteurs attribuent des scores aux espèces d’organismes benthiques en fonction de

leur mobilité (depuis 1 pour une espèce tubicole fixe jusqu’à 4 pour une espèce errante créant

un système de terriers), de leur position dans le sédiment et de leur mode de remaniement


[209]

sédimentaire (depuis 1 pour une espèce épibenthique modifiant uniquement l’interface eau-

sédiment jusqu’à 5 pour une espèce appartenant au groupe des régénérateurs). Ceci leur

permet de dériver un indice per capita (BPi) propre à chaque espèce d’après la formule

suivante :

BPi = √ Mi Ri

où Bi est la biomasse la biomasse individuelle moyenne, Mi le score relatif à la mobilité et Ri

le score relatif à la position dans le sédiment et le mode de remaniement sédimentaire, pour

l’espèce considérée.

Cet indice per capita est ensuite multiplié par l’abondance de l’espèce considérée

mesurée et l’obtention d’un indice synthétique pour l’ensemble de la communauté (BPc)

s’effectue ensuite simplement via le calcul de la somme de ces contributions spécifiques.

En l’état, cet indice ne tient pas compte : (1) des interactions intraspécifiques

(Braeckman et al. 2010) et interspécifiques (Maire et al. 2010), connues pour moduler

l’intensité du remaniement sédimentaire, (2) des interactions interspécifiques, liées à l’impact

de certaines espèces ingénieure susceptibles, à l’instar des phanérogames marines, de

modifier leur environnement physique et par conséquent l’activité de remaniement

sédimentaire des autres espèces présentes et, (3) de l’impact des facteurs environnementaux,

comme démontré dans le troisième chapitre de ce manuscrit. Pour ce qui est du deuxième

point, ceci tend à valider la démarche de quantification des phénomènes de restriction du

remaniement sédimentaire sur sa composante verticale induits par des espèces

« ingénieures » à différentes densités, à partir d’études in-situ ou mécanistiques grâce à

l’utilisation de la nouvelle méthode, comme proposée dans la partie IV de cette synthèse. A

terme, dans des milieux particuliers comme les herbiers de phanérogames, une telle

démarche aboutirait à l’établissement, si possible, de coefficients de corrections, reflétant

l’ampleur de la restriction du remaniement sédimentaire. Ces coefficients pourraient alors

être introduits dans l’équation permettant de calculer le BPc donnant ainsi à cet indice la

capacité à déterminer plus précisément, à partir de la composition faunistique, un potentiel

de bioturbation d’une communauté dans des habitats comme les herbiers.


[210]

V. Bilan

Figure 2 : Représentation schématique de l’articulation des principaux résultats présentés

dans ce manuscrit et des perspectives qui en découlent. Les boîtes blanches indiquent les

principales actions menées, les boîtes grises les résultats et les bleues les perspectives

envisageables (en bleu foncé celles correspondent à celles à court terme liées à l’exploitation

de données déjà acquises, et en bleu clair les perspectives à plus long terme).

Références bibliographiques

[211]

Références bibiographiques


[212]




Aller R. C. and J. Y. Aller. 1998. The effect of biogenic irrigation intensity and solute

exchange on diagenetic reaction rates in marine sediments. J. Mar. Res., 52, 905-936.

Amouroux, J. M., A. Grémare and J. Amouroux. 1989. Modelling of consumption and

assimilation in Abra alba (Mollusca, Bivalvia). Mar. Ecol. Prog. Ser., 51, 87-97.







Auby, I. 1991. Contribution à l'étude des herbiers de Zostera noltii dans le bassin d'Arcachon

: Dynamique, production et dégradation, macrofaune associée. PhD Thesis, Université

Bordeaux 1, 162 pp.

Auby, I. and P. –J. Labourg. 1996. Seasonal dynamics of Zostera noltii Hornem. in the bay of

Arcachon (France). J. Sea Res., 34(4), 269-277.

Auby, I., C. –A. Bost, H. Budzinski, S. Dalloyau, A. Desternes, A. Belles, G. Trut, M. Plus,

C. Pere, L. Couzi, C. Feigne, J. Steinmetz. 2011. Régression des herbiers de zostères dans

le Bassin d’Arcachon : état des lieux et recherche des causes.

Bachelet, G., X. de Montaudouin, I. Auby and P-J. Labourg. 2000. Seasonal changes in

macrophyte and macrozoobenthos assemblages in three coastal lagoons under varying

degrees of eutrophication. ICES J. Mar. Sci., 57, 1495-1506.




Barbier, E. D., S. D. Hacker, C. Kennedy, E. W. Koch, A. C. Stier and B. R. Silliman. 2011.

The value of estuarine and coastal services. Ecol. Monogr., 81(2), 169-193.

Berkenbusch, K. and A. A. Rowden. 1999. Factors influencing sediment turnover by the

burrowing ghost shrimp Callianassa filholi (Decapoda: Thalassinidea). J. Exp. Mar. Biol.

Ecol., 238, 283–292.

Berkenbusch, K. and A. A. Rowden. 2007. An examination of the spatial and temporal

generality of the influence of ecosystem engineers on the composition of associated

assemblages. Aquat. Ecol., 41, 129-147.

Berkenbusch, K., A. A. Rowden and T. E. Myers. 2007. Interactions between seagrasses and

burrowing ghost shrimps and their influence on infaunal assemblages. J. Exp. Mar. Biol.

