07.08.2013 Views

lien - Ecobio - Université de Rennes 1

lien - Ecobio - Université de Rennes 1

lien - Ecobio - Université de Rennes 1

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

N° Ordre : 3392<br />

THÈSE<br />

présentée<br />

DEVANT L’UNIVERSITÉ DE RENNES 1<br />

pour obtenir<br />

le gra<strong>de</strong> <strong>de</strong> : DOCTEUR DE L’UNIVERSITÉ <strong>de</strong> RENNES 1<br />

Mention : BIOLOGIE<br />

par<br />

Ju<strong>lien</strong> Cucherousset<br />

UMR 6553 CNRS ECOBIO<br />

Equipe Biologie <strong>de</strong>s Populations et <strong>de</strong> la Conservation<br />

Campus <strong>de</strong> Beaulieu<br />

Ecole Doctorale Vie - Agro - Santé<br />

U.F.R. Sciences <strong>de</strong> la Vie et <strong>de</strong> l’Environnement<br />

Rôle fonctionnel <strong>de</strong>s milieux temporairement inondés pour l’ichtyofaune<br />

dans un écosystème sous contraintes anthropiques :<br />

approches communautaire, populationnelle et individuelle<br />

Soutenue le 14 décembre 2006 <strong>de</strong>vant la commission d’Examen<br />

COMPOSITION DU JURY<br />

Baglinière, J.-L. Directeur <strong>de</strong> Recherche, INRA, <strong>Rennes</strong> Examinateur<br />

Copp, G.H. Directeur <strong>de</strong> Recherche, CEFAS, Lowestoft Rapporteur<br />

Eybert, M.C. Chargée <strong>de</strong> Recherche, CNRS, <strong>Université</strong> <strong>de</strong> <strong>Rennes</strong> 1 Directrice<br />

Feunteun, E. Professeur, MNHN, Dinard Examinateur<br />

Marmonier, P. Professeur, <strong>Université</strong> <strong>de</strong> <strong>Rennes</strong> 1 Examinateur<br />

Pont, D. Directeur <strong>de</strong> Recherche, CEMAGREF, Aix-en-Provence Rapporteur


They are so many kinds of fishes,<br />

occupying such diverse habitats, evolving<br />

in such complex ways, and with such<br />

complicated ecological traits, that a total<br />

synthesis of un<strong>de</strong>rstanding fishes will<br />

always elu<strong>de</strong> ichthyologists and ecologists.<br />

William J. Matthews (1998)<br />

in Patterns of Freshwater Fish Ecology


REMERCIEMENTS<br />

Je tiens à remercier les membres du jury, qui m’ont fait l’honneur d’évaluer ce travail<br />

<strong>de</strong> thèse ainsi que pour l’ensemble <strong>de</strong>s commentaires avisés et <strong>de</strong>s discussions constructives<br />

qu’ils ont apportés : Jean-Luc Baglinière, Gordon H. Copp, Eric Feunteun, Pierre Marmonier<br />

et Didier Pont ainsi que Marie-Christine Eybert qui a encadré ce travail durant ces trois<br />

années.<br />

Je tiens aussi à remercier les différents organismes financeurs <strong>de</strong> cette thèse, à savoir<br />

le FEDER, la DIREN Pays <strong>de</strong> la Loire, la Région Pays <strong>de</strong> la Loire, l’Agence <strong>de</strong> l’Eau Loire -<br />

Bretagne et le Parc naturel régional <strong>de</strong> Brière (PNRB) ainsi que le PnrB pour son soutien<br />

logistique lors <strong>de</strong> la conduite <strong>de</strong>s opérations <strong>de</strong> terrain.<br />

La Commission Syndicale <strong>de</strong> Gran<strong>de</strong> Brière Mottière, le Syndicat Mixte<br />

d’Aménagement Hydraulique du Brivet, l’AAPMMA <strong>de</strong> PontChâteau et l’Association <strong>de</strong>s<br />

Pêcheurs <strong>de</strong> Gran<strong>de</strong> Brière Mottière ont apporté leurs autorisations et participations.<br />

Les membres <strong>de</strong> mon comité <strong>de</strong> thèse, et notamment Jean-Marc Roussel, Didier<br />

Montfort et Pierre-Yves Le Bail, ont su à plusieurs reprises vali<strong>de</strong>r mais aussi et surtout<br />

orienter mon travail et ma réflexion, qu’ils soient chaleureusement remerciés ici.<br />

Ce travail <strong>de</strong> thèse a nécessité un lourd investissement humain sur le terrain. Je tiens à<br />

remercier ici les très nombreuses personnes, et notamment celles du PnrB <strong>de</strong> l’Institut<br />

National <strong>de</strong> la Recherche Agronomique <strong>de</strong> <strong>Rennes</strong> et <strong>de</strong> l’<strong>Université</strong> <strong>de</strong> <strong>Rennes</strong> 1, pour leur<br />

ai<strong>de</strong> technique, logistique et scientifique.<br />

Pendant plusieurs mois, Eric Le Mitouard, Vincent Thoby et Matthieu Buoro ont été<br />

présents à mes côtés et sans eux ce travail n’aurait pu être possible. Il n’aurait pas été possible<br />

sans mes collègues <strong>de</strong> laboratoire et interlocuteurs nationaux et internationaux avec qui les<br />

discussions ont notamment permis d’enrichir ce travail.<br />

Mais surtout, cette thèse n’aurait jamais pu ni naître ni être sans Jean-Marc Paillisson,<br />

Alexandre Carpentier et Jean-Patrice Damien. Merci les gars, vraiment merci pour tout.


Préambule<br />

SOMMAIRE<br />

Chapitre 1 Introduction générale 1<br />

Chapitre 2 Matériel et métho<strong>de</strong>s général 15<br />

Partie I Utilisation fonctionnelle <strong>de</strong>s milieux temporairement inondés<br />

par la communauté piscicole 37<br />

Chapitre 3 Etat <strong>de</strong>s lieux <strong>de</strong> la communauté piscicole 39<br />

Chapitre 4 Approche synchronique <strong>de</strong> l’utilisation <strong>de</strong>s milieux temporairement<br />

inondés 51<br />

Chapitre 5 Approche diachronique <strong>de</strong> l’utilisation <strong>de</strong>s milieux temporairement<br />

inondés 71<br />

Chapitre 6 Rôle <strong>de</strong> la tolérance physiologique dans la variabilité inter-spécifique<br />

<strong>de</strong>s processus d’émigration 87<br />

Chapitre 7 Rôle <strong>de</strong>s facteurs environnementaux dans la variabilité inter- et intraspécifique<br />

<strong>de</strong>s processus d’immigration et d’émigration 103<br />

Partie II Réponses fonctionnelles aux contraintes anthropiques <strong>de</strong><br />

trois espèces d’intérêt majeur <strong>de</strong> la communauté piscicole 127<br />

Chapitre 8 Utilisation et sélection <strong>de</strong> l’habitat par le poisson chat (Ameiurus melas)<br />

129<br />

Chapitre 9 Effets <strong>de</strong> la <strong>de</strong>struction systématique <strong>de</strong>s individus capturés par la<br />

pêcherie sur la régulation du poisson chat 145<br />

Chapitre 10 Réponse <strong>de</strong> l’anguille européenne (Anguilla anguilla) à la mise en place<br />

<strong>de</strong> réserves <strong>de</strong> pêche 155<br />

Chapitre 11 Estimation <strong>de</strong>s taux <strong>de</strong> survie et i<strong>de</strong>ntification <strong>de</strong>s sources <strong>de</strong> mortalité<br />

<strong>de</strong>s brochets (Esox lucius) issus <strong>de</strong> programmes <strong>de</strong> soutien <strong>de</strong><br />

population 177<br />

Chapitre 12 Evolution du régime alimentaire et utilisation <strong>de</strong>s milieux<br />

temporairement inondés par les juvéniles <strong>de</strong> brochet 193<br />

Chapitre 13 Discussion générale et perspectives 215<br />

Chapitre 14 Bibliographie générale 245<br />

Coordonnées <strong>de</strong>s co-auteurs 266<br />

Résumé 268


PREAMBULE<br />

Créé en 1970, le Parc naturel régional <strong>de</strong> Brière a pour vocation prioritaire <strong>de</strong> protéger et <strong>de</strong><br />

valoriser les marais briérons dans lesquels s’épanouissait une diversité biologique<br />

exceptionnelle qui a justifié leur inscription à la convention internationale <strong>de</strong> Ramsar en 1995.<br />

Basé sur un concept précurseur en matière d’aménagement et <strong>de</strong> protection <strong>de</strong><br />

l’environnement, le Parc naturel régional <strong>de</strong> Brière veut être un territoire d’expérimentation,<br />

<strong>de</strong> connaissance et <strong>de</strong> transferts d’informations favorables au développement durable <strong>de</strong>s<br />

zones humi<strong>de</strong>s briéronnes. Cette volonté s’exprime particulièrement pour le patrimoine<br />

naturel et culturel représenté par la faune piscicole.<br />

Les travaux scientifiques ponctuels menés en trente ans se sont révélés insuffisants<br />

pour avoir une connaissance générale du peuplement, <strong>de</strong> sa répartition, <strong>de</strong> l’influence <strong>de</strong> la<br />

gestion <strong>de</strong>s milieux, <strong>de</strong> son exploitation halieutique mais également pour proposer aux acteurs<br />

locaux <strong>de</strong>s mesures <strong>de</strong> gestion raisonnées <strong>de</strong> la faune piscicole. De plus, les interrogations<br />

récurrentes sur la raréfaction d’espèces à forte valeur halieutique et patrimoniale comme le<br />

Brochet (Esox Lucius) ou l’Anguille (Anguilla anguilla), ainsi que la perception d’un<br />

bouleversement récent et profond <strong>de</strong>s milieux aquatiques ont motivé la mise en œuvre en<br />

2004 d’une importante investigation scientifique.<br />

Ce travail d’expertise a été confié à l’<strong>Université</strong> <strong>de</strong> <strong>Rennes</strong> 1 UMR CNRS <strong>Ecobio</strong><br />

6553, par le Parc naturel régional <strong>de</strong> Brière qui, par ailleurs, pilote le comité local <strong>de</strong> suivi <strong>de</strong><br />

l’étu<strong>de</strong>. La présente thèse s’inscrit pleinement dans cette étu<strong>de</strong> soutenue financièrement par<br />

l’Europe, le Ministère <strong>de</strong> l’Environnement et du Développement Durable, l’Agence <strong>de</strong> l’Eau<br />

Loire-Bretagne et la Région <strong>de</strong>s Pays <strong>de</strong> la Loire.<br />

Fruit d’une fédération <strong>de</strong>s volontés locales, cette étu<strong>de</strong> est le préalable indispensable à<br />

une gestion intégrée favorable à l’ichtyofaune. Enfin, elle contribue plus largement à la<br />

connaissance <strong>de</strong> <strong>de</strong>ux espèces menacées, l’Anguille et le Brochet, ainsi qu’à celle du<br />

fonctionnement <strong>de</strong> la communauté piscicole dans les marais et les milieux temporairement<br />

inondés.<br />

Le Directeur du Parc naturel régional<br />

<strong>de</strong> Brière<br />

Bernard GUIHENEUF.


Introduction générale<br />

Vue aérienne <strong>de</strong>s marais <strong>de</strong> Brière inondés en 1977<br />

Chapitre 1


Chapitre 1<br />

1.1. Approche conceptuelle du fonctionnement <strong>de</strong>s milieux temporairement<br />

inondés au sein d’écosystèmes aquatiques<br />

Les écosystèmes aquatiques, et plus particulièrement les zones humi<strong>de</strong>s, sont parmi les<br />

milieux présentant les plus fortes productivités et diversités floristiques et faunistiques. Ils<br />

possè<strong>de</strong>nt par conséquent d’importantes valeurs économiques et sociales (Tockner &<br />

Standford 2002, Williams 2006) mais sont aussi, paradoxalement, les milieux les plus<br />

menacés et dégradés <strong>de</strong> la planète (Dynesius & Nilsson 1994, Oberdorff et al. 2002).<br />

Plusieurs définitions ont été apportées pour le terme « zone humi<strong>de</strong> » et nous retiendrons ici<br />

<strong>de</strong>ux d’entre elles issues <strong>de</strong> concept écologique :<br />

(1) « les zones humi<strong>de</strong>s sont <strong>de</strong>s étendues <strong>de</strong> marais, <strong>de</strong> fagnes, <strong>de</strong> tourbières ou<br />

d’eaux naturelles ou artificielles, permanentes ou temporaires, où l’eau est stagnante ou<br />

courante, douce, saumâtre ou salée, y compris <strong>de</strong>s étendues d’eau marine dont la profon<strong>de</strong>ur à<br />

marée basse n’excè<strong>de</strong> pas six mètres » provenant <strong>de</strong> la Convention Ramsar (1971),<br />

(2) « les zones humi<strong>de</strong>s sont <strong>de</strong>s terres <strong>de</strong> transition entre les systèmes terrestres et<br />

aquatiques, la nappe étant habituellement soit à la surface, soit à proximité ou alors le terrain<br />

étant recouvert d’une couche d’eau peu profon<strong>de</strong> » provenant <strong>de</strong> Cowardin et al. (1979).<br />

Ces définitions prennent donc en compte un large éventail <strong>de</strong> milieux telles les zones<br />

estuariennes, les côtes ouvertes, les plaines inondables, les marais d’eau douce, les lacs, les<br />

tourbières ou encore les forêts marécageuses (Fustec & Lefeuvre 2000, Williams 2006). Ceci<br />

implique également que les zones humi<strong>de</strong>s sont et/ou comportent <strong>de</strong>s milieux<br />

temporairement inondés (MTI) qui sont le plus souvent situés en périphérie et qui possè<strong>de</strong>nt<br />

<strong>de</strong>s fonctions biologiques et écologiques indissociables <strong>de</strong>s autres compartiments <strong>de</strong><br />

l’écosystème, ces fonctions s’exprimant le long d’un continuum spatial et temporel (Euliss et<br />

al. 2004, Williams 2006).<br />

Tous ces milieux subissent les conséquences passées et présentes d’une pression<br />

anthropique croissante due au développement démographique <strong>de</strong> la population humaine et à<br />

une <strong>de</strong>man<strong>de</strong> accrue en énergie hydroélectrique, en eau, en production agricole et forestière<br />

ou en ressources halieutiques (Harding et al. 1998, Fustec & Lefeuvre 2000, Tockner &<br />

Stanford 2002). Les effets majeurs sont une augmentation du niveau <strong>de</strong> pollution et<br />

d’eutrophisation (Williams 2006), une fragmentation, voire une disparition <strong>de</strong> l’habitat (Byers<br />

2002, Nilsson et al. 2005), ainsi qu’une modification <strong>de</strong>s régimes thermiques et hydrologiques<br />

naturels (Poff et al. 1997, Bunn & Arthington 2002, Humphries & Baldwin 2003). Les effets<br />

2


Introduction générale<br />

<strong>de</strong> cette pression anthropique s'exercent bien évi<strong>de</strong>mment aussi sur la faune et flore <strong>de</strong> ces<br />

milieux, avec notamment une diminution <strong>de</strong> la biodiversité (Fustec & Lefeuvre 2000)<br />

associée à une raréfaction <strong>de</strong> la ressource halieutique (Williams 2006), <strong>de</strong>s introductions<br />

d’espèces invasives et la facilitation <strong>de</strong> leur installation (Dukes & Mooney 1999). Par<br />

exemple, l’altération <strong>de</strong> la fréquence, <strong>de</strong> l’amplitu<strong>de</strong> et <strong>de</strong> la durée <strong>de</strong>s phénomènes <strong>de</strong> crues<br />

et d’assèchements, ainsi que <strong>de</strong> leur niveau <strong>de</strong> prédictibilité, ont profondément modifié les<br />

caractéristiques naturelles <strong>de</strong> ces milieux et leur fonctionnement et altéré la diversité <strong>de</strong>s<br />

organismes vivant dans ces milieux (e.g. Bunn & Arthington 2002, Lytle & Poff 2004).<br />

Les MTI sont très largement distribués sur notre planète et rencontrés sur l’ensemble<br />

<strong>de</strong>s continents et dans tous les types <strong>de</strong> climats. En terme <strong>de</strong> diversité, les MTI comprennent<br />

un large éventail d’habitats à la fois dulçaquicoles et marins (voir revue complète dans<br />

Williams 2006). L’élément clé utilisé pour définir les MTI est la nature cyclique <strong>de</strong>s<br />

assèchements qu’ils subissent, qui conduira à classer les MTI en différentes catégories en<br />

fonction du régime d’inondation, c’est-à-dire <strong>de</strong> la prédictibilité et <strong>de</strong> la durée <strong>de</strong> l’inondation<br />

(Williams 2006, Figure 1.1).<br />

Régime d’inondation<br />

Jan.<br />

10 ans<br />

10 ans<br />

10 ans<br />

Jan.<br />

Ephémère<br />

Episodique<br />

Intermittent<br />

Saisonnier<br />

2 ans<br />

Semi permanent<br />

2 ans<br />

Prédictibilité et durée <strong>de</strong> l’inondation<br />

Se remplit uniquement après <strong>de</strong>s précipitations imprévisibles et par<br />

écoulement. La zone inondée s’assèche le jour suivant l’inondation et<br />

ne supporte que rarement <strong>de</strong>s organismes aquatiques macroscopiques.<br />

Asséché 9 années sur 10, avec <strong>de</strong>s inondations rares et très irrégulières<br />

qui peuvent durer jusqu’à quelques mois.<br />

Alternance <strong>de</strong> pério<strong>de</strong>s humi<strong>de</strong>s et sèches, mais à une fréquence plus<br />

faible que dans les milieux inondés saisonnièrement. L’inondation peut<br />

durer pendant plusieurs mois ou années.<br />

Alternance <strong>de</strong> pério<strong>de</strong>s humi<strong>de</strong>s et sèches chaque année, en relation<br />

avec le cycle <strong>de</strong>s saisons. L’inondation est habituellement annuelle et<br />

prédictible. Elle dure plusieurs mois, et assez longuement pour<br />

permettre à <strong>de</strong>s organismes (plantes et animaux) <strong>de</strong> réaliser la phase<br />

aquatique <strong>de</strong> leur cycle <strong>de</strong> vie.<br />

Inondation prédictible, accompagnée <strong>de</strong> variations <strong>de</strong> niveau d’eau.<br />

L’apport annuel d’eau est supérieur à la perte annuelle (pas<br />

d’assèchement) 9 années sur 10. La majorité <strong>de</strong>s organismes vivant<br />

dans ces milieux ne supportent pas la déshydratation.<br />

Figure 1.1. Schéma simplifié <strong>de</strong> la classification <strong>de</strong>s milieux temporairement inondés issu <strong>de</strong><br />

Williams (2006).<br />

3


Chapitre 1<br />

Lors <strong>de</strong>s <strong>de</strong>rnières déca<strong>de</strong>s, la compréhension du fonctionnement <strong>de</strong>s MTI au sein <strong>de</strong>s<br />

écosystèmes aquatiques, a augmenté <strong>de</strong> manière significative grâce à l’application <strong>de</strong><br />

nouveaux concepts théoriques (Ward et al. 2001). Alors que le « River Continuum Concept »<br />

développé par Vannote et al. (1980) s’intéressait à la dimension longitudinale et continue <strong>de</strong>s<br />

changements physiques et biologiques le long <strong>de</strong>s cours d’eau, le « Flood Pulse Concept» <strong>de</strong><br />

Junk et al. (1989) intégrait, quant à lui, la dimension latérale, se focalisant sur les échanges<br />

d’eau, <strong>de</strong> nutriments et d’organismes entre cours principaux et milieux inondés. Ainsi ce<br />

nouveau concept, qui est <strong>de</strong>venu un paradigme en écologie <strong>de</strong>s écosystèmes aquatiques<br />

(Tockner et al. 2000), s’intéresse principalement aux effets <strong>de</strong> la prédictibilité, <strong>de</strong> la durée et <strong>de</strong><br />

l’expansion spatiale <strong>de</strong>s inondations pour comprendre le rôle fonctionnel <strong>de</strong>s MTI (Junk et al.<br />

1989, Ward et al. 2001). Tockner et al. (2000) y ont par la suite ajouté la notion <strong>de</strong> « Flow<br />

Pulse » en s’intéressant tout particulièrement aux phases d’expansion et <strong>de</strong> contraction <strong>de</strong>s<br />

eaux se produisant avant l’atteinte du niveau <strong>de</strong> débor<strong>de</strong>ment.<br />

Plus récemment, Junk & Wantzen (2003) ont présenté une vision synthétique, fonction<br />

<strong>de</strong> la prédictibilité <strong>de</strong>s milieux et une mise à jour du flood pulse concept, qui permet <strong>de</strong> mieux<br />

comprendre le fonctionnement <strong>de</strong> ces milieux et <strong>de</strong> leurs communautés animales et végétales<br />

(Figure 1.2). Ainsi, dans les systèmes stabilisés tels que ceux <strong>de</strong>s lacs régulés ou <strong>de</strong>s cours<br />

d’eau sans zones inondables, le niveau d’eau 0 est présent la majeure partie du temps,<br />

permettant ainsi la mise en place bien définie <strong>de</strong> communautés littorales, pélagiques et<br />

benthiques qui sont parfaitement adaptées aux conditions environnementales et optimisées par<br />

l’utilisation <strong>de</strong>s ressources trophiques locales (Figure 1.2, gauche). Des inondations<br />

occasionnelles et particulièrement brutales sont <strong>de</strong>s évènements catastrophiques qui ne<br />

permettent pas l’utilisation <strong>de</strong>s ressources disponibles pour la faune et la flore dans les zones<br />

épi littorales. Mais dans les systèmes soumis à <strong>de</strong>s phénomènes d’inondations rythmés et<br />

prévisibles, les organismes sont adaptés aux changements périodiques <strong>de</strong> niveau d’eau et<br />

bénéficient ainsi <strong>de</strong> ressources d’origines variées (Figure 1.2, droite). La flore et la faune se<br />

distribuent le long du gradient d’inondation et les communautés littorales, pélagiques et<br />

benthiques maintiennent toujours leurs places respectives au sein <strong>de</strong> l’écosystème quel que<br />

soit le niveau d’eau. Cependant, <strong>de</strong>s niveaux d’eau extrêmement bas poussent les espèces<br />

benthiques et pélagiques soit à migrer dans <strong>de</strong>s zones plus profon<strong>de</strong>s soit à estiver dans ces<br />

milieux. Les espèces terrestres ont quant à elles développé <strong>de</strong>s mécanismes et stratégies pour<br />

survivre aux inondations.<br />

4


+1<br />

0<br />

L 0<br />

O 0<br />

P 0<br />

O +1<br />

O 0<br />

P -1<br />

P 0<br />

L -1<br />

L 0<br />

L +1<br />

Introduction générale<br />

Système stabilisé Systèmes à rythmicité<br />

Niveau d’eau prédictible Niveau d’eau<br />

L = Littoral<br />

O = Pélagique<br />

P = Benthique<br />

Utilisation limitée <strong>de</strong>s<br />

ressources épilittorales lors<br />

<strong>de</strong>s crues catastrophiques<br />

non prévisibles<br />

Figure 1.2. Schéma du fonctionnement <strong>de</strong>s écosystèmes hydrologiquement stabilisés (gauche) et<br />

soumis à <strong>de</strong>s phénomènes d’inondations rythmés et prévisibles (droite) à un niveau d’eau normal<br />

(0), extrêmement haut (+1) et extrêmement bas (-1) pour les communautés littorales (L),<br />

pélagiques (O) et benthiques (P).Adapté d’après Junk & Wantzen (2003).<br />

Euliss et al. (2004) ont <strong>de</strong>puis développé le « Wetland Continuum Concept » qui a<br />

pour objectif <strong>de</strong> prendre en compte, à une échelle multidimensionnelle, l’influence <strong>de</strong>s<br />

variations climatiques et hydrologiques dans le fonctionnement <strong>de</strong>s zones humi<strong>de</strong>s. Ce<br />

modèle permet <strong>de</strong> prédire les réponses <strong>de</strong>s communautés <strong>de</strong> plantes, d’invertébrés,<br />

d’amphibiens et d’oiseaux à ces changements en fonction du niveau <strong>de</strong> relation hydrologique<br />

du milieu avec les précipitations et avec les eaux souterraines. Parce que ces milieux ont un<br />

fort intérêt patrimonial et écologique et sont soumis aux effets <strong>de</strong>s activités anthropiques, la<br />

prise en compte <strong>de</strong> ces concepts peut fournir in fine un outil fiable pour la mise en place <strong>de</strong><br />

mesures <strong>de</strong> gestion et/ou <strong>de</strong> restauration <strong>de</strong> leur diversité faunistique et floristique et pour<br />

l’évaluation <strong>de</strong> l’efficacité <strong>de</strong> ces mesures (Middleton 2000, Middleton 2002, Tockner et al.<br />

2000, Euliss et al. 2004, Juradja et al. 2004, Angeler & Alvarez-Cobelas 2005).<br />

Les concepts précé<strong>de</strong>mment évoqués ont aussi permis <strong>de</strong> mettre en évi<strong>de</strong>nce les<br />

différentes adaptations que les organismes, utilisant ou vivant dans ces milieux, ont<br />

développé sur les plans morphologiques, anatomiques, physiologiques, <strong>de</strong> traits d’histoire <strong>de</strong><br />

vie et/ou comportementales, pour répondre aux conditions environnementales<br />

particulièrement contraignantes que représentent les phénomènes <strong>de</strong> crues et d’assèchements<br />

périodiques (Junk et al. 1989, Lake 2003, Lytle & Poff 2004). Les adaptations mises en place<br />

au niveau <strong>de</strong>s organismes confèrent ainsi aux communautés vivant dans ces milieux <strong>de</strong>s<br />

caractéristiques structurelles et fonctionnelles particulières (Junk et al. 1989, Wellborn et al.<br />

5<br />

+1<br />

0<br />

-1


Chapitre 1<br />

1996, Casanova & Brock 2000, Zimmer et al. 2000, Trexler et al. 2005). Ces caractéristiques<br />

sont variables à <strong>de</strong>s échelles locales et régionales en réponse à la forte variabilité <strong>de</strong> patrons<br />

temporels d’inondations et d’assèchements (Wellborn et al. 1996, Lytle & Poff 2004).<br />

Les adaptations mises en place par les espèces pour exploiter les MTI bordant <strong>de</strong>s<br />

milieux permanents, peuvent être interprétées comme la résultante d’un tra<strong>de</strong>-off entre <strong>de</strong>s<br />

coûts et <strong>de</strong>s bénéfices liés à l’utilisation <strong>de</strong>s MTI (Lytle & Poff 2004). Ce tra<strong>de</strong>-off est un <strong>de</strong>s<br />

mécanismes contrôlant la coexistence <strong>de</strong>s espèces et la structuration <strong>de</strong>s communautés<br />

(Kneithel & Chase 2004). Pour <strong>de</strong> nombreuses espèces utilisant les MTI, l’une <strong>de</strong>s phases<br />

critiques qui <strong>de</strong>vra être surmontée est celle <strong>de</strong> l’assèchement (Humphries & Baldwin 2003).<br />

Ainsi, les coûts pourront provenir par exemple <strong>de</strong> la mise en place <strong>de</strong> mécanismes <strong>de</strong><br />

persistance à un assèchement ou <strong>de</strong> la nécessité <strong>de</strong> gagner une zone <strong>de</strong> refuge. A l’inverse, les<br />

bénéfices pourront provenir d’un accès précoce à <strong>de</strong>s conditions environnementales<br />

attractives et favorables telles que la présence <strong>de</strong>s ressources trophiques souvent abondantes<br />

dans <strong>de</strong>s habitats peu utilisés (e.g. Sih 1997, Brocks et al. 2003). In fine, en terme <strong>de</strong> fitness<br />

individuelle, les coûts liés à l’utilisation <strong>de</strong> ces milieux proviendront principalement du risque<br />

<strong>de</strong> mortalité accrue par prédation ou stress physiologique alors que les gains proviendront<br />

d’une augmentation <strong>de</strong> fécondité accrue, <strong>de</strong> la croissance <strong>de</strong>s juvéniles et <strong>de</strong> la survie<br />

(Wellborn et al. 1996). Cependant, le <strong>de</strong>gré <strong>de</strong> prédictibilité <strong>de</strong>s fluctuations <strong>de</strong> niveau d’eau,<br />

aura <strong>de</strong>s conséquences variables sur les communautés et les réponses évolutives <strong>de</strong>s espèces<br />

en terme d’histoire <strong>de</strong> vie (Lytle 2001). Ainsi, par exemple l’altération <strong>de</strong>s régimes<br />

d’inondation (durée, fréquence et prédictibilité) par les activités anthropiques peut<br />

potentiellement conduire à une modification du tra<strong>de</strong>-off où les coûts d’une adaptation<br />

pourront l’emporter sur les bénéfices <strong>de</strong> l’organisme et éventuellement avoir d’importantes<br />

répercussions sur le fonctionnement <strong>de</strong>s populations et <strong>de</strong>s communautés exploitant ces<br />

milieux (Lytle & Poff 2004).<br />

L’étu<strong>de</strong> <strong>de</strong>s MTI, si elle prend en compte l'approche bio évolutive <strong>de</strong> la réponse <strong>de</strong>s<br />

organismes aux conditions environnementales fortement fluctuantes, doit aussi intégrer les<br />

concepts développés en écologie du paysage. En effet, l’écologie du paysage et sa démarche<br />

spatialisée ont permis d'apporter <strong>de</strong> nouveaux concepts pour l'étu<strong>de</strong> <strong>de</strong>s écosystèmes<br />

aquatiques et d’attribuer <strong>de</strong>s rôles fonctionnels spécifiques aux milieux inondables (e.g.<br />

Amoros & Bornette 2002, Townsend et al. 2003). Ces principaux concepts, synthétisés dans<br />

Wiens (2002), prédisent que:<br />

(1) les patchs d’habitat qui forment la mosaïque paysagère possè<strong>de</strong>nt <strong>de</strong>s qualités<br />

fonctionnelles différentes,<br />

6


Introduction générale<br />

(2) les frontières entre patchs jouent un rôle dans les flux <strong>de</strong> matière et les échanges<br />

d’organismes qui sont influencés par leur configuration et leur perméabilité,<br />

(3) les patchs et leur(s) rôle(s) fonctionnel(s) dépen<strong>de</strong>nt du contexte structurel dans<br />

lequel ils se situent,<br />

(4) la connectivité entre les milieux influence les mouvements d’individus, <strong>de</strong> matière,<br />

<strong>de</strong> nutriments et d’énergie ainsi que la transmission <strong>de</strong>s perturbations,<br />

(5) les composantes individuelles et leur variabilité ont un rôle important et,<br />

(6) l’échelle spatiale et temporelle <strong>de</strong> perception et d’analyse est un facteur clé pour la<br />

compréhension <strong>de</strong>s mécanismes étudiés.<br />

Ainsi, le recours à une approche paysagère <strong>de</strong> la structuration <strong>de</strong> l’habitat s’est révélé<br />

nécessaire pour interpréter les relations espèces - habitats en décrivant spatialement les<br />

patrons d’inondation, en quantifiant la disponibilité en habitat en fonction <strong>de</strong> ces patrons et en<br />

i<strong>de</strong>ntifiant les caractéristiques hydrologiques importantes pour le maintien <strong>de</strong> la diversité<br />

biologique et <strong>de</strong> l’habitat (Johnson & Gage 1997, Tockner et al. 2000, Wiens 2002). Enfin,<br />

ces milieux évoluant spatialement au cours <strong>de</strong> l’inondation, leur étu<strong>de</strong> bénéficie <strong>de</strong> l’apport<br />

d’une double approche temporelle: (1) synchronique, par comparaison <strong>de</strong>s processus<br />

écologiques se déroulant en différents lieux à un instant donné <strong>de</strong> l’inondation et (2)<br />

diachronique, par analyse <strong>de</strong>s processus écologiques se déroulant au cours d’une phase<br />

d’inondation dans un lieu donné.<br />

L’écologie <strong>de</strong>s communautés apporte une approche complémentaire qui s’intéresse à<br />

différents niveaux biologiques d’organisation en mettant en place <strong>de</strong> manière conjointe<br />

<strong>de</strong>ux types d’approches distinctes (Putman 1994 in Williams 2006). La première, dite<br />

« réductionniste », étudie les individus appartenant à une seule espèce, et les facteurs<br />

contrôlant le fonctionnement <strong>de</strong> cette espèce, avant d’établir une comparaison avec les autres<br />

espèces évoluant dans le même écosystème. Au final, les relations entre les différentes<br />

populations sont recherchées afin <strong>de</strong> reconstruire une image globale du fonctionnement <strong>de</strong> la<br />

communauté. La secon<strong>de</strong>, dite « holistique », a quant à elle pour objectif d’étudier d’emblée<br />

la communauté dans son ensemble (Williams 2006). Ces <strong>de</strong>ux approches souffrent chacune <strong>de</strong><br />

certains désavantages. Par exemple, l’approche réductionniste a tendance à ne se focaliser que<br />

sur les espèces prédominantes et sur les facteurs principaux alors que l’approche holistique<br />

entraîne souvent une trop gran<strong>de</strong> simplification <strong>de</strong>s mécanismes écologiques, ne permettant<br />

qu’une mesure qualitative ou semi-quantitative <strong>de</strong>s processus. C’est pourquoi ces <strong>de</strong>ux<br />

7


Chapitre 1<br />

approches doivent être menées conjointement dans l’étu<strong>de</strong> du fonctionnement <strong>de</strong> nombreux<br />

systèmes écologiques tels que les MTI (Williams 2006).<br />

1.2. Mécanismes fonctionnels <strong>de</strong> structuration <strong>de</strong> l’ichtyofaune en relation<br />

avec les milieux temporairement inondés<br />

Parmi l’ensemble <strong>de</strong>s organismes vivants utilisant les MTI, les poissons présentent<br />

certaines particularités. En effet, à part à <strong>de</strong> très rares exceptions et contrairement à d’autres<br />

vertébrés comme les oiseaux ou les amphibiens, les poissons se voient dans l’impossibilité<br />

d’utiliser certains types <strong>de</strong> MTI, à savoir ceux totalement isolés, alimentés exclusivement par<br />

les précipitations et/ou les eaux souterraines (e.g. Euliss et al. 2004). De plus, les poissons ne<br />

possè<strong>de</strong>nt pas, sous nos latitu<strong>de</strong>s, <strong>de</strong> mécanismes <strong>de</strong> rési<strong>lien</strong>ce ou <strong>de</strong> persistance à un<br />

assèchement complet (Sayer 2005, Williams 2006). Ils doivent coloniser les MTI lors <strong>de</strong>s<br />

pério<strong>de</strong>s d’inondation à partir <strong>de</strong>s d’habitats aquatiques permanents (adjacents aux MTI) qui<br />

leur servent <strong>de</strong> refuge lors <strong>de</strong>s pério<strong>de</strong>s d’assèchement (Magoulik & Kobza 2003). Ces<br />

processus ont été rendus possibles grâce à un certain nombre d’adaptations comportementales,<br />

morphologiques et <strong>de</strong> traits d’histoires <strong>de</strong> vie (Welcomme 1979, Matthews & Marsh-Mattews<br />

2003, Lytle & Poff 2004, DeAngelis et al. 2005, Winemiller 2005). Pour les poissons, comme<br />

pour les autres organismes strictement aquatiques, l’utilisation <strong>de</strong>s MTI est donc facultative<br />

dans le sens où ils doivent les coloniser volontairement. Les mouvements associés à cette<br />

colonisation peuvent d’ailleurs être assimiler à <strong>de</strong>s migrations (Lucas & Baras 2001).<br />

Comme pour les autres organismes, l’utilisation <strong>de</strong> ces milieux par les poissons tient à<br />

leur capacité <strong>de</strong> pouvoir optimiser les bénéfices par rapport aux coûts. En effet, ces milieux<br />

peuvent être considérés d’un côté comme particulièrement contraignants par rapport aux<br />

milieux permanents pour l’ichtyofaune puisque l’assèchement est la résultante finale d’une<br />

diminution <strong>de</strong> la surface et du volume d’eau qui entraîne une augmentation <strong>de</strong>s extrêmes <strong>de</strong>s<br />

paramètres physico-chimiques, tels que la température, le pH ou la concentration en<br />

dioxygène (Matthews 1998, Magoulik & Kobza 2003, Hanson et al. 2005). De l’autre côté,<br />

ces milieux sont très productifs et les bénéfices pour les poissons proviennent généralement<br />

d’un accès à <strong>de</strong>s ressources trophiques importantes et disponibles, favorisant certains sta<strong>de</strong>s<br />

<strong>de</strong> vie et donc une augmentation intrinsèque <strong>de</strong> la fécondité (Magoulik & Kobza 2003, Lytle<br />

& Poff 2004). Ainsi, certaines espèces optimiseront leur reproduction en tirant <strong>de</strong>s bénéfices<br />

<strong>de</strong>s caractéristiques <strong>de</strong> substrat <strong>de</strong> pontes, <strong>de</strong> hauteur d’eau et d’abondance en nourriture plus<br />

8


Introduction générale<br />

favorables au développement <strong>de</strong>s oeufs et <strong>de</strong>s sta<strong>de</strong>s larvaires que celles existant dans les<br />

milieux permanents. En contrepartie, les géniteurs mais également les futurs alevins,<br />

s’exposent à <strong>de</strong>s conditions environnementales potentiellement défavorables et à <strong>de</strong>s risques<br />

accrus <strong>de</strong> mortalité par prédation, principalement durant la phase d’exondation (e.g. Kushlan<br />

1976, Souchon 1983, Wright & Shoesmith 1988, Capone & Kushlan 1991). Basé sur le Flood<br />

Pulse Concept, King et al. (2003) ont ainsi développé un modèle d’optimisation <strong>de</strong>s<br />

conditions environnementales lors <strong>de</strong>s pério<strong>de</strong>s d’inondation pour maximiser le recrutement<br />

<strong>de</strong>s poissons, préconisant un couplage entre les régimes thermique et d’inondation,<br />

l’inondation <strong>de</strong>vant être prédictible et maintenue pendant une durée <strong>de</strong> l’ordre du mois sur <strong>de</strong>s<br />

surfaces <strong>de</strong> gran<strong>de</strong>s tailles.<br />

Les adaptions mises en place par les poissons ainsi que les caractéristiques <strong>de</strong>s<br />

évènements d’inondation (fréquence, durée et d’amplitu<strong>de</strong>) entraînent <strong>de</strong>s réponses propres <strong>de</strong><br />

chaque espèce conduisant à <strong>de</strong>s structurations particulières <strong>de</strong>s communautés piscicoles<br />

(Welcomme 1979, Wellborn et al. 1996, Trexler et al. 2005, Matthews & Marsh-Matthews<br />

2002). Ainsi, le fonctionnement d'une communauté est influencé par la réponse <strong>de</strong> chacune<br />

<strong>de</strong>s espèces aux conditions environnementales biotiques et abiotiques auxquelles elles sont<br />

soumises et aux caractéristiques intrinsèques à la communauté (Hanson et al. 2005, Mills et<br />

al. 2004, Jackson et al. 2001).<br />

Dans les écosystèmes aquatiques dulçaquicoles, une gran<strong>de</strong> diversité <strong>de</strong> questions<br />

biologiques et écologiques a été abordée et cela dans toute une gamme très large <strong>de</strong> MTI<br />

(Tableau 1.1). Ainsi, le rôle <strong>de</strong>s paramètres environnementaux dans les relations<br />

poissons/MTI a été mise en évi<strong>de</strong>nce par <strong>de</strong> nombreux auteurs, afin <strong>de</strong> comprendre le<br />

fonctionnement et la composition <strong>de</strong>s communautés piscicoles exploitant les MTI (Tableau<br />

1.1). En parallèle, les processus dynamiques <strong>de</strong> colonisation et d’émigration <strong>de</strong> ces milieux<br />

par l’ichtyofaune, qui traduisent le patron temporel d’utilisation <strong>de</strong>s MTI, ont fait l’objet <strong>de</strong><br />

quelques étu<strong>de</strong>s (e.g. Hohausova 2000, Hohausova et al. 2003, Poizat & Crivelli 1997). La<br />

synthèse <strong>de</strong>s étu<strong>de</strong>s portant sur les communautés piscicoles en relation avec les MTI (Tableau<br />

1.1) montre une fragmentation <strong>de</strong>s résultats à la fois dans le temps et dans l’espace. Ce<br />

travail <strong>de</strong> thèse va donc définir une série <strong>de</strong> problématiques permettant <strong>de</strong> fournir <strong>de</strong>s<br />

éléments à une réflexion globale sur l’ensemble <strong>de</strong>s mécanismes d’utilisation <strong>de</strong>s MTI par<br />

l’ichtyofaune.<br />

9


Tableau 1.1. Récapitulatif <strong>de</strong>s principaux paramètres <strong>de</strong> la communauté piscicole, du site d’étu<strong>de</strong> et <strong>de</strong>s paramètres environnementaux utilisés lors <strong>de</strong> l’étu<strong>de</strong><br />

<strong>de</strong>s relations entre ichtyofaune et milieux temporairement inondés.<br />

Descripteurs <strong>de</strong> la<br />

communauté<br />

Composition<br />

[a, b, f, j, l, n, p]<br />

Abondance absolue<br />

[b, l, n, o]<br />

Guil<strong>de</strong> écologique [i, n]<br />

Paramètres du site Nature du site<br />

Positionnement longitudinal<br />

amont-aval [a, g, j]<br />

Hydrologie (fréquence, durée<br />

et/ou amplitu<strong>de</strong>) [b, d, e, m]<br />

Positionnement latéral, connectivité<br />

et isolement [d, f, g, h, n]<br />

Groupe taxonomique [i, n] Précipitation [c]<br />

Recrutement [e]<br />

Structure d’âge [e, l, p, q]<br />

Synchronisation <strong>de</strong>s régimes<br />

<strong>de</strong> précipitation et inondation [e]<br />

Surface, taille et volume<br />

[d, e, f, g, h, i, k, l]<br />

Régime alimentaire [j, m] Statut trophique [i]<br />

Tolérance physiologique [l] Hétérogénéité du paysage [k]<br />

Zones lotiques<br />

[a, e, l, p]<br />

Zones lentiques<br />

[c, d, m, n]<br />

Lagunes marécageuses<br />

[j, o, b, q]<br />

Mouilles <strong>de</strong> rivières<br />

[g, k]<br />

Lacs et mares isolés<br />

[h, i, f]<br />

Paramètres environnementaux<br />

Abtiotiques Biotiques<br />

Profon<strong>de</strong>ur<br />

[a, f, i, j, k, t, n, o]<br />

Qualité chimique <strong>de</strong> l’eau<br />

[a, c, d, e, f, h, k, l, p]<br />

Température<br />

[e, f, h, k, n, o, p]<br />

Substrat [i, n]<br />

Végétation aquatiques<br />

et terrestres [d, i, k, l, m, n]<br />

Prédation [d, l]<br />

Ressources trophiques<br />

[c, p]<br />

Pâturage, élevage et gestion<br />

végétation [d, n]<br />

Références [nombre d’espèces piscicoles dans la communauté étudiée] : a Ostrand & Wil<strong>de</strong> (2002) [16] ; b Trexler et al. (2005) [29] ; c Hanson et al. (2005)<br />

[1] ; d Baber et al. (2002) [15]; e King et al. (2003) [11]; f Snodgrass et al. (1996) [12]; g Taylor (1997) [11]; h Ol<strong>de</strong>n et al. (2001) [24] ; i Hatzenbeler et al.<br />

(2000) [33]; j Hoeinghaus et al. (2003) [60]; k Magoulick (2000) [19]; l Ostrand & Wil<strong>de</strong> (2001) [5], m Chick et al. (2004) [17] n Carpentier et al. (2004) [10], o<br />

Poizat & Crivelli (1997) [22], p Hohausova et al. (2003) [14], q DeAngelis et al. (2005).