Ecol., 341, 70-84.


[213]

Bertics, V. J. and Ziebis, W. 2010. Bioturbation and the role of microniches for sulfate

reduction in coastal marine sediments. Environmental Microbiology. Vol. 12 (11), pp. 3022-

3034.

Blackburn, N. J. and R. J. Orth. 2013. Seed burial in eelgrass Zostera marina: The role of

infauna. Mar. Ecol. Prog. Ser., 474, 135-145.

Blair, N. E., L. A. Levin, D. DeMaster, G. Plaia. 1996. The short term fate of fresh algal

carbon in continental slope sediments. Limnol. Oceanogr., 41,1208–1219.

Blanchet, H. 2004. Structure et fonctionnement des peuplements benthiques du bassin

d’Arcachon. Phd thesis, University Bordeaux 1, 224 pp.

Blanchet, H., X. de Montaudouin, A. Lucas and P. Chardy. 2004. Heterogeneity of

macrozoobenthic assemblages within a Zostera noltii seagrass bed : diversity, abundance,

biomass and structuring factors. Est. Coast. Shelf Sci., 61, 111-123.

Blanchet, H., X. de Montaudouin, P. Chardy and G. Bachelet. 2005. Structuring factors and

recent changes in subtidal macrozoobenthic communities of a coastal lagoon, Arcachon

Bay (France). Est. Coast. Shelf. Sci., 64, 561-576.




Oceanogr. Ser., 70, 455-492.

Boström, C. and E. Bonsdorff. 1997. Community structure and spatial variation of benthic

invertebrates associated with Zostera marina (L.) beds in the northern Baltic Sea. J. Sea.

Res., 37, 153-166.

Boström, C., E. Bonsdorff, P. Kangas and A. Norkko. 2002. Long-term changes of a

brackish-water eelgrass (Zostera marina) community indicate effects of coastal

eutrophication. Est. Coast. Shelf Sci., 55, 795-804.

Boström, C., E. L. Jackson and C. A. Simenstad. 2006. Seagrass landscapes and their effects

on associated fauna: A review. Est. Coast. Shelf Sci., 68, 383-403.

Boucher, J. -M. 1968. Etude océanographique des chenaux du bassin d'Arcachon. PhD Thesis,

Université Bordeaux 1, 168 pp.

Bouchet, V. M. P., P.-G. Sauriau, J.-P. Debenay, F. Mermillod-Blondin, S. Schmidt, J.-C.

Amiard and B. Dupas. 2009. Influence of the mode of macrofauna-mediated bioturbation

on the vertical distribution of living foraminifera: First insight from axial

tomodensitometry. J. Exp. Mar. Biol. Ecol., 371, 20-33.



Boudreau, B. P. 1986b. Mathematics of tracer mixing in sediment: II. Non-local mixing and

biological conveyor-belt phenomena. Am. J. Sci., 286, 199-238.


[214]

Boudreau, B. P. 1997. A one-dimensional model for bed-boundary layer particle exchange. J.

Mar. Syst., 11, 279-303.

Boudreau, B. P. and D. M. Imboden. 1987. Mathematics of tracer mixing in sediments. III.

The theory of nonlocal mixing within sediments. Am. J. Sci., 287, 693-719.

Bouma T. J., V. Ortells and T. Ysebaert. 2009. Comparing biodiversity effects among

ecosystem engineers of contrasting strength: macrofauna diversity in Zostera noltii and

Spartina anglica vegetations. Helgol. Mar. Res., 63, 3-18.




Brenchley, G. A. 1982. Mechanisms of spatial competition in marine soft-bottom



seagrasses: A review. Est. Coast. Shelf Sci., 79, 354-366.

Cacabelos, E., J. Moreira, A. Lourido and J. S. Troncoso. Ecological features of Terebellida

fauna (Annelida, Polychaeta) from Ensenada de San Simon (NW Spain). Anim. Biodivers.

Conserv., 34(1), 141-150.

Cadée, G. C. 1976. Sediment reworking by Arenicola marina on tidal flats in the Dutch

Wadden Sea. Neth. J. Sea Res., 10, 440-460.



52, 199-210.

Castel, J., P-J. Labourg, V. Escaravage, I. Auby and M. E. Garcia. 1989. Influence of seagrass

beds and Oyster parks on the abundance and biomass patterns of meio- and macrobenthos

in tidal flats. Est. Coast. Shelf Sci., 28, 71-85.

Chauvaud, L., F. Jean, O. Ragueneau and G. Thouzeau. 2000. Long-term variation of the bay

of Brest ecosystem : Benthic-pelagic coupling revisited. Mar. Ecol. Prog. Ser., 200, 35-48.

Choi, J., F. François-Carcaillet and B. P. Boudreau. 2002. Lattice-automaton bioturbation

simulator (LABS): implementation for small deposit feeders. Comput. Geosci., 28, 213-

222.



Clifton, R. J., P. G. Watson, J.T. Davey and P. E. Frickers. 1995. A study of processes

affecting the uptake of contaminants by intertidal sediments, using radioactive tracers:

7Be, 137Cs and unsupported 210Pb. Est. Coast. Shelf Sci., 41, 459-474.

Cloern, J. E. 1982. Does the benthos control phytoplankton biomass in south San Francisco

bay? Mar. Ecol. Prog. Ser., 9, 191-202.


[215]

Conley, D. J., S. Björk, E. Bonsdorff, J. Carstensen, G. Destouni, B. G. Gustafson, S.

Hietanen, M. Kortekaas, H. Kuosa, H. E. M. Meier, B. Müller-Karulis, K. Nordberg, A.

Norkko, G. Nürnberg, H. Pitkänen, N. N. Rabalais, R. Rosenberg, O. P. Savchuk, C. P.

Slomp, M. Voss, F. Wulff and L. Zillén. 2009. Hypoxia-related processes in the Baltic Sea.

Environ. Sci. Technol., 43(10), 3412-3420.

Crank, J. 1975. The Mathematics of Diffusion. Oxford University Press, Oxford.

Creel, L. 2003. Ripple effect: population and coastal regions. Population reference bureau,

Measure communication, www.prb.org.

D’Andrea, A. F., Lopez G. R., Aller R. C. 2004. Rapid physical and biological particle

mixing on an intertidal sandflat. J. Mar. Res., 62, 67–92.

Darwin, C. 1881. The Formation of Vegetable Mould Through the Action of Worms With

Observation of Their Habits. John Murray.

Dauvin, J.-C., T. Ruellet, E. Thiebaut, F. Gentil, N. Desroy, A.-L. Janson, S. Duhamel, J.

Jourde and S. Simon. 2007. The presence of Melinna palmata (Annelida: Polychaeta) and

Ensis directus (Mollusca: Bivalvia) related to sedimentary changes in the bay of Seine

(English Channel, France). Cah. Biol. Mar., 48: 391-401.

Davis, R. B. 1974. Stratigraphic effects of tubificids in profundal lake sediments. Limnol.

Oceanogr. 19, 466–488.

Davison, C. 1891. On the amount of sand brought up by lobworms to the surface. Geol. Mag.

8, 489-493.

De Wit, R., L. J. Stal, B. A. Lomstein, R. A. Herbert, H. van Gemerden, P. Viaroli, V. U.

Cecherelli, F. Rodrìguez-Valera, M. Bartoli, G. Giordani, R. Azzoni, B. Schaub, D. T.

Welsh, A. Donnelly, A. Cifuentes, J. Antòn, K. Finster, L. B. Nielsen, A. G. Underlien

Pedersen, A. Turi Neubauer, M. A. Colangelo and S. K. Heijs 2001. ROBUST: The ROle

of BUffering capacities in Stabilising coastal lagoon ecosystems. Cont. Shelf Res., 21,

2021–2041.

Delefosse, M. and E. Kristensen. 2012. Burial of Zostera marina seeds in sediment inhabited

by three polychaetes: Laboratory and field studies. J. Sea. Res., 71, 41-49.

Delmotte, S., F. J. R. Meysman, A. Ciutat, A. Boudou, S. Sauvage and M. Gérino. 2007.

Cadmium Transport in Sediments by Tubificid Bioturbation: an Assessment of Model

Complexity. Geochim. Cosmochim. Acta., 71, 844-862.

Den Hartog, C. and Kuo, J., 2006. Taxonomy and biogeography of seagrasses. In: Larkum

A.W.D. (Editor), Seagrasses: Biology, Ecology and Conservation. Springer, pp. 1-23.

Do, V. T., X. de Montaudouin, N. Lavesque, H. Blanchet and H. Guyard. 2011. Seagrass

colonization: Knock-on effects on zoobenthic community, populations and individual

health. Est. Coast. Shelf Sci., 85, 458-469.

Do, V. T., H. Blanchet, X. de Montaudouin and N. Lavesque. 2013. Limited consequences of

seagrass decline on benthic macrofauna and associated biotic indicators. Estuar. Coast.,

published online.


[216]

Dorgan, K. M., P. A. Jumars, B. Johnson and B. P. Boudreau. 2006. Macrofaunal burrowing:

The medium is the message. Oceanogr. Mar. Biol., 44, 85-121.

Duarte, C. M. 2002. The future of seagrass meadows. Environ. Conserv., 29 (2), 192-206.

Duarte, C. M. and C. L. Chiscano. 1999. Seagrass biomass and production : a reassessment.

Aquat. Bot., 65, 159-174.

Duchêne J. –C. and C. Nozais 1994. Light influence on larval emission and vertical

swimming in the terebellid worm Eupolymnia nebulosa (Montagu, 1818). Mem. Mus.

Natl. Hist. Nat. Ser. A. Zool., 162, 405–412.

Duchêne, J. C. and H. Queiroga. 2001. Use of an intelligent CCD camera for the study of

endogenous vertical migration rhythms in first zoeae of the crab Carcinus maenas. Mar.

Biol. 139, 901–909.

Duchêne, J. C. and R. Rosenberg. 2001. Marine benthic faunal activity patterns on a sediment

surface assessed by video numerical tracking. Mar. Ecol. Prog. Ser., 223,113-119.

Duport, E., G. Stora, P. Tremblay and F. Gilbert. 2006. Effects of population density on the

sediment mixing induced by the gallery-diffusor Hediste (Nereis) diversicolor O.F. Müller,

1776. J. Exp. Mar. Biol. Ecol., 336, 33-41.

Duport, E., F. Gilbert, J. –C. Poggiale, K. Dedieu, C. Rabouille and G. Stora. 2007. Benthic

macrofauna and sediment reworking quantification in contrasted environments in the Thau

Lagoon. Est. Coast. Shelf Sci., 72, 522-533.