1.3. Objectifs <strong>de</strong> l’étu<strong>de</strong><br />

Introduction générale<br />

Ce travail <strong>de</strong> thèse a donc pour modèle biologique l’ichtyofaune, dont les particularités liées à<br />

l’utilisation <strong>de</strong>s MTI ont été évoquées précé<strong>de</strong>mment. De plus, les connaissances écologiques<br />

concernant le modèle poisson et sa relation avec les MTI sont considérées comme<br />

relativement rares et fragmentées en comparaison avec celles acquises sur <strong>de</strong> nombreux autres<br />

modèles végétaux et animaux (Baber et al. 2002, Williams 2006).<br />

La présente thèse se déroulera dans le contexte général et théorique précé<strong>de</strong>mment<br />

évoqué, c’est-à-dire :<br />

- une prise en compte du rôle <strong>de</strong>s activités anthropiques sur le fonctionnement <strong>de</strong>s<br />

écosystèmes aquatiques,<br />

- une démarche d’intégration <strong>de</strong> différentes échelles spatiales et temporelles dans<br />

l’appréciation <strong>de</strong> certaines relations poissons/MTI (concepts et outils d’écologie du paysage)<br />

- une analyse se basant conjointement sur <strong>de</strong>s approches à différents niveaux<br />

d’organisation : communauté, population et individu.<br />

- une réflexion <strong>de</strong> type bio évolutive pour comprendre les adaptations mises en place<br />

par les espèces piscicoles pour optimiser l’utilisation spatiale et temporelle <strong>de</strong>s MTI,<br />

- une prise en compte d’une diversité <strong>de</strong> facteurs environnementaux à la fois<br />

abiotiques et biotiques influençant cette utilisation ainsi que les processus dynamiques<br />

d’échanges associés.<br />

Ainsi, après une présentation générale du site d’étu<strong>de</strong>, une synthèse <strong>de</strong>s métho<strong>de</strong>s<br />

utilisées sera réalisée (Chapitre 2). En effet, la diversité <strong>de</strong>s questions abordées ici, ainsi que<br />

les différentes échelles spatiales et temporelles utilisées, a nécessité l’utilisation <strong>de</strong><br />

nombreuses métho<strong>de</strong>s d’échantillonnage et <strong>de</strong> suivi (pêche électrique, piégeage passive et<br />

enquête halieutique) ainsi que <strong>de</strong> certaines techniques <strong>de</strong> pointes (Système d’Information<br />

Géographique, marquage individuel par « Passive Integrated Transpon<strong>de</strong>r tag », télémétrie et<br />

analyse <strong>de</strong>s isotopes stables).<br />

L’objectif général <strong>de</strong> cette thèse est <strong>de</strong> comprendre le fonctionnement d’une<br />

communauté piscicole ainsi que <strong>de</strong> certaines <strong>de</strong>s espèces la composant dans un écosystème<br />

spatialement très variable, à savoir les marais <strong>de</strong> Brière situés au nord <strong>de</strong> l’estuaire <strong>de</strong> la Loire<br />

(Loire-Atlantique, France). Dans cet écosystème, très largement dépendant <strong>de</strong>s activités<br />

anthropiques, nous analyserons les relations entre la communauté piscicole et les MTI puis<br />

11


Chapitre 1<br />

concentrerons cette approche sur l’analyse du fonctionnement <strong>de</strong> trois espèces d’intérêt<br />

majeur, en relation étroite avec les MTI. Plus précisément, la présente thèse sera articulée<br />

autour <strong>de</strong>s <strong>de</strong>ux questions suivantes :<br />

(A) quelles sont les espèces <strong>de</strong> la communauté piscicole qui utilisent les MTI et<br />

comment ces processus se déroulent-ils dans le temps et l’espace ? (Partie I)<br />

Afin <strong>de</strong> présenter la communauté piscicole qui sera étudiée au cours <strong>de</strong> cette thèse, un<br />

bilan <strong>de</strong> sa composition actuelle en comparaison avec celle observée il y a plus <strong>de</strong> 30 ans sera<br />

réalisé (Chapitre 3).<br />

Ensuite, une analyse comparera la communauté présente dans les milieux permanents<br />

à celle utilisant les MTI et examinera plus particulièrement sa distribution spatiale au sein <strong>de</strong><br />

ces milieux. Ainsi, une première approche (synchronique) comparera ces processus pour le<br />

<strong>de</strong>ux grands types <strong>de</strong> MTI présents dans le site d’étu<strong>de</strong> à large échelle à un moment précis du<br />

cycle hydrologique (Chapitre 4) alors qu’une secon<strong>de</strong> approche (diachronique) analysera ces<br />

même processus à fine échelle au cours d’un cycle complet d’inondation (Chapitre 5).<br />

Enfin, les processus dynamiques d’échanges entre milieux permanents et différents<br />

types <strong>de</strong> MTI, seront analysés. Dans un premier temps, les patrons d’émigration lors <strong>de</strong> la<br />

phase d’exondation et le rôle <strong>de</strong>s caractéristiques intrinsèques <strong>de</strong>s espèces, à savoir leur<br />

tolérance physiologique, dans ce processus seront analysés (Chapitre 6). Dans un second<br />

temps, le rôle <strong>de</strong>s facteurs environnementaux dans les processus d’immigration et<br />

d’émigration sera analysé au niveau <strong>de</strong> différentes espèces et différents sta<strong>de</strong>s <strong>de</strong> vies<br />

(Chapitre 7).<br />

(B) comment certaines espèces i<strong>de</strong>ntifiées comme d’intérêt majeur au sein <strong>de</strong> la<br />

communauté, fonctionnent-elles dans cet écosystème, plus particulièrement au regard <strong>de</strong>s<br />

caractéristiques <strong>de</strong> l’habitat et <strong>de</strong>s mesures <strong>de</strong> gestion auxquelles elles sont soumises ?<br />

(Partie II).<br />

Le fonctionnement <strong>de</strong> la population <strong>de</strong> l’espèce invasive dominant actuellement la<br />

communauté sera d’abord étudié. Dans une première approche, l’utilisation et la sélection <strong>de</strong><br />

l’habitat (approche à différentes échelles spatiales) <strong>de</strong> cette espèce seront analysées. Cette<br />

analyse sera conduite en prenant en compte l’évolution <strong>de</strong>s MTI au cours du siècle <strong>de</strong>rnier en<br />

liaison avec les modifications d’usage <strong>de</strong> ces marais par l’homme (Chapitre 8). Dans une<br />

secon<strong>de</strong> approche, les effets <strong>de</strong> la <strong>de</strong>struction systématique par les pêcheurs <strong>de</strong> cette espèce,<br />

12


Introduction générale<br />

seule mesure <strong>de</strong> gestion actuellement utilisée, sur sa population (niveau d’abondance et<br />

structure d’âge) seront étudiés (Chapitre 9).<br />

Ensuite, l’analyse se portera sur la seule espèce catadrome <strong>de</strong> la communauté qui<br />

utilise le site d’étu<strong>de</strong> comme zone <strong>de</strong> croissance. Dans le site d’étu<strong>de</strong>, cette espèce, vulnérable<br />

et exploitée, a bénéficié d’un programme <strong>de</strong> gestion novateur, à savoir la restauration<br />

d’habitat dont <strong>de</strong>s MTI accompagnée d’une régulation <strong>de</strong>s pratiques halieutiques. Nous<br />

analyserons donc le fonctionnement <strong>de</strong> cette population, notamment en terme <strong>de</strong> production<br />

<strong>de</strong> géniteurs, vis-à-vis <strong>de</strong> ce programme <strong>de</strong> gestion (Chapitre 10).<br />

Enfin, le fonctionnement d’une espèce vulnérable, exploitée avec un cycle <strong>de</strong> vie<br />

fortement inféodé aux MTI et dont la population fait l’objet <strong>de</strong> programme <strong>de</strong> soutien sera<br />

analysé. Cette approche sera réalisée grâce à un suivi individuel <strong>de</strong>s juvéniles issus <strong>de</strong> ces<br />

programmes <strong>de</strong> soutien. Dans un premier temps, la survie <strong>de</strong> ces individus dans les MTI sera<br />

quantifiée et les différentes causes <strong>de</strong> mortalité i<strong>de</strong>ntifiées (Chapitre 11). Dans un second<br />

temps, les sources <strong>de</strong> variabilité lors <strong>de</strong>s changements <strong>de</strong> MTI et <strong>de</strong> l’évolution ontogénique<br />

du régime alimentaire seront analysées en relation avec les caractéristiques <strong>de</strong> chaque<br />

individu ayant survécu à ces programmes <strong>de</strong> soutien <strong>de</strong> population (Chapitre 12).<br />

Le chapitre 13 « Discussion Générale et Perspectives » viendra synthétiser et<br />

rassembler l’ensemble <strong>de</strong>s résultats obtenus au niveau <strong>de</strong> la communauté et <strong>de</strong>s espèces<br />

d’intérêt majeur en les mettant en relation avec les concepts théoriques d’écologie<br />

fonctionnelle, <strong>de</strong> biologie <strong>de</strong>s communautés et <strong>de</strong> la conservation évoqués en introduction,<br />

tout en dégageant <strong>de</strong>s perspectives <strong>de</strong> recherche et <strong>de</strong> gestion pour ce type d’écosystème<br />

dulçaquicole. Au final, une approche prédictive analysera les conséquences possibles <strong>de</strong>s<br />

changements globaux futurs sur le fonctionnement <strong>de</strong>s communautés piscicoles en relation<br />

avec les MTI (Chapitre 13).<br />

13


Chapitre 1<br />

Une prairie temporairement inondée et son canal adjacent au niveau <strong>de</strong> leur point <strong>de</strong><br />

connexion dans les marais <strong>de</strong> Brière<br />

14


Matériel et métho<strong>de</strong>s général<br />

Chapitre 2<br />

Echantillonnage par pêche électrique <strong>de</strong>s milieux temporairement inondés <strong>de</strong> Brière


Chapitre 2<br />

Le premier objectif <strong>de</strong> matériel et métho<strong>de</strong>s général est <strong>de</strong> présenter <strong>de</strong> manière détaillée la<br />

zone d’étu<strong>de</strong>, incluant <strong>de</strong>s aspects géographiques, hydrologiques et administratifs qui ne sont<br />

pas présentés par la suite dans les différentes sous-parties mais qui sont <strong>de</strong>s facteurs ayant<br />

conditionné la mise en place et la réalisation <strong>de</strong> ce travail <strong>de</strong> thèse. Le second objectif est <strong>de</strong><br />

présenter les différents outils méthodologiques retenus en fonction <strong>de</strong> leur pertinence pour<br />

répondre aux différents objectifs <strong>de</strong> cette thèse ainsi que <strong>de</strong> synthétiser l’ensemble <strong>de</strong>s<br />

opérations <strong>de</strong> terrain réalisées durant la pério<strong>de</strong> 2004-2006.<br />

Ainsi, le présent chapitre fournit une image globale <strong>de</strong> la zone d’étu<strong>de</strong>s et <strong>de</strong>s<br />

techniques d’échantillonnages et <strong>de</strong> suivis et calendrier <strong>de</strong>s opérations et justifie les choix<br />

méthodologiques. Cette partie ne se substitue pas à la présentation faite du « Matériel et<br />

métho<strong>de</strong>s » <strong>de</strong> chaque sous-partie où ces différents aspects ne seront plus abordés.<br />

2.1. Site d’étu<strong>de</strong><br />

2.1.1. Localisation géographique et <strong>de</strong>scription du site d’étu<strong>de</strong><br />

Les marais <strong>de</strong> Brière, parcourus par le Brivet (<strong>de</strong>rnier affluent sur la rive droite <strong>de</strong> la Loire),<br />

sont situés sur la faça<strong>de</strong> atlantique, entre les estuaires <strong>de</strong> la Vilaine et <strong>de</strong> la Loire (Figure 2.1).<br />

Ils s’intègrent dans un vaste ensemble <strong>de</strong> zones humi<strong>de</strong>s à fort intérêt patrimonial : le Golfe<br />

du Morbihan et l’estuaire <strong>de</strong> la Vilaine au Nord-ouest, les marais salants <strong>de</strong> Guéran<strong>de</strong> et du<br />

Mès à l’Ouest, l’estuaire <strong>de</strong> la Loire et le Lac <strong>de</strong> Grand-lieu au Sud (PnrB 2000).<br />

Le Brivet a une longueur <strong>de</strong> 30 km et est alimenté par un bassin versant <strong>de</strong> 80000<br />

hectares. Les marais <strong>de</strong> Brière, qui sont inscrits dans le territoire du Parc naturel régional <strong>de</strong><br />

Brière, sont constitués <strong>de</strong> 17000 ha <strong>de</strong> zones humi<strong>de</strong>s et composés <strong>de</strong> trois gran<strong>de</strong>s parties<br />

(Figure 2.1): (1) le marais <strong>de</strong> Gran<strong>de</strong> Brière Mottière, d’une superficie <strong>de</strong> 7000 ha, (2) les<br />

marais <strong>de</strong> Donges, d’une superficie <strong>de</strong> 7800 ha, et (3) un ensemble <strong>de</strong> petits marais adjacents<br />

disséminés sur le territoire qui couvrent une surface <strong>de</strong> 2200 ha (Figure 2.1).<br />

16


France<br />

La Vilaine<br />

Océan Atlantique<br />

Marais <strong>de</strong> Gran<strong>de</strong> Brière Mottière<br />

Marais <strong>de</strong> Donges<br />

Marais périphériques<br />

<strong>Rennes</strong><br />

Loire<br />

La Loire<br />

Parc naturel régional <strong>de</strong> Brière<br />

Matériel et metho<strong>de</strong>s général<br />

5 Km<br />

Le Brivet<br />

N<br />

Réseau hydrographique primaire<br />

Figure 2.1. Localisation géographique et cartographie <strong>de</strong>s marais <strong>de</strong> Brière incluant le réseau<br />

hydrographique primaire et les trois principales parties : le marais <strong>de</strong> Gran<strong>de</strong> Brière Mottière, les<br />

marais <strong>de</strong> Donges et les marais périphériques.<br />

17


Chapitre 2<br />

2.1.2. Origine, évolution et conditions climatiques<br />

Sur le plan morphologique, le bassin brièron se présente comme un grand marais d’où émerge<br />

quelques îlots et une ligne rocheuse orientée nord nord-ouest/sud sud-est. A la fin du<br />

secondaire, <strong>de</strong> nombreuses transgressions marines ont déposé une gran<strong>de</strong> quantité <strong>de</strong><br />

sédiments creusés par la suite par les cours d’eau lors <strong>de</strong>s régressions successives. Environ<br />

11000 ans avant notre ère, une <strong>de</strong>rnière transgression est venue déposer un cordon littoral qui<br />

a fermé les marais <strong>de</strong> Brière. Il y a environ 5000 ans, l’océan a rompu cette digue naturelle et<br />

envahi la dépression, submergeant et condamnant les forêts qui la recouvraient. La cuvette<br />

s’est colmatée peu à peu, et la tourbe s’est formée en surface à partir <strong>de</strong>s débris végétaux.<br />

C’est la raison pour laquelle les sols <strong>de</strong>s marais <strong>de</strong> Brière sont constitués aujourd’hui d’un<br />

humus <strong>de</strong> type hydromorphe, caractérisé par une forte proportion <strong>de</strong> matière organique peu<br />

transformée, une activité microbienne <strong>de</strong> décomposition freinée et <strong>de</strong>s conditions<br />

d’anaérobioses par saturation du milieu en eau.<br />

Durant les <strong>de</strong>rniers siècles, les variations <strong>de</strong> niveau d’eau mais surtout les activités<br />

liées à l’exploitation du marais (curage, extraction <strong>de</strong> la tourbe, exploitation du roseau,<br />

élevage) ont façonné ce milieu. Depuis ces cinquante <strong>de</strong>rnières années, le déclin <strong>de</strong> ces<br />

activités humaines a entraîné une évolution rapi<strong>de</strong> <strong>de</strong>s paysages qui s’est traduite par un<br />

envahissement <strong>de</strong> la roselière (Phragmites australis) au détriment <strong>de</strong>s prairies et <strong>de</strong>s plans<br />

d’eau (Bernard & Rolland 1990, Eybert et al. 1998). Ainsi, la surface en eau libre s’est réduite<br />

<strong>de</strong> 700 ha à 270 ha entre 1948 et 1980 et les prairies naturelles localisées actuellement à la<br />

périphérie du marais ont subi une forte récession passant d’une superficie représentant plus <strong>de</strong><br />

80 % <strong>de</strong> la surface <strong>de</strong>s marais <strong>de</strong> Brière en 1940 à environ 15% <strong>de</strong> cette surface actuellement<br />

(Bernard & Rolland 1990, Eybert et al. 1998, Figure 2.2). Aujourd’hui, le marais <strong>de</strong> Gran<strong>de</strong><br />

Brière Mottière est constitué d’une immense roselière (5500 ha) entrecoupée <strong>de</strong> prairies<br />

humi<strong>de</strong>s (1000 ha) et ponctuée d’une multitu<strong>de</strong> <strong>de</strong> plans d’eau appelés piar<strong>de</strong>s ou copis (300<br />

ha) et <strong>de</strong> canaux (130 Km <strong>de</strong> canaux principaux et secondaires). Les marais <strong>de</strong> Donges ainsi<br />

que les petits marais adjacents sont principalement constitués <strong>de</strong> prairies drainées par <strong>de</strong>s<br />

canaux (Figure 2.2). Cette importante et rapi<strong>de</strong> évolution <strong>de</strong> la physionomie générale <strong>de</strong>s<br />

marais <strong>de</strong> Brière a eu pour conséquence : une baisse <strong>de</strong> la richesse floristique et en particulier<br />

<strong>de</strong>s communautés amphibies et prairiales, une perte <strong>de</strong> diversité et d’abondance <strong>de</strong>s<br />

communautés d’amphibiens et <strong>de</strong> reptiles ainsi qu’un appauvrissement <strong>de</strong> la communauté<br />

d’oiseaux d’eau (Eybert et al. 1998). Cependant, les conséquences <strong>de</strong> ces changements sur<br />

l’ichtyofaune n’ont pas été mesurées et restent très mal connues.<br />

18


Habitat:<br />

Prairie<br />

Roselière<br />

Piar<strong>de</strong>s<br />

Matériel et metho<strong>de</strong>s général<br />

Réseau hydrographique:<br />

Primaire<br />

Secondaire<br />

Figure 2.2. Cartographie <strong>de</strong>s différents types <strong>de</strong> milieux temporairement inondés (prairie,<br />

roselière et piar<strong>de</strong>) et du réseau hydrographique <strong>de</strong>s marais <strong>de</strong> Brière.<br />

Sur la base <strong>de</strong>s relevés météorologiques, les marais <strong>de</strong> Brière se caractérisent, du fait<br />

<strong>de</strong> leur localisation géographique, par un climat océanique aux hivers doux et humi<strong>de</strong>s, et aux<br />

étés tempérés. Les amplitu<strong>de</strong>s thermiques saisonnières sont relativement faibles et les vagues<br />

<strong>de</strong> froid, <strong>de</strong> même que les températures caniculaires sont rares. Les précipitations (moyennes<br />

<strong>de</strong> 750 mm d’eau par an) ont lieu toute l’année avec un maximum à la fin <strong>de</strong> l’automne et au<br />

début <strong>de</strong> l’hiver, et un minimum en été. Lors <strong>de</strong>s 25 <strong>de</strong>rnières années, l’évolution du niveau<br />

d’eau montre <strong>de</strong> fortes variabilités interannuelles à la fois au niveau <strong>de</strong> l’amplitu<strong>de</strong> et <strong>de</strong> la<br />

durée <strong>de</strong> la phase d’inondation, mais aussi <strong>de</strong> la pério<strong>de</strong> d’étiage (Figure 2.3).<br />

Niveau d’eau (côte NGF en cm)<br />

160<br />

140<br />

120<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

-20<br />

*<br />

1980<br />

1981<br />

1982<br />

1983<br />

1984<br />

1985<br />

1986<br />

1987<br />

1988<br />

1989<br />

1990<br />

1991<br />

Année<br />

Figure 2.3. Evolution du niveau d’eau mensuel (côte Niveau Général Français, NGF en cm) dans<br />

les marais <strong>de</strong> Brière <strong>de</strong> 1980 à 2005 au niveau <strong>de</strong> la station <strong>de</strong> Fédrun (source : Parc naturel<br />

régional <strong>de</strong> Brière).<br />

19<br />

1992<br />

1993<br />

1994<br />

1995<br />

1996<br />

1997<br />

1998<br />

1999<br />

2000<br />

2001<br />

2002<br />

2003<br />

2004<br />

2005


Chapitre 2<br />

2.1.3. Caractéristiques <strong>de</strong>s milieux permanents et temporairement inondés<br />

[0]Les milieux permanents appartiennent à quatre gran<strong>de</strong>s catégories : le Brivet (19 mètres <strong>de</strong><br />

large et 1.6 mètres <strong>de</strong> profon<strong>de</strong>ur, en moyenne), les canaux drainant <strong>de</strong>s marais <strong>de</strong> Donges (12<br />

m par 0.9 m), les canaux du réseau primaire (24 m par 1.1 m) et du réseau secondaire <strong>de</strong><br />

Gran<strong>de</strong> Brière Mottière (9 m par 1.1 m). Ils se caractérisent tous par une végétation aquatique<br />

très éparse localisée le long <strong>de</strong>s berges et un substrat tourbeux. Le cycle hydrologique annuel<br />

entraîne l’inondation et l’assèchement <strong>de</strong> gran<strong>de</strong>s surfaces <strong>de</strong> MTI. Les tailles <strong>de</strong>s surfaces<br />

inondées sont unes <strong>de</strong> caractéristiques du site d’étu<strong>de</strong> et lui confèrent sa qualification <strong>de</strong><br />

milieu spatialement très fluctuant. En effet, dans les marais <strong>de</strong> Gran<strong>de</strong> Brière Mottière, 97.1%<br />

<strong>de</strong> la surface totale est composée <strong>de</strong> MTI (Figure 2.2). Le cycle hydrologique annuel résulte<br />

<strong>de</strong>s précipitations et <strong>de</strong> la gestion anthropique <strong>de</strong>s niveaux d’eau (système <strong>de</strong> vannage). Il peut<br />

être décomposé en 3 phases (Figure 2.4) : (1) l’inondation, en général, <strong>de</strong>s mois <strong>de</strong> novembre<br />

et décembre aux mois <strong>de</strong> février et mars, (2) l’exondation <strong>de</strong> février à mi-juin, et (3) une<br />

phase d’étiage <strong>de</strong> juillet à octobre (Figure 2.3). Ce cycle hydrologique et les caractéristiques<br />

topologiques très diversifiées <strong>de</strong>s différents MTI conduisent à <strong>de</strong>s patrons d’inondations et<br />

d’exondations très variables en termes d’amplitu<strong>de</strong> et <strong>de</strong>s durées (Figure 2.4). Ainsi, les<br />

prairies inon<strong>de</strong>nt pour une côte <strong>de</strong> 0.45 Niveau Général Français NGF (i.e. <strong>de</strong> mi-décembre à<br />

mi-juin pour une profon<strong>de</strong>ur maximale <strong>de</strong> l’ordre <strong>de</strong> 45 cm), les roselières inon<strong>de</strong>nt pour une<br />

côte <strong>de</strong> 0.40 NGF (i.e. <strong>de</strong> début décembre à fin juin pour une profon<strong>de</strong>ur maximale <strong>de</strong> 50 cm),<br />

alors que les piar<strong>de</strong>s (milieux semi-permanents) inon<strong>de</strong>nt en général pour une côte d’environ<br />

0 NGF (i.e. d’octobre à septembre pour une profon<strong>de</strong>ur maximale du mètre).<br />

Cote NGF (cm)<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

-20<br />

Juil.<br />

Août<br />

Niveau d’eau<br />

Seuil d’inondation<br />

Etiage<br />

Sept.<br />

Oct.<br />

Nov.<br />

Déc.<br />

Janv.<br />

Inondation maximale<br />

PRAIRIES<br />

ROSELIERES<br />

Févr.<br />

Date<br />

Figure 2.4. Représentation schématique du cycle <strong>de</strong> l’eau au cours d’une année hydrologique sur<br />

les différents types <strong>de</strong> milieux temporairement inondés <strong>de</strong>s marais <strong>de</strong> Brière. La cote NGF est une<br />

cote <strong>de</strong> référence nationale (Niveau Général Français, en cm).<br />

20<br />

PIARDES<br />

Mars<br />

Avr.<br />

Mai<br />

Juin<br />

Juil.<br />

Août<br />

Etiage<br />

Sept.<br />

Oct.<br />

Nov.


Matériel et metho<strong>de</strong>s général<br />

Les marais <strong>de</strong> Brière, qui ont été façonnés par l’homme au cours <strong>de</strong>s siècles précé<strong>de</strong>nts,<br />

présentent certaines caractéristiques particulières au niveau <strong>de</strong>s jonctions physiques entre<br />

milieux permanents et MTI. En effet, les canaux ont été creusés et sont toujours entretenus<br />

mécaniquement par les différents gestionnaires du site. Les matériaux issus du curage (i.e.<br />

tourbes délitées, vases et argiles) sont en général stockés le long <strong>de</strong>s canaux (bourrelet <strong>de</strong><br />

curage) et forment au fil du temps <strong>de</strong>s structures nettement plus élevées que les MTI. Ils sont<br />

dénommées localement « dosses » et ne sont jamais inondés. Le long <strong>de</strong> ces dosses, <strong>de</strong>s points<br />

<strong>de</strong> connections sont creusés afin <strong>de</strong> permettre les échanges hydrologiques entre les milieux<br />

permanents et les MTI (Figure 2.5).<br />

Roselière<br />

Prairie<br />

Plan d’eau<br />

Canal<br />

Bourrelet <strong>de</strong> curage<br />

Jonctions entre canal / MTI<br />

Figure 2.5. Représentation schématique <strong>de</strong> la configuration <strong>de</strong>s milieux temporairement<br />

inondés <strong>de</strong>s marais <strong>de</strong> Brière et <strong>de</strong>s jonctions entre milieux permanents et temporairement<br />

inondés.<br />

2.1.4. Statuts législatifs <strong>de</strong>s marais et acteurs <strong>de</strong> la gestion <strong>de</strong> l’eau et <strong>de</strong><br />

l’ichtyofaune<br />

L’historique <strong>de</strong>s relations d’usage et <strong>de</strong> gestions <strong>de</strong> ce milieu par l’homme est à la fois<br />

complexe et particulier et nécessite une brève présentation <strong>de</strong>s statuts législatifs <strong>de</strong>s marais<br />

ainsi que <strong>de</strong>s différents acteurs participants à cette gestion. Cette présente thèse n’aurait pu se<br />

dérouler sans l’accord et parfois la participation <strong>de</strong> ces différents acteurs, le plus souvent<br />

détenteurs <strong>de</strong> droits <strong>de</strong> pêche ainsi que du Parc naturel régional <strong>de</strong> Brière, organisme<br />

financeur <strong>de</strong> cette thèse.<br />

21


Chapitre 2<br />

Statuts <strong>de</strong>s marais<br />

Le marais <strong>de</strong> Gran<strong>de</strong> Brière Mottière est dit « indivis », c’est-à-dire qu’il n’est pas partagé<br />

matériellement et qu’il appartient en commun à plusieurs personnes, lesquelles exercent leur<br />

droit conjointement. En 1461, les lettres patentes <strong>de</strong> François II, Duc <strong>de</strong> Bretagne<br />

reconnaissaient la libre jouissance <strong>de</strong>s ressources du marais pour les habitants <strong>de</strong>s 21<br />

communes riveraines. Les lettres patentes <strong>de</strong> Louis XVI en 1784 reconnaissaient la propriété<br />

collective <strong>de</strong> la Gran<strong>de</strong> Brière Mottière et le droit accordé aux habitants <strong>de</strong>s paroisses<br />

riveraines « d’extraire la tourbe [...], <strong>de</strong> faire pacager le bétail, <strong>de</strong> récolter les foins ». Ce statut<br />

original a permis au cours <strong>de</strong>s siècles <strong>de</strong> préserver le marais du morcellement et <strong>de</strong> projets<br />

d’aménagement ou d’assèchement. Les eaux <strong>de</strong> cette propriété indivise sont classées en eaux<br />

libres <strong>de</strong> <strong>de</strong>uxième catégorie piscicole et relèvent du domaine privé. Les marais <strong>de</strong> Donges et<br />

les petits marais jouxtant le marais indivis <strong>de</strong> Gran<strong>de</strong> Brière appartiennent à <strong>de</strong>s propriétaires<br />

privés. Les eaux y sont également classées en eaux libres <strong>de</strong> <strong>de</strong>uxième catégorie piscicole et<br />

relèvent aussi du domaine privé.<br />

Acteurs <strong>de</strong> la gestion <strong>de</strong> l’eau et <strong>de</strong> la communauté piscicole<br />

La Commission Syndicale <strong>de</strong> Gran<strong>de</strong> Brière Mottière, créée par ordonnance <strong>de</strong> Louis Philippe<br />

en 1838, est le principal responsable <strong>de</strong> la gestion <strong>de</strong> l’eau et <strong>de</strong>s usages dans le marais indivis.<br />

Elle est composée <strong>de</strong> 21 représentants élus par les conseils municipaux <strong>de</strong>s 21 communes<br />

concernées. Les pêcheurs sont représentés par l’Association <strong>de</strong>s Pêcheurs <strong>de</strong> Gran<strong>de</strong> Brière<br />

(association loi 1901). La responsabilité <strong>de</strong> la gestion hydraulique dans les marais privés <strong>de</strong><br />

Donges incombe au Syndicat Mixte pour l’Aménagement Hydraulique du Bassin du Brivet.<br />

C’est un Etablissement Public <strong>de</strong> Coopération Intercommunale qui regroupe 23 communes.<br />

Deux Associations Agréées pour la Pêche et la Protection du Milieu Aquatique (AAPPMA),<br />

en échange du droit <strong>de</strong> pêche, ont l’obligation <strong>de</strong> gérer les ressources piscicoles : « La Carpe<br />

Pontchâtelaine » et « La Brème Trignacaise ». Dans les marais privés périphériques ces<br />

marais, la pêche est réservée aux propriétaires privés alors que la portion du Brivet située<br />

entre la limite transversale <strong>de</strong> la mer et la limite <strong>de</strong> salure <strong>de</strong>s eaux est intégrée au Domaine<br />

Public Maritime. Le Parc naturel régional <strong>de</strong> Brière est associé aux prises <strong>de</strong> décision <strong>de</strong> ces<br />

différents organismes à travers son appartenance à différentes commissions et la réalisation<br />

d’étu<strong>de</strong>s.<br />

22


2.2. Méthodologie <strong>de</strong> suivi <strong>de</strong> l’ichtyofaune<br />

Matériel et metho<strong>de</strong>s général<br />

2.2.1. Echantillonnage <strong>de</strong> la communauté piscicole dans les milieux permanents<br />

et les milieux temporairement inondés<br />

L’objectif <strong>de</strong> cet échantillonnage est <strong>de</strong> pouvoir obtenir <strong>de</strong>s informations fiables sur<br />

l’ichtyofaune à différentes échelles spatiales (macro- et micro- habitat), temporelles<br />

(synchroniques et diachroniques) et dans différents types <strong>de</strong> milieux à <strong>de</strong>s moments clés du<br />

cycle hydrologique. Il existe <strong>de</strong> nombreux outils pour échantillonner les poissons d’eau douce,<br />

l’un <strong>de</strong>s plus couramment utilisé étant la métho<strong>de</strong> <strong>de</strong> la pêche électrique (voir revue dans<br />

Cowx 1996). En effet, l’efficacité et la sélectivité <strong>de</strong> cette métho<strong>de</strong> varient en fonction <strong>de</strong><br />

nombreux paramètres tels que la taille <strong>de</strong>s poissons, les caractéristiques du site échantillonné<br />

(conductivité, profon<strong>de</strong>ur ou courant) ou <strong>de</strong> l’appareillage utilisé (nature du courant électrique,<br />

voltage ou taille <strong>de</strong> l’ano<strong>de</strong>). Cependant, <strong>de</strong> nombreuses adaptations méthodologiques et<br />

techniques ont été développées pour optimiser son efficacité et ainsi la qualité <strong>de</strong><br />

l’échantillonnage (voir la revue <strong>de</strong> Cowx 1996). De plus, la pêche électrique engendre une<br />

mortalité relativement faible et permet d’échantillonner à la fois les individus actifs ou passifs.<br />

L’échantillonnage a donc été conduit à l’ai<strong>de</strong> d’un appareillage <strong>de</strong> type EFKO F.E.G. 8000<br />

(400-600 V et 6-10 A) équipé d’une ano<strong>de</strong> <strong>de</strong> 30 cm <strong>de</strong> diamètre, taille offrant le meilleur<br />

compromis en terme <strong>de</strong> sélectivité sur la taille <strong>de</strong>s poissons (Copp & Penaz 1988, Persat &<br />

Copp 1989, Garner 1997, Carpentier et al. 2004)<br />

Cet échantillonnage a été réalisé en fonction <strong>de</strong> la configuration saisonnière <strong>de</strong>s<br />

milieux étudiés, <strong>de</strong>s précautions d’efficacité et <strong>de</strong> sécurité associées à la pêche électrique et<br />

<strong>de</strong>s objectifs <strong>de</strong> cette thèse. Dans un premier temps, l’objectif était d’avoir une approche<br />

globale <strong>de</strong> la communauté piscicole à l’échelle du site d’étu<strong>de</strong> dans les milieux permanents et<br />

les MTI (approche synchronique). Pour la plupart <strong>de</strong>s sites, il est impossible d’échantillonner<br />

les milieux permanents <strong>de</strong> manière efficace et non biaisée en même temps que les MTI en<br />

raison <strong>de</strong> leur profon<strong>de</strong>ur trop importante à cette pério<strong>de</strong>. Ainsi les milieux permanents,<br />

refuges estivaux <strong>de</strong> toute l’ichtyofaune, ont été échantillonnés entre juillet et août (i.e. au<br />

début d’étiage). A cette phase du cycle hydrologique, les niveaux d’eau permettent encore la<br />

navigation et l’accès aux sites et leur profon<strong>de</strong>ur permet un échantillonnage fiable. Les MTI<br />

ont, quant à eux, été échantillonnés au début du mois d’avril, c’est-à-dire avant la date <strong>de</strong><br />

reproduction <strong>de</strong> la plupart <strong>de</strong>s espèces et au milieu <strong>de</strong> la phase d’inondation, sur une pério<strong>de</strong><br />

réduite dans le temps, limitant ainsi au maximum tout effet <strong>de</strong> changement temporel <strong>de</strong> la<br />

composition <strong>de</strong> la communauté piscicole (Crain et al. 2004). Dans un second temps, l’objectif<br />

23


Chapitre 2<br />

était <strong>de</strong> suivre ces paramètres à une plus fine échelle spatiale en comparant un milieu<br />

permanent et son MTI adjacent pour réaliser <strong>de</strong>s comparaisons entre dates. Ces <strong>de</strong>ux types <strong>de</strong><br />

milieux ont donc été échantillonnés à pas <strong>de</strong> temps régulier tout au long <strong>de</strong> l’inondation<br />

jusqu’à l’assèchement du MTI (approche diachronique).<br />

Différentes approches <strong>de</strong> pêche électriques existent dont la plus fréquemment utilisée<br />

est la pêche électrique par épuisement <strong>de</strong> stock (Cowx & Lamarque 1990); cependant cette<br />

approche nécessite beaucoup <strong>de</strong> temps et <strong>de</strong> moyens humains et matériels et n’est utilisable<br />

que dans <strong>de</strong>s milieux facilement délimitables par <strong>de</strong>s filets pour empêcher la fuite <strong>de</strong>s<br />

poissons (e.g. Lobon-Cervia & Utrilla, 1993, Perrow et al. 1996, Laffaille et al. 2005), ce qui<br />

est pratiquement impossible dans les MTI très ouverts <strong>de</strong>s marais <strong>de</strong> Brière. L’approche <strong>de</strong>s<br />

Echantillonnages Ponctuels d’Abondances (EPA) par pêche électrique a été donc retenue<br />

(voir Nelva et al. 1979) pour plusieurs raisons : (1) elle est rapi<strong>de</strong>, possè<strong>de</strong> le meilleur rapport<br />

coût / efficacité et le moins <strong>de</strong> biais du fait <strong>de</strong> son efficacité pour capturer toute les espèces et<br />

la plupart <strong>de</strong>s sta<strong>de</strong>s <strong>de</strong> vie (Persat & Copp 1989), (2) elle est très bien adaptée pour<br />

l’échantillonnage <strong>de</strong> la communauté dans <strong>de</strong>s milieux peu profonds et végétalisés (Perrow et<br />

al. 1996) et (3) elle fournit <strong>de</strong>s données quantitatives et reproductibles permettant <strong>de</strong>s<br />

comparaisons spatiales et temporelles <strong>de</strong>s communautés piscicoles, à la fois inter- et intrasites<br />

(Copp 1989), dans la mesure où un minimum <strong>de</strong> 25 points ou EPA est réalisé par site<br />

échantillonné (e.g. Copp & Penaz 1988, Persat & Copp 1989 mais voir aussi Garner 1997).<br />

De plus, la technique <strong>de</strong>s lancers d’ano<strong>de</strong> ont été employés. En effet, un lancer à quelques<br />

mètres <strong>de</strong>vant les expérimentateurs permet <strong>de</strong> surprendre au maximum les poissons et ainsi <strong>de</strong><br />

minimiser les biais d’échantillonnage (Persat & Copp 1989). Ainsi, l’unité <strong>de</strong> base <strong>de</strong> ce plan<br />

d’échantillonnage correspond au point d’EPA (micro-habitat) alors que le site (milieu<br />

permanent ou MTI) correspondant à un minimum <strong>de</strong> 25 EPA et représente l’unité supérieure<br />

(macro-habitat). Dans l’approche synchronique, une soixantaine <strong>de</strong> sites (milieux permanents<br />

et <strong>de</strong> MTI), distribués sur l’ensemble <strong>de</strong> la zone d’étu<strong>de</strong> ont été échantillonnés ponctuellement<br />

en 2004 et 2005 à l’ai<strong>de</strong> <strong>de</strong> cette métho<strong>de</strong> (Tableau 2.1). L’échantillonnage s’est déroulé en<br />

été pour les milieux permanents et au printemps pour les MTI dans une fenêtre <strong>de</strong> temps<br />

restreinte (quinzaine <strong>de</strong> jours) afin d’obtenir une image fiable <strong>de</strong> la communauté piscicole à<br />

<strong>de</strong>ux moments clés du cycle hydrologique. En raison <strong>de</strong> certaines contraintes <strong>de</strong> terrain,<br />

certains sites n’ont pu être échantillonnés qu’une seule année (Tableau 2.1). Les milieux<br />

permanents ont été choisis <strong>de</strong> manière à couvrir l’ensemble <strong>de</strong> la zone d’étu<strong>de</strong> et ce en<br />

fonction <strong>de</strong> différents <strong>de</strong>scripteurs (Brivet, Marais <strong>de</strong> Gran<strong>de</strong> Brière Mottière et marais <strong>de</strong><br />

Donges ainsi que le réseau hydrographique primaire et secondaire, Figure 2.6). Les MTI<br />

24


Matériel et metho<strong>de</strong>s général<br />

échantillonnés étaient tous adjacents aux milieux permanents échantillonnés afin <strong>de</strong> comparer<br />

les communautés dans les <strong>de</strong>ux types d’habitats avec toutefois une subdivision <strong>de</strong>s MTI en<br />

fonction <strong>de</strong> leur nature : prairie, roselière ou plan d’eau. Les MTI échantillonnés se situent<br />

tous en Gran<strong>de</strong> Brière Mottière, seule zone où ces milieux sont présents en gran<strong>de</strong> quantité et<br />

suffisamment inondés (Figure 2.6). Dans l’approche diachronique, l’échantillonnage a été<br />

mené dans un milieu permanent et son MTI adjacent <strong>de</strong> février à juin 2006 tous les 15 jours.<br />

Cela a été rendu possible car la profon<strong>de</strong>ur <strong>de</strong> ce canal appartenant au réseau secondaire<br />

(Figure 2.6) n’excédait pas les 80 centimètres durant toute la phase d’inondation.<br />

Tableau 2.1. Liste <strong>de</strong>s sites échantillonnés par pêche électrique <strong>de</strong> 2004 à 2006. Voir détails <strong>de</strong> la<br />

localisation géographique <strong>de</strong>s sites dans la Figure 2.6.<br />

Milieux permanents Milieux temporairement inondés (GBM)<br />

N° Nom Lieu 1 Année N° Nom Type 2 Année 2<br />

1 Loncé Brivet 2004-2005 a Fougères Fédrun P 2004-2005<br />

2 Rozé Brivet 2004-2005 b Fougères Est R 2004-2005<br />

3 Boisman Brivet 2004-2005 c Fougères Ouest R 2004-2005<br />

4 La Guesne Brivet 2004-2005 d Rozé Sud P 2004-2005<br />

5 L’Angle Brivet 2004-2005 e Rozé Nord P 2004-2005<br />

6 Besné Brivet 2004-2005 f Cinq mètres <strong>de</strong> Bert R 2004-2005<br />

7 Pont Brûlé Donges 2004-2005 g Cinq mètres Sud R 2004-2005<br />

8 Curée à Monsieur Donges 2004-2005 h Lan<strong>de</strong>s P 2004-2005<br />

9 Priory sud Donges 2004-2005 i Tréhé P 2004-2005<br />

10 Priory nord Donges 2004-2005 j Bois <strong>de</strong> l’Isle R 2004-2005<br />

11 Martigné Donges 2005 k Pertuit <strong>de</strong>s Oies R 2004<br />

12 Les Lan<strong>de</strong>s GBM 2004-2005 l Butte <strong>de</strong>s Pierres Est P 2004-2005<br />

13 Tréhé GBM 2004-2005 m Butte <strong>de</strong>s Pierres Ouest P 2004-2005<br />

14 Bois <strong>de</strong> l’Isle GBM 2004-2005 n Butte à la Nonne P 2004-2005<br />

15 Bréca GBM 2004-2005 o Chenal Neuf PE 2004<br />

16 Quatre canaux GBM 2004-2005 p Grands Charreaux P 2004-2005<br />

17 Gouzien GBM 2004-2005 q Gouzien PE 2004-2005<br />

18 Chenal neuf GBM 2004-2005 r Gros fossés PE 2004-2005<br />

19 Réserve du Sud GBM 2004-2005 s Tranchée PE 2004-2005<br />

20 Cinq mètres Sud GBM 2004-2005 t Trou vasou PE 2004-2005<br />

21 Cinq mètres Nord GBM 2005 u Réserve sud plat P 2004-2005<br />

22 Cinq mètres Bert GBM 2004-2005 v Réserve sud roselière R 2004-2005<br />

23 Trignac GBM 2004-2005 w Réserve sud piar<strong>de</strong> PE 2004-2005<br />

24 Canal <strong>de</strong> Rozé GBM 2004-2005 x Réserve nord plat P 2004-2005<br />

25 Fougères Fédrun GBM 2004-2005 y Réserve nord roselière R 2004-2005<br />

26 Fougères Constant GBM 2004-2006 z Réserve nord piar<strong>de</strong> PE 2004-2005<br />

27 Canal du Nord GBM 2004-2005 α Yvonic PE 2005<br />

28 Butte <strong>de</strong>s Pierres GBM 2004-2005 β Julot PE 2005<br />

29 Réserve du Nord GBM 2004-2005 γ Européenne PE 2005<br />

30 Trou vasou GBM 2005 δ Prise du Coin PE 2005<br />

1 Zone dans laquelle le site se situe : Brivet = cours principal du Brivet, Donges = Marais <strong>de</strong> Donges et GBM =<br />

Marais <strong>de</strong> Gran<strong>de</strong> Brière Mottière, 2 Type <strong>de</strong> milieu temporairement inondé : P= prairie, R = roselière et PE =<br />

piar<strong>de</strong> ou plan d’eau.<br />

25


N<br />

2 Km<br />

i<br />

15<br />

j k<br />

13 14<br />

h<br />

12<br />

γ<br />

17<br />

q<br />

21<br />

δ<br />

m<br />

s<br />

27<br />

28 29<br />

r y<br />

x<br />

z<br />

l<br />

n<br />

16<br />

u v<br />

β<br />

α<br />

30<br />

w 19<br />

23<br />

18 p<br />

o<br />

g f<br />

22<br />

20<br />

t<br />

c<br />

24<br />

e<br />

d<br />

a<br />

26<br />

25<br />

b<br />

1<br />

2<br />

8<br />

3<br />

7<br />

Habitat:<br />

Prairie<br />

Roselière<br />

10<br />

4<br />

Sites échantillonnés:<br />

9<br />

5<br />

6<br />

11<br />

Réseau hydrographique:<br />

Primaire<br />

Secondaire<br />

Milieu permanent Milieu temporairement inondé<br />

Figure 2.6. Cartographie <strong>de</strong> la zone d’étu<strong>de</strong> (nature <strong>de</strong>s milieux temporairement inondés et réseau hydrographique) et localisation <strong>de</strong>s sites échantillonnés par pêche<br />

électrique. Voir le 2.1 pour la toponymie <strong>de</strong>s sites échantillonnés.


Matériel et metho<strong>de</strong>s général<br />

Lors <strong>de</strong>s l’échantillonnage <strong>de</strong>s MTI, en plus <strong>de</strong> la caractérisation <strong>de</strong> la communauté<br />

piscicole colonisant ces milieux, la distribution spatiale <strong>de</strong>s poissons dans ces milieux a été<br />

étudiée. Ainsi, le protocole développé par Carpentier et al. (2004) a été employé. Dans chaque<br />

MTI échantillonné, un minimum <strong>de</strong> 25 EPA a été réalisé et distribué aléatoirement dans tous<br />

les habitats disponibles. La série d’EPA était réalisée jusqu’à la limite accessible du front<br />

d’inondation. Le nombre total d’EPA correspondait à une <strong>de</strong>nsité constante <strong>de</strong> point<br />

d’échantillonnage dans chaque site et chaque type <strong>de</strong> milieux. De plus, les coordonnées GPS<br />

<strong>de</strong> chaque EPA ont été enregistrées afin <strong>de</strong> pouvoir ensuite calculer la distance minimale entre<br />

le point échantillonné et le milieu permanent adjacent. Le calcul <strong>de</strong> cette distance était adapté<br />

à la configuration <strong>de</strong> chaque MTI et <strong>de</strong>s jonctions avec le milieu permanent, comme cela est<br />

illustré dans la Figure 2.7.<br />

Milieu permanent<br />

Connections entre milieux<br />

permanents et<br />

temporairement inondés<br />

Point d’échantillonnage<br />

ponctuel d’abondance (EPA)<br />

Milieu<br />

temporairement<br />

inondé<br />

Milieu<br />

temporairement<br />

inondé<br />

Distance entre EPA et le<br />

milieu permanent adjacent<br />

Figure 2.7. Représentation schématique du protocole d’échantillonnage <strong>de</strong>s milieux<br />

temporairement inondés par pêche électrique (approche EPA) et <strong>de</strong> calcul <strong>de</strong> la distance entre<br />

chaque point d’échantillonnage et le milieu permanent adjacent.<br />

27


Chapitre 2<br />

2.2.2. Suivi <strong>de</strong>s échanges <strong>de</strong> poissons entre milieux permanents et<br />

temporairement inondés<br />

L’objectif <strong>de</strong> ce suivi est <strong>de</strong> quantifier tous les échanges se déroulant entre <strong>de</strong>s milieux<br />

permanents et différents MTI pour toutes les espèces <strong>de</strong> la communauté et durant toute la<br />

pério<strong>de</strong> d’inondation et d’exondation. Les métho<strong>de</strong>s utilisées pour quantifier les flux <strong>de</strong><br />

poissons entre milieux permanents et MTI tels que <strong>de</strong>s annexes, <strong>de</strong>s plaines inondées ou <strong>de</strong>s<br />

bras morts sont nombreuses. Par exemple, Borcherding et al. (2002) et Heermann &<br />

Borcherding (2006) utilisent l’observation directe du passage <strong>de</strong>s poissons à l’ai<strong>de</strong> d’une<br />

planche blanche, lestée sur le fond. Elle présente l’avantage <strong>de</strong> ne pas être intrusive, dans le<br />

sens où les poissons ne sont pas gênés ni retardés pendant leur déplacement. Cependant, elle<br />

présente <strong>de</strong> nombreux désavantages car elle ne permet pas <strong>de</strong> mesurer précisément les<br />

poissons, <strong>de</strong> distinguer <strong>de</strong>ux individus appartenant à <strong>de</strong>s espèces morphologiquement très<br />

proches, <strong>de</strong> connaître leur niveau <strong>de</strong> maturité puisque les poissons ne sont pas capturés, <strong>de</strong><br />

suivre les échanges nocturnes ou encore <strong>de</strong> suivre les flux pendant une longue pério<strong>de</strong><br />

continue. Enfin, cette métho<strong>de</strong> n’est utilisable que dans <strong>de</strong>s eaux dont la transparence est<br />

élevée, ce qui n’est pas le cas dans la zone d’étu<strong>de</strong>.<br />

Dans d’autres étu<strong>de</strong>s, <strong>de</strong>s techniques plus intrusives ont été utilisées comme <strong>de</strong>s filets<br />

maillants avec lesquels les poissons sont capturés par les opercules (Lusk et al. 2001). Cette<br />

métho<strong>de</strong> présente le désavantage d’entraîner un taux <strong>de</strong> mortalité élevé si les engins ne sont<br />

pas relevés fréquemment et ne permet donc pas un suivi continu pendant une longue pério<strong>de</strong>.<br />

De plus, la sélectivité <strong>de</strong> ces engins ne permet pas la capture <strong>de</strong>s plus petits individus. La<br />

pêche électrique couplée à un marquage massif a aussi été utilisée pour suivre ces échanges ;<br />

les <strong>de</strong>ux types <strong>de</strong> milieux étant alors échantillonnés à <strong>de</strong>s pas <strong>de</strong> temps réguliers et les flux<br />

estimés entre <strong>de</strong>ux campagnes d’échantillonnage (e.g. Hohausova 2000, Lusk et al. 2001).<br />

Cependant, cette technique est lour<strong>de</strong> d’un point <strong>de</strong> vue humain et en matériel et ne permettait<br />

pas un suivi quotidien <strong>de</strong>s échanges car <strong>de</strong>s pêches électriques répétées dans <strong>de</strong>s intervalles <strong>de</strong><br />

temps courts modifiaient nécessairement le fonctionnement <strong>de</strong> la communauté. Enfin, on peut<br />

citer la possibilité <strong>de</strong> suivre ce type d’échanges à l’ai<strong>de</strong> <strong>de</strong> techniques plus pointues telles que<br />

le marquage individuel couplé à un suivi par télémétrie <strong>de</strong>s poissons marqués (Lucas & Baras,<br />

2000). Cependant, ces techniques, très coûteuses, ne permettent pas <strong>de</strong> suivre l’ensemble <strong>de</strong> la<br />

communauté composée par un grand nombre d’individus <strong>de</strong> différentes espèces et <strong>de</strong> classes<br />

<strong>de</strong> tailles très contrastées.<br />

28


Matériel et metho<strong>de</strong>s général<br />

La métho<strong>de</strong> permettant un suivi à long terme (plusieurs mois) et continu (jour et nuit)<br />

<strong>de</strong>s échanges <strong>de</strong> poissons entre milieux permanents et MTI adjacents avec un <strong>de</strong>s plus faible<br />

taux <strong>de</strong> mortalité si certaines précautions sont prises est celle <strong>de</strong>s pièges passifs tels les<br />

verveux ou les nasses (Poizat & Crivelli 1997, Borcherding et al. 2002, Hladik & Kubecka<br />

2003, Hohausova et al. 2003). De plus, grâce à leurs tailles très adaptables, ces pièges peuvent<br />

être disposés dans <strong>de</strong>s milieux possédant <strong>de</strong>s configurations <strong>de</strong> largeur et <strong>de</strong> profon<strong>de</strong>ur très<br />

variables. Enfin, ces pièges permettent <strong>de</strong> bloquer toute la masse d’eau et donc <strong>de</strong> capturer<br />

l’ensemble <strong>de</strong>s poissons allant d’un milieu à un autre. Cette technique a donc été retenue pour<br />

ce travail <strong>de</strong> thèse. De plus, les mailles <strong>de</strong>s pièges utilisés (5-mm) permettent la capture<br />

exhaustive <strong>de</strong> l’ensemble <strong>de</strong>s classes <strong>de</strong> tailles <strong>de</strong> poissons se déplaçant entre ces milieux. Les<br />

pièges ont été disposés au niveau <strong>de</strong>s connections uniques entre milieux permanents et MTI.<br />

Les pièges ont été relevés la plupart du temps tous les matins, excepté pendant les pério<strong>de</strong>s <strong>de</strong><br />

très faibles déplacements où la relève ne s’est effectuée que tous les <strong>de</strong>ux jours, afin <strong>de</strong> limiter<br />

au maximum la mortalité et <strong>de</strong> permettre d’avoir une image fine et continue <strong>de</strong>s déplacements.<br />

De plus, afin <strong>de</strong> ne pas entraver les échanges naturels <strong>de</strong> poissons entre ces <strong>de</strong>ux milieux, le<br />

piégeage a été réalisé en permanence dans les <strong>de</strong>ux sens. Chaque poisson capturé a été<br />

anesthésié, déterminé, mesuré et parfois pesé, la maturité vérifiée à l’ai<strong>de</strong> d’une légère<br />

pression sur l’abdomen puis chaque individu a ensuite été relâché dans le sens <strong>de</strong> son<br />

déplacement.<br />

Le site pour le suivi <strong>de</strong>s échanges <strong>de</strong>vait présenter les trois caractéristiques suivantes :<br />

(1) possé<strong>de</strong>r dans une zone restreinte l’ensemble <strong>de</strong>s MTI caractéristiques <strong>de</strong>s marais <strong>de</strong><br />

Brière (i.e. prairie, roselière et piar<strong>de</strong>) pour permettre un suivi représentatif <strong>de</strong> la zone d’étu<strong>de</strong>,<br />

un relevé quotidien (proximité <strong>de</strong>s engins afin d’optimiser le temps <strong>de</strong> déplacement d’un<br />

engin à un autre) et <strong>de</strong> minimiser les effets <strong>de</strong> la structuration géographique <strong>de</strong> la<br />

communauté pour établir <strong>de</strong>s comparaisons entre sites, (2) être d’accessibilité réduite, pour<br />

éviter tout risque <strong>de</strong> vandalisme éventuel et permettre ainsi un piégeage continu (jour/nuit) à<br />

l’ai<strong>de</strong> d’engins réservés uniquement aux étu<strong>de</strong>s scientifiques (aspects législatifs et interactions<br />

avec la gar<strong>de</strong>rie et les pêcheurs amateurs) et (3) avoir un niveau <strong>de</strong> perturbation par l’activité<br />

halieutique <strong>de</strong> la communauté piscicole et <strong>de</strong> ces échanges très faible. Les suivis <strong>de</strong>s<br />

déplacements ont donc été réalisés dans un espace protégé (Réserve du Sud, Figure 2.8), situé<br />

en plein cœur <strong>de</strong>s marais <strong>de</strong> Gran<strong>de</strong> Brière Mottière, lors <strong>de</strong> 2 années (2004 et 2005). En 2004,<br />

l’objectif a été <strong>de</strong> suivre les processus d’émigration pendant la phase d’exondation <strong>de</strong> quatre<br />

MTI alors qu’en 2005, l’objectif a été <strong>de</strong> suivre les processus d’immigration et d’émigration<br />

pendant toute la phase d’inondation. De manière exceptionnelle, l’accès à cette zone <strong>de</strong><br />

29


Chapitre 2<br />

réserve intégrale nous a été autorisé par la Commission Syndicale <strong>de</strong> Gran<strong>de</strong> Brière Mottière<br />

pour la durée <strong>de</strong> ce travail <strong>de</strong> thèse.<br />

Canal<br />

temporaire<br />

N<br />

50 m<br />

Canal<br />

permanent<br />

(*)<br />

Canal permanent<br />

Prairie<br />

Milieu temporairement inondé<br />

Roselière<br />

Piar<strong>de</strong><br />

Emigration<br />

Plan d’eau<br />

Piège passif<br />

Canal<br />

permanent<br />

Immigration<br />

Berge<br />

Figure 2.8. Cartographie d’un secteur <strong>de</strong> la réserve sud <strong>de</strong> Gran<strong>de</strong> Brière Mottière et<br />

représentation <strong>de</strong>s dispositifs <strong>de</strong> piégeage pour le suivi <strong>de</strong>s échanges <strong>de</strong> poissons entre les milieux<br />

permanents et divers milieux temporairement inondés (prairie, roselière et piar<strong>de</strong>). (*) Zone suivie<br />

uniquement en 2004.<br />

30


2.2.3. Acquisition <strong>de</strong> paramètres environnementaux<br />

Matériel et metho<strong>de</strong>s général<br />

En plus <strong>de</strong>s données concernant la communauté piscicole, <strong>de</strong>s données environnementales ont<br />

été collectées et analysées afin <strong>de</strong> comprendre leur rôle dans les processus d’utilisation <strong>de</strong>s<br />

MTI et d’échanges entre ces milieux et les milieux permanents. Le niveau d’eau, les<br />

précipitations et la température <strong>de</strong> l’air ont été enregistrés quotidiennement et nous ont été<br />

fournis par le Parc naturel régional <strong>de</strong> Brière. La température <strong>de</strong> l’eau a, quant à elle, été<br />

enregistrée en 2005 et 2006 à l’ai<strong>de</strong> <strong>de</strong> son<strong>de</strong>s enregistreuses (Sensor StowAway TidBit,<br />

Onset Computer Corporation) sur une base d’un enregistrement toutes les quinze minutes<br />

dans les MTI et dans les canaux. Enfin, lors <strong>de</strong> certaines campagnes <strong>de</strong> pêches électriques, les<br />

paramètres physico-chimiques <strong>de</strong> l’eau, à savoir la concentration en dioxygène (en mg.L -1 ), le<br />

pH, la conductivité (en µS.cm -1 ) et la transparence <strong>de</strong> l’eau (distance du disque <strong>de</strong> Secchi en<br />

cm) ont été mesurés. Cependant, puisque la plupart <strong>de</strong> ces paramètres varient fortement au<br />

cours <strong>de</strong> la journée, ces mesures n’ont été réalisées que lorsqu’une standardisation <strong>de</strong> l’heure<br />

<strong>de</strong> mesure était rendue possible, c’est-à-dire lors d’approches diachroniques ou continues.<br />

Lors <strong>de</strong> ces approches, un suivi <strong>de</strong> l’évolution <strong>de</strong> la ressource en zooplancton a été effectué<br />

toujours dans la même tranche horaire, c’est-à-dire entre 11h et 12h, ceci afin <strong>de</strong> réduire les<br />

biais liés aux migrations journalières du zooplancton. En effet, cette ressource constitue un<br />

élément clé dans l’ontogenèse du développement alimentaire <strong>de</strong> nombreuses espèces <strong>de</strong><br />

poisson qui la consomment aux sta<strong>de</strong>s les plus jeunes voire même en tant qu’adultes. Un<br />

volume <strong>de</strong> 15L d’eau a été prélevé dans la colonne d’eau puis filtré (maille <strong>de</strong> 0.1 mm). Les<br />

échantillons ont été conservés dans l’alcool à 70°C, puis décomptés (Copepoda et Cladocera)<br />

au laboratoire sous loupe binoculaire à l’ai<strong>de</strong> d’une boîte <strong>de</strong> Dollfus.<br />

31


Chapitre 2<br />

2.2.4. Outils complémentaires pour l’analyse du fonctionnement <strong>de</strong> certaines<br />

espèces dans les marais <strong>de</strong> Brière<br />

Pour compléter l’approche communautaire précé<strong>de</strong>mment évoquée, ce travail <strong>de</strong> thèse s’est<br />

aussi focalisé sur le fonctionnement <strong>de</strong>s populations <strong>de</strong> certaines espèces <strong>de</strong>s marais <strong>de</strong> Brière,<br />

en relation avec les programmes mis en place pour les gérer. En plus <strong>de</strong>s données collectées<br />

sur ces espèces lors <strong>de</strong> l’étu<strong>de</strong> <strong>de</strong> la communauté piscicole, <strong>de</strong>ux autres outils <strong>de</strong> suivi ont été<br />

spécifiquement mis en place afin <strong>de</strong> répondre à cet objectif.<br />

2.2.4.1. Enquête sur les pratiques halieutiques<br />

Dans un premier temps, ce travail <strong>de</strong> thèse s’est intéressé aux fonctionnements d’espèces<br />

exploitées et une étu<strong>de</strong> <strong>de</strong>s pratiques halieutiques a donc été menée afin <strong>de</strong> comprendre le rôle<br />

<strong>de</strong> cette activité sur le fonctionnement <strong>de</strong>s populations <strong>de</strong> poisson chat (Ameiurus melas) et<br />

d’anguille (Anguilla anguilla). Acteur incontournable, le mon<strong>de</strong> halieutique <strong>de</strong>s marais <strong>de</strong><br />

Gran<strong>de</strong> Brière Mottière a ainsi été sollicité afin <strong>de</strong> caractériser la pêcherie, exclusivement<br />

composée d’amateurs. Ainsi, en 2005, une enquête halieutique a été réalisée (Thoby 2005) en<br />

s’attachant à trois aspects :<br />

1) l’analyse <strong>de</strong>s ventes <strong>de</strong> cartes <strong>de</strong> pêche par la Commission Syndicale <strong>de</strong><br />

Gran<strong>de</strong> Brière Mottière. Celle-ci permet <strong>de</strong> quantifier le nombre <strong>de</strong> pêcheurs. Cette<br />

analyse renseigne également sur les techniques utilisées (carte <strong>de</strong> pêche différente<br />

pour chaque type d’engin),<br />

2) la mise en place d’un questionnaire (anonyme, simple et synthétique, Figure<br />

2.9) concernant principalement la pêche aux engins pratiquée dans le marais indivis <strong>de</strong><br />

Gran<strong>de</strong> Brière. L’objectif premier a été d’obtenir un aperçu global sur les techniques<br />

<strong>de</strong> pêche utilisées, les pério<strong>de</strong>s et fréquences <strong>de</strong>s pêches, les espèces et sta<strong>de</strong>s ciblés<br />

et/ou capturés, les prélèvements et la localisation géographique <strong>de</strong> l’activité<br />

halieutique,<br />

3) un suivi <strong>de</strong> pêcheurs afin <strong>de</strong> compléter et affiner les informations recueillies<br />

par le biais du questionnaire. Concrètement, ce volet <strong>de</strong> l’enquête a consisté à<br />

accompagner les pêcheurs qui le souhaitaient lors <strong>de</strong> leurs sorties afin <strong>de</strong> caractériser<br />

qualitativement leurs captures. Ainsi, les poissons capturés ont été anesthésiés,<br />

déterminés, mesurés et pesés, puis remis aux pêcheurs, et ce dans le respect <strong>de</strong> leurs<br />

habitu<strong>de</strong>s.<br />

32


Figure 2.9. Questionnaire <strong>de</strong>stiné aux pêcheurs <strong>de</strong> Gran<strong>de</strong> Brière Mottière.<br />

Matériel et metho<strong>de</strong>s général<br />

2.2.4.2. Suivi individuel <strong>de</strong>s juvéniles <strong>de</strong> l’année <strong>de</strong> brochet (Esox lucius)<br />

issus <strong>de</strong> ces programmes <strong>de</strong> soutien<br />

La viabilité <strong>de</strong> la population naturelle <strong>de</strong> cette espèce semble être compromise et malgré <strong>de</strong>s<br />

programmes <strong>de</strong> soutien mise en place <strong>de</strong>puis <strong>de</strong> nombreuses années, les effectifs <strong>de</strong> cette<br />

espèce sont toujours faibles, comme nous le verrons à travers l’étu<strong>de</strong> <strong>de</strong> la communauté<br />

piscicole. Ainsi, le premier objectif <strong>de</strong> ce suivi a été <strong>de</strong> quantifier la survie <strong>de</strong> ces individus et<br />

d’i<strong>de</strong>ntifier les différentes causes <strong>de</strong> mortalité au niveau individuel lors <strong>de</strong> leur lâcher dans<br />

différents MTI. Le second objectif a été <strong>de</strong> comprendre les processus <strong>de</strong> déplacements entre<br />

MTI <strong>de</strong> ces individus lors <strong>de</strong> leur développement ontogénique et leur changement <strong>de</strong> régime<br />

alimentaire. Cette expérimentation s’est déroulée en 2005 parallèlement au suivi <strong>de</strong>s échanges<br />

<strong>de</strong> poissons afin d’utiliser la même zone d’étu<strong>de</strong> et le matériel <strong>de</strong> piégeage précé<strong>de</strong>mment<br />

évoqués.<br />

L’étu<strong>de</strong> <strong>de</strong>s déplacements <strong>de</strong>s poissons d’eau douce est un enjeu majeur pour <strong>de</strong><br />

nombreux scientifiques et permet <strong>de</strong> répondre à <strong>de</strong> nombreuses problématiques écologiques<br />

(voir la revue <strong>de</strong> Lucas & Baras 2001). Depuis <strong>de</strong>s siècles, les scientifiques ont développé <strong>de</strong><br />

nombreuses techniques pour étudier les mouvements <strong>de</strong> poissons dans leur milieu naturel<br />

(Lucas & Baras 2000). Pour y parvenir, le marquage individuel est obligatoire pour mesurer<br />

ces déplacements, et là encore, <strong>de</strong> nombreuses métho<strong>de</strong>s ont été mises en place (Lucas &<br />

Baras 2000). L’une <strong>de</strong>s plus prometteuses est le marquage à l’ai<strong>de</strong> <strong>de</strong> « Passive Integrated<br />

33


Chapitre 2<br />

Transpon<strong>de</strong>r » PIT-tag (Prentice et al. 1990) car elle permet le marquage d’individus <strong>de</strong><br />

petites tailles grâce à <strong>de</strong>s marques <strong>de</strong> taille adaptée (2.1 x 11.5 mm) et leur localisation à<br />

l’ai<strong>de</strong> d’une antenne mobile (Roussel et al. 2000, Cucherousset et al. 2005). Initialement<br />

développée pour l’étu<strong>de</strong> <strong>de</strong>s salmonidés dans les cours d’eau, le champ d’application <strong>de</strong> cette<br />

technique a été étendu, par le biais <strong>de</strong> ce travail <strong>de</strong> thèse, aux milieux lentiques peu profonds,<br />

à savoir les MTI, en l’appliquant à un modèle biologique particulier qu’est le brochet, et plus<br />

particulièrement les juvéniles <strong>de</strong> l’année (0+). Dans un premier temps, l’applicabilité <strong>de</strong> cet<br />

outil dans ce type d’écosystème avec ce modèle biologique a été évaluée pour ensuite<br />

répondre aux objectifs précé<strong>de</strong>mment évoqués.<br />

L’analyse <strong>de</strong>s isotopes stables a été utilisée afin <strong>de</strong> comprendre l’évolution <strong>de</strong>s<br />

régimes alimentaires au cours du développement ontogénique <strong>de</strong>s juvéniles <strong>de</strong> l’année <strong>de</strong><br />

brochet. Pour connaître le régime alimentaire <strong>de</strong>s poissons en milieu naturel et en déduire leur<br />

niveau trophique, <strong>de</strong> nombreuses métho<strong>de</strong>s ont été utilisées, telle que l’analyse <strong>de</strong>s contenus<br />

stomacaux. Cependant, ces métho<strong>de</strong>s impliquent, le plus souvent, le sacrifice <strong>de</strong>s individus et,<br />

elles ne donnent qu’une image très ponctuelle <strong>de</strong> leur régime alimentaire. L’analyse <strong>de</strong>s<br />

isotopes stables est <strong>de</strong>venue, au cours <strong>de</strong>s vingt <strong>de</strong>rnières années, un outil d’investigation<br />

utilisé en écologie. Cette technique permet, <strong>de</strong> manière non invasive à partir d’échantillons <strong>de</strong><br />

tissus (e.g. nageoire), <strong>de</strong> connaître le niveau trophique d’un individu en fonction <strong>de</strong> sa<br />

signature isotopique et <strong>de</strong> celle <strong>de</strong>s proies potentielles présentes dans le milieu naturel. Cela<br />

est rendu possible du fait que certains éléments naturels présentent une variation <strong>de</strong> leur<br />

masse atomique due à une différence <strong>de</strong> leur nombre <strong>de</strong> neutrons. Cette propriété n’affecte<br />

pas leurs caractéristiques physiques et chimiques mais permet leur i<strong>de</strong>ntification par<br />

spectrométrie <strong>de</strong> masse du ratio isotopique, puisque chaque forme porte une signature unique.<br />

Les éléments qui ne se dégra<strong>de</strong>nt pas dans le temps (non radioactifs) sont les isotopes stables.<br />

Traditionnellement utilisée en géochimie, l’analyse <strong>de</strong>s isotopes stables est utilisée en<br />

écologie pour mesurer les flux d’énergie au sein <strong>de</strong>s écosystèmes, grâce à leurs propriétés<br />

d’enregistrer les caractéristiques <strong>de</strong> la source et les phénomènes <strong>de</strong> fragmentation. Les <strong>de</strong>ux<br />

principaux éléments utilisés en écologie sont le carbone (C) et l’azote (N). Dans la nature, la<br />

forme principale <strong>de</strong> carbone est le 12 C (98.89%) avec une petite proportion <strong>de</strong> 13 C (1.11%)<br />

alors que la forme principale <strong>de</strong> l’azote et le 14 N (99.64%) avec une faible proportion <strong>de</strong> 15 N<br />

(0.36%). Ces proportions changent en fonction du niveau trophique, les niveaux les plus<br />

élevés s’enrichissant <strong>de</strong> la forme la plus lour<strong>de</strong> (i.e. 13 C ou 15 N). Leurs caractéristiques<br />

thermodynamiques et leurs masses atomiques variables permettent alors <strong>de</strong> mesurer le niveau<br />

d’enrichissement d’un échantillon et d’interpréter son niveau trophique en comparaison avec<br />

34


Matériel et metho<strong>de</strong>s général<br />

les sources potentielles, c’est-à-dire ici les proies (voir revue détaillée et détails techniques<br />

dans Jardine et al. 2003). De plus, cette technique a permis <strong>de</strong> nombreuses avancées<br />

notamment dans le domaine <strong>de</strong> l’écologie <strong>de</strong>s poissons (voir synthèse dans Dufour &<br />

Ger<strong>de</strong>aux 2001).<br />

A partir d’échantillons <strong>de</strong> nageoires et <strong>de</strong> l’ensemble <strong>de</strong>s proies potentielles <strong>de</strong>s<br />

juvéniles <strong>de</strong> l’année <strong>de</strong> brochet (Esox lucius), à savoir le zooplancton, les macro-invertébrés et<br />

le poisson fourrage, nous avons analysé les changements <strong>de</strong> niveau trophique <strong>de</strong> ces individus<br />

entre leur zone <strong>de</strong> naissance, l’aire <strong>de</strong> nourricerie et la zone <strong>de</strong> refuge pendant l’étiage à l’ai<strong>de</strong><br />

<strong>de</strong> l’analyse <strong>de</strong>s isotopes stables couplée au marquage individuel PIT tag. Ce protocole,<br />

couplé aux recaptures <strong>de</strong>s mêmes individus marqués lors <strong>de</strong> leurs changements <strong>de</strong> milieux, a<br />

permis <strong>de</strong> suivre les changements d’habitats et <strong>de</strong> régime alimentaire et surtout <strong>de</strong> mesurer les<br />

sources <strong>de</strong> variabilités interindividuelles.<br />

2.3. Synthèse <strong>de</strong>s opérations <strong>de</strong> terrain<br />

Au cours <strong>de</strong> ce travail <strong>de</strong> thèse, 119 sites ont été échantillonnés en pêche électrique en 10<br />

semaines pleines, pour un total <strong>de</strong> 4183 EPA réalisés et 26842 poissons capturés entre avril<br />

2004 et juin 2006. Concernant le piégeage passif, 19190 individus ont été capturés dans les 8<br />

pièges utilisés qui ont été relevés 139 fois pour une durée totale <strong>de</strong> piégeage <strong>de</strong> 9 mois en<br />

2004 et 2005. De plus, concernant l’enquête sur les pratiques halieutiques, 75 questionnaires<br />

ont été remplis et analysés et <strong>de</strong>s pêcheurs ont été accompagnés lors <strong>de</strong> 17 sorties. Enfin,<br />

concernant le suivi individuel <strong>de</strong>s juvéniles <strong>de</strong> l’année, 297 individus ont été marqués et<br />

l’antenne portable a été utilisée une vingtaine <strong>de</strong> fois pour obtenir leur localisation.<br />

L’ensemble <strong>de</strong> ces données a été utilisé afin <strong>de</strong> répondre aux différents objectifs <strong>de</strong> ce travail<br />

<strong>de</strong> thèse et ce, en fonction du niveau d’organisation utilisé (communauté, population et<br />

individu). Le tableau 2.2 synthétise l’ensemble <strong>de</strong>s opérations <strong>de</strong> terrain qui sont ventilées<br />

pour chaque chapitre <strong>de</strong> cette thèse.<br />

35


Tableau 2.2. Synthèse <strong>de</strong>s opérations <strong>de</strong> terrain conduites dans les marais <strong>de</strong> Brière et menées pour chacun <strong>de</strong>s chapitres <strong>de</strong> ce travail <strong>de</strong> thèse.<br />

Partie Chapitre Outils 1 Zone 2 Milieu 3 Année Durée 4 Op. 5 Effort 6 Captures<br />

I<br />

II<br />

3 Pêche électrique MB MP-MTI 2004 5 4 1565 EPA 11949 poissons<br />

4 Pêche électrique GBM MP-MTI<br />

2004<br />

2005<br />

10 4 1821 EPA 10468 poissons<br />

5 Pêche électrique PR MP-MTI 2006 1.5 4 895 EPA 589 poissons<br />

6 Piégeage passif RS MTI 2004 12 2 8 pièges 6321 poissons<br />

7 Piégeage passif RS MTI 2005 24 2 8 pièges 8670 poissons<br />

8 Pêche électrique MB MP-MTI 2004 5 4 1305 EPA<br />

1603 poissonschats<br />

9<br />

Pêche électrique, piégeage<br />

passif & enquête<br />

GBM MP 2004 3 4+2<br />

401 EPA<br />

& 8 pièges<br />

2268 poissonschats<br />

10<br />

Pêche électrique, piégeage<br />

passif & enquête<br />

GBM MP<br />

2004<br />

2005<br />

32 4+1<br />

914 EPA<br />

& 8 pièges<br />

1868 anguilles<br />

11 Piégeage passif & PIT RS MTI<br />

2004<br />

2005<br />

4 2 1 piège 192 brochetons<br />

12 Piégeage passif & PIT RS MTI 2005 24 2 4 pièges 297 brochetons<br />

1 PIT = Passive Integrated Transpon<strong>de</strong>r technologie ; 2 MB = marais <strong>de</strong> Brière, GBM = marais <strong>de</strong> Gran<strong>de</strong> Brière Mottière, RS = réserve sud <strong>de</strong> Gran<strong>de</strong> Brière<br />

Mottière, PR = plat <strong>de</strong> Rozé ; 3 MP = milieu permanent, MTI = milieu temporairement inondé ; 4 Durée en semaines ; 5 Nombre d’opérateurs lors du suivi ; 6<br />

EPA = point d’échantillonnage ponctuel d’abondance, ind = individus, PCH = poisson chat, ANG = anguille, BRO = brochet.


Partie I<br />

Utilisation fonctionnelle <strong>de</strong>s milieux temporairement<br />

inondés par la communauté piscicole<br />

Le premier objectif <strong>de</strong> ce travail <strong>de</strong> thèse est <strong>de</strong> décrire la communauté piscicole<br />

<strong>de</strong>s marais <strong>de</strong> Brière à l’étiage, pério<strong>de</strong> du cycle hydrologique où les poissons sont<br />

confinés dans le réseau <strong>de</strong> canaux, mais également lors <strong>de</strong> la crue printanière dans<br />

les MTI. Dans un second temps, l’approche portera sur l’utilisation comparée <strong>de</strong>s<br />

<strong>de</strong>ux grands types <strong>de</strong> milieux (milieux permanents et MTI) par la communauté<br />

piscicole à un moment précis <strong>de</strong> la pério<strong>de</strong> d’inondation, puis ensuite durant toute<br />

la phase <strong>de</strong> crue printanière. L’accent sera mis sur l’étu<strong>de</strong> du processus <strong>de</strong><br />

diffusion <strong>de</strong>s poissons dans les MTI à partir <strong>de</strong>s connexions au milieu permanent.<br />

Finalement, les patrons d’échanges <strong>de</strong> poissons entre milieux permanents et MTI,<br />

ainsi que les mécanismes structurant ces échanges lors du cycle hydrologique<br />

seront examinés afin <strong>de</strong> bien cerner les fonctions <strong>de</strong>s milieux permanents et <strong>de</strong>s<br />

MTI pour la communauté piscicole en place.<br />

Une prairie temporairement inondée dans les marais <strong>de</strong> Brière


Chapitre 3<br />

Etat <strong>de</strong>s lieux <strong>de</strong> la communauté piscicole<br />

Contrairement à <strong>de</strong> nombreux autres écosystèmes aquatiques français, les marais<br />

<strong>de</strong> Brière n’ont pas fait l’objet d’un suivi régulier <strong>de</strong> la faune piscicole, tel que<br />

cela est pratiqué dans le cadre du Réseau Hydrobiologique et Piscicole (RHP) mis<br />

en place par le Conseil Supérieur <strong>de</strong> la Pêche à l’échelle nationale. Au<br />

commencement <strong>de</strong> cette étu<strong>de</strong>, le niveau <strong>de</strong> connaissance <strong>de</strong> la faune piscicole<br />

locale était donc pour le moins fragmenté, à la fois dans le temps et dans l’espace<br />

et manquait clairement d’une actualisation <strong>de</strong>s données à l’échelle <strong>de</strong> l’ensemble<br />

<strong>de</strong> la zone d’étu<strong>de</strong>. Ainsi, en guise <strong>de</strong> présentation <strong>de</strong> la faune piscicole <strong>de</strong>s marais<br />

<strong>de</strong> Brière, ce chapitre retrace les principales modifications observées dans la<br />

composition du peuplement piscicole au cours <strong>de</strong> ces 30 <strong>de</strong>rnières années. Ce<br />

travail est basé sur une comparaison <strong>de</strong>s données acquises lors <strong>de</strong> ce travail <strong>de</strong><br />

thèse à celles provenant <strong>de</strong>s seuls travaux antérieurs portant sur l’ensemble <strong>de</strong> la<br />

zone d’étu<strong>de</strong>, en l’occurrence ceux <strong>de</strong> Maillard (1972) publiés, à l’époque, dans la<br />

revue Penn ar Bed.


Chapitre 3<br />

Première campagne d’échantillonnage <strong>de</strong> la communauté piscicole <strong>de</strong>s marais <strong>de</strong> Brière par<br />

pêche électrique en 2004<br />

40


Biodiversité piscicole <strong>de</strong> Brière<br />

Le peuplement piscicole dulçaquicole <strong>de</strong>s marais <strong>de</strong> Brière :<br />

évolution durant les 30 <strong>de</strong>rnières années et composition actuelle 1<br />

Résumé : Les marais <strong>de</strong> Brière, situés le long <strong>de</strong> la faça<strong>de</strong> atlantique aux portes <strong>de</strong><br />

Saint-Nazaire, ont été retenus à la convention RAMSAR <strong>de</strong> 1995 en raison <strong>de</strong> leur richesse<br />

floristique et faunistique. Malgré tout, comparativement à d’autres organismes animaux et<br />

végétaux <strong>de</strong>s marais <strong>de</strong> Brière, la faune piscicole est mal connue et aucune étu<strong>de</strong> récente n’a<br />

été réalisée sur l’ensemble <strong>de</strong>s marais. Le peuplement piscicole a donc été échantillonné et<br />

comparé à celui établi vers 1970 puis 1985. Il ressort que le peuplement actuel est dominé par<br />

le poisson-chat (44.7 % <strong>de</strong>s effectifs) et est composé <strong>de</strong> 20 espèces, contre 19 en 1972. Trois<br />

espèces autochtones ont disparu lors <strong>de</strong> cette pério<strong>de</strong> (le chevaine, l’épinochette et la vandoise)<br />

alors que <strong>de</strong>ux sont apparues (la bouvière et l’ablette). Parallèlement, trois espèces<br />

allochtones ont été introduites (la gambusie, le carassin argenté et le pseudorasbora). Nous<br />

discutons <strong>de</strong> manière très générale <strong>de</strong>s causes pouvant expliquer une telle composition et<br />

évolution du peuplement ainsi que <strong>de</strong>s différentes perspectives <strong>de</strong> gestion.<br />

1 Ce chapitre est adapté <strong>de</strong> l’article : Cucherousset, J., Carpentier, A., Eybert, M.-C., Paillisson, J.-M. & Damien,<br />

J.-P. 2005. Le peuplement piscicole dulçaquicole <strong>de</strong>s marais <strong>de</strong> Brière. Penn ar Bed. 195 : 15-24.<br />

41


Chapitre 3<br />

3.1. Introduction<br />

Les marais du Brivet, plus généralement dénommés Marais <strong>de</strong> Brière (Figure 3.1), sont situés<br />

sur la faça<strong>de</strong> atlantique aux portes <strong>de</strong> Saint Nazaire et formés <strong>de</strong> dépressions en étroite<br />

interdépendance, occupant une surface <strong>de</strong> 19 000 ha <strong>de</strong> terrains inondables alimentés par un<br />

bassin versant <strong>de</strong> 80 000 ha. La création d’un Parc naturel régional en 1970, qui occupe<br />

actuellement une surface <strong>de</strong> 49 000 ha et les différents inventaires au titre <strong>de</strong>s ZNIEFF<br />

« Zones Naturelles d’Intérêt Ecologique Floristique et Faunistique » et <strong>de</strong> la Directive<br />

Oiseaux (inventaire ZICO « Zones Importantes pour la Conservation <strong>de</strong>s Oiseaux ») ont<br />

permis <strong>de</strong> désigner cette zone humi<strong>de</strong> en 1995 en tant que site RAMSAR.<br />

Prairie<br />

Roselière<br />

<strong>Rennes</strong><br />

N<br />

France<br />

Figure 3.1. Cartographie <strong>de</strong>s réseaux <strong>de</strong> canaux et du recouvrement végétal <strong>de</strong>s zones<br />

temporairement inondées (prairie vs roselière) dans les marais <strong>de</strong> Brière (Source : Parc naturel<br />

régional <strong>de</strong> Brière).<br />

42<br />

2 Km


Biodiversité piscicole <strong>de</strong> Brière<br />

D’intérêt majeur pour l’avifaune (Constant et al. 1987, Eybert et al. 1998), ce marais<br />

présente également un fort intérêt pour la faune piscicole, et plus précisément l’anguille<br />

européenne (Anguilla anguilla) et le brochet (Esox lucius), espèces patrimoniales toutes <strong>de</strong>ux<br />

classées vulnérables dans la liste rouge <strong>de</strong>s espèces menacées <strong>de</strong> France (Fiers et al. 1997).<br />

L’anguille est une espèce migratrice dont la phase juvénile se déroule en eau douce. La<br />

localisation géographique <strong>de</strong>s marais <strong>de</strong> Brière en fait une zone <strong>de</strong> croissance privilégiée pour<br />

cette espèce. Le brochet, dont la reproduction se déroule préférentiellement sur <strong>de</strong>s prairies<br />

temporairement inondées trouve dans les marais <strong>de</strong> Brière une large surface d’habitat<br />

correspondant à ses exigences écologiques. Le façonnement <strong>de</strong>s marais briérons et leur<br />

richesse biologique sont le fruit d’une histoire très particulière entre l’homme et son milieu<br />

par l’influence <strong>de</strong> ses activités que ce soit pour se loger (fauche <strong>de</strong>s roseaux pour les toits <strong>de</strong><br />

chaume), se chauffer (extraction <strong>de</strong> la tourbe) ou se nourrir (chasse, pêche et élevage sur les<br />

prairies inondables). Cependant <strong>de</strong>puis le milieu du XX ème siècle, la régression <strong>de</strong> bon nombre<br />

<strong>de</strong> ces usages a favorisé l’expansion du roseau (Phragmites australis) et du carex (Carex elata)<br />

contribuant à la fermeture et l’uniformisation progressive du milieu (Figure 3.2). Ainsi en<br />

Gran<strong>de</strong> Brière Mottière (7000 ha), le taux <strong>de</strong> recouvrement par le roseau est passé <strong>de</strong> 5 à 85 %<br />

<strong>de</strong> la surface totale du marais <strong>de</strong>puis 1850. Inversement, la surface <strong>de</strong> prairies inondables est<br />

quant à elle passée <strong>de</strong> 83 à 10 %, le territoire restant étant occupé par <strong>de</strong>s eaux libres (Bernard<br />

& Rolland 1990, Figure 3.2). La préservation et la restauration <strong>de</strong> ce patrimoine écologique et<br />

culturel très riche constituent l’un <strong>de</strong>s enjeux majeurs du Parc. Si l’avifaune a fait l’objet d’un<br />

suivi relativement régulier (Constant 1972, Constant et al. 1987, Eybert et al. 1998) en<br />

revanche, le peuplement piscicole manque d’une actualisation <strong>de</strong>s données.<br />

Abondancerelative Abondancerelative (%)<br />

100%<br />

80%<br />

60%<br />

40%<br />

20%<br />

0%<br />

Eau libre Prairie Roselière<br />

1850-1900 1940-1948 1980-1985<br />

Pério<strong>de</strong><br />

Figure 3.2. Evolution <strong>de</strong> la proportion <strong>de</strong> chaque type <strong>de</strong> milieu temporairement inondé lors <strong>de</strong>s<br />

150 <strong>de</strong>rnières années. La surface totale représente 7000 ha (d’après Bernard & Rolland 1990).<br />

43


Chapitre 3<br />

En effet, la <strong>de</strong>rnière étu<strong>de</strong> du peuplement piscicole sur l’ensemble <strong>de</strong>s marais <strong>de</strong> Brière<br />

a été réalisée en 1972 par Y. Maillard, qui soulignait déjà à l’époque le manque <strong>de</strong><br />

connaissance sur la faune piscicole (« les publications antérieures traitant <strong>de</strong> notre faune<br />

ichtyologique régionale ne fournissent aucune information sur le bassin du Brivet »). Depuis,<br />

seules quelques étu<strong>de</strong>s ponctuelles (dans le temps et l’espace) ont été réalisées. Ainsi, le<br />

Conseil Supérieur <strong>de</strong> la Pêche (1985) a échantillonné le peuplement piscicole sur la zone du<br />

Brivet (partie Est <strong>de</strong>s marais <strong>de</strong> Brière). Puis Feunteun & Constant (1992) et Eybert et al.<br />

(1997) ont étudié les «réserves» <strong>de</strong> pêche et chasse <strong>de</strong> Gran<strong>de</strong> Brière Mottière créées en 1973<br />

et 1989 par la Commission Syndicale <strong>de</strong> Gran<strong>de</strong> Brière Mottière, gestionnaire du marais<br />

indivis. Plus récemment, Carpentier et al. (2001 & 2002) ont échantillonné quelques zones<br />

temporairement inondées dans le cadre d’un programme <strong>de</strong> réouverture du milieu par<br />

gyrobroyage <strong>de</strong> cariçaies. Ainsi, un important déficit <strong>de</strong> connaissances sur la faune piscicole<br />

restait à combler sur les marais <strong>de</strong> Brière, compte tenu <strong>de</strong>s profonds changements<br />

écologiques qu’ils ont subis.<br />

C’est dans ce contexte qu’une étu<strong>de</strong> a été confiée en 2004 à l’<strong>Université</strong> <strong>de</strong> <strong>Rennes</strong> 1<br />

par le Parc naturel régional <strong>de</strong> Brière afin <strong>de</strong> caractériser le peuplement piscicole et d’analyser<br />

son état actuel en comparaison avec son statut passé. Cet article a donc pour objectif <strong>de</strong><br />

dresser un bilan <strong>de</strong>s connaissances acquises sur le peuplement piscicole <strong>de</strong>s marais <strong>de</strong> Brière<br />

et <strong>de</strong> le mettre en relation avec l’évolution <strong>de</strong> la structure <strong>de</strong> l’habitat et <strong>de</strong>s mesures <strong>de</strong><br />

restauration actuellement entreprises. Les poissons ont été échantillonnés en 2004 par pêche<br />

électrique qui permet la capture et l’i<strong>de</strong>ntification sans mortalité induite.<br />

3.2. Composition actuelle du peuplement piscicole dans les marais <strong>de</strong> Brière<br />

Le peuplement piscicole <strong>de</strong> Brière a été échantillonné pendant la pério<strong>de</strong> d’étiage (août) dans<br />

27 sites distribués <strong>de</strong> manière homogène sur l’ensemble du marais. Nous avons capturé 11<br />

949 individus appartenant aux différentes espèces reportées dans le tableau 3.1. En termes<br />

d’effectif, le peuplement est nettement dominé par le poisson-chat (Ameiurus melas) qui<br />

représente en moyenne 44.7 % <strong>de</strong>s captures. Les <strong>de</strong>ux espèces <strong>de</strong> brèmes communes<br />

(Abramis brama) et bor<strong>de</strong>lières (Blicca bjoerkna) composent à elles <strong>de</strong>ux 22.7 % <strong>de</strong>s<br />

individus échantillonnés. Suivent ensuite le gardon (Rutilus rutilus), le rotengle (Scardinius<br />

erythrophthalmus) et la gambusie (Gambusia holbrooki) qui représentent respectivement 12.0<br />