Eklöf, J. S., T. van der Heide, S. Donadi, E. M. van der Zee, R. O’Hara and B. K. Eriksson.

2011. Habitat-mediated facilitation and counteracting ecosystem engineering interactively

influence ecosystem responses to disturbance. Plos one, 6(8), e23229.

Fauchald, K and P. A. Jumars. 1979. The diet of worms: a study of polychaete feeding guilds.

Oceanogr. Mar. Ann. Rev., 17, 193-284.

Fonseca M. S. and J. S. Fisher. 1986. A comparison of canopy friction and sediment

movement between four species of seagrass with reference to their ecology and restoration.




45, 295-319.

François, F., M. Gérino, G. Stora, J. P. Durbec and J. C. Poggiale. 2002. Functional approach

to sediment reworking by gallery-forming macrobenthic organisms: modeling and

application with the polychaete Nereis diversicolor. Mar. Ecol. Prog. Ser., 229, 127-136.

Fraterrigo, J. M. and J. A Rusak. 2008. Disturbance-driven changes in the variability of

ecological patterns and processes. Ecol. Lett., 11, 756-770.

Fredriksen, S., A. de Backer, C. Boström and H. Christie. 2010. Infauna from Zostera marina

L. meadows in Norway. Differences in vegetated and unvegetated areas. Mar. Biol. Res.,

6, 189-200.


[217]

Ganthy, F. 2011. Rôle des herbiers de zostères (Zostera noltii) sur la dynamique sédimentaire

du Bassin d’Arcachon. PhD thesis, University Bordeaux 1, 284 pp.

Ganthy, F., A. Sottolichio and R. Verney. 2013. Seasonal modification of tidal flat sediment

dynamics by seagrass meadows of Zostera noltii (Bassin d’Arcachon, France). J. Mar.

Syst., 109-110, S233-S240.

Garcia, C. 2010. Approche fonctionnelle des communautés benthiques du bassin oriental de la

Manche et du sud de la mer du Nord. Phd thesis, University Lille 1, 399 pp.

Gardner, L. R., P. Sharma and W. S. Moore. 1987. A regeneration model for the effect of

bioturbation by fiddler crabs on 210Pb profiles in salt marsh sediments. J. Environ.

Radioact., 5, 25−36.




46, 531-547.

Gérino, M., M. Frignani, C. Mugnai, L. G. Bellucci, D. Prevedelli, A. Valentini, A. Castelli,

S. Delmotte and S. Sauvage. 2007. Bioturbation in the Venice Lagoon: Rates and

relationship to organisms. Acta Oecol., 32, 14-25.

Giangrande, A., M. Montresor, A. Cavallo and M. Licciano. 2002. Influence of Naineris

laevigata (Polychaeta: Orbiniidae) on vertical grain size distribution, and dinoflagellate

resting stages in the sediment. J. Sea. Res., 47, 97-108.





Biol. Ecol., 285/286, 251–263.

Gilbert, F., S. Hulth, V. Grossi, J. –C. Poggiale, G. Desrosiers, R. Rosenberg, M. Gérino, F.

François-Carcaillet, E. Michaud and G. Stora. 2007. Sediment reworking by marine

benthic species from the Gullmar Fjord (Western Sweden): Importance of faunal

biovolume. J. Exp. Mar. Biol. Ecol., 348, 133-144.

Glé, C., Y. Del Amo, B. Sautour, P. Laborde and P. Chardy. 2008. Variability of nutrients and

phytoplankton primary production in a shallow macrotidal coastal ecosystem (Arcachon

Bay, France). Est. Coast. Shelf. Sci., 76, 642-656.

Goldberg, E. D. and M. Koide. 1962. Geochronological studies of deep sea sediments by the

ionium/thorium method. Geochim. Cosmochim. Acta., 26, 417-450.

Grall, J. et L. Chauvaud. 2002. Marine eutrophication and benthos: the need for new

approaches and concepts. Glob. Change Biol., 8, 813-830.


261.


[218]

Green, E.P. and Short, F.T., 2003. World atlas of seagrasses. University of California Press.






Heck Jr., K. L., G. Hays and R. J. Orth. 2003. Critical evaluation of the nursery role

hypothesis for seagrass meadows. Mar. Ecol. Prog. Ser., 253, 123-136.

Herman, P. M. J., J. J. Middelburg and C. H. R. Heip. 2001. Benthic community and sediment

processes on an intertidal flat: results from the ECOFLAT project. Cont. Shelf Res., 21,

2055-2071.

Hewitt, J. E. and S. F. Trush. 2009. Do species’ abundances become more spatially variable

with stress? T. O. Ecol. J., 2, 37-46.

Hollertz K and J. –C. Duchêne. 2001. Burrowing behaviour and sediment reworking in the

heart urchin Brissopsis lyrifera Forbes (Spatangoida). Mar. Biol., 139, 951–957.

Hovel, K. A., M. S. Fonseca, D. L. Myer, W. J. Kenworthy and P. E. Whitfield. 2002. Effects

of seagrass landscape structure, structural complexity and hydrodynamic regime on

macrofaunal densities in North Carolina seagrass beds. Mar. Ecol. Prog. Ser., 243, 11-24.

Hughes, T. G. 1973. Deposit feeding in Abra tenuis (Bivalvia: tellinacea). J. Zool., 171(4),

499-512.


Mar. Biol. Ecol., 20, 137-15.

Hughes, R. G., D. Lloyd, L. Ball and D. Emson. 2000. The effects of the polychaete Nereis

diversicolor on the distribution and transplanting success of Zostera noltii. Helgol. Mar.

Res., 54, 129-136.

Huston, M. 1979. A general hypothesis of species diversity. Am. Nat., 113(1), 81-101. Mar.

Jones, C. G., J. H. Lawton and M. Shachak. 1994. Organisms as ecosystem engineers. Oikos,

69: 373-386.

Jones, C. G., J. H. Lawton and M. Shachak. 1997. Positive and negative effects of organisms

as physical ecosystem engineers. Ecology, 78(7), 1946-1957.

Josefson, A. B., T. L. Forbes and R. Rosenberg. 2002. Fate of phytodetritus in marine

sediments: functional importance of macrofaunal community. Mar. Ecol. Prog. Ser., 230,

71-85.

Karlson, K., E. Bonsdorff and R. Rosenberg. 2007. The impact of benthic macrofauna for

nutrient fluxes from Baltic Sea sediments. Ambio, 36(2-3), 161-167.


[219]

Koch, E.W., Ackerman, J.D., Verduin, J.J. and van Keulen, M., 2006. Fluid Dynamics in

seagrass ecology-from molecules to ecosystems. In: Larkum A.W.D., Orth R.J. and Duarte

C.M. (Editors), Seagrasses: Biology, Ecology and Conservation, pp. 193-225.





Mar. Ecol. Prog. Ser., 446, 285-302.

Laruelle, G. G., P. Regnier, O. Ragueneau, M. Kempa, B. Moriceau, S. Ni Longphuirt, A.

Leynaert, G. Thouzeau and L. Chauvaud. 2009. Benthic-pelagic coupling and the seasonal

silica cycle in the Bay of Brest (France) : new insights from a coupled physical-biological

model. Mar. Ecol. Prog. Ser., 385, 15-32.

Lazzari, P., C. Solidoro, V. Ibello, S. Salon, A. Teruzzi, K. Béranger, S. Colella and A. Crise.

2012. Seasonal and inter-annual variability of plankton chlorophyll and primary production

in the Mediterranean Sea: A modeling approach. Biogeosciences, 9, 217-233.

Lecroart P., Schmidt S., Jouanneau J. -M., Weber O. 2005. Be-7 and Th-234 as tracers of

sediment mixing on seasonal time scale at the water-sediment interface of the Thau

Lagoon. Radioprotection, Suppl. 1, vol. 40, S661-S667.

Lecroart, P., S. Schmidt and J. M. Jouanneau. 2007a. Numerical estimation of the error of the


Lecroart P., Schmidt S., Anschutz P., Jouanneau J. -M. 2007b. Modeling sensitivity of

biodiffusion coefficient to seasonal bioturbation, J. Mar. Res., 65, 417-440.



Leorri, E., R. E. Martin and B. P. Horton. 2009. Field experiments on bioturbation in salt

marshes (Bombay hook national wildlife refuge, Smyrna, DE, USA): implication for sea-

level studies. J. Quarternary Sci., 24, 139-149.

Levinton, J. S. 1991. Variable feeding behavior in three species of Macoma (Bivalvia:

Tellinacea) as a response to water flow and sediment transport. Mar. Biol., 110, 375-383.

Levinton, J. S. 2001. Marine biology: Function, biodiversity, ecology (second edition).

Oxford university press, New York. 515 pp.

Levinton, J. and B. Kelaher. 2004. Opposing organizing forces of deposit-feeding marine


Lindahl, O., A. Belgrano, L. Davidsson and B. Hernroth. 1998. Primary production, climatic

oscillations, and physic-chemical processes: The Gullmar fjord time-series data set (1985-

1996). ICES J. Mar. Sci., 55, 723-729.




[220]

Lohrer, A. M., S. F. Thrush, L. Hunt, N. Hancock and C. Lundquist. 2005. Rapid reworking

of subtidal sediments by burrowing spatangoid urchins. J. Exp. Mar. Biol. Ecol. 321, 155–

169.

Lotze, H. K., H. S. Lenihan, B. J. Bourque, R. H. Bradbury, R. G. Cooke, M. C. Kay, S. M.

Kidwell, M. X. Kirby, C. H. Peterson and J. B. C. Jackson. 2006. Depletion, degradation

and recovery potential of estuarine and coastal seas. Science, 312, 1806-1809.

Mahaut, M.L. and G. Graf. 1987. A luminophore tracer technique for bioturbation studies.

Oceanol. Acta., 10, 323-328.



Ecol. Prog. Ser., 319, 135-153.










Mar. Biol., 151, 737-749.

Maire, O., P. Lecroart, F. J. R. Meysman, R. Rosenberg, J. C. Duchene and A. Grémare.


Aquat. Biol., 2, 219-238.





Meadows P. S., A. Meadows and J. M. H. Murray. 2012. Biological modifiers of marine

benthic seascapes: Their role as ecosystem engineers. Geomorphology, 157-158, 31-48.

Meysman, F. J. R., B. P. Boudeau and J. J. Middelburg. 2003. Relations between local,






Acta, 72, 3460-3478.


[221]



239-254.



Michaud, E., R. C. Aller and G. Stora. 2010. Sedimentary organic matter distributions,

burrowing activity, and biochemical cycling: Natural patterns and experimental artifacts.