%, 6.0 % et 5.7 % <strong>de</strong>s captures. Les 14 autres espèces représentent individuellement moins <strong>de</strong><br />

5 % <strong>de</strong>s effectifs du peuplement et 9 % lorsqu’on les regroupe. Il est intéressant <strong>de</strong> noter que<br />

44


Biodiversité piscicole <strong>de</strong> Brière<br />

les <strong>de</strong>ux espèces emblématiques <strong>de</strong> la Brière, à savoir l’Anguille (Anguilla anguilla) et le<br />

Brochet (Esox lucius), représentent respectivement seulement 3.0 % et 0.1 % <strong>de</strong>s individus<br />

échantillonnés.<br />

Tableau 3.1. Liste <strong>de</strong>s espèces piscicoles capturées dans les marais <strong>de</strong> Brière lors <strong>de</strong>s principales<br />

campagnes d’échantillonnage <strong>de</strong> 1972 à 2004. Sources : Maillard (1972), Conseil Supérieur <strong>de</strong> la<br />

Pêche (1985) et la présente étu<strong>de</strong> (2004). + : espèce présente, - : espèce non recensée.<br />

Présence<br />

Nom scientifique Nom vernaculaire Origine 1972 1985 2004<br />

Abramis brama Brème commune Autochtone + + +<br />

Alburnus alburnus Ablette Autochtone - - +<br />

Ameiurus melas Poisson-chat Introduite + + +<br />

Anguilla anguilla Anguille européenne Autochtone + + +<br />

Blicca bjoerkna Brème bor<strong>de</strong>lière Autochtone + + +<br />

Carassius gibelio Carassin argenté Introduite - + +<br />

Ctenopharyngodon i<strong>de</strong>lla Amour blanc Introduite - - +<br />

Cyprinus carpio Carpe commune Introduite + + +<br />

Esox lucius Brochet Autochtone + + +<br />

Gambusia holbrooki Gambusie Introduite - - +<br />

Gasterosteus aculeatus Epinoche Autochtone + + +<br />

Pungitius pungitius Epinochette Autochtone + + -<br />

Lepomis gibbosus Perche soleil Introduite + + +<br />

Leuciscus cephalus Vandoise Autochtone + - -<br />

Leuciscus leuciscus Chevaine Autochtone + + -<br />

Micropterus salmoi<strong>de</strong>s Black-bass Introduite + + +<br />

Perca fluviatilis Perche franche Autochtone + + +<br />

Pseudorasbora parva Pseudorasbora Introduite - - +<br />

Rho<strong>de</strong>us amarus Bouvière Autochtone - - +<br />

Rutilus rutilus Gardon Autochtone + + +<br />

Scardinius erythrophthalmus Rotengle Autochtone + + +<br />

San<strong>de</strong>r lucioperca Sandre Introduite + + +<br />

Tinca tinca Tanche Autochtone + + +<br />

45


Chapitre 3<br />

L’analyse <strong>de</strong> la distribution spatiale <strong>de</strong>s poissons-chats fait ressortir que cette espèce est<br />

plus abondante dans un réseau assez <strong>de</strong>nse <strong>de</strong> canaux <strong>de</strong> petite taille, <strong>de</strong> faible profon<strong>de</strong>ur et<br />

bordés <strong>de</strong> roselière. L’ensemble <strong>de</strong> ces caractéristiques définit typiquement l’abandon <strong>de</strong><br />

l’exploitation <strong>de</strong>s marais <strong>de</strong> Brière provoquant une fermeture globale du milieu par une<br />

réduction <strong>de</strong> la section hydraulique et un envahissement progressif par les roselières.<br />

La prospection <strong>de</strong> l’ichtyofaune ne s’est pas bornée qu’aux canaux et plans d’eau <strong>de</strong>s<br />

marais <strong>de</strong> Brière mais a également inclus, en avril 2004, les zones temporairement inondées<br />

représentées par les prairies et les roselières. En effet, occupant une surface très importante en<br />

Brière, ces milieux sont inondés en général <strong>de</strong> la fin <strong>de</strong> l’automne à la fin du printemps et<br />

jouent un rôle fondamental dans le déroulement du cycle <strong>de</strong> vie <strong>de</strong> nombreuses espèces <strong>de</strong><br />

poissons. Ces milieux constituent vraisemblablement un refuge contre la prédation pour les<br />

cyprinidés, <strong>de</strong>s zones <strong>de</strong> croissance pour les anguilles, <strong>de</strong>s nourriceries pour les alevins mais<br />

également <strong>de</strong>s zones <strong>de</strong> reproduction pour le brochet ou la carpe.<br />

La comparaison <strong>de</strong>s communautés <strong>de</strong> poissons sur les prairies et dans les roselières<br />

montre une richesse spécifique similaire dans ces <strong>de</strong>ux milieux mais une <strong>de</strong>nsité piscicole<br />

supérieure dans les roselières provenant essentiellement <strong>de</strong> la forte abondance du poisson-chat<br />

dans ce milieu. Ainsi, à l’échelle globale <strong>de</strong>s marais briérons, l’augmentation <strong>de</strong> l’abondance<br />

du poisson-chat peut être mise en relation avec l’abandon <strong>de</strong>s pratiques traditionnelles<br />

d’exploitation du marais. En effet, l’expansion <strong>de</strong>s roselières semble être, d’après nos<br />

résultats, un élément susceptible <strong>de</strong> favoriser la prolifération <strong>de</strong>s poissons-chats (Chapitre 8).<br />

3.3. Evolution du peuplement piscicole <strong>de</strong>puis 30 ans<br />

Composé <strong>de</strong> 17 espèces en 1972, le peuplement piscicole en comprend actuellement 20<br />

(Tableau 3.1) : 11 espèces autochtones et 9 allochtones. Maillard (1972) n’avait échantillonné<br />

que 5 espèces allochtones (carpe (Cyprinus carpio), poisson-chat, perche soleil (Lepomis<br />

gibbosus), sandre (San<strong>de</strong>r lucioperca) et black-bass (Micropterus salmoi<strong>de</strong>s)) pour 12<br />

autochtones (Figure 3.3). Trois espèces autochtones ont disparu au cours <strong>de</strong> ces trente<br />

<strong>de</strong>rnières années : le chevaine (Leuciscus leuciscus) et l’épinochette (Pungitius pungitius)<br />

n’ont pas été échantillonnés <strong>de</strong>puis 1985, la vandoise (Leuciscus cephalus) <strong>de</strong>puis 1972, ces<br />

<strong>de</strong>ux <strong>de</strong>rnières espèces, sensibles à la qualité <strong>de</strong> l’eau, ne trouvent certainement plus les<br />

conditions favorables à leur développement dans les eaux <strong>de</strong> Brière. A l’inverse, la bouvière<br />

(Rho<strong>de</strong>us amarus) et l’ablette (Alburnus aburnus), espèces autochtones, ont été nouvellement<br />

échantillonnées en 2004. Cependant cette <strong>de</strong>rnière semble avoir fait l’objet d’un déversement<br />

acci<strong>de</strong>ntel et n’a été échantillonnée que <strong>de</strong> manière très localisée.<br />

46


Abondancerelative (%)<br />

100%<br />

80%<br />

60%<br />

40%<br />

20%<br />

0%<br />

Autochtone Introduite<br />

n=5 n=6<br />

n=12 n=11<br />

1972 1985 2004<br />

Biodiversité piscicole <strong>de</strong> Brière<br />

n=9<br />

n=11<br />

Année<br />

Figure 3.3. Evolution <strong>de</strong> la composition du peuplement piscicole (espèces autochtones et<br />

introduites) dans les marais <strong>de</strong> Brière (n = nombre d’espèce).<br />

Pendant la même pério<strong>de</strong>, <strong>de</strong>s espèces introduites, plus tolérantes comme le carassin<br />

(Carassius gibelio), le pseudorasbora (Pseudorasbora parva), l’amour blanc<br />

(Ctenopharyngodon i<strong>de</strong>lla) et la gambusie, ont fait leur apparition. L’apparition du poissonchat<br />

en Brière a été datée par Maillard (1972) <strong>de</strong> 1929, celle <strong>de</strong> la perche soleil du début du<br />

XXème siècle, celle du sandre <strong>de</strong> 1967 et enfin celle du black-bass <strong>de</strong> 1960. Ces <strong>de</strong>ux<br />

<strong>de</strong>rnières espèces ont été introduites pour la pêche sportive. L’observation du carassin, cousin<br />

<strong>de</strong> la carpe et du poisson rouge, semble plus récente, cette espèce étant décrite pour la<br />

première fois en Brière en 1985 par le Conseil Supérieur <strong>de</strong> la Pêche. Le carassin est<br />

considéré comme potentiellement indésirable du fait <strong>de</strong> sa capacité <strong>de</strong> prolifération (Keith &<br />

Allardi 2001). La gambusie, originaire du sud <strong>de</strong>s Etats-Unis, a été introduite en Europe pour<br />

lutter contre les moustiques (Keith & Allardi 2001). Son arrivée en Brière semble s’être<br />

produite à la fin <strong>de</strong>s années 90 puisque, non répertoriée en 1985, elle a été capturée pour la<br />

première fois en 2000 (Carpentier et al. 2001). Le pseudorasbora, poisson originaire d’Asie, a<br />

été introduit acci<strong>de</strong>ntellement en Europe (Roumanie) en 1960 par le biais d’alevinages <strong>de</strong><br />

carpes, et s’est ensuite propagé en France, <strong>de</strong> façon volontaire et/ou acci<strong>de</strong>ntelle, ainsi que<br />

grâce à sa capacité à coloniser <strong>de</strong> nouveaux habitats. Ce poisson est classé comme nuisible à<br />

l’échelle mondiale principalement à cause <strong>de</strong>s risques <strong>de</strong> compétition avec d’autres espèces et<br />

notamment la consommation <strong>de</strong> leurs pontes (Keith & Allardi 2001). Il a été observé pour la<br />

première fois dans les marais <strong>de</strong> Brière en 2004. Enfin l’amour blanc, originaire d’Asie, a été<br />

introduit en France en 1957 et est commercialisé pour lutter contre l’envahissement <strong>de</strong>s<br />

milieux dulcicoles par la végétation aquatique. Cette espèce a cependant peu d’intérêt dans la<br />

limitation <strong>de</strong>s algues filamenteuses <strong>de</strong> gran<strong>de</strong> taille mais aurait par contre un rôle important<br />

47


Chapitre 3<br />

dans la <strong>de</strong>struction <strong>de</strong>s frayères d’autres espèces. Cependant, aucune reproduction n’a été<br />

jusqu’à présent signalée en milieu naturel en France (Keith & Allard 2001). L’amour blanc a<br />

été capturé pour la première fois dans les marais <strong>de</strong> Brière en 2004, sans que l’on connaisse à<br />

ce jour son niveau d’abondance qui semble cependant très limité.<br />

3.4. Discussion et perspectives <strong>de</strong> gestion<br />

La diversité du peuplement <strong>de</strong> poissons peuplant les marais briérons est, comme dans<br />

l’ensemble <strong>de</strong>s zones humi<strong>de</strong>s, le produit d’introductions successives : quatre espèces dont<br />

l’arrivée est antérieure aux années 70, notées par Maillard (1972), <strong>de</strong>ux espèces dont les<br />

arrivées sont postérieures et signalées par le Conseil Supérieur <strong>de</strong> la Pêche et cette étu<strong>de</strong>.<br />

L’augmentation sensible du nombre d’espèces <strong>de</strong> poissons que nous avons observée <strong>de</strong>puis<br />

ces trente <strong>de</strong>rnières années provient donc <strong>de</strong> l’arrivée <strong>de</strong> ces espèces allochtones qui, pour la<br />

plupart, manifestent une gran<strong>de</strong> amplitu<strong>de</strong> écologique. Cette augmentation est aussi le produit<br />

<strong>de</strong> l’arrivée <strong>de</strong> <strong>de</strong>ux espèces autochtones, malgré la disparition <strong>de</strong> trois autres probablement<br />

en liaison avec la diminution <strong>de</strong> la qualité <strong>de</strong> l’eau. Ainsi, même si nous constatons que<br />

l’ichtyofaune <strong>de</strong>s marais briérons gar<strong>de</strong> une biodiversité relativement élevée, l’évolution <strong>de</strong> sa<br />

composition et la faible représentation d’espèces <strong>de</strong> fort intérêt patrimonial comme le brochet<br />

et l’anguille, nous amènent à nous poser <strong>de</strong> sérieuses questions quant à son avenir.<br />

L’extension <strong>de</strong>s roselières, la diminution consécutive <strong>de</strong>s surfaces prairiales inondables,<br />

l’existence d’écluses qui freinent certainement la remontée <strong>de</strong>s anguilles et d’autres facteurs<br />

non évalués tel que l’impact <strong>de</strong>s pratiques halieutiques, qui agissent probablement en synergie,<br />

semblent donc constituer <strong>de</strong>s menaces réelles pour le peuplement piscicole <strong>de</strong>s marais<br />

briérons.<br />

Les pratiques agricoles, par la pollution, les pompages ou l’aménagement <strong>de</strong>s berges<br />

qu’elles engendrent, ont souvent été désignées comme <strong>de</strong>s facteurs importants affectant les<br />

milieux aquatiques et contribuant <strong>de</strong> façon significative à la dégradation <strong>de</strong>s peuplements<br />

piscicoles. Cependant, dans les systèmes mis en place et entretenus par l’homme, tels que les<br />

marais brièrons, la déprise agricole a eu pour effet <strong>de</strong> fermer le milieu en laissant se<br />

développer le roseau et ainsi conduire à uniformisation progressive <strong>de</strong>s habitats. Cette<br />

expansion a vraisemblablement eu pour effet <strong>de</strong> favoriser, <strong>de</strong> manière corrélative la<br />

diminution du brochet et l’expansion du poisson-chat qui semble être un facteur limitant la<br />

diversité du peuplement piscicole. Il reste désormais à savoir si les programmes <strong>de</strong><br />

réhabilitation <strong>de</strong>s prairies inondables et la remise en pratique du pâturage qui font partie <strong>de</strong>s<br />

48


Biodiversité piscicole <strong>de</strong> Brière<br />

objectifs affichés dans la charte 2000-2010 du Parc naturel régional <strong>de</strong> Brière auront un effet<br />

significatif sur la richesse et la diversité du peuplement piscicole comme cela a été le cas pour<br />

l’avifaune (Eybert et al. 1997). Cependant, l’arrivée récente d’espèces invasives en Brière<br />

(telles que le carassin ou le pseudorasbora), dont les exigences écologiques, les capacités<br />

d’expansion et le niveau d’interaction avec les autres espèces par prédation, transmission <strong>de</strong><br />

parasites ou <strong>de</strong> maladies et compétition (e.g. Gozlan et al. 2005) sont totalement méconnus,<br />

risque <strong>de</strong> poser <strong>de</strong> nouveaux problèmes dans un futur plus ou moins proche.<br />

De même, l’introduction récente <strong>de</strong> l’écrevisse <strong>de</strong> Louisiane (Procambarus clarkii) a<br />

probablement affecté le peuplement piscicole en modifiant par exemple la composition <strong>de</strong> la<br />

communauté d’invertébrés aquatiques ou <strong>de</strong> macrophytes comme cela a été montré en<br />

Espagne par Rodriguez et al. (2005). De même il est important <strong>de</strong> noter l’absence dans notre<br />

échantillonnage <strong>de</strong> nombreuses espèces amphihalines citées par Maillard (1972), certainement<br />

liée aux changements dans la gestion hydraulique <strong>de</strong> la zone d’étu<strong>de</strong>. Ainsi, malheureusement,<br />

les dysfonctionnements <strong>de</strong>s marais <strong>de</strong> Brière du point <strong>de</strong> vue piscicole, décrits déjà en 1972<br />

par Maillard « comme si la Brière était <strong>de</strong>venue <strong>de</strong> moins en moins capable d’entretenir une<br />

faune ichtyologique abondante et prospère», sont toujours d’actualité.<br />

Remerciements<br />

Nous remercions le Parc naturel régional <strong>de</strong> Brière pour son ai<strong>de</strong> logistique lors <strong>de</strong> la conduite<br />

<strong>de</strong>s opérations <strong>de</strong> terrain et le FEDER, la DIREN Pays <strong>de</strong> la Loire, la Région Pays <strong>de</strong> la Loire,<br />

l’Agence <strong>de</strong> l’Eau Loire - Bretagne et le Parc naturel régional <strong>de</strong> Brière pour leur soutien<br />

financier. De même, nous tenons à remercier toutes les personnes nous ayant permis la<br />

réalisation <strong>de</strong>s opérations <strong>de</strong> terrain à savoir Messieurs les Prési<strong>de</strong>nts <strong>de</strong> la Commission<br />

Syndicale <strong>de</strong> Gran<strong>de</strong> Brière Mottière, du Syndicat Mixte d’Aménagement Hydraulique du<br />

Brivet et <strong>de</strong> l’AAPPMA <strong>de</strong> Pont-Château.<br />

49


Chapitre 3<br />

Des résultats obtenus, il ressort que le peuplement piscicole <strong>de</strong>s marais <strong>de</strong> Brière a<br />

subi <strong>de</strong> profonds changements en terme <strong>de</strong> composition spécifique au cours <strong>de</strong>s<br />

30 <strong>de</strong>rnières années. En plus <strong>de</strong> l’évolution paysagère forte notée lors du siècle<br />

<strong>de</strong>rnier, une modification <strong>de</strong> la gestion hydrologique du marais, et plus<br />

particulièrement <strong>de</strong> son niveau <strong>de</strong> connectivité avec l’estuaire <strong>de</strong> la Loire qui<br />

limite au minimum les échanges hydrologiques et biologiques, a fait perdre au<br />

peuplement piscicole un grand nombre <strong>de</strong> ses espèces amphihalines.<br />

Concrètement, <strong>de</strong>s espèces telles que le flet, les gobies ou les mulets sont absentes<br />

<strong>de</strong> notre échantillonnage. Seule l’anguille, espèce catadrome, est encore présente.<br />

En termes d’espèces strictement dulçaquicoles, le peuplement a aussi subi <strong>de</strong>s<br />

changements avec notamment une augmentation du nombre d’espèces allochtones<br />

(5 à 9) et une stabilisation du nombre d’espèces autochtones. La population <strong>de</strong><br />

brochet, autre espèce emblématique <strong>de</strong>s marais <strong>de</strong> Brière, apparaît aussi comme<br />

proportionnellement très faible. De manière plus nette encore, une espèce<br />

allochtone, le poisson chat, domine le peuplement et certaines hypothèses<br />

relatives à l’introduction <strong>de</strong> l’écrevisse <strong>de</strong> Louisiane et à la modification <strong>de</strong> la<br />

nature <strong>de</strong>s MTI ont été émises. Dans la suite <strong>de</strong> ce travail, nous nous proposons<br />

d’analyser les modalités d’utilisation <strong>de</strong>s MTI par les espèces décrites ici et les<br />

conséquences possibles sur la composition <strong>de</strong> la communauté piscicole. En effet,<br />

les MTI sont exploités lors <strong>de</strong> l’inondation par la communauté piscicole qui est<br />

confinée dans les milieux permanents à l’étiage. Les patrons <strong>de</strong> cette utilisation<br />

(sélectivité d’habitat et distribution spatiale <strong>de</strong>s poissons au sein <strong>de</strong>s MTI) seront<br />

étudiés avec une double approche, synchronique (à large échelle à un moment<br />

précis du cycle d’inondation, Chapitre 4) puis approche diachronique (à fine<br />

échelle lors d’un cycle complet d’inondation, Chapitre 5).<br />

50


Chapitre 4<br />

Approche synchronique <strong>de</strong> l’utilisation <strong>de</strong>s milieux<br />

temporairement inondés<br />

L’abondance <strong>de</strong>s différentes espèces dans les MTI est comparée à celle<br />

inventoriée dans le réseau <strong>de</strong> canaux et ce, à l’échelle <strong>de</strong> l’ensemble <strong>de</strong>s marais <strong>de</strong><br />

Brière afin d’examiner le niveau <strong>de</strong> colonisation (abondance et distribution<br />

spatiale) <strong>de</strong> ces milieux par l’ichtyofaune. Ces paramètres sont aussi analysés afin<br />

d’étudier une éventuelle variabilité interspécifique. La comparaison <strong>de</strong> ces<br />

processus est aussi analysée dans les <strong>de</strong>ux principaux types <strong>de</strong> MTI, à savoir les<br />

prairies et les roselières. Cette approche est d’autant plus pertinente que les<br />

connaissances sur ces processus sont très limitées dans la littérature. De plus, les<br />

différentes étu<strong>de</strong>s disponibles ne prennent en général en considération que la seule<br />

communauté piscicole <strong>de</strong>s MTI, sans tenir compte du fait qu’elle provient<br />

nécessairement <strong>de</strong> la communauté « totale » présente dans les milieux<br />

permanents.


Chapitre 4<br />

Echantillonnage d’une prairie temporairement inondée<br />

Echantillonnage d’une roselière temporairement inondée<br />

52


MTI et poissons – approche synchronique<br />

Selective use and spatial distribution of fish onto temporary<br />

floo<strong>de</strong>d habitats 2<br />

Résumé : Les milieux temporairement inondés (MTI) ont un rôle fondamental pour<br />

l’accomplissement du cycle <strong>de</strong> vie (reproduction, croissance et nourricerie) <strong>de</strong> nombreuses<br />

espèces <strong>de</strong> poissons. Cependant, les principaux facteurs qui influencent l’utilisation <strong>de</strong> ces<br />

milieux par l’ichtyofaune restent mal connus. En tant que zone transitionnelle entre<br />

écosystèmes terrestres et aquatiques, les MTI ont été particulièrement exposés aux effets<br />

néfastes <strong>de</strong> nombreuses activités anthropiques, dont les conséquences sur les MTI ont<br />

rarement été analysées. Dans cette étu<strong>de</strong>, nous avons pour objectifs <strong>de</strong> comparer la<br />

composition <strong>de</strong> la communauté piscicole utilisant les MTI à celle présente dans les milieux<br />

permanents (MP) adjacents et d’examiner les patrons <strong>de</strong> distribution <strong>de</strong> l’ichtyofaune dans les<br />

MTI. Dans les marais <strong>de</strong> Brière où cette étu<strong>de</strong> a été conduite, les MTI sont composés <strong>de</strong><br />

prairies (milieu naturel) et <strong>de</strong> roselières (milieu envahi) alors que la communauté piscicole est<br />

composée d’espèces natives et d’espèces introduites. Dans cette étu<strong>de</strong>, nous allons donc<br />

profiter <strong>de</strong> cette opportunité afin d’étudier les processus d’utilisation <strong>de</strong>s MTI par<br />

l’ichtyofaune en nous intéressant aussi au contexte <strong>de</strong>s invasions biologiques. Tous les MTI<br />

que nous avons échantillonné étaient colonisés par <strong>de</strong>s poissons et les caractéristiques <strong>de</strong> la<br />

communauté piscicole (abondance et composition) étaient très différentes entre les MP et les<br />

MTI à cause <strong>de</strong> patrons d’utilisation <strong>de</strong>s MTI variables entre espèces. Nos résultats ont aussi<br />

démontré que la distance au MP est le principal facteur qui influence la distribution spatiale<br />

<strong>de</strong>s poissons dans les MTI, les abondances étant plus fortes dans les zones proches <strong>de</strong>s MP.<br />

Nous avons aussi observé que les espèces introduites <strong>de</strong> poisson étaient plus abondantes et<br />

colonisaient <strong>de</strong> plus gran<strong>de</strong>s distances dans les roselières que dans les prairies, alors que les<br />

espèces natives ne présentaient pas <strong>de</strong> différences entre les <strong>de</strong>ux types <strong>de</strong> MTI. Ces résultats<br />

permettent d’améliorer la compréhension <strong>de</strong>s mécanismes écologiques se déroulant dans les<br />

écosystèmes spatialement variables et confirment l’existence possible du processus <strong>de</strong><br />

facilitation entre plantes et espèces piscicoles introduites.<br />

2 Ce chapitre est adapté d’un article intitulé « Cucherousset, J., Carpentier, A. & Paillisson, J.-M. Selective use<br />

and spatial distribution of fish onto temporary floo<strong>de</strong>d habitats » soumis à la revue « Canadian Journal of<br />

Fisheries and Aquatic Sciences ».<br />

53


Chapitre 4<br />

Abstract: Temporary floo<strong>de</strong>d habitats (TFH) are crucial for many freshwater fish<br />

species that use them as spawning, feeding or nursery areas <strong>de</strong>pending upon their respective<br />

life cycle. However, it remains unclear as to the i<strong>de</strong>ntification of the major factors that<br />

influence the use of TFH by fish. As transitional zones between aquatic and terrestrial zones,<br />

TFH have particularly suffered from many human disturbances during recent <strong>de</strong>ca<strong>de</strong>s. The<br />

effects of these disturbances on TFH use have been rarely investigated. In the present study,<br />

we aim to compare the composition of the fish community that colonized TFH with the<br />

community inhabiting the adjacent permanent habitats (PH) and to examine the distributional<br />

patterns of fish species onto the TFH. In the Brière marsh we studied, some TFH are natural<br />

(grassland) while the others are inva<strong>de</strong>d (reed bed) and the fish community is composed of<br />

native and introduced species. Here, we also examined the use of TFH by fish in the context<br />

of biological invasions. Our results <strong>de</strong>monstrated that all TFH were colonized by fish and that<br />

fish community characteristics (abundance and composition) were different between PH and<br />

TFH as a consequence of variable propensity to use TFH among species. We also observed<br />

that the distance to PH was the main factor affecting the spatial abundance of fish in TFH,<br />

with greater abundance observed close to PH. Regarding TFH type (grassland vs. reed bed)<br />

and fish species origin (native vs. introduced), we found that introduced fish species were<br />

more abundant and colonized further distance in the reed beds than in the grassland, whereas<br />

native fish species exhibited no differences in habitat use between the two types of TFH. Our<br />

results improved the un<strong>de</strong>rstanding of ecological mechanisms occurring in this spatially<br />

varying environment and confirmed the existence of possible facilitative interactions between<br />

invasive plant and introduced fish species.<br />

54


4.1. Introduction<br />

MTI et poissons – approche synchronique<br />

Temporary floo<strong>de</strong>d habitats (hereafter called TFH), <strong>de</strong>fined here as habitats that are<br />

seasonally inundated and connected to permanents habitats (hereafter called PH) by surface<br />

waters, have been reported as playing fundamental functions for a broad range of plant and<br />

animal communities (see review in Williams 2006). Specifically, TFH are key habitats for<br />

many fish species (e.g. Snodgrass et al. 1996, Magoulick 2000, King et al. 2003, DeAngelis et<br />

al. 2005, Trexler et al. 2005). In<strong>de</strong>ed, <strong>de</strong>pending on their life history, some fish species are<br />

totally restricted to PH whereas other species use TFH as spawning, feeding or nursery areas.<br />

Based on their behavioural responses to flooding regime (magnitu<strong>de</strong>, frequency and duration),<br />

fish species that use these habitats have been classified by Welcomme (1979) into two groups:<br />

(1) fish that avoid environmental fluctuations by migrating to and from the main river channel,<br />

using PH as refuges (see review in Magoulick & Kobza 2003), and (2) fish that are able to<br />

survive wi<strong>de</strong>ly fluctuating conditions (Welcomme 1979). Although many species belong to<br />

this latter category, very few temperate species are adapted to survive a complete loss of<br />

water from their habitat (Sayer 2005, Williams 2006), and almost fish species will have to<br />

colonize TFH. Several studies have already addressed the ecological processes that lead to the<br />

relationships between fish and TFH, such as: fish community composition in the TFH,<br />

changes in fish community structure throughout the flooding period, and the intensity of fish<br />

movements between TFH and the adjacent PH (e.g. Poizat & Crivelli 1997, Magoulick 2000,<br />

Baber et al. 2002, Hohausova et al. 2003, Carpentier et al. 2004, Crain et al. 2004). However,<br />

it remains unclear which factors influence the use of TFH by fish (Baber et al. 2002). For<br />

instance, the selective use of TFH by fish and the spatial distribution of fish on these habitats<br />

have been rarely addressed (but see Rehage & Trexler 2006). This question is of great interest,<br />

notably in the framework of some recent conceptual <strong>de</strong>velopments in freshwater ecology<br />

(Baber et al. 2002, Morris 2003), such as the mechanistic approach of the coexistence of<br />

animal species in harsh and spatially-varying environments (Chesson 2000).<br />

As ecological transition zones between aquatic and terrestrial ecosystems, TFH are<br />

particularly sensitive to various human disturbances and have suffered severe <strong>de</strong>gradations<br />

from resource exploitation and harvesting, urban and agricultural encroachment and global<br />

climate change (Tockner & Stanford 2002, Williams 2006). These drastic changes have<br />

undoubtedly modified the fish community function in these habitats (e.g. Pont et al. 2006).<br />

Furthermore, these human-caused environmental disturbances and habitat modifications are<br />

known to facilitate fish invasions greatly by unbalancing native fish faunas in favour of more<br />

55


Chapitre 4<br />

tolerant introduced species (Moyle & Light 1996, Marchetti et al. 2004, Alcaraz et al. 2005).<br />

Nevertheless, the consequences of human activities on the use of TFH by fish with regards to<br />

the species status (i.e. native or introduced) have been rarely addressed to our knowledge (but<br />

see Kobza et al. 2004). The Brière marsh (Northwest France, 7000 ha), a freshwater<br />

seasonally pulsed wetland characterized by the presence of vast TFH that flood and dry out<br />

every year, is an i<strong>de</strong>al site to investigate all of these questions. In<strong>de</strong>ed, this ecosystem has<br />

been subjected to drastic changes in its landscape and fish community during the last century.<br />

Reed beds that are mainly composed of the invasive Phragmites australis have largely<br />

colonized the wetland to the <strong>de</strong>triment of natural grasslands. Currently, reed beds cover 85%<br />

of the wetland surface area, whereas they covered only 2% at the beginning of the last century<br />

(Bernard & Rolland 1990). These drastic changes came from profound modifications in<br />

human activities, notably strong <strong>de</strong>creases of grazing, mowing and peat exploitation (Bernard<br />

& Rolland 1990). Furthermore, the local fish community is currently dominated by several<br />

introduced fish species (Chapitre 8).<br />

In the present study, we first compare the composition of the fish community that<br />

colonized TFH with the community inhabiting PH to analyse the attractiveness (species<br />

richness, abundance and occurrence) of TFH. Second, we analyse spatial distribution patterns<br />

of fish species that colonized TFH (i.e. the effect of the distance from PH on the presence of<br />

fish onto TFH). In<strong>de</strong>ed, if the distance from PH influences the spatial distribution of fish onto<br />

TFH, we would expect a <strong>de</strong>crease in the presence probability of fish when the distance to PH<br />

increases as a response to higher risks of colonizing areas far from the refuge habitats.<br />

Conversely, if this variable does not affect the spatial distribution of fish onto TFH, higher<br />

risks would be are counterbalanced by increasing benefits. Furthermore, by taking advantages<br />

of the presence of two contrasted types of TFH (grasslands and reed beds) and fish origins<br />

(native and introduced) in the study site, we explore these mechanisms with regards to the<br />

context of biological invasions.<br />

4.2. Material and methods<br />

4.2.1. Study area and fish sampling<br />

The study area is <strong>de</strong>tailed in the Chapitre 2 (see Figure 4.1).<br />

56


Water level (French General Level, cm)<br />

N<br />

1 Km<br />

120<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

-20<br />

Sept.<br />

2003<br />

MTI et poissons – approche synchronique<br />

Brière<br />

Marsh<br />

Canal<br />

Nov. Jan. March May Jul. Sept. Dec. Jan. March May<br />

2003 2004 2004 2004 2004 2004 2005 2005 2005 2005<br />

Reed bed<br />

Grassland<br />

Sampled sites:<br />

Permanent<br />

habitat<br />

Reed bed<br />

Grassland<br />

Figure 4.1. Map of the Brière marsh showing the canal network, the two types of temporary<br />

floo<strong>de</strong>d habitats (grassland and reed bed) and the localization of sites sampled in 2004 and 2005<br />

(top). Water level (French General Level, cm) in the Brière marsh from September 2003 to August<br />

2005 with the sampling period of temporary floo<strong>de</strong>d (■) and permanent (■) habitats (bottom). The<br />

horizontal dotted line represents the minimal threshold of water level to inundate the temporary<br />

floo<strong>de</strong>d habitats (source: Parc naturel régional <strong>de</strong> Brière).<br />

57<br />

Jul.<br />

2005


Chapitre 4<br />

The fish community was sampled in two periods of 2004 and 2005 that were dictated<br />

by the seasonal configuration of habitats in relation to the flooding pattern, and health and<br />

safety issues associated with the sampling method (electrofishing). The sampling was<br />

<strong>de</strong>signed to obtain a global picture of the fish community in the whole wetland (7000 ha) at<br />

two key seasons (Figure 4.1). Eleven grasslands and seven reed beds were sampled in 2004<br />

and 2005 during the recession stage of water levels (Figure 4.1). Despite some differences in<br />

flooding pattern between years (Figure 4.1), fish sampling operations were conducted exactly<br />

at the same dates in 2004 and 2005 (i.e. from April 5-16 in 2004 and from April 4-16 in 2005,<br />

respectively) to limit a possible bias arising from a seasonal variations in the use of TFH by<br />

the fish community (see for instance Crain et al. (2004)). Because electrofishing is less<br />

effective in <strong>de</strong>ep waters, PH (n = 13, adjacent to one or two sampled TFH) were sampled in<br />

summer 2004 and 2005, when water level were lower. 665 PAS were performed in PH (i.e. on<br />

average 25.6 PAS per site ± 3.8 S.D.) and 1156 PAS in TFH (i.e. 32.1 PAS per site ± 6.0 S.D.)<br />

in total in 2004 and 2005. Fish sampling approach is <strong>de</strong>tailed in the Chapitre 2.<br />

4.2.2. Data analysis<br />

Since spawning periods differ between fish species, it was not possible to study accurately the<br />

fish community for all life stages in PH in summer. Therefore, young-of-the-year (YOY)<br />

individuals were removed from the dataset based on the distribution of specific size-classes<br />

(Carpentier et al. 2004, Chapitre 8). As sampling in TFH was conducted in early spring, all<br />

fish species (except pike Esox lucius) have not yet reproduced and all individuals sampled in<br />

TFH were thus used for further analyses. Fish abundance was expressed in catch per unit<br />

effort (CPUE; i.e. the number of individuals caught per PAS).<br />

Preliminary investigations <strong>de</strong>monstrated that no differences occurred in species<br />

richness and mean fish abundance (CPUE) of the entire fish community and also of each<br />

dominant fish species (see results section below) in PH and the two types of TFH between<br />

years (Wilcoxon-signed ranks test, P > 0.05). Therefore, no distinction in sampling year was<br />

inclu<strong>de</strong>d in the analyses. Comparisons of the fish community in the different habitats were<br />

performed at the site level (each TFH and its adjacent PH), whereas the investigation of the<br />

effect of the distance from PH on the abundance of fish in TFH was analysed at the PAS level.<br />

Since <strong>de</strong>viations from normality were found in the dataset of the fish community <strong>de</strong>scriptors,<br />

non-parametric Mann-Whitney U test were used to compare species richness and fish<br />

abundance between habitats. To <strong>de</strong>scribe the propensity of each species to use TFH relative to<br />

58


MTI et poissons – approche synchronique<br />

PH, the Ivlev’s electivity in<strong>de</strong>x E (Ivlev (1961) and see also its use in Zitek et al. (2004)) was<br />

calculated for each TFH and adjacent PH pair using the following formula: Ei = [TFHi (%) –<br />

PHi (%)]*[TFHi (%) + PHi (%)], where TFHi and PHi are the relative abundance of speciesi in<br />

the habitats. The resulting mean value was compared to a predicted mean value of 0 (neutral<br />

propensity) using the non-parametric Wilcoxon-signed rank test. When a statistical difference<br />

occurred, a positive mean Ei value indicates a “preference” for the type of TFH by the speciesi<br />

whereas a negative mean Ei value indicated an “avoidance” of the type of TFH by the speciesi.<br />

Because fish abundances in TFH were particularly low at the PAS level, they were<br />

transformed into presence/absence data. Consequently, the effect of the distance from PH on<br />

the presence of fish on TFH was investigated using logistic regression (i.e. presence/absence<br />

response curve; <strong>de</strong>tails in Trexler & Travis (1993)). We estimated the probability of a fish<br />

species or a group of species (i.e. native and introduced species) being present (p) as a<br />

function of a measured variable, x (see <strong>de</strong>tails for instance in Pont et al. (2005)) using the<br />

f ( x)<br />

⎛ p(<br />

x)<br />

⎞<br />

e<br />

log⎜ ⎟ = f ( x)<br />

⇔ p(<br />

x)<br />

= x<br />

⎝1<br />

− p(<br />

x)<br />

⎠<br />

1+<br />

e<br />

formula: f ( )<br />

where f(x) is a linear combination of the<br />

variable x. This equation can be exten<strong>de</strong>d to inclu<strong>de</strong> more than one explanatory variable. In<br />

the present study, we tested the effect of the distance to PH and the type of TFH (i.e.<br />

grassland and reed bed). We tested differences between the mo<strong>de</strong>ls (with variables and<br />

intercept) and the global null mo<strong>de</strong>l (intercept only) using the likelihood-ratio test based on<br />

the difference in <strong>de</strong>viance (D = -2 log (likelihood ratio)) between the two mo<strong>de</strong>ls. The<br />

significance of reduction in <strong>de</strong>viance was tested by comparing the observed value with a χ 2<br />

distribution. All statistical analyses were performed using STATISTICA software (released<br />

4.5, StatSoft Inc., Tulsa, OK). A rejection level of P > 0.05 was used in all tests.<br />

59


Chapitre 4<br />

4.3. Results<br />

4.3.1. Fish communities in permanent and temporary floo<strong>de</strong>d habitats<br />

In PH, we caught a total of 9496 individuals belonging to 15 species. The mean species<br />

richness was 8.73 (± 1.31 S.D.) species per PH (range: 6 to 11 species) and the mean fish<br />

abundance was 14.62 (± 8.67 S.D.) CPUE (range: 3.4 to 37.4). Seven species largely<br />

dominated the fish community in PH: black bullhead (Ameiurus melas), common bream<br />

(Abramis brama), mosquitofish (Gambusia holbrooki), roach (Rutilus rutilus), rudd<br />

(Scardinius erythrophalmus), European eel (Anguilla anguilla) and pumpkinseed (Lepomis<br />

gibbosus). They accounted for 99.1% of the total number of fish sampled (Table 4.1). Bream,<br />

rudd, roach and eel are native species whereas black bullhead, mosquitofish and pumpkinseed<br />

are introduced species. The others species accounted for 0.90% of the total number of fish<br />

sampled (see <strong>de</strong>tails in Table 4.1).<br />

All sampled TFH contained fish, resulting in the capture of 927 individuals belonging<br />

to the 12 aforementioned species (Table 4.1). The mean species richness in TFH was 4.66 (±<br />

2.10 S.D.) species per TFH (range: 1 to 9 species) and the mean fish abundance was 0.81 (±<br />

0.77 S.D.) CPUE (range: 0.02 to 3.0). The mean species richness was 1.9 times lower in TFH<br />

than in PH (Mann-Whitney test, U = 884, P < 0.001, n = 62) and the mean fish abundance<br />

was 18 times lower in TFH than in PH (Mann-Whitney test, U = 936, P < 0.001, n = 62).<br />

Nevertheless, no differences were found in mean species richness (Mann-Whitney test, U =<br />

131, P = 0.451, n = 36) and in mean fish abundance (Mann-Whitney test, U = 121, P = 0.284,<br />

n = 36) between grasslands and reed beds. Only three species sampled in PH have not been<br />

found in TFH (Perca fluviatilis, Gasterosteus aculeatus and San<strong>de</strong>r lucioperca). Six species<br />

largely dominated the fish community in the two types of TFH (i.e. black bullhead, bream,<br />

mosquitofish, rudd, pumpkinseed and eel) and represented 93.2% of the total catch (Table<br />

4.1). These species were used for further analyses.<br />

60


Table 4.1. Fish species (scientific name and co<strong>de</strong>), number of individuals (N) and percentage of site occupied (%) in the permanent habitats (n = 26) and in the<br />

two types of temporary floo<strong>de</strong>d habitats, i.e. grasslands (n=22) and reed beds (n = 14), sampled in 2004 and 2005 in the Brière marsh.<br />

Species Permanent habitats Temporary floo<strong>de</strong>d habitats<br />

Grasslands Reed beds<br />

Scientific name N Occurrence N Occurrence N Occurrence<br />

Ameiurus melas* 3903 100 % 105 68 % 138 100 %<br />

Abramis brama § 2013 100 % 95 55 % 22 50 %<br />

Gambusia holbrooki* 1211 100 % 9 23 % 32 43 %<br />

Rutilus rutilus 1040 100 % 19 36 % 11 36 %<br />

Scardinius erythrophalmus 564 100 % 222 95 % 86 100 %<br />

Lepomis gibbosus* 398 92 % 18 32 % 63 57 %<br />

Anguilla anguilla 282 100 % 41 59 % 33 71 %<br />

Carassius gibelio* 47 50 % 8 36 % 1 7 %<br />

Cyprinus carpio* 18 42 % 8 18 % 2 7 %<br />

Esox lucius 7 19 % 8 18 % 2 14 %<br />

Perca fluviatilis 5 19 % 0 0 % 0 0 %<br />

Gasterosteus aculeatus 4 12 % 0 0 % 0 0 %<br />

Tinca tinca 2 8 % 0 0 % 3 14 %<br />

Micropterus salmoi<strong>de</strong>s* 1 4 % 0 0 % 1 7 %<br />

San<strong>de</strong>r lucioperca* 1 4 % 0 0 % 0 0 %<br />

* <strong>de</strong>notes introduced species based on Keith and Allardi (2001); § some Blicca bjoerkna individuals were grouped with Abramis brama since they were not reliably<br />

i<strong>de</strong>ntifiable in the field.