Est. Coast. Shelf. Sci., 90, 21-34.

Murphy, E. A. K. and K. M. Dorgan. 2011. Burrow extension with a proboscis: mechanics of

burrowing by the glycerid Hemipodus Simplex. J. Exp. Biol., 2014, 1017-1027.

Officer, C. B., T. J. Smayda and R. Mann. 1982. Benthic filer feeding: A natural

eutrophication control. Mar. Ecol. Prog. Ser., 9, 203-210.

Olabarria, C., M. Incera, J. Garrido and F. Rossi. 2010. The effect of wrack composition and

diversity on macrofaunal assemblages in intertidal sediments. J. Exp. Mar. Biol. Ecol., 396,

18-26.

Orth, R. J. 1977. The importance of sediment stability in seagrass communities. In: Coull, B.

C. (ed). Ecology of marine benthos. Univ. South Carolina Press, Columbia, 281-300.

Orth R. J., T. J. B. Carruthers, W. C. Dennison, C. M. Duarte, J. W. Fourqurean, K. L. Heck

jr., A. R. Hughes, G. A. Kendrick, W. J. Kenworthy, S. Olyarnik, F. T. Short, M. Waycott

and S. L. Williams. 2006. A global crisis for seagrass ecosystems. BioScience, 56(12),

987-996.

Oyenakan, J.A. 1988. Population dynamics and secondary production in Melinna palmata

(Polychaeta: Ampharetidae). Mar. Biol., 98, 247-251.



229-331.

Persson, A. and R. Rosenberg. 2003. Impact of grazing and bioturbation of marine benthic

deposit feeders on dinoflagellate cysts. Harmful Algae, 2, 43-50.

Philippart, C. J. M. 1994. Interactions between Arenicola marina and Zostera noltii on a tidal

flat in the Wadden Sea. Mar. Ecol. Prog. Ser., 111, 251-257.




101.

Pitkänen, H., J. Lehtoranta and A. Räike. 2001. Internal nutrient fluxes counteract decreases

in external load: The case of the estuarial eastern Gulf of Finland, Baltic Sea. Ambio, 30(4-

5), 195-201.


[222]

Plus, M., S. Dalloyau, G. Trut, I. Auby, X. de Montaudouin, E. Emery, C. Noël and C. Viala.

2010. Long-term evolution (1988-2008) of Zostera spp. Meadows in Arcachon Bay (Bay

of Biscay). Est. Coast. Shelf Sci., 87(2), 357-366.

Quintina C. O., M. Tang and E. Kristensen. 2007. Simultaneous study of particle reworking,

irrigation transport and reaction rates in sediment bioturbated by the polychaetes

Heteromastus and Marenzellaria. J. Exp. Mar. Biol. Ecol., 352, 392-406.

Rasmussen, E. 1973. Systematics and ecology of the Isefjord marine fauna (Denmark).

Ophelia, 11, 1-495.

Reed, D. C., K. Huang, B. P. Boudreau and F. J. R. Meysman. 2006. Steady-state tracer

dynamics in a lattice-automaton model of bioturbation. Geochim. Cosmochim. Acta, 70,

5855-5867.

Reise K. 2002. Sediment mediated species interactions in coastal waters. J. Sea. Res., 48, 127-

141.

Rhoads, D. C. 1963. Rates of sediment reworking by Yoldia limatula in Buzzards Bay,

Massachusetts, and Long Island Sound. J. Sediment Petrol, 33, 723–727.

Rhoads, D. C. 1967. Biogenic reworking of intertidal and subtidal sediments in Barnstable

Harbor and Buzzards Bay, Massachusetts. J. Geol., 75, 461−476.


Biol. Ann. Rev., 12, 263-300.





Rhoads, D. C. and J. D. Germano. 1982. Characterization of Organism-Sediment remations

using Sediment Profile Imaging: An efficient method of remote ecological monitoring of

the seafloor (RemotsTM

System). Mar. ecol. Prog. Ser., 8, 115-128.

Rice, D. L. 1986. Early diagenesis in bioadvective sediments: relationships between the

diagenesis of beryllium-7, sediment reworking rates, and the abundance of conveyorbelt

deposit-feeders. J. Mar. Res., 44, 149-184.

Rice, D. L., T. S. Bianchi and E. H. Roper. 1986. Experimental studies of sediment reworking

and growth of scoloplos spp. (Orbiniidae: Polychaeta). Mar. Ecol. Prog. Ser., 30, 9-19.

Richter, R., 1952. Fluidal-texture in Sediment-Gesteinen und ober Sedifluktion überhaupt.

Notizbl. Hess. L.-Amt. Bodenforsch. 3, 67-81.

Riisgård, H. U. and P. Kamermans. 2001. Switching between deposit and suspension feeding

in coastal zoobenthos. In: Ecological comparisons of sedimentary shores, Reise K. (Ed).

Ecological studies 151, Springer-Verlag, Heidelberg. pp. 73-101.

Robbins, J. A. 1986. A model for particle-selective transport of tracers in sediments with

conveyor-belt deposit feeders. J. Geophys. Res. 91, 8542-8558.


[223]

Roberge, J. –M. and P. Angelstam. 2004. Usefulness of the umbrella species concept as a

conservation tool. Conserv. Biol., 18(1),76-85.

Rosenberg, R. 1993. Suspension feeding in Abra alba (Mollusca). Sarsia, 78, 119-121.