Chapitre 4<br />

The numerical importance of these six dominant species was different between PH and<br />

each of the two types of TFH (grasslands (Spearman correlation, Rs = 0.314, P = 0.544, n =<br />

6); reed beds (Rs = 0.257, P = 0.622)). Mean abundance of black bullhead was significantly<br />

higher in reed beds than in grasslands (Mann-Whitney test, U = 71, P = 0.007, n = 36) and the<br />

mean abundance of pumpkinseed was marginally higher in reed beds than in grasslands<br />

(Mann-Whitney test, U = 106, P = 0.082, n = 36). No significant differences were observed<br />

for the other species (Mann-Whitney test, U > 115, P > 0.126, n = 36, Figure 4.2).<br />

Mean abundance (CPUE)<br />

0.5<br />

0.4<br />

0.3<br />

0.2<br />

0.1<br />

0<br />

NS<br />

**<br />

Rudd Black<br />

bullhead<br />

NS<br />

NS<br />

NS<br />

Bream Eel Pumpkinseed<br />

NS<br />

Mosquitofish<br />

Figure 4.2. Mean abundance (CPUE = number of individuals per PAS) of the six most abundant<br />

species sampled in the temporary floo<strong>de</strong>d grasslands (, n = 22) and reed beds (, n = 14) in 2004<br />

and 2005 in the Brière marsh. Errors-bars are standard error. NS: no significant and ** P < 0.01<br />

for Mann-Whitney U test of between temporary floo<strong>de</strong>d habitats comparisons. See <strong>de</strong>tails in Table<br />

4.1 for species co<strong>de</strong>s.<br />

4.3.2. Preference and avoidance of temporary floo<strong>de</strong>d habitats<br />

Investigations based on the Ivlev’s electivity in<strong>de</strong>x yiel<strong>de</strong>d three types of fish species<br />

response (Figure 4.3). The first type correspon<strong>de</strong>d to species that were indifferent to use TFH:<br />

The eel for the two types of TFH (Wilcoxon signed ranks test, Z = 0.157, P = 0.875, n = 22<br />

grasslands and Z = 1.162, P = 0.245, n = 14 reed beds) and the pumpkinseed and the black<br />

bullhead in reed beds (Z < 0.910, P > 0.363, n = 14, Figure 4.3). The second type was species<br />

that avoi<strong>de</strong>d the TFH: pumpkinseed and black bullhead in grasslands (Wilcoxon signed ranks<br />

test, Z > 2.330, P < 0.020, n = 22) and the mosquitofish and the bream in the two types of<br />

TFH (Z > 3.758, P < 0.001, n = 22 grasslands and Z > 2.897, P < 0.004, n = 14 reed beds,<br />

Figure 4.3). The third type was preference for TFH. The rudd was the only species belonging<br />

62


MTI et poissons – approche synchronique<br />

to the third response type. This species preferentially used the two types of TFH compared to<br />

its relative abundance measured in PH (Z = -3.323, P < 0.001, n = 22 grasslands and Z = -<br />

2.794, P = 0.005, n = 14 reed beds, Figure 4.3).<br />

Ivlev’s electivity in<strong>de</strong>x<br />

1.0<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

-0.2<br />

-0.4<br />

-0.6<br />

-0.8<br />

-1.0<br />

*** **<br />

*** **<br />

Rudd Black<br />

bullhead<br />

*<br />

NS<br />

*<br />

NS<br />

NS<br />

NS<br />

Bream Eel Pumpkinseed<br />

*** **<br />

Mosquitofish<br />

Figure 4.3 Selectivity for the two types of temporary floo<strong>de</strong>d habitats based on Ivlev’s in<strong>de</strong>x of<br />

electivity by the six dominant species in grasslands (, n = 22) and reed beds (, n = 14). Fish<br />

were sampled in 2004 and 2005 in the Brière marsh. Error-bars are standard errors. NS: no<br />

significant, * P < 0.05, ** P < 0.01 and *** P < 0.001 for Wilcoxon-signed rank test. See <strong>de</strong>tails<br />

in Table 4.1 for species co<strong>de</strong>s.<br />

4.3.3. Effect of the distance from permanent habitats on the presence of fish onto<br />

temporary floo<strong>de</strong>d habitats<br />

Universally, the presence probability curve of the six most abundant species always <strong>de</strong>creased<br />

as the distance to the PH increased in both grasslands and reed beds, but this trend was not<br />

similar for all species between the two types of TFH (Figure 4.4 and Table 4.2). The distances<br />

from PH colonized by introduced fish species (black bullhead, pumpkinseed and mosquitofish)<br />

onto TFH (data being either analysed at the species level or at the species group level) were<br />

greater in reed beds than in grasslands. For native species, the bream colonized a greater<br />

distance in grasslands than in reed beds, and colonization patterns of the rudd, the eel and the<br />

native species group were similar between the two types of TFH. In grasslands, the rudd<br />

colonized the furthest, followed by the black bullhead, the bream, the pumpkinseed, the eel<br />

and the mosquitofish. In reed beds, the black bullhead and the rudd colonized the furthest,<br />

followed by the pumpkinseed, the eel, the bream and the mosquitofish (Figure 4.4).<br />

63


Chapitre 4<br />

Presence probability<br />

0.25<br />

0.2<br />

0.15<br />

0.1<br />

0.05<br />

0<br />

0.25<br />

0.2<br />

0.15<br />

0.1<br />

0.05<br />

0<br />

0.25<br />

0.2<br />

0.15<br />

0.1<br />

0.05<br />

0<br />

0.4<br />

0.35<br />

0.3<br />

0.25<br />

0.2<br />

0.15<br />

0.1<br />

0.05<br />

0<br />

Eel<br />

Bream<br />

Rudd<br />

Native<br />

0 50 100 150 200 250 300 0 50 100 150 200 250 300<br />

Distance to the permanent ditch (meters)<br />

Black bullhead<br />

Mosquitofish<br />

Pumpkinseed<br />

Introduced<br />

Figure 4.4. Probability curves (logistic regression) of fish species presence onto temporary floo<strong>de</strong>d<br />

habitats relative to the distance to permanent habitats (in meters). Mo<strong>de</strong>ls were performed for the six<br />

dominant fish species and for the two species groups (i.e. native (left) and introduced (right) and in<br />

grasslands (full lines) and reed beds (dotted lines)).See <strong>de</strong>tails in Table 4.1 for co<strong>de</strong>s and status of<br />

fish species and in Table 4.2 for variables of the mo<strong>de</strong>ls and statistics.<br />

64


MTI et poissons – approche synchronique<br />

Table 4.2. Results of logistic regressions mo<strong>de</strong>lling the presence of fish onto temporary floo<strong>de</strong>d<br />

habitat in relation to the distance to permanent habitats and the type of temporary floo<strong>de</strong>d habitat<br />

(mo<strong>de</strong>l information and p-values). The mo<strong>de</strong>ls were performed for the six dominant species<br />

recor<strong>de</strong>d and for the two species groups (native and introduced) on a set of 1156 PAS performed<br />

on grasslands (n = 22) and reed beds (n = 14) in 2004 and 2005 in the Brière marsh.<br />

4.4. Discussion<br />

Mo<strong>de</strong>l fitting information and testing global null hypothesis β = 0<br />

Species<br />

Intercept<br />

only<br />

Intercept and<br />

covariates<br />

χ² for covariates<br />

Ana -252.569 -245.285 7.284 with 2 d.f. (P = 0.026)<br />

Abb -176.879 -157.916 18.963 with 2 d.f. (P < 0.0001)<br />

Sce -425.672 -421.263 4.409 with 2 d.f. (P = 0.1103)<br />

Native -550.685 -538.330 12.355 with 2 d.f. (P = 0.021)<br />

Amm -417.710 -397.894 19.816 with 2 d.f. (P < 0.001)<br />

Gaa -88.487 -79.469 9.018 with 2 d.f. (P = 0.011)<br />

Leg -189.866 -162.150 27.716 with 2 d.f. (P < 0.0001)<br />

Introduced -488.381 -449.536 38.845 with 2 d.f. (P < 0.001)<br />

P-values<br />

Species Intercept Habitat Distance<br />

Ana


Chapitre 4<br />

71% (Snodgrass et al. 1996, Baber et al. 2002). This result suggests that TFH in the Brière<br />

marsh are relatively attractive for fish, at least in terms of presence. Nevertheless, because this<br />

study was conducted at a large spatial scale to compare a large set of TFH, future<br />

investigations should also incorporate the temporal variation that might occur throughout the<br />

floo<strong>de</strong>d period.<br />

The comparison of the fish community found in TFH to that of adjacent PH has been<br />

poorly investigated in the literature (but see Trexler et al. 2000, Rehage & Trexler 2006). We<br />

found that the two fish communities were highly different in terms of species richness, ranks<br />

of species and total fish abundance, suggesting that TFH are used by a particular range of<br />

species. Furthermore, we found a variety of responses (abundance, selectivity and distribution)<br />

to the type of TFH among species. We also found that fish species exhibited variable<br />

selectivity in terms of use of the two types of TFH, certainly in relation to differences in<br />

functional roles played by TFH for these species. For example, the rudd reproduces in shallow<br />

vegetated areas from April to June (Keith & Allardi 2001). By analogy with the terminology<br />

<strong>de</strong>veloped in Williams (2006) for invertebrate organisms, this fish species could be<br />

consi<strong>de</strong>red as an “obligate” species in term of use of TFH, since it must colonize these<br />

habitats to accomplish a key phase of its life cycle. In contrast, the bream (another cyprinid<br />

species) reproduces later in the season and in <strong>de</strong>eper water along canal banks (Keith & Allardi<br />

2001). This species can thus be consi<strong>de</strong>red as a “facultative” species in term of use of TFH.<br />

This species potentially colonizes these habitats because they provi<strong>de</strong> good opportunities in<br />

terms of food availability and support a limited number of predators (Snodgrass et al. 1996,<br />

Corti et al. 1999), at least until a certain threshold of water level (Kushlan 1976). Nevertheless,<br />

our investigation did not allow us to quantify the proportion of individuals of each species<br />

that colonize TFH, and it is thus not possible to <strong>de</strong>termine the precise facultative/obligate use<br />

of TFH by each fish species.<br />

Our <strong>de</strong>termination of the spatial distribution of fish species on the two types of TFH<br />

confirm the results observed previously by others (Carpentier et al. 2004, Rehage & Trexler<br />

2006). In<strong>de</strong>ed, these authors have <strong>de</strong>monstrated that the distance from the PH was a major<br />

factor controlling the abundance of fish onto TFH. Here, we exten<strong>de</strong>d these results to two<br />

types of TFH. We showed the existence of a similar pattern of fish probability presence on<br />

TFH among species. In<strong>de</strong>ed, we measured a <strong>de</strong>crease in the fish presence probability along a<br />

distance gradient from PH, but this <strong>de</strong>crease was variable among types of TFH. Furthermore,<br />

we <strong>de</strong>monstrated that only a narrow and restricted area, close to PH, was well colonized by<br />

fish, followed by a rapid <strong>de</strong>crease in fish presence probability. The use of TFH could be<br />

66


MTI et poissons – approche synchronique<br />

interpreted as a tra<strong>de</strong>-off that would occur when physiological, behavioral or ecological traits<br />

of an organism (e.g. colonization of TFH) that confers advantage for performing one function<br />

(e.g. reproduction or growth) simultaneously confers a disadvantage for performing another<br />

function (e.g. physiological maintenance) due to the reduction of habitats in space, time and<br />

quality during drought (Chase & Leibold 2003, Magoulick & Kobza 2003, Matthews &<br />

Marsh-Matthews 2003, Chapitre 6).<br />

This interpretation might help un<strong>de</strong>rstanding the spatial distribution patterns of fish<br />

onto TFH. In<strong>de</strong>ed, the availability of resources (i.e. food or spawning substratum) do not<br />

<strong>de</strong>crease when moving away from PH (unpublished data) and fish would increase their<br />

benefits by colonizing attractive habitats occupied by low fish <strong>de</strong>nsity (i.e. far from PH). By<br />

doing so, these individuals would access less exploited resources and would increase their<br />

fitness (e.g. growth rate). At the same time, by exploiting areas far away from PH (refuges),<br />

these fish increase the risk of being trapped in isolated pools as the season progresses. Such<br />

trapping would incur adverse costs, such as low survival rates. Our findings show that the<br />

probability of presence of fish rapidly reached low values as the distance to the PH increased.<br />

Consequently, the costs related to the exploitation of TFH far from the adjacent canal are<br />

probably high and may be linked to the harsh conditions that exist in TFH of the Brière marsh.<br />

The timing and duration of the flooding event is very limited, variable among years and the<br />

amplitu<strong>de</strong> of water level onto grasslands and reed beds is relatively low. The temporal<br />

window during which fish can access these habitats is very restricted, and we suggest that all<br />

these conditions might severely impact the use of TFH for fish species. It is obvious that<br />

further investigations are nee<strong>de</strong>d to better un<strong>de</strong>rstand and predict the role of flooding<br />

conditions on the use of TFH by fish and on the community functioning (for instance, see<br />

Trexler et al. (2005) and Rehage & Trexler (2006)).<br />

Freshwater habitats have been the recipient of many fish invasions (García-Berthou et<br />

al. 2005) and the variable success of introduced fish species have been attributed to a variety<br />

of factors (e.g. Trexler et al. 2000, Marchetti et al. 2004). In the present study, we found<br />

convergent results regarding this issue: (1) the abundance of introduced fish species (black<br />

bullhead and pumpkinseed) was higher in reed beds than in grasslands, whereas it was not the<br />

case for native fish species; and (2) the distance to PH colonized by introduced species was<br />

more exten<strong>de</strong>d in reed beds than in grassland, whereas it was not the case for native species.<br />

These results are consistent with the invasional meltdown hypothesis of Simberloff & Von<br />

Holle (1999) whereby the presence of particular invasive species (here reed beds) facilitates<br />

the subsequent establishment of other non-indigenous species (here introduced fish species),<br />

67


Chapitre 4<br />

maybe by modifying habitat characteristics in a way that favours introduced species.<br />

Nevertheless, the biological mechanisms creating this facilitative interaction remain poorly<br />

un<strong>de</strong>rstood partly because studies on the interactions between invasive plant species and fish<br />

species are rare. For example, Laffaille et al. (2005) suggested that the invasive sea couch<br />

grass (Elymus athericus) would likely disturb trophic functions of natural salt marshes by<br />

changing the availability and accessibility of the main prey item consumed by fish species. As<br />

a possible positive interaction between fish and habitat, we can propose that in the present<br />

study, some fish species, and more particularly the black bullhead and the pumpkinseed, may<br />

benefit from some advantages brought by reed beds, in terms of food availability (e.g.<br />

invertebrate community, see Bedford & Powell (2005)) favored by large amounts of <strong>de</strong>tritus<br />

produced (Masson & Bryant 1975, Ailstock et al. 2001). Other characteristics of introduced<br />

species, such as their high physiological tolerance (Marchetti et al. 2004), may provi<strong>de</strong> them<br />

significant advantages to the harsher environmental conditions of reed beds. This mechanism<br />

might facilitate introduced fish species such as the mosquitofish and the black bullhead that<br />

are able to tolerate higher water temperature or able to face lower oxygen levels by using<br />

aerial surface respiration (e.g. Otto 1973, McKinsey & Chapman 1998, Sargent & Galat<br />

2002), at least during the lowering water level phase. Further investigations are nee<strong>de</strong>d to<br />

clarify the biological mechanisms implicated in the facilitative interactions between inva<strong>de</strong>d<br />

habitats and introduced fish species and to monitor the cumulative effects of future species<br />

introduction that might modify these interactions (Trexler et al. 2000).<br />

Acknowledgments<br />

We are grateful to the numerous persons for assistance during field work, and particularly J.-P.<br />

Damien. We also thank J. Trexler and A. Crivelli for comments on an earlier draft and J.<br />

Walters for the post-edition of the English style. We thank the Parc naturel régional <strong>de</strong> Brière<br />

for logistic support, the PnrB, FEDER, DIREN (MEDD), Région Pays <strong>de</strong> la Loire and<br />

Agence <strong>de</strong> l’Eau from Loire-Bretagne for financial support.<br />

68


MTI et poissons – approche synchronique<br />

Cette approche synchronique permet <strong>de</strong> souligner, sur l’ensemble <strong>de</strong> la zone<br />

d’étu<strong>de</strong>, les fortes disparités existantes entre la communauté piscicole dans les<br />

MTI et celle observée dans les milieux permanents. En effet, seul un nombre<br />

limité <strong>de</strong>s espèces <strong>de</strong> poissons utilisent <strong>de</strong> manière significative les MTI, leurs<br />

abondances y étant toutefois plus faibles que dans les milieux permanents. Par<br />

ailleurs l’attractivité <strong>de</strong>s MTI est variable selon les espèces et la nature <strong>de</strong>s MTI<br />

(roselière/prairie). Les autres espèces présentent <strong>de</strong>s abondances extrêmement<br />

faibles dans les MTI pour <strong>de</strong>ux raisons : soit ces espèces sont minoritaires dans la<br />

communauté et il est donc normal <strong>de</strong> ne pas les capturer dans les MTI, soit ces<br />

espèces sont normalement présentes dans la communauté mais n’utilisent pas ou<br />

peu les MTI. Concernant les espèces exploitant les MTI, ces résultats démontrent<br />

qu’outre le fait que seulement une surface limitée <strong>de</strong>s MTI est colonisée par les<br />

poissons, l’utilisation <strong>de</strong> ces <strong>de</strong>rniers est différente selon le statut <strong>de</strong> l’espèce<br />

(native vs. introduite). Les espèces natives ne présentent pas <strong>de</strong> différences en<br />

termes d’abondance et <strong>de</strong> distribution spatiale entre les types <strong>de</strong> MTI. A l’opposé,<br />

les espèces introduites montrent une interaction positive avec les roselières : leur<br />

abondance y est plus forte et la surface inondée exploitée est plus importante que<br />

dans les prairies naturelles. Ces résultats ont cependant été obtenus à un moment<br />

précis du cycle d’inondation. On peut logiquement penser que la relation entre le<br />

statut <strong>de</strong>s espèces et la nature du MTI est un phénomène dont l’ampleur évolue<br />

certainement au cours <strong>de</strong> l’inondation, mais dont la tendance reste la même. Ce<br />

n’est, par contre, peut être pas le cas en ce qui concerne l’abondance <strong>de</strong> poissons<br />

et la distribution spatiale <strong>de</strong> ceux-ci dans les MTI. Ainsi, <strong>de</strong> nouvelles questions<br />

se posent suite à ce travail : comment, au cours d’un cycle hydrologique complet,<br />

les trois paramètres - richesse spécifique, abondance et distribution spatiale <strong>de</strong>s<br />

poissons - évoluent-ils au sein <strong>de</strong>s MTI ? Qu’en est-il par rapport à la dynamique<br />

<strong>de</strong> ces mêmes <strong>de</strong>scripteurs <strong>de</strong>s poissons présents dans les milieux permanents ?<br />

Finalement, à quelles variables environnementales peut-on relier l’évolution <strong>de</strong><br />

ces paramètres ?<br />

69


Chapitre 4<br />

Une prairie temporairement inondée en périphérie <strong>de</strong>s marais <strong>de</strong> Brière<br />

70


Chapitre 5<br />

Approche diachronique <strong>de</strong> l’utilisation <strong>de</strong>s milieux<br />

temporairement inondés<br />

Comme nous l’avons vu dans le chapitre précé<strong>de</strong>nt, les paramètres <strong>de</strong><br />

composition spécifique, d’abondance et <strong>de</strong> distribution spatiale <strong>de</strong>s poissons dans<br />

les MTI, sont dépendants à la fois <strong>de</strong> la nature du MTI et <strong>de</strong> la communauté<br />

piscicole présente dans le milieu permanent adjacent. Au cours d’un cycle<br />

complet d’inondation, les facteurs environnementaux qui caractérisent ces <strong>de</strong>ux<br />

types <strong>de</strong> milieux évoluent fortement. Le but ici est donc <strong>de</strong> décrire l’évolution <strong>de</strong><br />

ces paramètres <strong>de</strong> la communauté piscicole et <strong>de</strong> déterminer comment les<br />

changements environnementaux influencent cette évolution, à la fois dans les<br />

milieux permanents et dans les MTI. Une telle approche n’est cependant pas<br />

réalisable à l’échelle <strong>de</strong> l’ensemble du site d’étu<strong>de</strong> et a donc été conduite dans une<br />

zone restreinte comportant un MTI (i.e. une prairie inondable) et un milieu<br />

permanent adjacent (i.e. un canal). Le protocole utilisé dans l’étu<strong>de</strong> précé<strong>de</strong>nte a<br />

été réutilisé ici et répété tous les quinze jours sur les <strong>de</strong>ux milieux. L’approche<br />

retenue ici sera donc qualifiée <strong>de</strong> diachronique (mêmes sites échantillonnés à<br />

plusieurs reprises au cours du cycle complet <strong>de</strong> crue <strong>de</strong> printemps).


Chapitre 5<br />

La prairie <strong>de</strong> Rozé <strong>de</strong> Mars à Juillet 2006<br />

72


MTI et poissons – approche diachronique<br />

How do fish exploit temporary waters throughout a flooding<br />

episo<strong>de</strong>? 3<br />

Résumé: Les milieux temporairement inondés jouent un rôle fondamental pour <strong>de</strong><br />

nombreuses espèces <strong>de</strong> poissons, et <strong>de</strong> nombreuses étu<strong>de</strong>s ont déjà abordé cette<br />

problématique. Cependant, une approche intégrative examinant à la fois les espèces <strong>de</strong> la<br />

communauté piscicole <strong>de</strong>s milieux permanents concernées par la colonisation <strong>de</strong> ces milieux,<br />

le moment où <strong>de</strong> cette colonisation et les modalités <strong>de</strong> distribution spatiales <strong>de</strong>s poissons dans<br />

les milieux temporairement inondés tout au long <strong>de</strong> la phase d’inondation n’a pas été encore<br />

conduite. Dans cette étu<strong>de</strong>, nous avons échantillonné une prairie temporairement inondée et<br />

son canal permanent adjacent par pêche électrique et suivi l’évolution <strong>de</strong> certains paramètres<br />

environnementaux biotiques et abiotiques dans ces <strong>de</strong>ux habitats <strong>de</strong> Février à Mai 2006 dans<br />

les marais <strong>de</strong> Brière. Au total, nous avons capturé 589 individus appartenant à 12 espèces et<br />

observé que la composition <strong>de</strong> la communauté piscicole était systématiquement différente<br />

entre les <strong>de</strong>ux habitats. La richesse spécifique, la fréquence d’occurrence et l’abondance <strong>de</strong><br />

poissons a constamment évolué dans les <strong>de</strong>ux habitats tout au long <strong>de</strong> l’étu<strong>de</strong>, du fait <strong>de</strong><br />

timings d’immigration et d’émigration variables entre les espèces. L’abondance maximale <strong>de</strong><br />

poisson a été mesurée à la mi-Avril lorsque l’abondance <strong>de</strong> zooplancton et la température <strong>de</strong><br />

l’eau augmentaient. En moyenne, les poissons colonisaient 60% <strong>de</strong> l’habitat inondé<br />

disponible. La distance maximale colonisée a été observée au début du mois d’Avril, lorsque<br />

le niveau d’eau était le plus élevé. Le découplage entre les régimes hydrologiques et<br />

thermiques a donc causé la désynchronisation du pic d’abondance et <strong>de</strong> distribution spatiale<br />

dans la prairie inondée. Ces résultats posent la question <strong>de</strong> l’accessibilité <strong>de</strong>s milieux<br />

temporairement inondés et <strong>de</strong> leur niveau <strong>de</strong> connectivité avec les milieux permanents qui<br />

peuvent perturber l’utilisation <strong>de</strong>s milieux temporairement inondés par l’ichtyofaune.<br />

3 Ce chapitre est adapté d’un article intitulé « Cucherousset, J., Carpentier, A. & Paillisson, J.-M. How do fish<br />

exploit temporary waters throughout a flooding episo<strong>de</strong>? » soumis dans la revue « Fisheries Managment and<br />

Ecology ».<br />

73


Chapitre 5<br />

Abstract: Temporary waters, i.e. aquatic habitats that cyclically dry out, constitute<br />

significant areas for key-phases of the life cycle of many fish species and several studies have<br />

already focused on the environmental biology of fish in temporary waters. Nevertheless, an<br />

integrated study has not been un<strong>de</strong>rtaken to <strong>de</strong>termine which species in adjacent permanent<br />

habitat colonise the temporary waters, when this phenomenon occurs, and how the fish<br />

distribute themselves within the temporary waters during the flooding period. In the present<br />

study, the fish assemblages in a temporarily floo<strong>de</strong>d grassland and its adjacent permanent<br />

canal were examined by electrofishing and monitored along with biotic and abiotic<br />

environmental variables in these two habitats from February to May 2006 in the Brière marsh<br />

(Northwest France). In total, 589 fish belonging to 12 species were captured. Fish assemblage<br />

composition always differed between the two habitats for each sampling date, and fish species<br />

richness, frequency of occurrence and <strong>de</strong>nsity varied markedly within the two habitats during<br />

the flooding season as a result of the immigration and emigration sequences of species.<br />

Maximum <strong>de</strong>nsity values were observed in mid-April, corresponding with elevated<br />

zooplankton <strong>de</strong>nsity and water temperature values. Fish were on average found to colonise<br />

floo<strong>de</strong>d area within 60% of the total distance between the adjacent canal and flood limit. The<br />

maximal distance colonized was observed in early April, i.e. when water level was the highest.<br />

A lack of synchronisation in the flood and water temperature regimes resulted in a mismatch<br />

between fish <strong>de</strong>nsity and spatial extent within the floo<strong>de</strong>d grassland. These results raise the<br />

question of accessibility of temporary waters and connectivity with the permanent habitat that<br />

affect fish species <strong>de</strong>nsity and spatial distribution in temporary waters.<br />

74


5.1. Introduction<br />

MTI et poissons – approche diachronique<br />

Temporary waters, i.e. aquatic habitats that cyclically dry out, are highly diverse (e.g. tidal<br />

salt marshes, river floodplains and oxbow ponds) and represent significant components of the<br />

global landscape (Williams 2006). These habitats play fundamental ecological functions for<br />

numerous plant and animal organisms (see review in Williams 2006), and particularly for fish<br />

(e.g. Welcomme 1979, Copp 1989, Magoulick 2000, King et al. 2003, DeAngelis et al. 2005,<br />

Trexler et al. 2005). In<strong>de</strong>ed, temporary waters constitute significant areas for reproduction,<br />

nursery, feeding and refuges against predators at particular periods <strong>de</strong>pending on species life<br />

histories and flooding conditions (Copp 1989, Snodgrass et al. 1996, Poizat & Crivelli 1997,<br />

Corti et al. 1999, Baber et al. 2002, Ostrand & Wil<strong>de</strong> 2002, King et al. 2003). Nevertheless,<br />

studies examining the use of these habitats by fish species in relation to the particular<br />

environmental features of temporary waters are few in number and relatively sparse (Baber et<br />

al. 2002, Hohausova et al 2003, Williams 2006). Some studies have investigated the<br />

composition and structure of fish assemblages in temporarily floo<strong>de</strong>d habitats at a given time,<br />

occasionally including their changes over time (e.g. Copp 1989, Baber et al. 2002, King et al.<br />

2003, Crain et al. 2004, Chapitre 8). On the other hand, some authors have more precisely<br />

analysed fish exchanges between temporary waters and the adjacent permanent habitats (e.g.<br />

Poizat & Crivelli 1997, Hohausova 2000, Hohausova et al. 2003, Chapitre 6). In parallel, the<br />

role played by abiotic environmental parameters on the use of temporary waters by fish<br />

species has been addressed (e.g. Magoulick 2000, Ostrand & Wil<strong>de</strong> 2001, Baber et al. 2002,<br />

Hohausova et al. 2003) and more recently, Carpentier et al. (2004) have studied the spatial<br />

distribution of fish in a temporarily floo<strong>de</strong>d grassland. Nevertheless, an integrative study has<br />

not been un<strong>de</strong>rtaken to examine how fish exploit temporarily floo<strong>de</strong>d habitats throughout a<br />

flooding period. In the present study, the following questions were examined: (1) Wich fish<br />

species of an adjacent permanent habitat colonize the temporary habitat?, (2) When do these<br />

species colonize the temporary waters throughout a flooding episo<strong>de</strong>?, (3) Where do these<br />

species distribute themselves within the temporary waters?, and (4) Do some biotic and<br />

abiotic environmental variables regulate the exploitation of this habitat by fish species during<br />

the flooding episo<strong>de</strong>?<br />

75


Chapitre 5<br />

5.2. Material and methods<br />

5.2.1. Study area<br />

The study was conducted throughout the flooding period in 2006 (February/May) in a typical<br />

sector of the Gran<strong>de</strong> Brière Mottière marsh that inclu<strong>de</strong>s a permanent canal section (the<br />

“fougères” canal, belonging to the secondary level of the marsh network, see <strong>de</strong>tails in<br />

Chapitre 2) and an adjacent grassland (the “Rozé” grassland, Figure 5.1). The canal section<br />

was 390 m long and 13.2 m (± 2.4 S.D.) width with a low (8.4 % ± 9.9) vegetation cover. The<br />

grassland was 5.2 ha area and was more <strong>de</strong>nsely vegetated (72.6 % ± 27.8) than the canal. At<br />

the maximal flood level (April, 2), water level reached 66.5 cm (± 5.8) in the canal and 30.6<br />

cm (± 8.7) in the grassland (Figure 5.2). During flooding, fish can only colonise the grassland<br />

from the permanent canal via five limited connecting points (3-5 m width) opened in the bank<br />

(Figure 5.1). In<strong>de</strong>ed, this bank, which resulted from digging and dredging operations (peat<br />

and mud stocked along the canal), stayed unfloo<strong>de</strong>d throughout the flooding period (see<br />

<strong>de</strong>tails in Chapitre 2).<br />

5.2.2. Fish sampling and monitoring of environmental parameters<br />

Field sampling was un<strong>de</strong>rtaken once every two weeks from early February (i.e. just after the<br />

last frost) when small shallow watered areas began to form in the grassland to late May, at the<br />

end of the flood event, when only a few narrow and inaccessible pools were still floo<strong>de</strong>d<br />

(Figure 5.2). Fish sampling was conducted simultaneously in the two habitats using the point<br />

abundance sampling (PAS) approach (see <strong>de</strong>tails in Chapitre 2). In the floo<strong>de</strong>d grassland,<br />

sampling was conducted from each connecting point until the maximum water’s edge (i.e.<br />

extent of flooding, Figure 5.1). In total, 225 PAS were performed in the permanent canal (25<br />

PAS per sampling date) and 670 in the floo<strong>de</strong>d grassland (74.4 ± 16.8 PAS per sampling date).<br />

In early spring, all fish species (except pike Esox lucius) had not yet reproduced and youngof-the<br />

year represented only 1.7% of the fish sampled throughout the study. Consequently, all<br />

size classes were pooled for analyses.<br />

Three <strong>de</strong>scriptors were used to characterise the fish assemblage in the two habitats:<br />

standardised species richness (according to sampling effort, S’), frequency of occurrence<br />

(proportion of all PAS in which a fish species was captured, in %) and <strong>de</strong>nsity (total and for<br />

each species, expressed in Catch Per Unit Effort (CPUE), i.e. number of individuals.PAS -1 ).<br />

To <strong>de</strong>scribe the dispersal of fish colonising the floo<strong>de</strong>d grassland from the permanent canal,<br />

76


MTI et poissons – approche diachronique<br />

we calculated the linear distance (in metres) between each PAS and the nearest connecting<br />

point with the canal using a Global Positioning System and a Geographical Information<br />

System (source: Parc naturel régional <strong>de</strong> Brière). On each sampling date, we reported the<br />

distance between the permanent canal and the water’s edge in the floo<strong>de</strong>d grassland (i.e. the<br />

maximal distance we sampled, Figure 5.1).<br />

Permanent canal<br />

Canal bank<br />

Floo<strong>de</strong>d grassland<br />

Point abundance sample<br />

Connecting point<br />

“Rozé”<br />

grassland<br />

“les fougères”<br />

canal<br />

N<br />

30 m<br />

Figure 5.1.Map of the Gran<strong>de</strong> Brière Mottière marsh and location of the study site, i.e. the “les<br />

fougères” canal and the “Rozé” floo<strong>de</strong>d grassland. The black dots () represent a series of fish<br />

sampling points (PAS, electrofishing Point Abundance Sampling method) performed on 19 April<br />

2006 in the permanent canal and the floo<strong>de</strong>d grassland. The white triangles () indicate the<br />

connecting points between the two habitats opened in the emerged bank (canal clearance<br />

materials).<br />

77


Chapitre 5<br />

For each fish sampling date, four environmental variables were measured in the two<br />

habitats. Water temperature was recor<strong>de</strong>d every 15 min from 10 a.m. to 2 p.m. using<br />

automatic sensors (Sensor StowAway TidBit, Onset Computer Corporation). We calculated<br />

the mean and the coefficient of variation of water temperature (in %). Water conductivity<br />

(μS.cm -1 ) and dissolved oxygen concentration (mg.L -1 ) were measured in the two habitats<br />

during the same time slot (11 to 12 a.m.). We also recor<strong>de</strong>d changes in food availability in the<br />

two habitats by monitoring the <strong>de</strong>nsity of zooplankton (mainly Cladocera and Copepoda),<br />

which constitutes the major food source for fish. In the same time, we filtered a 15 L volume<br />

of the water column (0.1-mm mesh) and samples were preserved in 70% ethanol.<br />

Zooplankton <strong>de</strong>nsities were measured in the laboratory using a Dollfus counting chamber and<br />

a binocular microscope, and expressed as ind.L -1 . All environmental parameters were always<br />

measured in the same locations in the middle parts of the canal and the grassland.<br />

Precipitation (mm)<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

20 Jan<br />

27 Jan<br />

3 Feb<br />

10 Feb<br />

17 Feb<br />

24 Feb<br />

3 Mar<br />

10 Mar<br />

17 Mar<br />

24 Mar<br />

31 Mar<br />

Grassland flooding threshold<br />

7 Apr<br />

Date<br />

Figure 5.2. Precipitation (vertical bars), daily water level (thick line) and air temperature (fine line)<br />

in the Gran<strong>de</strong> Brière Mottière marsh from 20 January to 20 June 2006. The triangles (▼) indicate<br />

sampling dates. The horizontal dotted line represents the minimal threshold of water level to flood<br />

the grassland.<br />

5.2.3. Statistical analyses<br />

The data were tested for normality and found not to comply, so non-parametric tests<br />

(Wilcoxon signed rank test, Spearman rank correlation and Kruskal-Wallis test) were<br />

performed for between-sites and between-dates comparisons in species compositions and<br />

environmental conditions. A rejection level of P ≥ 0.05 was used in all tests. Values are mean<br />

± standard <strong>de</strong>viation (S.D.).<br />

78<br />

14 Apr<br />

21 Apr<br />

28 Apr<br />

5 May<br />

12 May<br />

19 May<br />

26 May<br />

2 Jun<br />

9 Jun<br />

16 Jun<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

Water level (French General Level (NGF, cm)<br />

and air temperature (°C)


5.3. Results<br />

5.3.1. Succession of environmental variables<br />

MTI et poissons – approche diachronique<br />

Overall, water conductivity and dissolved oxygen concentration <strong>de</strong>creased in the two habitats<br />

during the flooding season, and water temperature simultaneously increased. Water<br />

conductivity and dissolved oxygen concentration slightly differed between the two habitats,<br />

whereas mean water temperatures did not (see Z-statistics in Figure 5.3). Nevertheless, the<br />

mean value of the coefficient of variation of water temperature was significantly higher<br />

(Wilcoxon-signed rank test, Z = 2.310, P = 0.021, n = 9) in the grassland (4.9 ± 3.9%) than in<br />

the canal (4.7 ± 3.3%). At the start of the survey, zooplankton <strong>de</strong>nsity was low, and<br />

afterwards largely increased in early May, corresponding to an increase in water temperature<br />

above 12°C (Figure 5.3). This pattern was similar in the two sites. Nevertheless, zooplankton<br />

<strong>de</strong>nsity was significantly lower in the permanent canal than in the floo<strong>de</strong>d grassland (Figure<br />

5.3).<br />

Conductivity (μS.cm -1 )<br />

Dissolved oxygen (mg.L -1 )<br />

6000<br />

5000<br />

4000<br />

3000<br />

2000<br />

1000<br />

0<br />

16<br />

14<br />

12<br />

10<br />

8<br />

6<br />

4<br />

2<br />

0<br />

8 Feb<br />

Z = 1.955<br />

P = 0.051<br />

22 Feb<br />

Z = 1.836<br />

P = 0.066<br />

8 Mar<br />

22 Mar<br />

5 Apr<br />

Floo<strong>de</strong>d grassland<br />

Permanent canal<br />

19 Apr<br />

3 May<br />

17 May<br />

31 May<br />

Figure 5.3. Environmental parameters in the floo<strong>de</strong>d grassland () and in the permanent canal ()<br />

from Feb 8 th to May 31 st Date<br />

Date<br />

2006: conductivity, mean water temperature (error bars are standard<br />

<strong>de</strong>viations), dissolved oxygen concentration and zooplankton <strong>de</strong>nsity. Results of between-sites<br />

comparisons (Wilcoxon signed rank test) are provi<strong>de</strong>d for each environmental parameter (n = 9<br />

dates).<br />

79<br />

Temperature (°C)<br />

Zooplankton <strong>de</strong>nsity (ind.L -1 )<br />

20<br />

18<br />

16<br />

14<br />

12<br />

10<br />

8<br />

6<br />

4<br />

2<br />

0<br />

120<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

8 Feb<br />

22 Feb<br />

Z = -1.125<br />

P = 0.260<br />

Z = 2.666<br />

P = 0.008<br />

8 Mar<br />

22 Mar<br />

5 Apr<br />

19 Apr<br />

3 May<br />

17 May<br />

31 May


Chapitre 5<br />

5.3.2. Fish assemblage composition succession during inundation<br />

In total, we sampled 589 individuals belonging to 12 species: 386 individuals (belonging to<br />

11 species) captured in the canal and 203 individuals (11 species) in the grassland (Table 5.1).<br />

The fish assemblage was dominated by the same six species in the two habitats (Table 5.1):<br />

common bream (Abramis brama), black bullhead (Ameiurus melas), roach (Rutilus rutilus),<br />

rudd (Scardinius erythrophthalmus), European eel (Anguilla anguilla), and Prussian carp<br />

(Carassius gibelio). The other species (relative abundance


Standardised species richness (S’)<br />

MTI et poissons – approche diachronique<br />

Standardised species richness patterns were similar in the two habitats during<br />

inundation: a continuous increase in the number of species followed by a <strong>de</strong>crease at the end<br />

of the flooding period in the grassland. The mean standardised species richness was<br />

significantly higher in the permanent canal than in the floo<strong>de</strong>d grassland (Figure 5.4). Fish<br />

frequency of occurrence was constant and low in the two habitats until April 5. On April 19, it<br />

strongly increased in the canal from 28% to 72% followed by high values (92-96%) until the<br />

end of the flood while a brief peak (25%) was observed in the grassland on May 3. Fish<br />

frequency of occurrence was significantly higher in the canal (52.9 ± 13.0%) than in the<br />

grassland (35.7 ± 7.2%, Figure 5.4). Patterns of fish <strong>de</strong>nsity were similar to those <strong>de</strong>scribed<br />

for the frequency of occurrence throughout the flooding period. Fish <strong>de</strong>nsity was higher in the<br />

canal (1.72 ± 1.50 CPUE) than in the grassland (0.28 ± 0.25 CPUE, Figure 5.4). The general<br />

increased values in the fish assemblage <strong>de</strong>scriptors in the two habitats throughout the study<br />

period coinci<strong>de</strong>d with the maximum water level of the flood (Figure 5.2), a large increase in<br />

water temperature and the start of increase in the zooplankton <strong>de</strong>nsity (Figure 5.3).<br />

0.45<br />

0.40<br />

0.35<br />

0.30<br />

0.25<br />

0.20<br />

0.15<br />

0.10<br />

0.05<br />

0.00<br />

8 Feb<br />

Z = -2.666<br />

p = 0.008<br />

Feb 22<br />

8 Mar<br />

22 Mar<br />

4 Apr<br />

19 Apr<br />

Date<br />

3 May<br />

17 May<br />

31 May<br />

Frequency of occurrence (%)<br />

100<br />

90<br />

80<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

8 8 Feb<br />

Floo<strong>de</strong>d grassland<br />

Permanent canal<br />

Z = -2.666<br />

P = 0.008<br />

Feb 22<br />

8 Mar<br />

22 Mar<br />

4 Apr<br />

19 Apr<br />

Date<br />

3 May<br />

17 May<br />

31 May<br />

Density (CPUE)<br />

4<br />

3.5<br />

3<br />

2.5<br />

2<br />

1.5<br />

1<br />

0.5<br />

0<br />

8 Feb<br />

Z = -2.666<br />

P = 0.008<br />

Figure 5.4. Descriptors of the fish assemblage present in the canal () and in the floo<strong>de</strong>d<br />

grassland () throughout the flooding period (8 February to 31 May 2006): standardised species<br />

richness (S’), frequency of occurrence and <strong>de</strong>nsity. Results of between-sites comparisons<br />

(Wilcoxon signed rank test) are provi<strong>de</strong>d for each <strong>de</strong>scriptor (n = 9 dates).<br />

At the species level, variations occurred in the fish <strong>de</strong>nsity between the two habitats<br />

(Figure 5.5): higher fish <strong>de</strong>nsities in the canal than in the grassland were found for common<br />

bream, black bullhead and roach (Wilcoxon signed rank test, Z < -2.197, P < 0.028 n = 9),<br />

slightly higher <strong>de</strong>nsities for eel (Wilcoxon signed rank test, Z = -1.955, P = 0.051, n = 9) and<br />

for Prussian carp (Wilcoxon signed rank test, Z = -1.753, P = 0.080, n = 9) whereas no<br />

significant difference was found for rudd (Wilcoxon signed rank test, Z = 0.178, P = 0.859, n<br />

81<br />

Feb 22<br />

8 Mar<br />

22 Mar<br />

4 Apr<br />

19 Apr<br />

Date<br />

3 May<br />

17 May<br />

31 May


Chapitre 5<br />

= 9). On the other hand, the species compositions were not correlated in the two habitats at<br />

each date (Spearman rank or<strong>de</strong>r correlation, 0.06 < r < 0.66, 0.136 < P < 0.919, n = 12 species<br />

Figure 5.5).We also found differences in the timing of use of the temporary habitat between<br />

species (Figure 5.5). For instance, rudd and eel colonised the grassland early and were<br />

recor<strong>de</strong>d in this habitat during the entire flooding event, whereas the other species generally<br />

colonised the grassland later (in mid April, at the maximum water level and when the water<br />

temperature increased strongly, Figure 5.3).<br />

Density (CPUE)<br />

Density (CPUE)<br />

Density (CPUE)<br />

1.5<br />

1.0<br />

0.5<br />

0.0<br />

0.5<br />

1.0<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0.0<br />

0.2<br />

1.0<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0.0<br />

0.2<br />

8 Feb<br />

(a) Common bream<br />

(c) Roach<br />

(b) Black bullhead<br />

(d) Rudd<br />

(d) European eel (e) Prussian carp<br />

22 Feb<br />

8 Mar<br />

22 Mar<br />

4 Apr<br />

19 Apr<br />

3 May<br />

17 May<br />

31 May<br />

Date Date<br />

1.0<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0.0<br />

0.2<br />

8 Feb<br />

22 Feb<br />

8 Mar<br />

22 Mar<br />

4 Apr<br />

Permanent canal<br />

Floo<strong>de</strong>d grassland<br />

Figure 5.5. Temporal patterns of fish <strong>de</strong>nsity (in Catch Per Unit Effort (CPUE, ind.PAS -1 ) of the<br />

six dominant fish species in the canal () and in the floo<strong>de</strong>d grassland () throughout the<br />

flooding period. Note a different y-scale for common bream (a). See Table 5.1 for species<br />

scientific names.<br />

82<br />

19 Apr<br />

3 May<br />

17 May<br />

31 May


5.3.3. Fish distribution in the grassland during inundation<br />

MTI et poissons – approche diachronique<br />

The mean distance colonized by fish in the grassland changed significantly between sampling<br />

dates (Kruskal-Wallis, KW = 59.4, P < 0.0001, n = 203 fish), as the relative distance to the<br />

water’s edge in the grassland (KW = 59.9, P < 0.0001, n = 203 fish). A strong <strong>de</strong>crease in<br />

both variables was observed from 19 April (Figure 5.6). During the first half of the flood, the<br />

mean distance where fish were sampled increased continuously in conjunction with the<br />

increase in water level (Figure 5.2), whereas the <strong>de</strong>nsity of fish in the grassland was still low<br />

(Figure 5.4). At this period, fish exploited, on average, 45.9 % (± 8.2 S.D.) of the available<br />

floo<strong>de</strong>d habitat (Figure 5.6). Fish colonised the maximal distance with the larger distributional<br />

range at the maximal water level, i.e. on 4 April (Figure 5.6). During the second part of the<br />

flood, as water level <strong>de</strong>creased (Figure 5.2) and water temperature increased (Figure 5.3), the<br />

distance to the permanent canal, the distributional range and the relative distance colonised by<br />

fish in the grassland strongly <strong>de</strong>creased (Figure 5.6). On average, fish exploited 17.8 % (± 4.7<br />

S.D.) of the available floo<strong>de</strong>d habitat during this period, being confined close to the<br />

connecting points.<br />

Distance to the permanent ditch (metres)<br />

160<br />

140<br />

120<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

n = 3<br />

ab<br />

8 Feb<br />

a<br />

n = 7<br />

22 Feb<br />

n = 5<br />

8 Mar<br />

ab<br />

n = 12<br />

a<br />

22 Mar<br />

n = 11<br />

a<br />

4 Apr<br />

Date<br />

Figure 5.6. Distance to the permanent canal (boxes, in meters) and relative distance to the water’s<br />

edge (solid line with solid circle, in %) colonised by fish in the grassland throughout the flooding<br />

season (8 Feb to 31 May 2006). The boxes indicate the interquartile (distributional range), the<br />

horizontal lines within the box represent the mean and the whiskers show maximum and minimum<br />

distance values. The small dotted line represents the water’s edge in the grassland. Two different<br />

letters indicate significant differences at P < 0.05 from pairwise comparisons of the distance to the<br />

permanent canal variable.<br />

83<br />

n = 70<br />

ab<br />

19 Apr<br />

n = 52<br />

b<br />

3 May<br />

n = 25<br />

b<br />

17 May<br />

n = 18<br />

b<br />

31 May<br />

100<br />

90<br />

80<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

Relative distance to the water’s edge (%)


Chapitre 5<br />

5.4. Discussion<br />

The present study showed that the exploitation of temporary floo<strong>de</strong>d habitats by fish was (1)<br />

highly variable over the course of the flooding season, (2) influenced by changes in the fish<br />

assemblage of the adjacent canal, (3) governed by environmental variables, which presented a<br />

synergistic succession during the flooding season, and (4) restricted in time and space relative<br />

to the duration of the flood and the area floo<strong>de</strong>d. The use of temporarily floo<strong>de</strong>d habitats can<br />

be interpreted as a tra<strong>de</strong>-off between benefits and costs provi<strong>de</strong>d by these habitats (Chapitre<br />

6). This tra<strong>de</strong>-off is likely to vary throughout the flooding period as a consequence of changes<br />

in environmental variables (here water level, zooplankton <strong>de</strong>nsity, temperature) that modify<br />

the ratio of costs from physiological stress (Magoulick & Kobza 2003, Matthews & Marsh-<br />

Matthews 2003) and predation pressure (Capone & Kushlan 1991, Kushlan 1976) vs benefits<br />

from growing conditions (Capone & Kushlan 1991) compared to the conditions of the<br />

permanent habitats. In<strong>de</strong>ed, Williams (2006) stated that the main advantages for fish of<br />

colonising temporary waters are the presence of abundant food resources, earlier spring<br />

spawning breeding possibilities (the water being often warmer than in adjacent permanent<br />

waters), and reduced predation notably by large fish. In the present study, although we did not<br />

find any differences in mean water temperature between the two habitats, the amplitu<strong>de</strong> of<br />

variation of temperature, the zooplankton resources (particularly when consi<strong>de</strong>red with<br />

respect to the number of fish), the vegetation coverage and consequently the oxygenation of<br />

water were higher in the floo<strong>de</strong>d grassland than in the adjacent permanent habitat, particularly<br />

during the first part of the flooding season. At this time, the grassland constituted a high<br />

quality and attractive habitat for fish, at least up to a certain distance from the canal.<br />

Afterwards and as the water level <strong>de</strong>clines a series of adverse conditions occur: a <strong>de</strong>crease in<br />

food availability (e.g. Capone & Kushlan 1991), an increase in predation pressure by birds<br />

(e.g. Kushlan 1976), and a <strong>de</strong>cline in water quality (e.g. Magoulick & Kobza 2003: low<br />

dissolved oxygen concentration and high water temperature). At this period, in the present<br />

study we observed that either fish present in the grassland moved nearer the connecting points<br />

(lower distributional range) or fish emigrated from the grassland, i.e. to <strong>de</strong>eper habitats being<br />

used as refuge in summer (Gozlan et al. 1998, Magoulick & Kobza 2003). So the floo<strong>de</strong>d<br />

grassland was only used during a restricted period of time. Consequently, fish should optimise<br />

their use by adapting the immigration and emigration timings (Chapitre 6) and also their<br />

distribution within the grassland with respect to changes in environmental conditions.<br />

Differences in species life histories and characteristics could then explain the observed<br />

84


MTI et poissons – approche diachronique<br />

seasonal variability in the exploitation of this temporary water by fish species (see Poizat &<br />

Crivelli 1997, Baber et al. 2002, Chapitre 6).<br />

In the present study, we observed that fish <strong>de</strong>nsity and dispersal were the highest in<br />

April. The maximum distance colonised in the grassland correspon<strong>de</strong>d to the maximum water<br />

level associated with a low water temperature (12°C, mid-April). Furthermore, this high number of fish was mainly composed<br />

by sexually mature cyprinids (i.e. common bream, rudd, Prussian carp, unpublished data),<br />

confirming the relationship between the spawning migration sequences and water temperature<br />

in this genus (Hladik & Kubecka 2003). As observed elsewhere (King et al. 2003), flood and<br />

water temperature regimes were <strong>de</strong>coupled in the present study, so environmental conditions<br />

were certainly far from the optimum for cyprinid recruitment. These results also raise the<br />

question of accessibility of temporary waters and their connectivity with the permanent<br />

habitat. In<strong>de</strong>ed, we observed <strong>de</strong>lays in colonisation of the grassland by many species<br />

compared with their pattern of presence in the canal and a restricted spatial distribution of fish<br />

in the grassland. Several authors suggest that a key role is played by the level of connectivity<br />

between the temporary and the permanent habitats for fish exchanges (e.g. Nicolas & Pont<br />

1995, Baber et al. 2002, Hohausova et al. 2003, Carpentier et al. 2004). In the present study,<br />

the grassland/canal connectivity that results from the canal digging plan applied in the Gran<strong>de</strong><br />

Brière Mottière marsh only represented 4% of the canal length (18 m along 450 m of canal).<br />

Baber et al. (2002) suggested that dredging operations could adversely affect the fish<br />

assemblages in temporary waters by reducing duration of inundation and connectivity. We<br />

then hypothesise that the presence of a bank in our study site probably limits the use of<br />

floo<strong>de</strong>d grassland by fish in both time and space. Nevertheless, further investigations are<br />

nee<strong>de</strong>d to un<strong>de</strong>rstand how variable flooding and connectivity conditions could influence the<br />

exploitation of temporarily floo<strong>de</strong>d habitats by fish throughout a flooding period.<br />

Acknowledgements<br />

We are particularly grateful to the numerous people for their assistance during field work. We<br />

thank R.E. Gozlan for constructive comments on an earlier draft. We thank the Parc naturel<br />

régional <strong>de</strong> Brière for logistic support, the PnrB, FEDER, DIREN (MEDD), Région Pays <strong>de</strong><br />

la Loire and Agence <strong>de</strong> l’Eau from Loire-Bretagne for financial support.<br />

85


Chapitre 5<br />

Cette approche diachronique souligne la forte évolution <strong>de</strong>s différents paramètres<br />

<strong>de</strong> la communauté piscicole au cours d’un cycle d’inondation, et ce en réponse<br />

aux changements <strong>de</strong> certains facteurs environnementaux. Nous avons aussi<br />

observé que l’utilisation temporelle et spatiale <strong>de</strong> la prairie temporairement<br />

inondée est relativement réduite du fait <strong>de</strong> fortes contraintes environnementales.<br />

En effet, la distribution spatiale <strong>de</strong> l’ichtyofaune est fortement dépendante du<br />

niveau d’eau alors que l’abondance <strong>de</strong>s différentes espèces est plutôt influencée<br />

par la température, sans que nous puissions établir <strong>de</strong> relation directe entre ces<br />

<strong>de</strong>ux paramètres <strong>de</strong> la communauté. De plus, lorsque les régimes thermiques et<br />

hydrologiques sont décalés dans le temps, le pic d’abondance <strong>de</strong>s espèces ne<br />

correspond pas à la distribution spatiale maximale, laissant penser que ces<br />

facteurs environnementaux vont limiter, qualitativement, l’utilisation <strong>de</strong>s MTI par<br />

l’ichtyofaune. A l’avenir, une étu<strong>de</strong> à long terme <strong>de</strong> ces phénomènes permettrait<br />

<strong>de</strong> quantifier les conséquences <strong>de</strong> la non adéquation temporelle <strong>de</strong> ces régimes sur<br />

l’utilisation <strong>de</strong>s MTI par la communauté piscicole, notamment en terme <strong>de</strong><br />

recrutement (quantité et qualité) pour les espèces dont la reproduction se déroule<br />

dans les MTI. Il serait aussi important <strong>de</strong> comparer dans le futur ces phénomènes<br />

également dans les roselières, où les conditions environnementales peuvent être<br />

différentes et où il est donc probable que les réponses <strong>de</strong> la communauté soient<br />

également différentes. Enfin, l’évolution <strong>de</strong>s paramètres caractérisant la<br />

communauté (composition, abondance et distribution spatiale), observée ici<br />

même, traduit <strong>de</strong>s phénomènes d’immigration et <strong>de</strong> d’émigration très variables<br />

entre espèces qu’il convient donc d’examiner.<br />

86


Chapitre 6<br />

Rôle <strong>de</strong> la tolérance physiologique dans la variabilité<br />

interspécifique <strong>de</strong>s processus d’émigration<br />

Comme nous l’avons montré dans le chapitre précé<strong>de</strong>nt, l’utilisation <strong>de</strong>s MTI par<br />

l’ichtyofaune est un processus dynamique dans le temps et l’espace. Sous nos<br />

latitu<strong>de</strong>s, les poissons qui utilisent ces milieux ne possè<strong>de</strong>nt pas <strong>de</strong> mécanisme <strong>de</strong><br />

rési<strong>lien</strong>ce à un assèchement complet du milieu et ils sont donc obligés <strong>de</strong> migrer<br />

en fonction du cycle hydrologique. Ces flux <strong>de</strong> poissons sont potentiellement<br />

déterminés par les caractéristiques <strong>de</strong>s espèces en réaction aux conditions<br />

environnementales. Nous allons donc plus particulièrement analyser les patrons<br />

d’émigration <strong>de</strong> poissons (i.e. du MTI vers le milieu permanent) lors <strong>de</strong><br />

l’exondation progressive <strong>de</strong> quatre MTI différents. Cette première approche va se<br />

faire au niveau spécifique en se basant sur les caractéristiques <strong>de</strong> tolérance<br />

physiologique <strong>de</strong>s espèces. En effet, si les conditions environnementales se<br />

dégra<strong>de</strong>nt dans les MTI lors <strong>de</strong> l’exondation et si l’utilisation <strong>de</strong>s MTI procure<br />

<strong>de</strong>s avantages, alors les espèces les plus tolérantes <strong>de</strong>vraient émigrer les <strong>de</strong>rnières.<br />

Nous avons testé cette hypothèse à l’ai<strong>de</strong> d’indicateurs <strong>de</strong> la tolérance<br />

physiologique issus <strong>de</strong> la littérature. Cette étu<strong>de</strong> a aussi pour objectif <strong>de</strong> mettre en<br />

relation <strong>de</strong>s processus écologiques complexes observés sur le terrain et <strong>de</strong>s<br />

résultats le plus souvent issus <strong>de</strong> travaux expérimentaux conduits en laboratoire.


Chapitre 6<br />

Vue aérienne <strong>de</strong> la Réserve du Sud fin Août 2005<br />

Relève <strong>de</strong>s verveux entre le plan d’eau et le canal permanent<br />

88


Tolérance physiologique et émigration<br />

Fish emigration from temporary wetlands during drought: the<br />

role of physiological tolerance 4<br />

Résumé: Les patrons d’émigration <strong>de</strong>s poissons lors <strong>de</strong> l’exondation <strong>de</strong> quatre milieux<br />

temporairement inondés ont été étudiés <strong>de</strong> mai à août 2004. Deux <strong>de</strong>s milieux étudiés ont subi<br />

un assèchement complet alors que les <strong>de</strong>ux autres ont seulement supporté une forte<br />

diminution du niveau d’eau mais tous ont montré <strong>de</strong>s changements significatifs en termes <strong>de</strong><br />

conditions physico-chimiques. Quel que soit le niveau d’assèchement, les patrons<br />

d’émigration <strong>de</strong>s six espèces les plus communes (Ameiurus melas, Anguilla anguilla,<br />

Scardinius erythrophthalmus, Lepomis gibbosus, Esox lucius et Gambusia holbrooki), qui<br />

représentent 98.6% <strong>de</strong>s captures, sont similaires entre les quatre milieux temporairement<br />

inondés. La pério<strong>de</strong> à laquelle l’émigration <strong>de</strong>s poissons <strong>de</strong> chaque espèce a lieu est fortement<br />

corrélée à la tolérance physiologique spécifique (donnée rapportée <strong>de</strong> la littérature),<br />

démontrant ainsi le <strong>lien</strong> étroit entre l’évolution <strong>de</strong> la qualité physico-chimique <strong>de</strong> l’eau et<br />

l’émigration <strong>de</strong>s poissons. A. melas et G. holbrooki (espèces invasives) émigrent en <strong>de</strong>rnier<br />

<strong>de</strong>s milieux temporairement inondés, ceci reflétant le niveau <strong>de</strong> tolérance élevé <strong>de</strong> ces espèces<br />

à la dégradation <strong>de</strong>s conditions environnementales <strong>de</strong>s milieux lors du processus d’exondation.<br />

Cette caractéristique pourrait fortement contribuer au succès d’invasion du marais par ces<br />

espèces.<br />

Abstract: Fish emigration patterns from four temporary wetlands exposed to drought<br />

were studied from May to August. Two wetlands became totally dry, and two experienced<br />

severe water level <strong>de</strong>cline and significant changes in physico-chemical characters.<br />

Irrespective of the <strong>de</strong>gree of <strong>de</strong>siccation, emigration patterns of the six most commonly<br />

trapped species, representing 98.6 % of the total abundance (Ameiurus melas, Anguilla<br />

anguilla, Scardinius erythrophthalmus, Lepomis gibbosus, Esox lucius and Gambusia<br />

holbrooki), were similar among wetlands. Emigration timing was highly correlated with<br />

published physiological tolerance levels for these species, <strong>de</strong>monstrating a tight linkage<br />

between water quality and emigration patterns. Two non-native species (A. melas and G.<br />

holbrooki) showed the latest emigration from the temporary habitats, reflecting a high level of<br />

tolerance to drought conditions that may contribute to their success as wetland inva<strong>de</strong>rs.<br />

4 Ce chapitre est adapté <strong>de</strong> l’article : Cucherousset, J., Paillisson, J.-M., Carpentier, A. & Chapman, L.J. 2007.<br />

Fish emigration from temporary wetlands during drought: the role of physiological tolerance. Archiv für<br />

Hydrobiologie. (In press).<br />

89


Chapitre 6<br />

6.1. Introduction<br />

A fundamental issue in ecology is the un<strong>de</strong>rstanding of patterns in animal movements and<br />

selective pressures un<strong>de</strong>rlying these patterns. This is of particular interest in harsh and<br />

spatially-varying environments (Chesson & Huntly 1997, Chesson 2000), where the <strong>de</strong>gree to<br />

which organisms can select alternative habitats through movement influences their ability to<br />

avoid environmental stressors. In many aquatic systems, drought and flooding are natural<br />

features (Matthews 1998, Humphries & Baldwin 2003, Lytle & Poff 2004) that affect both the<br />

temporal and spatial variability of the environment. In<strong>de</strong>ed, except where humans regulate<br />

flow regimes with dams and diversions, almost all freshwater ecosystems are subject to<br />

fluctuations in water level (Matthews 1998). In many aquatic systems, there are elements of<br />

seasonal predictability to the flow regime; however, the onset, magnitu<strong>de</strong>, and duration of<br />

flooding can vary from year to year such that organisms may require a high level of flexibility<br />

in their response (Lytle & Poff 2004). Flow regime generates temporary habitats (e.g.,<br />

floodplains, backwaters) that are exploited by a wi<strong>de</strong> range of organisms, including many fish<br />

species (see review in Williams 2006). These temporary waters serve as critical seasonal<br />

habitats for reproduction, juvenile <strong>de</strong>velopment, growth, and refuges against predation<br />

<strong>de</strong>pending on species life histories and flooding conditions (Snodgrass et al. 1996, Poizat &<br />

Crivelli 1997, Corti et al. 1999, Baber et al. 2002, Ostrand & Wil<strong>de</strong> 2002, King et al. 2003).<br />

As water level <strong>de</strong>clines in floo<strong>de</strong>d areas, habitat quality can concurrently <strong>de</strong>cline for a<br />

number of reasons including <strong>de</strong>creasing food availability (e.g. Capone & Kushlan 1991),<br />

increasing predation pressure (e.g. Kushlan 1976), and <strong>de</strong>clining water quality (e.g. high<br />

temperature, low dissolved oxygen, high pH, Magoulick & Kobza 2003). Fish are<br />

occasionally trapped in <strong>de</strong>siccating pools, resulting in high mortality (Lowe-McConnell 1964,<br />

1975, Kushlan 1976, Capone & Kushlan 1991, Chapman & Kramer 1991, Poizat & Crivelli<br />

1997); however, the majority of individuals emigrate from temporary habitats as waters fall<br />

(e.g. Poizat & Crivelli 1997, Magoulick & Kobza 2003). The <strong>de</strong>gree to which fish can access<br />

and select alternative habitats through immigration to, and emigration from, temporary<br />

habitats influences the <strong>de</strong>gree to which they can exploit the rich resources of flooding habitats,<br />

but also avoid environmental stressors such as high temperature, low dissolved oxygen, or<br />

low food availability. Emigration patterns vary among fish species (Poizat & Crivelli 1997)<br />

and are likely to be influenced by specific characteristics such as their physiological tolerance.<br />

For many fish species, physiological tolerance has been addressed through laboratory studies;<br />

90


Tolérance physiologique et émigration<br />

however, the efficiency of these experimentally <strong>de</strong>rived measures for predicting ecological<br />

processes in nature has been rarely addressed.<br />

In this study, we quantified the timing and patterns of fish emigration from four<br />

temporary wetlands exposed to drought. We focused on interspecific variations in movement<br />

and related this to literature-<strong>de</strong>rived physiological tolerance data to explore the potential role<br />

of physiological tolerance as a predictor of emigration patterns.<br />

6.2. Methods<br />

6.2.1. Study area<br />

We conducted fish surveys from spring to summer 2004 in four temporary wetlands (Sites A,<br />

B, C and D) that represented the range of habitats available in the marsh and were located in a<br />

700-ha area that could be continuously monitored (Figure 6.1 and <strong>de</strong>tails in Chapitre 2 for the<br />

study area). At the start of the survey, Site A was a 6.26-ha grassland with a mean water <strong>de</strong>pth<br />

of 16.3 cm (± 4.7 S.D.) and 49.0% (± 26.3 S.D.) vegetation cover composed of submerged<br />

Poaceae, sub-aquatic plants (e.g. Ranunculus sp.) and filamentous algae (Enteromorpha sp.).<br />

Site B was a 5.25-ha reed bed (Phragmites sp.) with a mean water <strong>de</strong>pth of 22.2 cm (± 14.8<br />

S.D.), a scattered reed cover (40.0% ± 35.1 S.D.) and a substratum consisting mainly of peat<br />

and woody <strong>de</strong>bris. Site C was a 2.13-ha water pond surroun<strong>de</strong>d by reed on the banks. Mean<br />

water <strong>de</strong>pth was 61.0 cm (± 12.6 S.D.), mean vegetation coverage along the banks was 33.2%<br />

(± 23.5 S.D.), and the substratum was composed of mud and peat. Site D was a 0.34-ha <strong>de</strong>ad<br />

end canal with a mean water <strong>de</strong>pth of 105.1 cm (± 40.0 S.D.). Banks were covered by<br />

scattered reed, and the substratum was composed of peat and mud. All sites experienced a<br />

gradual change in physical and chemical conditions: Sites A and B dried out totally during the<br />

survey period, whereas Sites C and D were subjected to a <strong>de</strong>cline in water level (Figure 6.1).<br />

During the trapping study, water transparency (Secchi disk) <strong>de</strong>creased from 55.2 (± 15.2 S.D.)<br />

to 17.8 (± 4.9 S.D.) cm; water pH increased from 8.4 (± 0.1 S.D.) to 9.5 (± 0.2 S.D.); and<br />

conductivity increased from 648 (± 8.4 S.D.) to 1501 (± 10.9 S.D.) μS·cm -1 at Site D. Daily<br />

water temperature at Site D, calculated from maximal daily air temperature during the study<br />

period, and the relationship between maximal daily air temperature and maximal daily water<br />

temperature obtained in 2005 (linear regression, R 2 = 0.86, P < 0.001), indicated a general<br />

increase in water temperature, with large daily fluctuations: 16.9 °C (May 31, 2004) to 32.7°C<br />

(August 1, 2004, Figure 6.1).<br />

91


Chapitre 6<br />

Waterlevel (cm)<br />

Site D<br />

France<br />

Brière<br />

Marsh<br />

Temporary wetland<br />

Site A<br />

Permanent ditch Fyke net<br />

110<br />

105<br />

100<br />

95<br />

90<br />

85<br />

80<br />

75<br />

27-May<br />

1-Jun<br />

6-Jun 6-Jun<br />

11-Jun<br />

16-Jun<br />

21-Jun<br />

26-Jun<br />

1-Jul<br />

6-Jul<br />

Date<br />

Site B<br />

Trapping start Trapping end<br />

11-Jul 11-Jul<br />

16-Jul 16-Jul<br />

Site C<br />

Water level Maximal water temperature<br />

Site A<br />

Site B<br />

21-Jul 21-Jul<br />

26-Jul 26-Jul<br />

Dow nstream<br />

trap<br />

Fish<br />

emigration<br />

Water pond<br />

31-Jul 31-Jul<br />

Site C<br />

Site D<br />

5-Aug 5-Aug<br />

10-Aug<br />

15-Aug<br />

N<br />

50 m<br />

Ditch<br />

Bank<br />

Upstream<br />

trap<br />

Figure 6.1. (top) Map of the study area, <strong>de</strong>scription of the trap setting and trapping period in the<br />

four temporary wetlands. (bottom) Water level and maximal water temperature at Site D (May to<br />

August 2004).<br />

92<br />

35<br />

30<br />

25<br />

20<br />

15<br />

Maximal water temperature(°C)<br />

temperature(°C)


6.2.2. Fish sampling procedure<br />

Tolérance physiologique et émigration<br />

Fish were trapped using fyke nets (see <strong>de</strong>tails in Chapitre 2 for fish sampling procedure). The<br />

survey commenced on May 28, 2004 for all of the four sites when water levels started to<br />

<strong>de</strong>crease continuously. Fyke nets were exposed continuously (overnight) during the entire<br />

period (Figure 6.1). Sites B, C and D were directly connected to the permanent canal. Site A<br />

was connected to Site C that was still floo<strong>de</strong>d when Site A totally dried out; it is likely that<br />

Site C was used as a first refuge by fish emigrating from Site A. Fish immigration into the<br />

wetland was permitted by the presence in each connection of an additional fyke net set in the<br />

opposite direction (Figure 6.1). Immigrating fish represented only 26.9 % of the total catches;<br />

emigration was the dominant movement mo<strong>de</strong>, and we therefore focused on emigration<br />

patterns during the <strong>de</strong>watering period. Sites A and B were monitored until June 28 and June<br />

18, 2004, respectively; and Sites C and D were monitored until August 3, 2004 (Figure 6.1).<br />

Trapping was conducted until all fish had left each temporary wetland. Complete emigration<br />

from Sites C and D was confirmed by electrofishing the remaining water (mean water <strong>de</strong>pth<br />

33.8 cm (± 9.54 S.D.) and 84.8 cm (± 29.3 S.D.) in Sites C and D, respectively) at the end of<br />

the trapping survey (August 4, 2004). No fish were caught in these two sites, whereas fish<br />

were recor<strong>de</strong>d in the adjacent permanent canal.<br />

6.2.3. Data analyses<br />

6.2.3.1. Trapping data<br />

To facilitate comparisons between sites, trapping data were standardized to daily capture rates<br />

for each site and each species and expressed in catch per unit of effort (CPUE expressed in<br />

fish.net -1 .day -1 ). Emigration start (ES) was calculated as the number of trapping days before<br />

the first individual was captured leaving a site, referenced to the beginning of the study.<br />

Emigration maximum (EMX) was calculated as the number of trapping days before the<br />

maximum CPUE value was reached. Emigration end (EE) was calculated as the number of<br />

trapping days before the last individual was captured. Two synthetic variables were then<br />

calculated to <strong>de</strong>scribe the timing of fish emigration from the temporary wetlands. First,<br />

emigration range (ER) was <strong>de</strong>fined as the difference between ES and EE. Second, we<br />

calculated the emigration moment (EM) as the number of days of trapping required to catch<br />

50% of the total number of fish captured over the survey. These variables were calculated for<br />

each species and expressed relative to the trapping duration for each site.<br />

93


Chapitre 6<br />

6.2.3.2. Physiological tolerance<br />

We used literature-<strong>de</strong>rived values as proxies of fish physiological tolerance. First, we used the<br />

physiological tolerance in<strong>de</strong>x (PTI), which represents the tolerance to changes in water<br />

quality (usually temperature, dissolved oxygen, turbidity and salinity) or to extreme<br />

conditions in water quality (Marchetti et al. 2004), following the classification of Halliwell et<br />

al. (1999) and with the addition of the extremely tolerant category from Marchetti et al.<br />

(2004), i.e. 1: intolerant, 2: mo<strong>de</strong>rately tolerant, 3: tolerant and 4: extremely tolerant. Second,<br />

we used the coefficient of water quality flexibility (WQF) after Verneaux (1981) with low<br />

values indicating a narrow range of acceptable water quality (see also Oberdorff et al. 2002).<br />

Third, because thermal tolerance is a well-documented variable of physiological tolerance in<br />

fishes, we used the temperature upper avoidance (TUA) <strong>de</strong>fined as the temperature at which<br />

fish spend less time in comparison to a control (<strong>de</strong>tailed in Giattinna & Garton 1982) and that<br />

<strong>de</strong>limits the upper boundary of the zone of thermal preference (Coutant 1977). These three<br />

parameters were gathered from published sources for the six most abundant species captured<br />

during the trapping period (Table 6.1).<br />

Table 6.1. Physiological tolerance in<strong>de</strong>x (PTI), coefficient of water quality flexibility (WQF) and<br />

temperature of upper avoidance (TUA in °C) of the six most commonly trapped species in<br />

wetlands of the Gran<strong>de</strong> Brière Mottière Marsh, France with bibliographical references.<br />

Species PTI WQF TUA (in°C)<br />

European eel (Anguilla anguilla) 3 a - 29 c<br />

Northern pike (Esox lucius) 2 a 5.5 b 23 d<br />

Mosquitofish (Gambusia holbrooki) 4 a - 35 e<br />

Black bullhead (Ameiurus melas) 4 a 6.5 b 35 f<br />

Pumpkinseed (Lepomis gibbosus) 2 a 5.5 b 25 g<br />

Rudd (Scardinius erythrophthalmus) 3 a 6 b 28 h<br />

- <strong>de</strong>notes missing value. a Marchetti et al. (2004) adapted from Halliwell et al. (1999); b Verneaux<br />

(1981); c Sadler (1979); d Casselman (1978); e Cherry et al. (1976); f Beltz et al. (1974); g Evans<br />

(1977) and h Elliot (1981).<br />

94


6.3. Results<br />

6.3.1. Fish assemblage<br />

Tolérance physiologique et émigration<br />

In total, 6,321 fish belonging to 12 species were captured emigrating from the four temporary<br />

wetlands (185, 346, 552 and 5,238 from Sites A, B, C and D, respectively). The fish<br />

assemblages were dominated by six species that represented 98.6% of the total number of fish<br />

captured. In or<strong>de</strong>r of <strong>de</strong>creasing abundance (based on mean relative abundance, Table 6.2),<br />

the six most sampled species were juvenile and adult black bullhead (Ameiurus melas,<br />

Rafinesque), sub-adult European eel (Anguilla anguilla, L.), juvenile and adult rudd<br />

(Scardinius erythrophthalmus, L), adult pumpkinseed (Lepomis gibbosus, L.), young-of-theyear<br />

Northern pike (Esox lucius, L.), and juvenile and adult mosquitofish (Gambusia<br />

holbrooki, Baird & Girard). Species differed markedly in their relative abundance among the<br />

four sites, e.g., from 10.3 to 98.4 % for the black bullhead and from 30.3 to 1.4 % for the<br />

European eel (Table 6.2). The other species captured in the survey (representing 1.4 % of the<br />

total number) were roach (Rutilus rutilus L.), common bream (Abramis brama, L.), prussian<br />

carp (Carassius gibelio, Bloch), tench (Tinca tinca, L.), common carp (Cyprinus carpio, L.),<br />

and pikeperch (San<strong>de</strong>r lucioperca, L.).<br />

Table 6.2. Number, life stage, and relative abundance of the six most abundant species monitored<br />

emigrating from the four temporary wetlands (Sites A, B, C and D) of the Gran<strong>de</strong> Brière Mottière<br />

Marsh, France. Other fish species are listed in the text and represented 0.6, 1.1, 13.8 and 0.09% of<br />

the total abundance in Sites A, B, C and D, respectively.<br />

Relative abundance per site (%)<br />

Species N Life stage A B C D<br />

Black bullhead 5530 Juveniles & adults 10.3 10.7 57.6 98.4<br />

European eel 320 Sub-adults 30.3 23.7 20.1 1.4<br />

Rudd 204 Juveniles & adults 8.1 53.5 0.7 0.0<br />

Pumpkinseed 78 Adults 8.6 6.9 6.2 0.1<br />

Northern pike 59 Young-of-the-year 24.3 1.2 1.6 0.01<br />

Mosquitofish 43 Juveniles & adults 17.8 2.9 0.0 0.0<br />

6.3.2. Fish emigration patterns among sites<br />

Emigration patterns were analysed for the six most abundant species (i.e. mean abundance per<br />

site >5%, Table 6.2). Figure 6.2 illustrates daily catch proportional to the total catch of each<br />

species and relative to the trapping duration in the four sites (i.e. 22, 32, 32, and 68 days for<br />

Sites A, B, C, and D, respectively). Despite some variation at the start of the survey when<br />

Sites A and B are compared to Sites C and D, the interspecific patterning of emigration was<br />

surprisingly similar. Pike were the first to emigrate, and did so with the shortest emigration<br />

95


Chapitre 6<br />

duration during the first third of the survey. European eel left temporary wetlands primarily<br />

during the first third of the survey but over a relatively long period. Pumpkinseed emigration<br />

was characterised by a short emigration duration that took place during the first half of the<br />

trapping survey. Migration timing of rudd ten<strong>de</strong>d to occur during the first half of the survey.<br />

Black bullhead exhibited <strong>de</strong>layed emigration that occurred primarily during the second half of<br />

the survey, and mosquitofish emigrated the last and over a brief period.<br />

Relative daily capture (%)<br />

Relative daily capture (%)<br />

Mosquitofish<br />

Site A Site B<br />

Black bullhead<br />

Rudd<br />

Pumpkinseed<br />

European eel<br />

Pike<br />

0 25 50 75 100<br />

Trapping duration (%)<br />

Site C Mosquitofish Site D<br />

Black bullhead<br />

Rudd<br />

Pumpkinseed<br />

European eel<br />

Pike<br />

0 25 50 75 100<br />

Trapping duration (%)<br />

Relative daily capture (%)<br />

Relative daily capture (%)<br />

Mosquitofish<br />

Black bullhead<br />

Rudd<br />

Pumpkinseed<br />

European eel<br />

Pike<br />

0 25 50 75 100<br />

Trapping duration (%)<br />

Mosquitofish<br />

Black bullhead<br />

Rudd<br />

Pumpkinseed<br />

European eel<br />

Pike<br />

0 25 50 75 100<br />

Trapping duration (%)<br />

Figure 6.2. Emigration patterns expressed relative to the trapping duration of the six studied fish<br />

species in the four temporary wetlands (Sites A, B, C and D). Sha<strong>de</strong>d areas were constructed by<br />

joining daily catch values proportional to the total catch per site and per species. No mosquitofish<br />

were captured in Sites C and D and no rudd in Site D (dashed lines).<br />

96


6.3.3. Species emigration ranking<br />

Tolérance physiologique et émigration<br />

Trapping data of the four sites were pooled to compare quantitatively the emigration among<br />

species using EM and ER (ES and EE, Figure 6). Pike ten<strong>de</strong>d to emigrate first (EM = 0.15)<br />

and relatively briefly (ER = 0.48), followed by pumpkinseed (EM = 0.16; ER = 0.32) and<br />

European eel (EM = 0.20) with the most exten<strong>de</strong>d range (ER = 0.84). The three other species<br />

emigrated later. Then, rudd left temporary wetlands (EM = 0.31) and their emigration was<br />

characterized by a relatively short range (ER = 0.45) whereas black bullhead emigrated later<br />

(EM = 0.53) and over a more exten<strong>de</strong>d time period (ER = 0.77). Mosquitofish was the last<br />

species to emigrate just before the complete drying out (EM = 0.87) and did so over a short<br />

period (ER = 0.28).<br />

Emigration moment<br />

1.0<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

ES<br />

Pumpkinseed<br />

Pike<br />

EMX EE<br />

Rudd<br />

Mosquitofish<br />

Black bullhead<br />

European eel<br />

0.0<br />

0 20 40 60 80 100<br />

Trapping duration (%)<br />

Figure 6.3. Emigration range (emigration start ES and emigration end EE) and emigration moment<br />

(EM) expressed relative to the trapping duration of the six studied fish species in the four<br />

temporary wetlands. Each box plot is limited by the first and third quartile of captures and the<br />

vertical line represents the emigration maximum (EMX). Sha<strong>de</strong>d boxes are used for non-native<br />

species and white boxes for native ones. See text for abbreviation <strong>de</strong>finitions.<br />

6.3.4. Physiological tolerance and emigration timing<br />

Irrespective of the physiological <strong>de</strong>scriptor (Table 6.1), the emigration moment (EM) was<br />

positively correlated with the physiological tolerance of the species (Figure 6.4). Later<br />

emigrating species were characterized by a higher physiological tolerance (Pearson<br />

correlation, r > 0.861, P < 0.028, Figure 6.4). Interestingly, two non-native species (black<br />

bullhead and mosquitofish) that had the highest physiological tolerance of the six species<br />

consi<strong>de</strong>red, emigrated the last (Figure 6.4).<br />

97


Chapitre 6<br />

Emigration moment (EM)<br />

1<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

1<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

Pike<br />

0<br />

5.4 5.6 5.8 6 6.2 6.4 6.6<br />

Coefficient of water quality flexibility<br />

1<br />

0.8<br />

0.6<br />

0.4<br />

R = 0.861<br />

p = 0.028<br />

0 1 2 3 4 5<br />

Physiological tolerance in<strong>de</strong>x<br />

R = 0.996<br />

p = 0.004<br />

Pumpkinseed<br />

R = 0.874<br />

p = 0.023<br />

Pumpkinseed<br />

Pike<br />

Rudd<br />

Rudd<br />

Mosquitofish<br />

Rudd<br />

European eel<br />

Black bullhead<br />

Black bullhead<br />

Mosquitofish<br />

Black bullhead<br />

0.2<br />

Pike<br />

Pumpkinseed<br />

European eel<br />

0<br />

20 22 24 26 28 30 32 34 36<br />

Temperature of upper avoidance (°C)<br />

Figure 6.4. Emigration patterns of fish species in the four temporary wetlands as a function of their<br />

physiological tolerance, i.e. the physiological tolerance in<strong>de</strong>x (PTI), the coefficient of water quality flexibility<br />

(WQF) and the temperature of upper avoidance (TUA in °C). Black symbols are used for non-native species and<br />

white symbols for native ones.<br />

98


6.4. Discussion<br />

Tolérance physiologique et émigration<br />

Our results <strong>de</strong>monstrated that for the six most abundant fish species, (1) emigration patterns<br />

differed among species, (2) interspecific patterning was similar among sites, and (3) there was<br />

a significant relationship between physiological tolerance and emigration timing. Matthews &<br />

Marsh-Matthews (2003) emphasized that little was known about the applicability of the<br />

results of laboratory tolerance studies to fish in the field un<strong>de</strong>r sustained drought conditions.<br />

This study suggests that literature data on fish species physiological tolerance from<br />

experimental studies are relevant to both un<strong>de</strong>rstanding and predicting fish emigration from<br />

temporary wetlands during drought.<br />

We might expect use of temporary wetlands to reflect a tra<strong>de</strong>-off between the benefits<br />

of temporary habitats and the costs of exploitation; and this tra<strong>de</strong>-off is likely to vary among<br />

species. Temporary shallow habitats are known to provi<strong>de</strong> good feeding and growth<br />

conditions for many fish (Neckles et al. 1990) and to provi<strong>de</strong> protection against predators for<br />

juveniles and smaller species (Schlosser 1987). However, benefits might be counterbalanced<br />

by costs linked to physiological stress due to the reduction of both available habitat and water<br />

quality during drought (Magoulick & Kobza 2003, Matthews & Marsh-Matthews 2003). Such<br />

tra<strong>de</strong>-offs occur when a physiological, behavioural or ecological trait of an organism (e.g.<br />

wetland use in the present study) that confers advantage for performing one function (e.g.<br />

growth performance) simultaneously confers a disadvantage for performing another function<br />

(e.g. physiological maintenance) (Chase & Leibold 2003).<br />

Our results suggest that the emigration patterns are quite predictable and that habitat<br />

selection behaviour is involved. Habitat selection (interpreted as an evolutionary strategy,<br />

Morris 2003) might be useful in interpreting habitat use in spatially varying environments<br />

such as temporary wetlands. Habitat selection generally refers to the non-random use of space<br />

or habitat that results from the voluntary movement of organisms (Kramer et al. 1997). It has<br />

been argued that in seasonally floo<strong>de</strong>d systems, habitat distributions reflect both habitat<br />

selection but also non-random movement (e.g. haphazard trapping in floodplain pools, or<br />

downstream flushing in high-gradient systems). In the low-gradient wetlands of the Gran<strong>de</strong><br />

Brière Mottière Marsh that we studied, fishes were clearly able to move from temporary to<br />

permanent waters, as evi<strong>de</strong>nced by the absence of fish in the remaining water (Sites C and D)<br />

at the end of the survey and our observations of directed movement to and from the sites.<br />

Therefore, we assume that emigration among sites is voluntary to some <strong>de</strong>gree, and therefore<br />

reflects an element of choice. If habitat selection is driving distribution in wetlands, then we<br />

99


Chapitre 6<br />

would expect each species to maximize its fitness by optimizing the time spent in temporary<br />

wetlands and to emigrate when use of the temporary wetlands does not benefit the fish in<br />

comparison to alternative habitats (e.g. permanent waters). This study suggests that the costs<br />

of staying in the temporary wetlands relate, at least in part, to the interaction between<br />

changing water quality conditions and the physiological tolerance of the inhabitants. The<br />

strong relationships observed between the species-specific emigration moments and<br />

physiological tolerance estimates suggest that tolerance indices are strong predictors of<br />

emigration patterns, and that environmental cues were effectively used by these species.<br />

Physiological tolerance has been reported to influence freshwater fish assemblage<br />

composition and patterns of mortality in response to extreme conditions (e.g. Smale & Rabeni<br />

1995, McKinsey & Chapman 1998, Ostrand & Wil<strong>de</strong> 2002). Kushlan (1974), in his study of a<br />

natural fish kill in the Big Cypress Swamp of Florida, classified fish species into three<br />

categories based on their survival. The mosquitofish fell into the most tolerant category, while<br />

centrarchids were eliminated rapidly supporting the results of our study. Physiological<br />

tolerance also has implications for persistence and impacts of non-native species (Kobza et al.<br />

2003). In the wetlands of the Gran<strong>de</strong> Brière Mottière Marsh, black bullhead and mosquitofish<br />

are non-native, well-established species. These two species were characterized by the highest<br />

emigration moments in our study that correlated with their high physiological tolerance. The<br />

capacity of mosquitofish and the black bullhead to exploit temporary wetlands may contribute<br />

to their wi<strong>de</strong>spread success as invasive species. Mosquitofish, because of their reputation as<br />

mosquito-control agents, have been stocked in many temperate and tropical areas throughout<br />

the world (Fuller et al. 1999), and they have been remarkably successful inva<strong>de</strong>rs with the<br />

greatest ecological impact of introduced poeciliid fishes (Courtenay & Meffe 1989, Fuller<br />

1999). They inhabit a wi<strong>de</strong> variety of habitats reflective of their broad physiological tolerance.<br />

For example, mosquitofish tolerate very high water temperatures (Otto 1973) and extremely<br />

low levels of dissolved oxygen (McKinsey & Chapman 1998). Successful exploitation of<br />

shallow hypoxic waters by mosquitofish has been attributed, in part, to their apparent<br />

proficiency at aquatic surface respiration (Lewis 1970, McKinsey & Chapman 1998). The<br />

black bullhead has been reported as a successful non-indigenous species (e.g. Wheeler 1978,<br />

Fuller et al. 1999). It may be favoured by its broad physiological tolerance range and its<br />

ability to perform aquatic surface respiration as reported by Sargent & Galat (2002). The way<br />

in which a species responds to resources, natural enemies, and the physical environment<br />

<strong>de</strong>termines its ability to inva<strong>de</strong> (Shea & Chesson 2002), and the role of high physiological<br />

100


Tolérance physiologique et émigration<br />

tolerance in the success and extent of fish inva<strong>de</strong>rs has already been recognized (Marchetti et<br />

al. 2004).<br />

In summary, the present study provi<strong>de</strong>s evi<strong>de</strong>nce for strong emigration patterns of<br />

fishes from temporary wetlands. Literature-<strong>de</strong>rived physiological tolerance indices serve as<br />

good predictors of interspecific variation in emigration timing. However, it should be noted<br />

that residual variance in the relationships between tolerance levels and emigration timing<br />

suggest that other factors interact with physico-chemical stressors to influence emigration<br />

timing at the individual and species level. For example, species characteristics including body<br />

size and life history traits can affect patterns of movement and survival in seasonally flooding<br />

systems such as the Florida Evergla<strong>de</strong>s (Kushlan 1974, Trexler et al. 2003). Other biotic<br />

factors such as food availability and predation pressure are known to vary during the<br />

progressive drought of the temporary wetlands (Capone & Kushlan 1991) and may interact<br />

with physico-chemical stressors to affect population dynamics. Future studies that explore the<br />

interactive effects of abiotic and biotic stressors will be extremely useful in increasing our<br />

un<strong>de</strong>rstanding of emigration patterns from temporary waters.<br />

Acknowledgments<br />

We thank the Parc naturel régional <strong>de</strong> Brière for logistic support, the FEDER, DIREN<br />

(MEDD), Région Pays <strong>de</strong> la Loire and Agence <strong>de</strong> l’Eau from Loire-Bretagne for financial<br />

support and particularly J.-P. Damien and E. Le Mitouard for assistance during field work.<br />

We are grateful to J. Ol<strong>de</strong>n, J.-M. Roussel, P. Marmonier and two anonymous referees for<br />

valuable comments on earlier drafts of the manuscript.<br />

101


Chapitre 6<br />

Il ressort <strong>de</strong> cette étu<strong>de</strong> que les patrons d’émigration <strong>de</strong>s six espèces les plus<br />

abondantes sont différents d’une espèce à l’autre, mais que d’un site à l’autre,<br />

l’ordre dans lequel ces émigrations ont lieu est similaire. De plus, nous avons<br />

montré qu’il existe une relation très forte entre ces patrons d’émigration et la<br />

tolérance physiologique décrite dans la littérature. Cette approche novatrice<br />

prouve que ce type <strong>de</strong> données peut être utilisé avec succès pour comprendre les<br />

processus d’échanges <strong>de</strong> poissons en relation avec l’utilisation <strong>de</strong>s MTI. Nous<br />

avons également montré que les caractéristiques propres <strong>de</strong>s espèces ont <strong>de</strong>s<br />

conséquences directes sur le fonctionnement <strong>de</strong> la communauté dans les MTI<br />

caractérisés par <strong>de</strong> fortes contraintes environnementales. Cependant, il sera<br />

nécessaire à l’avenir <strong>de</strong> confronter directement et plus précisément la tolérance<br />

physiologique <strong>de</strong>s espèces aux mesures directes <strong>de</strong>s paramètres<br />

environnementaux in situ, telles que la température ou la concentration en<br />

dioxygène. Ce type d’approche pourrait aussi être employé pour l’étu<strong>de</strong> <strong>de</strong>s<br />

processus d’émigration sur un pas <strong>de</strong> temps réduit puisque <strong>de</strong> fortes variations <strong>de</strong><br />

ces paramètres ont lieu au cours du cycle jour/nuit. Une analyse plus fine <strong>de</strong>s<br />

sources <strong>de</strong> variabilité <strong>de</strong> cette tolérance physiologique, notamment le rapport<br />

taille du corps / surface branchiale, pourrait également permettre d’affiner ces<br />

résultats. Enfin, <strong>de</strong> nouvelles questions se posent suite à ce travail : puisqu’il<br />

existe une variance au sein <strong>de</strong> chaque patron spécifique, quelle en est la (les)<br />

source(s) ? Peut-on relier cette variance aux différentes classes d’âges d’une<br />

espèce ou <strong>de</strong> plusieurs et aux différents rôles fonctionnels <strong>de</strong>s MTI pour ces<br />

différents sta<strong>de</strong>s <strong>de</strong> vie, voire pour les différentes espèces ? De plus, ces sources<br />

<strong>de</strong> variance sont-elles les mêmes lors <strong>de</strong> l’immigration et <strong>de</strong> l’émigration,<br />

processus qui se déroulent dans <strong>de</strong>s conditions environnementales très<br />

différentes ?<br />

102


Chapitre 7<br />

Rôle <strong>de</strong>s facteurs environnementaux dans la variabilité<br />

inter- et intra-spécifique <strong>de</strong>s processus d’immigration et<br />

d’émigration<br />

Dans le chapitre précé<strong>de</strong>nt, le rôle <strong>de</strong> la tolérance physiologique dans les<br />

processus d’échanges <strong>de</strong> poissons entre milieux permanents et temporairement<br />

inondés a été démontré dans une comparaison inter-spécifique. Cependant cette<br />

analyse n’a pas tenu compte <strong>de</strong> manière directe <strong>de</strong>s changements <strong>de</strong>s paramètres<br />

environnementaux ayant lieu dans les MTI. Dans le présent chapitre, l’analyse <strong>de</strong>s<br />

échanges sera réalisée en liaison à ces paramètres. Par ailleurs, la réponse d’un<br />

individu (ici le mouvement du poisson d’un milieu à un autre) à ces paramètres<br />

environnementaux peut être différente au cours du développement ontogénique<br />

(en d’autres termes, en fonction <strong>de</strong> l’âge du poisson). Nous posons l’hypothèse<br />

que l’âge <strong>de</strong>s individus peut être une source majeure <strong>de</strong> variabilité intra-spécifique<br />

<strong>de</strong>s processus d’échange <strong>de</strong> poissons entre milieux. C’est pourquoi les<br />

investigations présentées ci-après reposent sur un arrangement <strong>de</strong>s poissons <strong>de</strong><br />

chaque espèce en différents sta<strong>de</strong>s <strong>de</strong> développement. Ils sont au nombre <strong>de</strong> trois<br />

et sont définis en fonction du rôle fonctionnel <strong>de</strong>s MTI pour ces individus: adultes<br />

matures (MTI = zone <strong>de</strong> reproduction), juvéniles non matures (MTI = zone <strong>de</strong><br />

nourricerie) et juvéniles <strong>de</strong> l’année, (MTI = zone <strong>de</strong> naissance). Cette étu<strong>de</strong><br />

portant sur l’ensemble du cycle <strong>de</strong> crue printanière présentera les déplacements <strong>de</strong><br />

poissons dans les <strong>de</strong>ux sens (immigration et émigration <strong>de</strong> MTI) et les analysera<br />

en fonction <strong>de</strong> l’évolution <strong>de</strong>s facteurs environnementaux.