Rosenberg R. 2001. Marine benthic faunal successional stages and related sedimentary

activity. Sci. Mar., 65 (Suppl. 2), 107-119.

Rossi, F. and A. J. Underwood. 2002. Small-scale disturbance and increased nutrients as

influences on intertidal macrobenthic assemblages: experimental burial of wrack in

different intertidal environments. Mar. Ecol. Prog. Ser., 241, 29-39.



18, 1347-1363.

Salvo, F. 2010. Approche compare des populations naturelles et cultivées d’huître japonaise

Crassostrea gigas dans une lagune macro-tidale (Bassin d’Aracchon) : cycle biologique,

relations trophiques et effets sur le benthos. PhD Thesis. University Bordeaux 1, 510 pp.

Schiffers, K., L. R. Teal, J. M. J. Travis and M. Solan. An open source simulation model for

soil and sediment bioturbation. Plos one, 6(12), e280228.

Schmidt, S., J. –M. Jouanneau, O. Weber, P. Lecroart, O. Radakovitch, F. Glibert and D.

Jézéquel. 2007. Sedimentary processes in the Thau Lagoon (france): From seasonal to

century time scales. Est. Coast. Shelf Sci., 72, 534-542.

Schückel, U., M. Beck and I. Kröncke. 2012. Spatial variability in structural and functional

aspects of macrofauna communities and their environmental parameters in the Jade Bay

(Waden sea Lower Saxony, southern North Sea). Helgol. Mar. Res., published online.

Sgro, L., M. Mistri and J. Widdows. 2006. Impact of infaunal Manila clam, Ruditapes

philippinarum, on sediment stability. Hydrobiologia, 550, 175-182.

Shull, D. H. 2001. Transition-Matrix Model of Bioturbation and Radionuclide Diagenesis.

Limnol. Oceanogr., 46, 905-916.

Shull, D. H. and M. Yasuda .2001. Size-selective downward particle transport by cirratulid

polychaetes. J. Mar. Res., 59(3), 453-473.

Siebert, T. and G. M. Branch. 2005. Interactions between Zostera capensis, Callianassa

Kraussi and Upogebia africana: deductions from field surveys in Langebaan lagoon, South

Africa. Afr. J. Mar. Sci., 27(2), 345-356.

Siebert, T. and G. M. Branch. 2007. Influences of biological interactions on community

structure within seagrass beds and sandprawn-dominated sandflats. J. Exp. Mar. Biol.

Ecol., 340, 11-24.

Soetaert K, P. M. J. Herman, J. J. Middelburg, C. Heip, H. S. De Stigter, T. C. E. Van

Weering, E. Epping and W. Helder. 1996. Modelling 210Pb-derived mixing activity in

ocean margin sediments: diffusive versus non local mixing. J. Mar. Res., 54, 1207-1227.


[224]

Solan, M., B. J. Cardinale, A. L. Downing, K. A. M. Engelhardt, J. L. Ruesink and D. S.

Srivastava. 2004a. Extinction and ecosystem function in the marine benthos. Science, 306,

1177-1180.


and R. Rosenberg. 2004b. In Situ Quantification of Bioturbation Using Time-Lapse



Solan, M. and B. D. Wigham. 2005. Biogenic particle reworking and bacterial−invertebrate

interactions in marine sediments. In: Kristensen, E., R. R. Haese, J. E. Kostka (eds)

Interactions between macro- and microorganisms in marine sediments. American

Geophysical Union, Washington, DC, 105−124.

Stead, R. A. and R. J. Thompson. 2006. The influence of an intermittent food supply on the

feeding behaviour of Yoldia hyperborea (Bivalvia: Nuculanidae). J. Exp. Mar. Biol. Ecol.,

332, 37-48.

Suchanek, T. H., P. L. Colin, G. M. McMurty, C. S. Suchanek. 1986. Bioturbation and

redistribution of sediment radionuclides in Enewetak Atoll lagoon by callianassid shrimp:

biological aspects. Bull. Mar. Sci., 38, 144–154.

Summerson, H. C. and C. H. Peterson. 1984. Role of predation in organizing benthic

communities of a temperate-zone seagrass bed. Mar. Ecol. Prog. Ser., 15, 63-77.

Sun, M., R. C. Aller and C. Lee. 1991. Early diagenesis of chlorophyll-a in Long Island

Sound sediments: a measure of carbon flux and particle reworking. J. Mar. Res., 49, 379–

401.

Suykerbuyk, W., T. J. Bouma, T. van der Heide, C. Faust, L. L. Govers, W. B. J. T. Giesen,

D. J. de Jong and M. M. van Katwijk. 2012. Suppressing antagonistic bioengineering

feedbacks doubles restoration success. Ecol. Appl., 22(4), 1224-1231.

Taghon, G. L. and R. R. Greene. 1992. Utilization of deposited and suspended particulate

matter by benthic “interface” feeders. Limnol. Oceanogr., 37(7), 1370-1391.

Taghon, G. L., A. R. M. Nowell and P. A. Jumars. 1980. Induction of suspension feeding in

spionid polychaetes by high particulate fluxes. Science, 210, 562-564.

Teal, L. R., M. T. Bulling, E. R. Parker and M. Solan. 2008. Global patterns of bioturbation

intensity and mixed depth of marine soft sediments. Aquat. Biol., 2, 207-218.

Teal, L. R., E. R. Parker and M. Solan. 2012. Simultaneous quantification of in-situ infaunal

activity and pore-water metal concentrations: establishment of benthic ecosystem process-

function relations. Biogeosciences Discuss., 9, 8541-8570.

Tebble, N. 1966. British bivalve seashells. A handbook for identification. Trustees of the

British Museum (Natural History), London, 212 pp.

Tenore, K. R. and E. J. Chesney.1985. The effects of interactions of rate of food supply and

population density on the bioenergetics of the opportunistic polychaete, Capitella capitata

(type 1). Limnol. Oceanogr., 30(6), 1188-1195.


[225]

Thrush S. F. and P. K. Dayton. 2002. Disturbance to marine habitats by trawling and

dredging: Implications for marine biodiversity. Annu. Rev. Ecol. Syst., 33, 449-473.

Townsend, E. C. and M. S. Fonseca. 1998. Bioturbation as a potential mechanism influencing

spatial heterogeneity of North Carolina seagrass beds. Mar. Ecol. Prog. Ser., 169, 123-132.



Venturini, N., A. M. S. Pires-Vanin, M. Salhi, M. Bessonart and P. Muniz. 2011. Polychaete

response to fresh food supply at organically enriched coastal sites: Repercussion on

bioturbation potential and trophic structure. J. Marine Syst., 88, 526-541.

Volkenborn, N., S. I. C. Hedtkamp, J. E. E. van Beusekom and K. Reise. 2007. Effects of

bioturbation and biorrigation by lugworms (Arenicola marina) on physical and chemical

sediment properties and implications for intertidal habitat succession. Est. Coast. Shelf Sci.

74, 331-343.

Volkenborn, N., L. Polerecky, D. S. Wethey, T. H. DeWitt and S. A. Woodin. 2012.

Hydraulic activities by gost shrimp Neotrypaea californiensis induce oxic-anoxic

oscillations in sediment. Mar. Ecol. Prog. Ser., 455, 141-156.

Warwick, R.M. and K. R. Clarke. 1993. Increased variability as a symptom of stress in marine


Waycott, M., C. M. Duarte, T. J. B. Carruthers, R. J. Orth, W. C. Dennison, S. Olyarnik, A.

Calladine, J. W. Fourqurean, K. L. J. Heck, A. R. Hughes, G.A. Kendrick, W. J.

Kenworthy, F. T. Short and S. L. Williams. 2009. Accelerating loss of seagrasses across

the globe threaten coastal ecosystems. PNAS, 106(30), 12377-12381.

Welsh, D. T., M. Bartoli, D. Nizzoli, G. Castaldelli, S. A. Riou and P. Viaroli. 2000.

Denitrification, nitrogen fixation, community primary productivity and inorganic-N and

oxygen fluxes in an intertidal Zostera noltii meadow. Mar. Ecol. Prog. Ser., 208, 65-77.

Wesenbeeck van, B. K., J. van de Koppel, P. M. J. Herman, J. P. Bakker and T. J. Bouma.

2007. Biomechanical warfare in ecology; negative interactions between species by habitat

modification. Oikos, 116, 742-750.

Wheatcroft, R. A. 1991. Conservative tracer study of horizontal sediment mixing rates in a

bathyal basin, California borderland. J. Mar. Res. 49, 565–588.



Wheatcroft, R. A. 2006. Time-series measurements of macrobenthos abundance and sediment

bioturbation intensity on a flood-dominated shelf. Prog. Oceanogr., 71, 88-122.



Res., 48, 177-201.


[226]

Wheatcroft, R. A., I. Olmez and F. X. Pink. 1994. Particle bioturbation in Massachusetts Bay:

preliminary results using a new deliberate tracer technique. J. Mar. Res. 52, 1129–1150.

Widdows, J., A. Blauw, C. H. R. Heip, P. M. J. Herman, C. H. Lucas, J. J. Middelburg, S.

Schmidt, M. D. Brinsley, F. Twisk and H. Verbeek. 2004. Role of physical and biological

processes in sediment dynamics of a tidal flat in Westerschele Estuary, SW Netherland.


Wikander, P. B. 1980. Biometry and behaviour in Abra nitida (Müller) and A. longicallus

(Scacchi) (Bivalvia, Tellinacea). Sarsia, 65, 255-268.

Wikander, P. B. 1981. Quantitative aspects of deposit feeding in Abra nitida (Mueller) and

Abra longicallus (Scacchi) (Bivalvia, Tellinacea). Sarsia 66: 35-48.

Wilkie, L., M. T. O’Hare, I. Davidson, B. Dudley and D. M. Paterson. 2012. Particle trapping

and retention by Zostera noltii: A flume and field study. Aquat. Bot., 102, 15-22.

Wilson, W. H. 1991. Competition and predation in marine soft-sediment communities. Annu.

Rev. Ecol. Syst., 21, 221-241.

Worm, B., E. B. Barbier, N. Beaumont, J. E. Duffy, C. Folke, B. S. Halpern, J. B. C. Jackson,

H. K. Lotze, F. Micheli, S. R. Palumbi, E. Sala, K. A. Selkoe, J. J. Stachowicz and R.

Watson. 2006. Impacts of biodiversity loss on ocean ecosystem services. Science, 314,

787-790.

N° d’ordre : 3125 -...

Documents

Transcript of N° d’ordre : 3125 -...