Chapitre 7<br />

Installation d’une son<strong>de</strong> enregistreuse <strong>de</strong> température dans le plan d’eau<br />

Prélèvement <strong>de</strong> zooplancton dans la praire temporairement inondée<br />

104


Facteurs environnementaux, immigration et émigration<br />

Dynamical processes of temporary waters use by fish: species and<br />

life stages responses to environmental changes 5<br />

Résumé: Les milieux temporairement inondés sont <strong>de</strong>s habitats fondamentaux pour <strong>de</strong><br />

nombreuses espèces <strong>de</strong> poisons. Malgré <strong>de</strong>s conditions environnementales très fluctuantes, ils<br />

servent <strong>de</strong> zones <strong>de</strong> reproduction pour les adultes, <strong>de</strong> nourricerie pour les juvéniles et <strong>de</strong><br />

naissance pour les juvéniles <strong>de</strong> l’année. Les processus dynamiques d’utilisation <strong>de</strong> ces<br />

milieux par les poissons ont déjà été étudiés. Ces étu<strong>de</strong>s, conduites la plupart du temps au<br />

niveau spécifique, ont d’ailleurs démontré le rôle important <strong>de</strong>s paramètres environnementaux<br />

dans ces échanges. Dans la présente étu<strong>de</strong>, nous avons analysé ces processus pour les<br />

différents sta<strong>de</strong>s <strong>de</strong> vie <strong>de</strong>s espèces car il a été démontré que leurs sensibilités aux stimuli<br />

environnementaux évoluent au cours du développement ontogénétique, laissant présager<br />

l’existence d’une source potentielle <strong>de</strong> variabilité intra-spécifique <strong>de</strong> ces processus. Nous<br />

avons donc suivi les déplacements <strong>de</strong> poissons <strong>de</strong> mars à août 2005 dans les marais <strong>de</strong> Brière<br />

le long d’un gradient <strong>de</strong> milieux temporairement inondés et permanents qui se sont asséchés<br />

successivement durant l’étu<strong>de</strong>. Les poissons ont été capturés à l’ai<strong>de</strong> pièges verveux disposés<br />

à l’unique connexion entre chaque habitat et son milieu adjacent. En parallèle, certains<br />

paramètres environnementaux (niveau d’eau, température et abondance <strong>de</strong> zooplancton) ont<br />

été suivis. Basés sur la capture <strong>de</strong> 8670 individus appartenant à 15 espèces différentes, nos<br />

résultats démontrent que la composition <strong>de</strong> la communauté piscicole était différente entre les<br />

trois milieux temporairement inondés étudiés et que les déplacements (immigration et<br />

émigration) étaient très variables à la fois entre les espèces et entre les sta<strong>de</strong>s <strong>de</strong> vie. Ainsi,<br />

nous avons démontré que les immigrations <strong>de</strong> poissons étaient regroupées par espèce alors<br />

que les émigrations étaient regroupées par sta<strong>de</strong> vie. De plus, nous avons mis en évi<strong>de</strong>nce que<br />

ces déplacements étaient fortement influencés par les paramètres environnementaux mais que<br />

ces influences n’étaient pas les mêmes entre les espèces, les sta<strong>de</strong>s <strong>de</strong> vie et le sens <strong>de</strong>s<br />

déplacements.<br />

5 Ce chapitre est adapté d’un article en préparation écrit par Cucherousset, J., Carpentier, A. & Paillisson, J.-M.<br />

105


Chapitre 7<br />

Abstract: Temporary floo<strong>de</strong>d habitats, i.e. aquatic habitats that cyclically dry out, are<br />

key habitats for fishes. In<strong>de</strong>ed, these habitats, whose environmental conditions strongly varied<br />

during the flooding period, serve as spawning ground for adults, nursery area for juveniles<br />

and hatching ground for young-of-the-year individuals. Several studies have already been<br />

conducted on the dynamical processes related to temporary floo<strong>de</strong>d habitats use by fish and<br />

the role of environmental variables in these displacements but principally at the species level.<br />

Here, we investigated the patterns of immigration to, and emigration from, temporary floo<strong>de</strong>d<br />

habitats with respect to species life stages because fish responses to environmental variables<br />

are known to <strong>de</strong>velop during ontogeny. From March to August 2005, we monitored fish<br />

displacements in the Brière marsh (Northwest France). This was conducted along the<br />

permanent-temporary floo<strong>de</strong>d gradient, i.e. between a permanent canal and three temporary<br />

floo<strong>de</strong>d habitats that dried out successively in the season. Fish displacements were monitored<br />

using fyke nets that were set up at the unique connecting point between each temporary<br />

floo<strong>de</strong>d habitat and its adjacent habitat and environmental parameters (water level,<br />

temperature and zooplankton abundance) were monitored in parallel. In total, 8,670 fish<br />

belonging to 15 species were captured immigrating in or emigrating from the three temporary<br />

floo<strong>de</strong>d habitats. Our results <strong>de</strong>monstrated that fish assemblage composition differed between<br />

the three temporary floo<strong>de</strong>d habitats and that timing and pattern of fish displacements<br />

(immigration/emigration) were highly variable both among species and among life stages.<br />

More precisely, fish immigration timing and patterns were more closely arranged according to<br />

species whereas emigration processes were mainly arranged according to life stages. Fish<br />

displacements were largely influenced by the succession of environmental conditions<br />

throughout the flooding episo<strong>de</strong>, the responses being, however, variable among species, life<br />

stages, and directions of displacement.<br />

106


7.1. Introduction<br />

Facteurs environnementaux, immigration et émigration<br />

During the last <strong>de</strong>ca<strong>de</strong>s, much research has <strong>de</strong>monstrated the biological significance of<br />

temporary floo<strong>de</strong>d habitats, i.e. aquatic habitats that cyclically dry out, for a large number of<br />

plant and animal organisms (see review in Williams 2006). This is particularly the case for<br />

fish of various ecosystems (Welcomme 1979, Snodgrass et al. 1996, Magoulick 2000, King et<br />

al. 2003, DeAngelis et al. 2005, Trexler et al. 2005). For fish, temporary habitats play at least<br />

three principal functional roles according to species life history and life stages: (1) spawning<br />

grounds for adults, (2) hatching grounds for young-of-the-year individuals and (3) nursery<br />

areas for juveniles. In<strong>de</strong>ed, temporary floo<strong>de</strong>d habitats are known to provi<strong>de</strong> suitable substrata<br />

and adapted thermal conditions for spawning, sufficient food resources for larvae and<br />

juveniles, and refuge against large aquatic predators (Snodgrass et al. 1996, Poizat & Crivelli<br />

1997, Corti et al. 1999, Baber et al. 2002, Ostrand & Wil<strong>de</strong> 2002, King et al. 2003).<br />

Temporary floo<strong>de</strong>d habitats are also characterised by the strong variation of their environment<br />

due to flood regime during the year and the synergic changes of several environmental<br />

variables (e.g. temperature, oxygen concentration, pH, etc...).<br />

Facing these dynamic conditions, fish select habitats generally through migration to<br />

and from temporary floo<strong>de</strong>d habitats. In<strong>de</strong>ed, in temperate waters where fish do not normally<br />

possess mechanisms to persist during drought (Sayer 2005), the use of temporary floo<strong>de</strong>d<br />

habitats will occur where the necessary connection exists with a permanent water source and<br />

<strong>de</strong>pend upon dynamical processes of immigration and emigration as water level evolves<br />

(Williams 2006). The <strong>de</strong>gree to which fish can access and select alternative habitats through<br />

immigration to and emigration from temporary floo<strong>de</strong>d habitats influences the <strong>de</strong>gree to<br />

which they can exploit the resources of floo<strong>de</strong>d habitats (e.g. spawning substrate, food<br />

resources), but also avoid the environmental stressors occurring as water level <strong>de</strong>cline (e.g.<br />

high temperature, low dissolved oxygen, low food availability, or increased bird predation)<br />

(Kushlan 1976). It is, therefore, probable that patterns of fish exchanges between temporary<br />

floo<strong>de</strong>d habitats and the adjacent habitats vary among fish species and also among life stages.<br />

In previous investigations, fish immigration and emigration from temporary floo<strong>de</strong>d<br />

habitats has been found to correlate with some environmental variables (e.g. temperature,<br />

water level, food availability), and the species-specific patterns vary according to<br />

characteristics such as life history, physiological tolerance or body size (e.g. Poizat & Crivelli<br />

1997, Hohausova 2000, Hohausova et al. 2003, Trexler et al. 2003, Chapitre 6). However, to<br />

our knowledge, all investigations on fish exchanges between permanent and temporary<br />

107


Chapitre 7<br />

floo<strong>de</strong>d habitats focused on inter specific variations, regardless to the variable functions<br />

played by temporary floo<strong>de</strong>d habitats for the different life stages of a given species. In these<br />

previous studies, residual variance in the relationships was observed at the inter-specific level,<br />

suggesting that other factors influence fish displacements at the intra-specific level. This<br />

result encouraged us to investigate the patterns of immigration to and emigration from<br />

temporary floo<strong>de</strong>d habitats with respect to species life stages because fish responses to<br />

environmental variables are known to change during ontogeny (e.g. Coutant 1977, Robb &<br />

Abrahams 2003, Boët 2001, Matthews 1998). In the present study, we inten<strong>de</strong>d to test the<br />

hypothesis that the variations of fish individual responses to the environmental conditions due<br />

to the ontogenetic <strong>de</strong>velopment will generate the major part of the intra-specific variability in<br />

the pattern of use of temporary habitats. Concretely, the objectives of this study were to: (1)<br />

quantify the timing and patterns of displacements in and from three contrasted temporary<br />

habitats distributed along the permanent-temporary floo<strong>de</strong>d gradient during a flooding<br />

episo<strong>de</strong> for the three life stages of fish species; and (2) explore the potential role of some<br />

environmental variables (i.e. water level, temperature and food resources) as predictors of the<br />

three fish life stages displacements (i.e. young-of-the-year, juvenile and adult). If the<br />

influence of environmental parameters on displacements varies between life stages, then we<br />

would expect differences in fish displacement patterns and the variance in these patterns<br />

between life stages of a species might be either higher or similar to that found between<br />

species for similar life stages. Conversely, if the effects of the environmental variables on<br />

different life stages of a species do not influence displacements, then the species responses to<br />

the varying environmental constraints would represent the major driving forces, and the<br />

particular ecological requirements of each life stage would be of a secondary importance. By<br />

doing so, we use a functional approach of habitat selection in spatially varying environment to<br />

investigate the source of variance within a community and species responses.<br />

7.2. Materiel and methods<br />

7.2.1. Study site<br />

The study was carried out from March to August 2005 and we investigated fish movements<br />

along the permanent-temporary floo<strong>de</strong>d gradient, i.e. between a permanent canal and three<br />

temporary floo<strong>de</strong>d habitats (a temporary pond, a reed bed and a grassland) that dried out<br />

successively in the season and that constitute the range of habitats available in the marsh<br />

(Figure 7.1 and <strong>de</strong>tails in Chapitre 2).<br />

108


N<br />

100 m<br />

Grassland<br />

Facteurs environnementaux, immigration et émigration<br />

Reed bed<br />

Permanent<br />

canal<br />

Temporary<br />

pond<br />

Bank<br />

Emigration<br />

Immigration<br />

Figure 7.1. Description of the trap setting (Δ) to catch fish entering or leaving the three temporary<br />

floo<strong>de</strong>d habitats (grassland, reed bed and temporary pond) monitored from March to August 2005.<br />

The reed bed and the pond were directly connected to the permanent canal and the grassland was<br />

connected to the pond.<br />

At the start of the survey, based on multiple transects distributed through the sites, the<br />

permanent canal was 108.4 cm (± 37.2 S.D.) <strong>de</strong>pth, 19.6 (± 10.0) m width with a low<br />

vegetation coverage (8.3 ± 18.3 %), on average, only the banks being covered by reed<br />

(Phragmites australis). The pond was 2.13 ha and surroun<strong>de</strong>d by reed on the banks (mean<br />

vegetation coverage of 6.6 ± 22.1 %) and had a mean water <strong>de</strong>pth of 65.5 (± 17.2) cm. The<br />

reed bed was 5.25 ha with a mean water <strong>de</strong>pth of 21.2 (± 15.8) cm and a reed bed cover (33.1<br />

± 33.3) %. The grassland (0.47 ha) was the shallowest (mean water <strong>de</strong>pth of 17.3 (± 5.6) cm )<br />

and the more vegetated site (50.25 (± 65.8) %), vegetation cover being composed of<br />

submerged Poaceae, sub-aquatic plants (e.g. Ranunculus sp.) and filamentous algae<br />

(Enteromorpha sp.). In the four sites, the substrata were composed mainly of peat and mud.<br />

109<br />

Bank


Chapitre 7<br />

Importantly, fish can only colonize the temporary floo<strong>de</strong>d habitats via gaps in the banks (3-5<br />

m width, Figure 7.1). In<strong>de</strong>ed, these banks, which result from canal and pond dredging<br />

operations (peat and mud stocked along the canal and the pond), stayed unfloo<strong>de</strong>d throughout<br />

the flooding period. The survey started just after the last frost event (i.e. on March 15, when<br />

the water level was increasing) and finished when all sites either totally dried or disconnected<br />

from the permanent habitat. In<strong>de</strong>ed, the grassland and the reed bed dried out totally on June,<br />

15 and June, 26, respectively (i.e. 62 and 103 days of survey) whereas the pond became<br />

isolated from the permanent canal on 18 August (i.e. 156 days of survey). No more<br />

displacement between the water pond and the permanent canal was possible after this period,<br />

and the pond totally dried out at the end of September.<br />

7.2.2. Environmental variables<br />

The succession of mean water <strong>de</strong>pth in each site throughout the study period was calculated<br />

from the relation between the mean water level at the start of the study and daily records of<br />

the water level in the Gran<strong>de</strong> Brière Mottière marsh (source: Parc naturel régional <strong>de</strong> Brière).<br />

Water temperature was recor<strong>de</strong>d continuously (4 records per hour) in the three temporary<br />

floo<strong>de</strong>d habitats and the permanent canal sites using automatic sensors (Sensor StowAway<br />

TidBit, Onset Computer Corporation) disposed at a constant water <strong>de</strong>pth of 20 cm. Based on<br />

these data, the coefficient of variation of water temperature was calculated (mean daily water<br />

temperature ± S.D., expressed in %). We also recor<strong>de</strong>d changes in food availability in the four<br />

sites by monitoring zooplankton abundance (mainly Cladocera and Copepoda), which<br />

constitutes one of the major food sources for YOY fish. It was estimated by filtering (0.1-mm<br />

mesh) a 15L volume sampled in the water column. The procedure was repeated weekly in<br />

four locations in each site throughout the study period. Samples were preserved in 70%<br />

ethanol, and zooplankton abundance was estimated in the laboratory by counting the total<br />

number of individuals using a Dollfus counting chamber and a binocular microscope. A<br />

random sub-sample of the zooplankton samples (100 individuals) was dried at 60°C for 48h<br />

and weighted to assess the sample biomass, expressed in milligrams dry weight per liter of<br />

water (mg.L -1 ).<br />

110


7.2.3. Fish displacements<br />

Facteurs environnementaux, immigration et émigration<br />

Fish displacements were monitored using fyke nets (5-mm mesh, see technical <strong>de</strong>tails in<br />

Chapitre 2) that were set up at the unique connecting point between each temporary floo<strong>de</strong>d<br />

habitat and its adjacent habitat (Figure 7.1). We sampled fish immigrating and emigrating in<br />

and from the temporary floo<strong>de</strong>d habitats using two fyke nets per connecting point that were<br />

set up in the opposite directions. Fyke nets were fished continuously (overnight) during the<br />

entire study period, and fish catches were checked and processed daily (occasionally every<br />

two days during periods of less activity). Trapped fish were anaesthetised using eugenol (0.04<br />

mL·L -1 ), i<strong>de</strong>ntified, measured to the nearest mm and inspected to i<strong>de</strong>ntify mature individuals<br />

(gentle pressure on the abdomen). Fish were then released 10 m from the fyke nets according<br />

to their direction of movement.<br />

7.2.4. Data analyses<br />

7.2.4.1. Life stages partitioning and displacement patterns<br />

Based on observed size-class profile distribution, we <strong>de</strong>fined age-size relationships for each<br />

species. Then, based on the maturity of fish observed on the field, three life stages were<br />

<strong>de</strong>fined for almost all species: mature adults, immature juvenile fish and young-of-the-year<br />

(YOY) fish (individuals born during the study period). For these three life stages, temporary<br />

floo<strong>de</strong>d habitats function as spawning grounds (reproduction), nursery areas (growth) and<br />

incubation ground (hatching). In each temporary habitats, timing and pattern of fish<br />

displacements of the most abundant species (relative abundance >5%) were quantified by<br />

calculating five synthetic <strong>de</strong>scriptors based on daily catches in each direction (Chapitre 6).<br />

Immigration start (IS) was calculated as the number of trapping days before the first<br />

individual of a species was captured entering a site, referenced to the beginning of the study.<br />

Immigration moment (IM) was calculated as the number of days of trapping required to catch<br />

50% of the total number of fish captured over the survey. This latter value was limited by the<br />

first and third quartile of captures, which are the numbers of days of trapping required to<br />

catch 25% and 75% of the total number of fish captured over the survey, I25% and I75%<br />

respectively. Finally, immigration end (IE) was calculated as the number of trapping days<br />

before the last individual was captured. The same parameters were calculated for fish<br />

emigrating from each site (ES, E25%, EM, E75% and EE). In or<strong>de</strong>r to classify species life stages<br />

that exhibit similar displacement patterns, we first performed Correspon<strong>de</strong>nce Analyses (CA)<br />

on a series of contingency tables: one per habitat (the duration of trapping period being<br />

111


Chapitre 7<br />

variable between the three sites) and per fish displacement direction (fish abundance per week<br />

of survey for the dominant species life stages). Abundances of fish were log (x + 1)<br />

transformed prior to analysis. Then, we reported <strong>de</strong>ndrogram representations resulting from<br />

clustering procedures (Ward method applied to factorial coordinates) to show species life<br />

stages assemblages (ADE-4 software, Thioulouse et al. 1997).<br />

variables<br />

7.2.4.2. Relationships between fish displacements and environmental<br />

Non-parametric ANOVA were performed to test for differences in environmental variables<br />

between sites using the data collected during the same period, i.e. from March, 15 to June, 15.<br />

To relate fish displacements to the environmental variables, we firstly <strong>de</strong>fined a set of<br />

environmental variables based on mean weekly values: water temperature (T°), water level<br />

(WL), zooplankton <strong>de</strong>nsity (Zoo) and the changes in these variables (Δ T°, Δ WL and Δ Zoo,<br />

respectively) expressed as the difference between two consecutive weeks values. Correlated<br />

environmental <strong>de</strong>scriptors were removed prior to perform further analyses. Ordination<br />

analyses were then performed to investigate fish species/environmental parameters<br />

relationships.<br />

Preliminary Detren<strong>de</strong>d Correspon<strong>de</strong>nce Analyses (DCA) were performed for each<br />

data set in or<strong>de</strong>r to explore the response of species to environmental gradients and select the<br />

appropriate ordination mo<strong>de</strong>l (following Ter Braak & Šmilauer’s (2003) recommendations).<br />

When the maximum length of gradient (resulting from the DCA and expressed in standard<br />

<strong>de</strong>viation units of species turnover, SDs) is close to or upper 3 SDs, most species exhibits<br />

unimodal response to environmental gradients. In that case, a Canonical Correspon<strong>de</strong>nce<br />

Analysis (CCA) can be performed. For shorter gradients, species responses to environmental<br />

gradient are more linear, and the more suitable canonical analysis is the Redundancy Analysis<br />

(RDA) (see Ter Braak & Prentice 1988, Ter Braak 1986). In the present study, CCA were<br />

used in all cases except in two situations: immigration pattern of fish in the grassland and in<br />

the reed bed. Basically, in a CCA diagram, species are represented by points (corresponding<br />

to their approximate optima in the two-dimensional environmental subspace), and the<br />

environmental variables by arrows indicating their direction and rate of change through the<br />

subspace. Environmental variables with long arrows are more strongly correlated with the<br />

ordination axes than those with short arrows, and so more closely related to the pattern of<br />

species assemblage shown in the ordination diagram (Ter Braak 1986, 1996). The<br />

interpretation of a CCA is based on the biplot rule: each arrow <strong>de</strong>termines a direction (axis) in<br />

112


Facteurs environnementaux, immigration et émigration<br />

the diagram and species points can be projected orthogonally onto this axis. The or<strong>de</strong>r of the<br />

projection points corresponds approximately to the ranking of the species with respect to that<br />

environmental variable gradient (Ter Braak 1986). In a RDA diagram, both species and<br />

environmental variables are represented by arrows. Arrows pointing in roughly the same<br />

direction indicate a high positive correlation, arrows crossing at right angles indicate nearzero<br />

correlations and arrows pointing in opposite directions indicate high negative correlation.<br />

Moreover, species and environmental variables with long arrows are the most important in the<br />

analysis (Ter Braak & Prentice 1988). For each canonical analysis, the significance of the first<br />

axis and of all axes were tested using Monte Carlo permutation tests (n = 500 permutations).<br />

Ordination analyses and graphics were performed using CANOCO and CANODRAW<br />

(version 4.51, Ter Braak & Šmilauer (2003)).<br />

7.3. Results<br />

7.3.1. Environmental characteristics of the studied sites<br />

The permanent-temporary floo<strong>de</strong>d gradient we studied was characterised by two shallower<br />

sites (i.e. grassland and reed bed) that did not differ in mean water temperature, but exhibited<br />

higher coefficient of temperature variance (i.e. pond and permanent canal, Table 7.1).<br />

Zooplankton were significantly more abundant in the grassland and in the pond than in the<br />

permanent canal and the reed bed (Table 7.1). During the survey, a similar trend was observed<br />

in all sites throughout the water cycle: a <strong>de</strong>crease in water level accompanied by an increase<br />

in water temperature and a peak of zooplankton abundance occurring in the middle of the<br />

flooding period. Nevertheless, this peak occurred earlier in the permanent canal and later in<br />

the pond than in the two other sites (Figure 7.2).<br />

Table 7.1. Comparison of the environmental parameters (mean ± S.D.) measured in the permanent<br />

canal and the three temporary floo<strong>de</strong>d habitats from March, 15 to June, 15, 2005. For a given<br />

parameter, two different letters indicate significant difference between two sites (P < 0.05).<br />

Environmental Sites Non-parametric ANOVA<br />

parameters Canal Pond Reed bed Grassland KW P n<br />

Mean water<br />

temperature (°C)<br />

17.2 A<br />

(±3.4)<br />

16.8 A<br />

(±3.4)<br />

16.9 A<br />

(±3.3)<br />

16.5 A<br />

(±3.4)<br />

1.8 0.626 372<br />

Water temperature<br />

coefficient of<br />

variance (%)<br />

6.89 A<br />

(±3.24)<br />

8.3 A<br />

(±3.3)<br />

17.5 B<br />

(±6.2)<br />

17.2 B<br />

(±6.7)<br />

191.0


Zooplankton abundance (mg.L -1 ) Zooplankton abundance (mg.L -1)<br />

140<br />

120<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

2500<br />

2000<br />

1500<br />

1000<br />

500<br />

Permanent canal<br />

0<br />

15/3 30/3 14/4 29/4 14/5 29/5 13/6 28/6 13/7 28/7 12/8<br />

Figure 7.2. Water level (bol<strong>de</strong>d lines), water temperature (dotted lines) and the abundance of zooplankton (black bars) in the permanent canal and the three<br />

temporary floo<strong>de</strong>d habitats from March to August 2005. Note different y-axes between sites.<br />

120<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

80<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

90<br />

80<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

400<br />

350<br />

300<br />

250<br />

200<br />

150<br />

100<br />

50<br />

Reed bed<br />

Pond Grassland<br />

0<br />

15/3 30/3 14/4 29/4 14/5 29/5 13/6<br />

Date Date<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

Water level (cm) and water temperature (°C)<br />

Water level (cm) and water temperature (°C)


7.3.2. Fish assemblage<br />

Facteurs environnementaux, immigration et émigration<br />

In total, 8,670 fish belonging to 15 species were captured immigrating in or emigrating from<br />

the three temporary floo<strong>de</strong>d habitats (Table 7.2). The fish assemblage was dominated by eight<br />

species, each representing more than 5% of the total abundance of fish in at least one of the<br />

three temporary floo<strong>de</strong>d habitats. In or<strong>de</strong>r of <strong>de</strong>creasing abundance, we caught: rudd<br />

(Scardinius erythrophthalmus), common carp (Cyprinius carpio), mosquitofish (Gambusia<br />

holbrooki), European eel (Anguilla anguilla), black bullhead (Ameiurus melas), common<br />

bream (Abramis brama), roach (Rutilus rutilus) and pumpkinseed (Lepomis gibbosus).<br />

Additionally, 306 individuals belonging to 6 occasional species were caught (see Table 7.2<br />

for <strong>de</strong>tails). Species composition differed markedly among the three habitats. For instance, the<br />

fish assemblage was dominated by mosquitofish (65.0%) and rudd (17.6%) in the grassland,<br />

rudd (36.5%) and common carp (26.9%) in the reed bed, and common bream (49.1%) and<br />

rudd (16.6%) in the pond.<br />

Table 7.2. Details on the fish species caught immigrating to and emigrating from the three<br />

temporary floo<strong>de</strong>d habitats (grassland, reed bed and water pond) throughout the study period<br />

(March, 15 to August, 18, 2005). Bold values represent the most abundant species (relative<br />

abundance > 5%) in each habitat.<br />

Species Number (relative abundance in %)<br />

Name Co<strong>de</strong> Grassland Reed bed Pond<br />

Rudd (Scardinius erythrophthalmus) Sce 117 434 1133<br />

Common carp (Cyprinus carpio) Cyc 44 320 780<br />

Mosquitofish (Gambusia holbrooki) Gah 433 106 39<br />

European eel (Anguilla anguilla) Ana 35 163 16<br />

Black bullhead (Ameiurus melas) Amm 14 56 615<br />

Common bream (Abramis brama) Abb 1 1 3347<br />

Roach (Rutilus rutilus) Rur 0 0 583<br />

Pumpkinseed (Lepomis gibbosus) Leg 6 91 30<br />

Other species* - 16 19 271<br />

Total - 666 1190 6814<br />

* Others species (mean relative abundance < 5% in each of the three temporary floo<strong>de</strong>d habitats)<br />

were: Prussian carp (Carassius gibelio), pikeperch (San<strong>de</strong>r lucioperca), northern pike (Esox<br />

lucius), tench (Tinca tinca), Eurasian perch (Perca fluviatilis) and three spine stickleback<br />

(Gasterosteus aculeatus).<br />

115


Chapitre 7<br />

7.3.3. Life stages partitioning<br />

Because of the minimal number of individuals nee<strong>de</strong>d for analyses, the life stages of all<br />

species were not analysed in each site and for each displacement direction (Table 7.3). For<br />

three cyprinid species (Sce, Cyc and Rur) and Amm, the three life stages were observed<br />

throughout the study period and differentiated for further analyses (Table 7.3). For Abb, we<br />

subdivi<strong>de</strong>d individuals into only two groups: YOY and adults, since mature individuals were<br />

found in all classes of age >0 (Table 7.3). For Ana, a catadromous species that reproduces in<br />

the sea, the continental life-stage is only composed of juveniles. Nevertheless, we<br />

distinguished two different groups since drastic changes in diet (omnivory/piscivory) occurred<br />

during the ontogenetic <strong>de</strong>velopment of individuals (Table 7.3). For Leg, no YOY were caught<br />

during the survey because the species reproduce later in the season and consequently, two<br />

groups were partitioned. All Gah were gathered within the same group since this species has a<br />

short term life cycle (one year) with a maturity attained rapidly and a continuous<br />

ovoviviparous reproduction that occurred during the survey period. Consequently, different<br />

life stages were indistinguishable (Table 7.3).<br />

116


Table 7.3. Details on life stages of the most abundant fish species sampled in the three temporary floo<strong>de</strong>d habitats from March, 15 to August, 18, 2005): age<br />

classes, total number, body size (mean, S.D. and range, in mm), presence in the study sites, spawning (guild and period) and displacements analysed in the<br />

three temporary floo<strong>de</strong>d habitats. See <strong>de</strong>tails in Table 7.2 for co<strong>de</strong>s.<br />

Displacement analysable 4<br />

Species Age class<br />

Life<br />

stage 1 N<br />

Mean size<br />

(S.D.)<br />

Size range<br />

(min-max)<br />

Spawning 2<br />

guild (period)<br />

Life stage<br />

co<strong>de</strong> 3 Grassland Reed bed Pond<br />

Sce 0+ YOY 631 41.9 (5.9) 23-55 Phytophilous (Avril-June) Sce-H - E E<br />

1+ Juvenile 332 74.0 (9.5) 56-95 Sce-G I & E I & E I & E<br />

Cyc<br />

>1+<br />

0+<br />

Adult<br />

YOY<br />

722<br />

1138<br />

114.3 (12.4)<br />

75.2 (28.6)<br />

96-193<br />

21-147 Phytophilous (May-July)<br />

Sce-R<br />

Cyc-H<br />

I & E<br />

E E<br />

I & E I & E<br />

E<br />

1+ Juvenile 3 272.3 (272.3) 237-297 Cyc-G - - -<br />

>1+ Adult 3 506.3 (139.7) 402-665 Cyc-R - - -<br />

Gah All YOY-Adult 578 38.4 (9.5) 13-65 Ovoviviparous (June-Sept.) Gah I & E E -<br />

Ana < 350mm Juvenile 70 280.0 (57.5) 101-350 Pelagophilous (Sea) Ana-G1 E I & E -<br />

> 350 mm Juvenile 143 488.6 (102.5) 351-885 Ana-G2 I & E I & E -<br />

Amm 0+ YOY 93 37.0 (10.3) 26-50 Nest guar<strong>de</strong>r (June-July) Amm-H - - E<br />

1+ Juvenile 126 76.2 (11.0) 51-95 Amm-G - - I & E<br />

>1+ Adult 464 150.5 (26.5) 96-221 Amm-R - - I & E<br />

Abb 0+ YOY 2704 43.4 (7.4) 20-74 Phytolithophilous (May-June) Abb-H - - E<br />

>0+ Adult 645 130.8 (26.6) 75-265 Abb-R - - I & E<br />

Rur 0+ YOY 488 61.2 (7.0) 30-79 Phytolithophilous (April-June) Rur-H - - E<br />

1+ Juvenile 17 95.8 (9.9) 81-109 Rur-G - - E<br />

>1+ Adult 79 135.0 (21.8) 110-209 Rur-R - - E<br />

Leg 1+ Juvenile 85 46.2 (6.7) 26-64 Nest guar<strong>de</strong>r (May- August) Leg-G - I & E -<br />

>1+ Adult 42 90.8 (20.0) 66-154 Leg-R - I -<br />

1 YOY= Young of the Year; 2 from Balon (1975) and Keith & Allardi (2001), 3 H = Hatching, G = Growth and R = Reproduction, 4 I = immigration and E =<br />

emigration.<br />

117


7.3.4. Fish displacement patterns<br />

Overall, we found that some species (e.g. Ana) were characterized by exten<strong>de</strong>d displacement<br />

periods, whereas other species (e.g. Gah) exhibited limited displacement periods (Figure 7.3).<br />

As a general ten<strong>de</strong>ncy, emigrations lasted longer than immigrations (Figure 7.3). Some<br />

species (e.g. Leg and Rur) were early immigrants and emigrants, whereas others species (e.g.<br />

Amm and Gah) immigrated and emigrated later. We also found that timing and pattern of<br />

immigration and emigration varied among species according to their different life stages but<br />

irrespectively to the temporary floo<strong>de</strong>d habitats we studied. Nevertheless, YOY were always<br />

the last to emigrate, with very short emigration patterns. No clear trends were observed<br />

between the duration (exten<strong>de</strong>d vs restricted) of the displacements and the number of<br />

individuals caught (Figure 7.3). The clustering procedure allowed us to i<strong>de</strong>ntify species life<br />

stages assemblages (Figure 7.4). During immigration, displacement patterns ten<strong>de</strong>d to be<br />

combined according to species. In<strong>de</strong>ed, for a given species, immigration patterns of the two<br />

life stages (i.e. juveniles and adults) were very similar while differences in immigration<br />

patterns between species were more marked (Figure 7.3). At the opposite, emigration patterns<br />

ten<strong>de</strong>d to be regrouped according to life stages (i.e. YOY, juveniles and adults) (Figure 7.4).<br />

Whatever the fish species, juvenile individuals were in general early emigrants, followed by<br />

adult individuals. Afterwards, YOY individuals emigrated just before the total drying-out of<br />

each site (Figure 7.3). Interestingly, mosquitofish that were combined in a single life stage<br />

exhibited an emigration patterns close to that found for YOY individuals of the other fish<br />

species.


Immigration<br />

Emigration<br />

Number of individuals<br />

Number of individuals<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

450<br />

400<br />

50<br />

45<br />

40<br />

35<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

Grassland Reed bed Temporary pond<br />

IS IM IE<br />

Sce-R<br />

I 25%<br />

Sce-G<br />

Gah<br />

Ana-G2<br />

EM<br />

Ana-G1<br />

ES EE<br />

E 25%<br />

I 75%<br />

E 75%<br />

Cyc-H<br />

Sce-G<br />

Sce-R<br />

Ana-G2<br />

Gah<br />

1 2 3 4 5 6 7 8 9 10 11121314<br />

Time (week)<br />

80<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

350<br />

300<br />

200<br />

180<br />

160<br />

140<br />

120<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

Leg-G<br />

Gah<br />

1 2 3 4 5 6 7 8 9 10 11121314 15<br />

Time (week)<br />

Figure 7.3. Fish immigration and emigration patterns for the dominant life stages of each species in the three temporary floo<strong>de</strong>d habitats (grassland, reed bed<br />

and temporary pond). Each box plot is limited by the first (I25%) and third (I75%) quartile of captures and the vertical line represents the immigration moment<br />

(IM). The whiskers represent the immigration start (IS) and immigration end (IE). The equivalent <strong>de</strong>scriptors are calculated for the emigration (ES, E25%, EM,<br />

E75% and EE). Black boxes are for young-of-the-year individuals (H = hatching), grey boxes are for juveniles (G = growth) and white boxes are for adults (R =<br />

reproduction). See text for <strong>de</strong>tails on the calculations of these parameters and in Table 7.2 for species co<strong>de</strong>s.<br />

Ana-G2<br />

Sce-G<br />

Sce-R<br />

Ana-G1<br />

Leg-R<br />

Leg-G<br />

Cyc-H<br />

Sce-G<br />

Sce-R<br />

Sce-H<br />

Ana-G2<br />

Ana-G1<br />

180<br />

160<br />

140<br />

120<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

3000<br />

2500<br />

700<br />

600<br />

500<br />

400 400<br />

300 300<br />

200 200<br />

100 100<br />

0<br />

Amm-G<br />

Cyc-H<br />

Sce-G<br />

Sce-H<br />

Abb-R<br />

Amm-R<br />

Sce-R<br />

1 2 3 4 5 6 7 8 9 10111213141516171819202122<br />

Amm-G<br />

Abb-H<br />

Rur-H Sce-R<br />

Rur-H Sce-R<br />

Rur-H Sce-R<br />

Abb-R<br />

Amm-R<br />

Rur-R<br />

Sce-G<br />

Rur-G Amm-H<br />

Time (week)


Immigration<br />

Emigration<br />

Sce-R<br />

Sce-G<br />

Gah<br />

Ana-G2<br />

Gah<br />

Cyc-H<br />

Sce-R<br />

Ana-G2<br />

Sce-G<br />

Ana-G1<br />

Grassland Reed bed Temporary pond<br />

5<br />

-0.37 1.3<br />

0<br />

7<br />

-0.55 1.9<br />

0<br />

7<br />

-0.34 1.2<br />

Sce-R 0<br />

Sce-G<br />

Leg-R<br />

Leg-G<br />

Ana-G2<br />

Ana-G1<br />

Gah<br />

Sce-H<br />

Cyc-H<br />

Sce-R<br />

Ana-G2<br />

Sce-G<br />

Leg-G<br />

Ana-G1<br />

9<br />

-0.79 2.7<br />

0<br />

Figure 7.4. Dendrograms of weekly fish displacements (immigration/emigration) in the three temporary floo<strong>de</strong>d habitats (grassland, reed bed and temporary pond) exhibiting<br />

the different levels of clustering among the dominant life stages of the most abundant species. Data are log(x+1) transformed and the clustering procedure is applied on scores<br />

resulting from factorial analyses. Black boxes are for young-of-the-year individuals (H = hatchery), grey boxes are for juveniles (G = growth) and white boxes are for adults<br />

(R = reproduction). See <strong>de</strong>tails in Table 7.2 for species co<strong>de</strong>s<br />

Sce-G<br />

Sce-R<br />

Amm-R<br />

Amm-G<br />

Abb-R<br />

Amm-R<br />

Amm-G<br />

Rur-R<br />

Rur-G<br />

Abb-R<br />

Amm-H<br />

Sce-R<br />

Sce-G<br />

Cyc-H<br />

Sce-H<br />

Rur-H<br />

Abb-H<br />

6<br />

-0.21 0.84<br />

0<br />

14<br />

-0.42 1.8<br />

0


7.3.5. Effect of environmental variables on fish displacements<br />

Overall, from the series of ordination analyses (Table 7.4), we found that most fish<br />

immigrating to the temporary habitats respon<strong>de</strong>d to variations in water level (Figure 7.5). Fish<br />

immigrations correspon<strong>de</strong>d to the maximal water level, occurring early in the season when<br />

water temperature was still low but had begin to increase. In this period, zooplankton<br />

<strong>de</strong>nsities were still low and did not significantly influence fish emigration (Figure 7.5).<br />

Nevertheless, variable responses of fish to these environmental variables occurred among life<br />

stages. For instance, eel (G1 and G2) displacements were not related to the water level<br />

parameter (see RDA representations in the grassland and the reed bed). It was also the case<br />

for black bullhead (adults and juveniles) in the pond. The other species life stages ten<strong>de</strong>d to be<br />

more related to this variable. Regarding the emigration process, although some slight<br />

differences occurred between temporary habitats, YOY fish were strongly influenced by<br />

increases in water temperature, associated with a <strong>de</strong>crease in water level. Adults and juveniles<br />

that emigrated earlier in the study period ten<strong>de</strong>d to be more associated with higher water level<br />

and variations in zooplankton abundance (Figure 7.5).<br />

Table 7.4. Details on the ordination analyses investigating fish species<br />

displacements/environmental parameters relationship in the three temporary floo<strong>de</strong>d habitats<br />

(grassland, reed bed and pond) during the survey period (March, 15 to August, 18, 2005). RDA:<br />

Redundancy Analysis, and CCA: Canonical Correspon<strong>de</strong>nce Analysis.<br />

Displacement<br />

direction<br />

Immigration<br />

Emigration<br />

Temporary Ordination Total inertia Monte Carlo test<br />

floo<strong>de</strong>d analysis (two first axes)<br />

(P-values)<br />

habitat<br />

Species Species<br />

Environment abundance<br />

First axis All axis<br />

Grassland RDA 93.6% 60.2% 0.026 0.002<br />

Reed bed RDA 84.3% 60.2% 0.050 0.005<br />

Pond CCA 91% 24.1% 0.080 0.170<br />

Grassland CCA 96.5% 57.6% 0.002 0.002<br />

Reed bed CCA 93.7% 72.2% 0.002 0.001<br />

Pond CCA 88.6% 39.9% 0.002 0.002


Immigration<br />

Emigration<br />

1.0<br />

1.0<br />

0.8<br />

0.8<br />

0.6<br />

0.6<br />

0.4<br />

0.4<br />

0.2<br />

0.2<br />

0<br />

0<br />

-0.2<br />

-0.2<br />

-0.4<br />

-0.4<br />

-0.6<br />

-0.6<br />

-0.8<br />

-0.8<br />

T°<br />

Grassland Reed bed Temporary pond<br />

Zoo<br />

Ana-G2<br />

Δ T°<br />

WL<br />

Gah<br />

Δ WL<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

1 2 3 4<br />

Sce-R<br />

Sce-G<br />

-1.0<br />

-1.0<br />

-1.0 -1.0 -0.5 -0.5 0 0 0.5 0.5 1.0<br />

1.0<br />

1.0<br />

1.0<br />

0.8<br />

0.8<br />

0.6<br />

0.6<br />

0.4<br />

0.4<br />

WL<br />

0.2<br />

0.2<br />

0<br />

0<br />

-0.2<br />

-0.2<br />

-0.4<br />

-0.4<br />

-0.6<br />

-0.6<br />

-0.8<br />

-0.8<br />

Sce-R<br />

Ana-G2<br />

Sce-G<br />

Ana-G1<br />

Zoo<br />

Δ T°<br />

Gah<br />

Cyc_B Cyc_H<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

1 2 3 4<br />

T°<br />

-1.0<br />

-1.0<br />

-1.0 -1.0 -0.5 -0.5 0 0 0.5 0.5 1.0<br />

1.0<br />

Eigenvalue<br />

Eigenvalue<br />

Axis<br />

Axis<br />

1.0<br />

0.8<br />

0.6 0.6 0.6 0.6<br />

0.4<br />

0.2<br />

-0.4<br />

-0.6<br />

-0.8<br />

Ana-G2<br />

0<br />

WL<br />

-0.2<br />

Sce-R<br />

Leg-G<br />

Leg-R<br />

Sce-G<br />

Ana-G1<br />

Δ T°<br />

Figure 7.5. Results of the ordination analyses performed on fish immigration in and emigration from the three temporary floo<strong>de</strong>d habitats (grassland, reed bed and pond)<br />

throughout the flooding event. RDA: Redundancy Analysis, and CCA: Canonical Correspon<strong>de</strong>nce Analysis. Black boxes are for young-of-the-year individuals (H =<br />

hatchery), grey boxes are for juveniles (G = growth) and white boxes are for adults (R = reproduction). See <strong>de</strong>tails in Table 7.2 for species co<strong>de</strong>s and in Table 7.4 for statistics<br />

Δ Zoo<br />

T°<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

1 2 3 4<br />

-1.0<br />

-1.0 -1.0 -0.5 -0.5 0 0 0.5 0.5 1.0<br />

1.0<br />

1.0<br />

0.8<br />

0.6 0.6 0.6 0.6<br />

0.4<br />

0.2 WL<br />

0<br />

-0.2 -0.2 -0.2 -0.2<br />

-0.4<br />

-0.6<br />

-0.8<br />

Zoo<br />

Sce-R<br />

Ana-G2<br />

Ana-G1<br />

Leg-G<br />

Sce-G<br />

Δ Zoo<br />

Gah<br />

T°<br />

Cyc-B Cyc-H<br />

Sce-B Sce-H<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

1 2 3 4<br />

-1.0<br />

-1.0 -1.0 -0.5 -0.5 0 0 0.5 0.5 1.0<br />

1.0<br />

Eigenvalue<br />

Eigenvalue<br />

Axis<br />

Axis<br />

1.0<br />

1.0<br />

0.8<br />

0.8<br />

0.6<br />

0.6<br />

0.4<br />

0.4<br />

0.2<br />

0.2<br />

0<br />

0<br />

-0.2<br />

-0.2<br />

-0.4<br />

-0.4<br />

-0.6<br />

-0.6<br />

-0.8<br />

-0.8<br />

Abb-R<br />

WL<br />

Δ T°<br />

Sce-R<br />

Amm-G<br />

Sce-G<br />

Zoo<br />

T°<br />

Amm-R<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

1 2 3<br />

Axis<br />

4<br />

-1.0<br />

-1.0<br />

-1.0 -1.0 -0.5 -0.5 0 0 0.5 0.5 1.0<br />

1.0<br />

1.0<br />

1.0<br />

0.8<br />

0.8<br />

0.6<br />

0.6<br />

0.4<br />

0.4<br />

0.2<br />

0.2<br />

0<br />

0<br />

-0.2<br />

-0.2<br />

-0.4<br />

-0.4<br />

-0.6<br />

-0.6<br />

-0.8<br />

-0.8<br />

T° T°<br />

Sce-G<br />

Sce-H Sce-B<br />

Cyc-B Cyc-H<br />

Rur-B Rur-H<br />

Amm-R<br />

Amm-G<br />

Abb-H Abb-B Abb-R<br />

Amm-B Amm-H<br />

Sce-R<br />

Δ T°<br />

Zoo<br />

WL<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

Rur-R<br />

Rur-G<br />

0<br />

1 2 3<br />

Axis<br />

4<br />

-1.0<br />

-1.0<br />

-1.0 -1.0 -0.5 -0.5 0 0 0.5 0.5 1.0<br />

1.0<br />

Eigenvalue<br />

Eigenvalue


Discussion<br />

In the present study, we found that 1) fish assemblage composition differed between the three<br />

temporary floo<strong>de</strong>d habitats, 2) timing and patterns of fish displacements<br />

(immigration/emigration) were highly variable both among species and among life stages, 3)<br />

more precisely, fish immigration timing and patterns were more closely related to species<br />

whereas emigration processes were mainly related to life stages, and 4) fish displacements<br />

were largely influenced by the succession of environmental conditions during the flooding<br />

period (notably water level and water temperature), these responses being, however, variable<br />

among species, life stages, and senses of displacement (immigration and emigration).<br />

In the fish assemblages we studied, we found that the different cyprinids species,<br />

which reproduced during the survey period, exhibited variable <strong>de</strong>grees of exploitation of the<br />

temporary floo<strong>de</strong>d habitats. According to the reproductive guilds <strong>de</strong>fined by Balon (1975), we<br />

observed that phytophilous species (e.g. rudd and common carp) were in high <strong>de</strong>nsity in the<br />

shallow and vegetated sites (i.e. grassland and reed bed) while phytolithophilous species (e.g.<br />

bream and roach) dominated the fish assemblage of the <strong>de</strong>eper and less vegetated site (i.e.<br />

pond). Nevertheless, the reproductive success of these species was variable between sites<br />

(Figure 7.3). For instance, no YOY rudd was caught emigrating from the grassland, but<br />

several YOY common carp were captured. The result might indicate that the hydroperiod (i.e.<br />

flooding duration) and the thermal regime might have preclu<strong>de</strong>d to the successful<br />

reproduction of Sce in this site, <strong>de</strong>spite its colonisation by mature individuals (King et al.<br />

2003). Regarding the two nest guarding species (i.e. pumpkinseed and black bullhead, Balon<br />

(1975)), which reproduce later in the season (Keith & Allardi 2001, Santoul et al. 2005), no<br />

YOY were captured emigrating from the grassland and the reed bed and only few individuals<br />

were caught emigrating from the water pond. These results <strong>de</strong>monstrate that, in the Brière<br />

marsh, those two species reproduce mainly in permanent and semi-permanent habitats. The<br />

composition of the fish assemblage in the temporary floo<strong>de</strong>d habitat that dried out every year<br />

is highly <strong>de</strong>pen<strong>de</strong>nt on their environmental characteristics (vegetation cover, temperature and<br />

flooding duration), which attract some life stages of some species. Nevertheless, in the Brière<br />

marsh, the inundation period (a few months) of some temporarily floo<strong>de</strong>d habitats is relatively<br />

restricted compared to other aquatic ecosystems (Baber et al. 2002, Trexler et al. 2005,<br />

Rehage & Trexler 2006). Consequently, a limited time window is offered to fish for<br />

exploiting temporary floo<strong>de</strong>d habitats and this exploitation should be optimized by the<br />

different species to accomplish their life cycle.


Regarding fish displacement timing and patterns (immigration and emigration), our<br />

results <strong>de</strong>monstrated that these dynamical processes were variable among species and among<br />

life stages of a given species. In<strong>de</strong>ed, inter specific variability was the main source of variance<br />

for fish immigration, whereas the main source of variance for fish emigration was related to<br />

the <strong>de</strong>velopmental status. Immigration might be un<strong>de</strong>r the influence of specific<br />

characteristics, such as a certain threshold of water temperature nee<strong>de</strong>d to mature and to start<br />

the spawning migration (Hladick & Kubecka 2003). A significant increase of water<br />

temperature was observed in April. Factors acting at the species level and influencing fish<br />

species immigration need to be thoroughly investigated. After a <strong>de</strong>crease in water level,<br />

habitat conditions <strong>de</strong>teriorated (increase of water temperature, <strong>de</strong>crease of zooplankton<br />

<strong>de</strong>nsity…) and we observed that for almost species, adults ten<strong>de</strong>d to emigrate earlier than<br />

juveniles. YOY individuals only emigrated when water level was very low and water<br />

temperature very high. These intra-specific variations in emigration timing might be related to<br />

differences in physiological tolerance between juveniles and adults (e.g. Coutant 1977). For<br />

example, there is a growing recognition in the literature that smaller fish are less sensitive to<br />

hypoxia (e.g. Robb & Abrahams 2003) because of a negative allometric relationship for massspecific<br />

gill surface area, i.e. smaller individuals may have more efficient gas exchange with<br />

their environment than larger fish (Hugues 1984). Furthermore, small individuals are able to<br />

have an efficient utilization of oxygen at the water-surface interface (see in Sargent & Galat<br />

2002). It is clear that other factors might influence emigration timing at the individual and<br />

specific level but, for example, we did not find any clear evi<strong>de</strong>nce of the significance of<br />

variations in food abundance in the observed variance in emigration of fish. However, it<br />

should be noted the existence of residual variance at the life stage level, resulting from inter<br />

individual variability and this source of variations need to be thoroughly investigated at the<br />

individual level. Finally, we cite the need to better investigate the daily processes of<br />

temporary floo<strong>de</strong>d habitats use to un<strong>de</strong>rstand how environmental variables might influence<br />

these dynamical processes at a finer temporal scale.


Acknowledgements<br />

We are grateful to J.P. Damien, M. Buoro, V. Thoby, E. Le Mitouard, and all the numerous<br />

persons that participated to field work. We thank the Parc naturel régional <strong>de</strong> Brière for<br />

logistic support, the Pnrb, FEDER, DIREN (MEDD), Région Pays <strong>de</strong> la Loire and Agence <strong>de</strong><br />

l’Eau from Loire-Bretagne for financial support.<br />

Juvénile <strong>de</strong> l’année <strong>de</strong> carpe émigrant du plan d’eau


Les résultats <strong>de</strong> cette étu<strong>de</strong> montrent que les processus d’échanges <strong>de</strong> poissons<br />

entre milieux permanents et MTI ainsi que leurs sources <strong>de</strong> variabilité peuvent<br />

être perçus à différents niveaux d’organisation <strong>de</strong> la communauté piscicole, i.e.<br />

espèces et sta<strong>de</strong>s <strong>de</strong> vie. Ces processus sont variables entre MTI, cette variabilité<br />

étant dépendante <strong>de</strong>s conditions environnementales et <strong>de</strong> l’évolution au cours du<br />

cycle hydrologique dans chaque type <strong>de</strong> MTI . Afin <strong>de</strong> comprendre les sources <strong>de</strong><br />

variabilités intra-spécifiques, l’analyse au niveau <strong>de</strong>s sta<strong>de</strong>s <strong>de</strong> vie, définis ici en<br />

fonction du rôle <strong>de</strong>s MTI pour les individus (naissance, croissance et<br />

reproduction), s’est avérée pertinente. En effet, cette approche a permis <strong>de</strong><br />

regrouper <strong>de</strong>s individus appartenant à <strong>de</strong>s espèces différentes mais qui présentent<br />

<strong>de</strong>s réponses similaires aux paramètres environnementaux lors <strong>de</strong>s échanges entre<br />

milieux permanents et MTI. A l’échelle <strong>de</strong> ces sta<strong>de</strong>s <strong>de</strong> vie, il existe là encore<br />

une variabilité interindividuelle dont il apparaît nécessaire d’en étudier les sources<br />

et les conséquences sur les populations. Cette problématique sera abordée dans la<br />

partie II <strong>de</strong> ce travail <strong>de</strong> thèse qui va s’intéresser au fonctionnement <strong>de</strong> certaines<br />

espèces i<strong>de</strong>ntifiées dans la partie précé<strong>de</strong>nte comme d’intérêt majeur au niveau <strong>de</strong><br />

la communauté du site d’étu<strong>de</strong> et définies comme telles en fonction <strong>de</strong> leur rôle<br />

dans cette communauté (e.g. niveau d’abondance) mais aussi en fonction du rôle<br />

du site d’étu<strong>de</strong> et <strong>de</strong> ses MTI pour ces espèces (e.g. vis-à-vis <strong>de</strong> leur<br />

reproduction), du statut <strong>de</strong> ces espèces et <strong>de</strong>s mesures <strong>de</strong> gestion auxquelles elles<br />

sont soumises. Cette approche sera réalisée à différents niveaux d’organisation, à<br />

savoir populationnel et individuel, afin <strong>de</strong> venir compléter et affiner la perception<br />

du fonctionnement <strong>de</strong> la communauté établie dans la partie I <strong>de</strong> cette thèse.


Partie II<br />

Réponses fonctionnelles aux contraintes anthropiques<br />

<strong>de</strong> trois espèces d’intérêt majeur <strong>de</strong> la communauté<br />

piscicole<br />

Si la partie I <strong>de</strong> cette thèse s’est intéressée tout particulièrement à l’utilisation <strong>de</strong>s<br />

MTI par la communauté piscicole, elle a aussi permis d’i<strong>de</strong>ntifier certaines<br />

espèces <strong>de</strong> cette communauté comme « espèces d’intérêt majeur » par leurs <strong>lien</strong>s<br />

étroits avec les MTI et par leur participation globale au fonctionnement <strong>de</strong> cet<br />

écosystème. Sur la base <strong>de</strong> multiples critères, nous pouvons considérer que<br />

chacune <strong>de</strong>s espèces est d’intérêt pour le fonctionnement <strong>de</strong> la communauté<br />

piscicole. Toutefois, nous avons pris le parti d’en i<strong>de</strong>ntifier trois que nous<br />

considérons d’intérêts majeurs aux regards <strong>de</strong> leurs places privilégiées, i<strong>de</strong>ntifiées<br />

dans la première partie <strong>de</strong> cette thèse, dans les grands groupes d’espèces aux<br />

spécificités écologiques bien marquées. La première est le poisson chat, espèce<br />

invasive qui domine <strong>de</strong> nos jours la communauté piscicole. Il appartient au groupe<br />

<strong>de</strong>s espèces qui semblent exploiter <strong>de</strong> façon très complémentaire les <strong>de</strong>ux gran<strong>de</strong>s<br />

entités du site : les milieux permanents, voire semi permanents, pour la<br />

reproduction et les MTI à <strong>de</strong>s fins <strong>de</strong> croissance. La secon<strong>de</strong> est l’anguille, espèce<br />

classée vulnérable et exploitée localement. Cette espèce a la particularité<br />

d’appartenir au groupe <strong>de</strong>s espèces amphihalines, d’en être la <strong>de</strong>rnière<br />

représentante et <strong>de</strong> ne pas se reproduire dans le site d’étu<strong>de</strong>. Par ailleurs, la<br />

position <strong>de</strong>s marais <strong>de</strong> Brière au niveau <strong>de</strong> l’estuaire <strong>de</strong> la Loire fait <strong>de</strong> ces marais<br />

continentaux l’une <strong>de</strong>s zones <strong>de</strong> croissance privilégiée par l’espèce. La troisième<br />

est le brochet, espèce classé vulnérable, exploitée localement et emblématique <strong>de</strong>s<br />

marais <strong>de</strong> Brière. Cette espèce appartient au groupe <strong>de</strong>s espèces qui, contrairement<br />

aux <strong>de</strong>ux espèces précé<strong>de</strong>ntes, ont un cycle <strong>de</strong> vie fortement inféodé aux MTI car<br />

elles s’y reproduisent.


Dans cette partie, les réponses fonctionnelles à l’utilisation <strong>de</strong> l’habitat et aux<br />

programmes <strong>de</strong> gestions <strong>de</strong> ces trois espèces seront examinées à <strong>de</strong>s niveaux<br />

d’organisation populationnels et individuels afin <strong>de</strong> fournir <strong>de</strong>s éléments concrets<br />

pour optimiser la gestion <strong>de</strong> ces trois espèces. En effet, les marais <strong>de</strong> Brière sont<br />

<strong>de</strong>s milieux subissant <strong>de</strong>puis toujours <strong>de</strong> fortes contraintes anthropiques et voient<br />

leur diversité biologique menacée. Afin d’y remédier, <strong>de</strong>s options conservatoires<br />

déjà engagées dans le site d’étu<strong>de</strong>, et notamment pour le modèle poisson, sont<br />

fortement orientées par la place que représentent les MTI dans le fonctionnement<br />

<strong>de</strong> l’écosystème, et vers une politique <strong>de</strong> réouverture du milieu (restauration <strong>de</strong>s<br />

prairies naturelles). Ce positionnement <strong>de</strong>s gestionnaires locaux a également<br />

contribué à développer un travail spécifique sur ces trois modèles biologiques. Il<br />

est donc indispensable <strong>de</strong> faire une évaluation <strong>de</strong>s différentes mesures <strong>de</strong> gestion à<br />

la suite <strong>de</strong> l’analyse du fonctionnement <strong>de</strong> la communauté piscicole. De plus,<br />

l’étu<strong>de</strong> du fonctionnement <strong>de</strong> ces espèces d’intérêt majeur peut constituer une ai<strong>de</strong><br />

substantielle à la compréhension du fonctionnement global d’une communauté et<br />

à la définition <strong>de</strong> mesures complémentaires <strong>de</strong> la gestion <strong>de</strong>s habitats.<br />

Une « bosse » <strong>de</strong> chasse le long d’une roselière dans les marais <strong>de</strong> Gran<strong>de</strong> Brière<br />

Mottière<br />

128


Chapitre 8<br />

Utilisation et sélection <strong>de</strong> l’habitat par le poisson chat<br />

(Ameiurus melas)<br />

Le poisson-chat, espèce originaire d’Amérique du Nord, a été introduite en Brière<br />

en 1929 et domine actuellement <strong>de</strong> manière outrancière la communauté piscicole<br />

<strong>de</strong>s marais <strong>de</strong> Brière. Quelques hypothèses pouvant expliquer ce niveau<br />

d’abondance ont d’ailleurs été formulées dans la partie précé<strong>de</strong>nte. Afin <strong>de</strong><br />

compléter et tester ces hypothèses, nous abordons ici une analyse fine <strong>de</strong><br />

l’utilisation et <strong>de</strong> la sélection <strong>de</strong> l’habitat par le poisson-chat dans les marais <strong>de</strong><br />

Brière. Cette analyse est menée à <strong>de</strong>ux échelles spatiales différentes (micro- et<br />

macro-habitats) afin d’obtenir une image plus complète <strong>de</strong> la réponse <strong>de</strong> l’espèce<br />

aux différentes variables d’habitat. L’intérêt <strong>de</strong> cette étu<strong>de</strong> est d’autant plus<br />

important que cette espèce a été largement délaissée par la communauté<br />

scientifique à l’échelle européenne et ce, malgré l’expansion rapi<strong>de</strong> <strong>de</strong> son aire <strong>de</strong><br />

distribution et son classement en tant qu’espèce invasive dans <strong>de</strong> nombreux pays<br />

européens. Elle a aussi pour but d’appréhen<strong>de</strong>r certaines bases écologiques <strong>de</strong><br />

cette espèce en <strong>de</strong>hors <strong>de</strong> son aire originelle <strong>de</strong> répartition.


Chapitre 8<br />

Un adulte <strong>de</strong> poisson chat (Ameiurus melas) échantillonné dans les marais <strong>de</strong> Brière<br />

130


Sélection d’habitat du poisson chat<br />

Habitat use of an artificial wetland by the invasive catfish<br />

Ameiurus melas 6<br />

Résumé: Cette étu<strong>de</strong> analyse la distribution et la sélection d’habitat d’une espèce<br />

invasive (le poisson-chat, Ameiurus melas), dans les canaux et les milieux temporairement<br />

inondés adjacents, <strong>de</strong>s marais <strong>de</strong> Brière. Une approche multi-échelle (micro- et macro-habitat)<br />

a été utilisée afin d’analyser les niveaux d’abondance <strong>de</strong> l’espèce en fonction <strong>de</strong>s<br />

caractéristiques physiques du réseau hydrographique. Les juvéniles <strong>de</strong> l’année et les adultes<br />

<strong>de</strong> poisson-chat dominent largement la communauté piscicole dans les marais <strong>de</strong> Brière, mais<br />

leurs abondances sont très variables d’un site à l’autre. Bien que nous ayons trouvé quelques<br />

différences <strong>de</strong> sélectivité <strong>de</strong> l’espèce pour certaines variables <strong>de</strong> micro-habitats en fonction<br />

<strong>de</strong>s sta<strong>de</strong>s <strong>de</strong> développement (juvéniles <strong>de</strong> l’année et adultes), l’abondance <strong>de</strong> poisson-chat est<br />

corrélée positivement et <strong>de</strong> façon systématique à la dominance du roseau (Phragmites<br />

australis) dans le milieu. De plus, les poissons utilisent préférentiellement les roselières (par<br />

opposition aux prairies) lors <strong>de</strong> la pério<strong>de</strong> d’inondation. Ces résultats suggèrent que l’invasion<br />

<strong>de</strong>s marais <strong>de</strong> Brière par le poisson-chat a certainement été facilitée par l’expansion <strong>de</strong>s<br />

roselières, conséquence directe <strong>de</strong> la diminution <strong>de</strong>s activités agricoles lors <strong>de</strong>s <strong>de</strong>rnières<br />

décennies. Cette étu<strong>de</strong> représente un exemple inhabituel où, dans un contexte bien particulier,<br />

la diminution <strong>de</strong>s activités humaines a eu un effet inattendu, à savoir la facilitation <strong>de</strong><br />

l’invasion d’une espèce piscicole.<br />

6 Ce chapitre est adapté <strong>de</strong> l’article : Cucherousset, J., Paillisson, J.-M., Carpentier, A., Eybert, M.-C. & Ol<strong>de</strong>n,<br />

J.D. 2006. Use of an artificial wetland by the invasive catfish Ameirus melas. 15: 589-596.<br />

131


Chapitre 8<br />

Abstract: This study examined the distribution and habitat selection of the invasive<br />

black bullhead (Ameiurus melas (Rafinesque)) in the canals and surroun<strong>de</strong>d temporary<br />

floo<strong>de</strong>d habitats of an artificial wetland in western France. A multi-scale approach was used<br />

to quantify patterns of A. melas abundance in relation to physical habitat characteristics in the<br />

canal network. Young-of-the-year (YOY) and adult A. melas largely dominated the local fish<br />

assemblage, but were highly variable among sites. Although we found evi<strong>de</strong>nce for some<br />

fine-scale habitat differences for YOY and adult individuals, the abundance of A. melas was<br />

positively and consistently related to the dominance of reed beds. Furthermore, A. melas<br />

preferentially used reed beds as opposed to marsh meadows during the flooding period. The<br />

results from this study suggest that the invasion of A. melas has been facilitated by the<br />

expansion of reed beds associated with the diminution of agricultural pressure in recent<br />

<strong>de</strong>ca<strong>de</strong>s. This study represents an unusual example where human activities can have had an<br />

unexpected effect by facilitating an invasive fish species.<br />

132


8.1. Introduction<br />

Sélection d’habitat du poisson chat<br />

Invasive species constitute a major threat to biodiversity and ecosystem integrity and cause<br />

substantial economic damage (Vitousek et al. 1997, Kolar & Lodge 2001, Lee 2002, Pimentel<br />

2005). Freshwaters are consi<strong>de</strong>red one of the most severely <strong>de</strong>gra<strong>de</strong>d ecosystems in<br />

industrialized countries (Oberdorff et al. 2002) and have been the recipient of many invasions,<br />

in particular by fish species (see review in Copp et al. 2005 and García-Berthou et al. 2005).<br />

The success of non-indigenous fishes has been wi<strong>de</strong>ly studied and attributed to several factors,<br />

including specific life history traits (Ol<strong>de</strong>n et al. 2006), phylogeny (Alcaraz et al. 2005) and<br />

environmental characteristics including abiotic factors or human use (Moyle & Light 1996,<br />

Alcaraz et al. 2005).<br />

Environmental disturbance is known to facilitate fish invasions greatly (Moyle &<br />

Light 1996). Habitat modification and fragmentation (e.g. water pollution, construction of<br />

dams and water diversions) threaten native fish faunas whilst favouring the invasion of more<br />

tolerant non-native species (Marchetti et al. 2004). In the case of some artificial ecosystems,<br />

the <strong>de</strong>cline of traditional human activities can impact local species diversity by favoring the<br />

loss and fragmentation of habitat (Wolter 2001). Studies of non-native fish species in artificial<br />

environments are particularly rare, although such species often dominate these areas that are<br />

now common throughout the landscape. It is the case for the black bullhead (Ameiurus melas<br />

(Rafinesque)), which was introduced into France from North America in 1871, and is now<br />

wi<strong>de</strong>spread in Europe (Wheeler 1978), notably with recent records in Spain and Portugal<br />

(Elvira 1984, Doadrio 2001, Gante & Santos 2002). Despite its classification as a “species<br />

liable to cause biological disequilibrium” by the French legislation (article R. 232-3 Co<strong>de</strong><br />

rural, see Guevel 1997), few ecological studies have been conducted in its European nonnative<br />

range (but see Boët 1980, Copp 1993, Poizat & Crivelli 1997).<br />

In the present paper, we study the distribution and habitat selection of young-of-theyear<br />

(YOY) and adult A. melas in a man-ma<strong>de</strong> wetland in western France. We use an<br />

integrative approach (see Copp 1989, Poizat & Pont 1996) to examine the response of A.<br />

melas to habitat heterogeneity at two spatial scales: canal network level and local microhabitat.<br />

133


Chapitre 8<br />

8.2. Material and methods<br />

8.2.1. Study area<br />

The Brière marsh (see <strong>de</strong>tails in Chapitre 2) has rapidly changed during the last century as a<br />

result of a 3-year flood period during the Second World War and a continuous <strong>de</strong>crease of<br />

many agricultural activities. For instance, grazing <strong>de</strong>creased greatly (200 bovines and horses<br />

in the early 1980s versus 3000 animals in the early 1900s) and peat extraction ceased<br />

completely in the 1950s whereas 100000 hectoliters of peat were extracted in the 1880s. The<br />

combination of changes in human activities has favoured the <strong>de</strong>velopment and expansion of<br />

reed beds mainly composed of the common reed (Phragmites australis (Cav.) Trin. ex Steud.)<br />

(Bernard & Rolland 1990). During the last century, the surface area covered by reed beds<br />

increased from 2% to 85% of the total area, and marsh meadows, mainly composed of other<br />

Poaceae, <strong>de</strong>creased from 83% to 10%, the rest being covered by open waters (Bernard &<br />

Rolland 1990). These changes in human activities have played a significant role in the loss of<br />

native plant and animal species (Bernard & Rolland 1990, Eybert et al. 1998). At the same<br />

time, the Brière marsh has experienced several invasion events, and notably that of A. melas.<br />

Historical data indicate that A. melas was introduced in 1929 (Maillard 1972). It is now<br />

wi<strong>de</strong>ly distributed throughout the entire wetland.<br />

8.2.2. Fish sampling<br />

Fish sampling was performed by electrofishing (see <strong>de</strong>tails in Chapitre 2) in two<br />

periods of 2004 that were dictated by the pattern of floods, the seasonal configuration of<br />

habitats and safety precautions associated with electrofishing. All fish species inhabit the<br />

canal web during drought (hereafter called permanent habitats), and some of them colonize<br />

reed beds and meadows (hereafter called temporary habitats) during the flooding season.<br />

These latter habitats (10 reed beds and 12 meadows) were sampled in April (Figure 8.1)<br />

during the recession of water levels (average water <strong>de</strong>pth of 51.0 cm ± 19.1 SD) and before<br />

spawning of most species. Permanent habitats (25 canals, Figure 8.1) were sampled in August<br />

when the water level was low (average water <strong>de</strong>pth of 52.4 cm ± 29.0 SD). 30.0 PAS ± 3.8<br />

per temporary habitat and 26 PAS ± 4.1 per canal. We used size-class distribution profiles to<br />

distinguish YOY from adults for each fish species (Carpentier et al. 2004). A threshold of 70<br />

mm was used to distinguish the two age classes in A. melas. This fish-size threshold agrees<br />

with the findings of Hanchin et al. (2002). Based on specific length-weight relationships<br />

<strong>de</strong>fined in situ (unpublished data), the relative biomass of each fish species was calculated.<br />

134


Vegetation type:<br />

Reed bed<br />

Marsh meadow<br />

Sélection d’habitat du poisson chat<br />

Canal network<br />

Outflow<br />

Sampled sites:<br />

Canal section<br />

Reed bed<br />

Marsh meadow<br />

Figure 8.1. Map of the Brière marsh illustrating the canal network and the matrix of temporary<br />

habitats (marsh meadow and reed bed). Location of permanent habitats sampled in August 2004<br />

() and temporary habitats sampled in April 2004, where reed beds are represented as ● and<br />

marsh meadows as ○.<br />

135<br />

2 Km<br />

N


Chapitre 8<br />

8.2.3. Environmental characteristics of permanent habitats<br />

Canal size (DS) was classified into two categories: those belonging to the primary network<br />

(main river or large canal directly connected to the main river, mean width = 19.7 ± 6.0 m)<br />

and those belonging to the secondary network (canal connected to the primary network level,<br />

mean width = 8.9 ± 2.6 m). A set of <strong>de</strong>scriptors of the canal network in the vicinity of each<br />

sampled canal (‘canal network variables’) was extracted from a Geographical Information<br />

System (ArcView 3.2, Environmental Science Research Institute Inc., Redlands, CA) using<br />

land cover and hydrographical network information. These variables represented the <strong>de</strong>gree of<br />

connectivity of the canal network. The number of canal (NC), number of connections<br />

(crossroads between canals), length of canal belonging to the primary (LCI) and the<br />

secondary (LDII) network levels and total length of canals (TLC = LCI + LCII) were<br />

calculated for each canal at different radius values (buffer area, calculated at the central part<br />

of each site): 200, 400, 600, 800 and 1000 m. This range of radius values corresponds to the<br />

mean linear home range values (0.5 to 2.1 km, Sakaris et al. 2005) for the closely-related<br />

species brown bullhead Ameiurus nebulosus (Leseur).<br />

Four fine-scale (microhabitat) variables were measured at each PAS during fish<br />

sampling. Water <strong>de</strong>pth (WD) was transformed into two classes (0-50 cm and > 51 cm, after<br />

inspection of the distribution frequency of values), substratum (SU) was classified into two<br />

types (peat, not peat), vegetation cover (VC) was visually estimated as the percentage of area<br />

covered by plants and transformed into two classes (absence or presence) for analyses and<br />

adjacent habitat type (AH) was categorized as reed beds or marsh meadows.<br />

8.2.4. Data analysis<br />

The PAS from each canal were analyzed collectively at the canal network scale and<br />

punctually at the microhabitat scale (Copp 1989). At the canal network level, we conducted a<br />

series of preliminary analyses to quantify the <strong>de</strong>gree of multi-collinearity among continuous<br />

variables (length of canals were ln(x+1) transformed). Next, a Principal Components Analysis<br />

(PCA), was performed on the resulting in<strong>de</strong>pen<strong>de</strong>nt variables (NC_200, TLC_200, NC_600,<br />

TLC_600, 77.5% of the variation in the canal network <strong>de</strong>scriptors explained by the first two<br />

axes) to classify canals into two groups highlighted by the significant contribution of NC_600<br />

(inertia analysis procedure (Legendre & Legendre 1998), see values in Table 8.1). The<br />

resulting variable was called ‘Canal connectivity, CC’ (Table 8.1). At the same time, we<br />

analyzed the relationships between three fish <strong>de</strong>scriptors: i.e., abundance (Catch per Unit<br />

136


Sélection d’habitat du poisson chat<br />

Effort: CPUE = number of individuals/PAS), relative abundance (%) and relative biomass<br />

(%), for YOY and adults of A. melas using linear regression analysis.<br />

Table 8.1. Summary statistics for the final set of habitat variables used to mo<strong>de</strong>l Ameiurus melas<br />

abundance. Reported values are proportions (%) for the categories of the canal network and<br />

microhabitat variables.<br />

Habitat variables Co<strong>de</strong> Categories (relative proportion)<br />

Canal network scale (n = 25 canals)<br />

Canal Size CS<br />

Canal Connectivity 1 CC<br />

Microhabitat scale (n = 646 PAS)<br />

Water Depth WD<br />

Substratum SU<br />

Vegetation Cover VC<br />

Adjacent Habitat type AH<br />

Primary<br />

(60.4%)<br />

Low<br />

(68%)<br />

0-50 cm<br />

(55.7%)<br />

Peat<br />

(67.6%)<br />

Absence<br />

(63.6%)<br />

Reed bed<br />

(46.1%)<br />

Secondary<br />

(39.6 %)<br />

High<br />

(32.0%)<br />

> 51 cm<br />

(47.3%)<br />

Not Peat<br />

(32.4%)<br />

Presence<br />

(36.4%)<br />

Marsh meadow<br />

(53.9%)<br />

1 Canal Connectivity results from a PCA on in<strong>de</strong>pen<strong>de</strong>nt continuous canal network variables (see<br />

the data analysis section for more <strong>de</strong>tails). The two modalities are mainly <strong>de</strong>rived from NC_600<br />

values (inertia analysis procedure); low: ≤ 4 canals in a radius of 600 m around the sampled canals,<br />

high: > 4 canals.<br />

The relationships between the abundance of each size-class of A. melas (<strong>de</strong>pen<strong>de</strong>nt<br />

variable) and the habitat <strong>de</strong>scriptors at the two spatial scales (in<strong>de</strong>pen<strong>de</strong>nt variables) were<br />

analyzed separately using general linear mo<strong>de</strong>ls (GLM). In these mo<strong>de</strong>ls, a forward selection<br />

procedure was applied to the variables with a probability of entry of < 0.15 and to exclu<strong>de</strong> of<br />

> 0.20. We also tested for differences in the abundance of A. melas (CPUE of YOY and adults<br />

combined) and for other species in the assemblage (species CPUE and total richness) in the<br />

two temporary habitat types (marsh meadows vs reed beds). Assumptions of normality and<br />

homoscedasticity were tested using the Kolmogorov-Smirnov/Lilliefors test and the Levene’s<br />

test, respectively before performing a t-test. When <strong>de</strong>viations from normality were <strong>de</strong>tected,<br />

non-parametric Mann-Whitney tests were performed. The statistical analyses were performed<br />

using STATISTICA (StatSoft Inc., Tulsa, OK) and a critical alpha level of 0.05 was used for<br />

all tests.<br />

137


Chapitre 8<br />

8.3. Results<br />

8.3.1. Distributional patterns of A. melas in permanent habitats<br />

In total, 11921 individuals (9650 adults and 2271 YOY) belonging to 17 species were<br />

captured in permanent habitats (646 PAS), including 5334 individuals of A. melas (4524<br />

adults and 810 YOY) (Table 8.2). A. melas largely dominated the fish assemblage in terms of<br />

relative <strong>de</strong>nsity, relative abundance and relative biomass for both adult and YOY populations<br />

followed by bream spp. (silver bream Blicca bjoerkna and common bream Abramis brama),<br />

roach (Rutilus rutilus), rudd (Scardinius erythrophthalmus), mosquitofish (Gambusia<br />

holbrooki), pumpkinseed (Lepomis gibbosus) and A. anguilla (Table 8.2).<br />

On average, YOY and adults of A. melas accounted for 30.3% and 33.8% of the<br />

relative abundance and 20.6% and 35.2% of the relative biomass of the fish assemblage,<br />

respectively. However, adult relative <strong>de</strong>nsity varied greatly between sites (range 1.1% to<br />

84.2%). Because fish relative <strong>de</strong>nsity (CPUE) was highly correlated with relative abundance<br />

(%) and relative biomass (%) for both adults (linear regression, R 2 = 0.91, P < 0.001 and R 2 =<br />

0.81, P = 0.002, respectively) and YOY (linear regression, R 2 = 0.99, P < 0.001 and R 2 = 0.53,<br />

P = 0.042, respectively), we used relative <strong>de</strong>nsity (CPUE) values to investigate the<br />

relationships between A. melas abundance and the suite of environmental variables.<br />

138


Table 8.2. Details on fish species sampled in the 25 permanent habitats in the Brière marsh: origin (introduced vs native), total number, mean (standard <strong>de</strong>viation)<br />

abundance (CPUE), relative abundance (%) and relative biomass (%) for adult and YOY.<br />

N Abundance (CPUE) Relative abundance (%) Relative biomass (%)<br />

Species Origin 1 Adults YOY Adult YOY Adult YOY Adult YOY<br />

Black bullhead Ameiurus melas (Rafinesque) I 4524 810 7.35<br />

(8.7)<br />

Silver bream Blicca bjoerkna (Linnaeus)<br />

Common bream Abramis brama (Linnaeus) 2<br />

N<br />

N<br />

2015 696 3.05<br />

(2.8)<br />

Roach Rutilus rutilus (Linnaeus) N 1021 393 1.61<br />

(1.3)<br />

Rudd Scardinius erythrophthalmus (Linnaeus) N 589 130 0.94<br />

(0.9)<br />

Mosquitofish Gambusia holbrooki Girard I 675<br />

3<br />

-<br />

1.10<br />

(1.3)<br />

Pumpkinseed Lepomis gibbosus (Linnaeus) I 538 49 0.85<br />

(0.7)<br />

European eel Anguilla anguilla (Linnaeus) N 240 123 0.38<br />

(0.3)<br />

Others species 4 N - I 48 70 0.07<br />

(0.1)<br />

1.19<br />

(0.9 )<br />

0.92<br />

(0.7)<br />

0.79<br />

(0.7)<br />

0.26<br />

(0.3)<br />

33.81<br />

(24.4)<br />

22.76<br />

(17.7)<br />

14.82<br />

(11.1)<br />

7.15<br />

(5.5)<br />

- 3 7.38<br />

(5.8)<br />

0.19<br />

(0.3)<br />

0.46<br />

(0.5)<br />

0.13<br />

(0.2)<br />

9.97<br />

(13.3)<br />

3.50<br />

(3.7)<br />

0.61<br />

(1.1)<br />

30.30<br />

(21.9)<br />

23.57<br />

(17.4)<br />

19.64<br />

(15.3)<br />

6.73<br />

(8.0)<br />

35.20<br />

(24.3)<br />

21.72<br />

(15.2)<br />

12.36<br />

(10.7)<br />

4.39<br />

(3.5)<br />

- 3 0.12<br />

(0.1)<br />

1 N for native and I for introduced based on Keith & Allardi (2001); 2 Blicca bjoerkna and Abramis brama were grouped together since juveniles were not reliably<br />

i<strong>de</strong>ntifiable in the field; 3 All Gambusia holbrooki were grouped as adult; 4 Others species (mean relative abundance


Chapitre 8<br />

8.3.2. Associations between A. melas and environmental characteristics of<br />

permanent habitats<br />

At the canal network level, we found no significant relationship between abundance of YOY<br />

and adults and the two habitat <strong>de</strong>scriptors, although A. melas abundances were marginally<br />

lower in large canals (Table 8.3). At the microhabitat scale, YOY and adults were more<br />

abundant in locations with peat substratum (Table 8.4 and Figure 8.2). We found a number of<br />

subtle differences for YOY and adults. Whereas abundance of adults increased with the<br />

presence of adjacent reed beds (Figure 8.2), this relationship was not observed for YOY<br />

(Table 8.4). YOY abundance was negatively correlated with water <strong>de</strong>pth (Table 8.4 and<br />

Figure 8.2) whereas adult abundance was not (Table 8.4). The VC*WD and WD*SU<br />

interactions were significant for adults.<br />

Table 8.3. Results of the general linear mo<strong>de</strong>l for the abundance (CPUE) of young-of-the-year<br />

and adult Ameiurus melas (ln(x+1) transformed values) performed at the canal network level (n =<br />

25 canals). See Table 8.1 for <strong>de</strong>tails on habitat variables.<br />

Habitat variable MS df F P ß<br />

Depen<strong>de</strong>nt variable: Relative <strong>de</strong>nsity of adult A. melas<br />

CS 0.639 1 3.189 0.089 -0.191<br />

CC 0.007 1 0.034 0.855 -0.019<br />

CS*CC 0.059 1 0.295 0.593 -0.058<br />

Depen<strong>de</strong>nt variable: Relative <strong>de</strong>nsity of young-of-the-year A. melas<br />

CS 0.268 1 4.092 0.056 -0.124<br />

CC 0.024 1 0.372 0.549 -0.037<br />

CS*CC 0.101 1 1.547 0.227 0.076<br />

8.3.3. Use of temporary habitats by A. melas and the fish assemblage<br />

In total, we captured 411 individuals (131 A. melas) on the temporary habitats (659 PAS),<br />

belonging to 13 species. We found no difference in species richness between the two habitat<br />

types (marsh meadows vs reed beds, t-test, pooled variance t = - 0.488, df = 20, P > 0.05),<br />

whereas more fish (total abundance) were captured in reed beds (Mann-Whitney, U = 24.0, P<br />

= 0.018). Abundance (CPUE) of the fish assemblage without A. melas was similar between<br />

the two types of temporary habitats (Mann-Whitney, U = 38.5, P > 0.05), but abundance of A.<br />

melas was significantly higher in reed beds (Mann-Whitney, U = 20.5, P = 0.008) when<br />

compared with marsh meadows.<br />

140


Sélection d’habitat du poisson chat<br />

Table 8.4. Results of the general linear mo<strong>de</strong>l for the abundance (CPUE) of young-of-the-year and<br />

adult Ameiurus melas (ln(x+1) transformed values) performed at the microhabitat scale in the 25<br />

sampled canals (n = 646 PAS).<br />

Habitat variable MS df F P ß<br />

Depen<strong>de</strong>nt variable: Relative <strong>de</strong>nsity of adult A. melas<br />

Adjacent Habitat type (AH) 4.81 1 19.02


Chapitre 8<br />

8.4. Discussion<br />

In the present study, we found that the contemporary fish assemblage of the Brière marsh is<br />

largely dominated by A. melas. Our results provi<strong>de</strong> strong evi<strong>de</strong>nce that, this invasive species<br />

prefers reed bed habitats, which themselves have been rapidly invading the marsh because of<br />

the <strong>de</strong>cline of agricultural practices in this man-ma<strong>de</strong> ecosystem (Bernard & Rolland 1990).<br />

This suggests that A. melas may have benefited from the expansion of reed beds in view of (1)<br />

the positive relationships found between its abundance and the presence of adjacent reed beds,<br />

(2) the selection of peat substratum in small canals, and (3) the preferential use of reed beds<br />

(vs marsh meadows) during the flooding period. Although, some differences were found<br />

between YOY and adult individuals as a consequence of the ontogenetic shift in habitat use<br />

from YOY (pelagic habits) to adult (benthic habits) life-stages (e.g. Boët 2001). In general,<br />

YOY and adult abundances in permanent habitats were both positively correlated with trends<br />

<strong>de</strong>rived from the <strong>de</strong>velopment of reed beds. In<strong>de</strong>ed, reed beds are highly productive aquatic<br />

habitats. They likely affect habitat structure and food resources because of the large amounts<br />

of <strong>de</strong>tritus produced, which can remain standing for one or more years in the litter (see <strong>de</strong>tails<br />

in Mason & Bryant 1975). A. melas may preferentially occupy reed beds during the flooding<br />

season for trophic cues, notably the presence of a specific invertebrate assemblage associated<br />

with reed bed litter (Bedford & Powell 2005). Moreover, A. melas is a species with<br />

limnophilic habits (Gante & Santos 2002) and is tolerant to the harsh conditions imposed in<br />

many wetlands; therefore, it is likely to benefit from such changes in the abiotic environment.<br />

In the same way, we found that A. melas illustrates a strong affinity to reed beds as temporary<br />

floo<strong>de</strong>d habitats. Furthermore, A. melas abundance increase when <strong>de</strong>creasing water <strong>de</strong>pth<br />

(Brown et al. 1999) and the production of a large amount of litter by the reed beds might favor<br />

A. melas by limiting water <strong>de</strong>pth.<br />

While we are unsure of the specific mechanisms that have lead to the invasion success<br />

of A. melas, it is likely that the presence of other invasive animals have played a facilitative<br />

role. For example, we believe that the recent introduction of the invasive red swamp crayfish<br />

(Procambarus clarkii) in the Brière marsh may have favored the establishment of A. melas by<br />

modifying the physical characteristics of the habitat. In fact, P. clarkii is known to impact<br />

substratum composition (Rodríguez et al. 2005), which may favor limnophilic species like A.<br />

melas, which <strong>de</strong>pend on specific substratum conditions for spawning and nest construction<br />

(Braig & Johnson 2003). Moreover, P. clarkii might represent a more profitable prey item for<br />

A. melas compared to smaller invertebrates. Empirical evi<strong>de</strong>nce for such patterns would<br />

142


Sélection d’habitat du poisson chat<br />

provi<strong>de</strong> support for the invasional meltdown hypothesis of Simberloff & Von Holle (1999)<br />

whereby the presence of particular invasive species facilitate the subsequent establishment of<br />

other non-indigenous species. Future research addressing these questions is warranted.<br />

During the last 10 years, an active management plan has been un<strong>de</strong>rtaken in the Brière<br />

wetland to restore marsh meadows through clear-cutting operation and promoting the grazing<br />

of reed beds. These restoration activities have already positively affected the <strong>de</strong>nsity and<br />

species richness of waterfowl assemblage (Eybert et al. 1998). Because physical habitat can<br />

mediate biotic resistance to an inva<strong>de</strong>r and control invasion success and subsequent impacts<br />

(Byers 2002), it would be of interest to investigate whether the restoration of marsh meadows<br />

could increase the resistance of this ecosystem to future spread of A. melas. Furthermore, the<br />

study of the functional use of these temporary floo<strong>de</strong>d habitats and its seasonal variability<br />

(Crain et al. 2004) is nee<strong>de</strong>d to better evaluate their effects on fish assemblage composition.<br />

At the very least, we cite the need to increase our knowledge of the autoecology of A. melas<br />

to better un<strong>de</strong>rstand the ecological impacts of introduced freshwater fishes in man-ma<strong>de</strong><br />

environments that are now prevalent across the landscape.<br />

Acknowledgements<br />

We are grateful to E. Feunteun and three anonymous reviewers for valuable comments on<br />

earlier draft. We thank the Parc naturel régional <strong>de</strong> Brière for logistic, the PnrB, FEDER,<br />

DIREN, Région Pays <strong>de</strong> la Loire and Agence <strong>de</strong> l’Eau from Loire-Bretagne for financial<br />

support and particularly J.-P. Damien for assistance during field work. We thank B. Boisteau<br />

for his help on GIS analyses.<br />

143


Chapitre 8<br />

Les résultats obtenus lors <strong>de</strong> ce travail suggèrent que l’expansion <strong>de</strong>s roselières a<br />

fortement favorisé l’invasion <strong>de</strong> la zone d’étu<strong>de</strong> par le poisson-chat, le succès<br />

d’invasion est d’autant plus fort que les relations sont observées chez les juvéniles<br />

et les adultes. Ces résultats fournissent donc une approche nouvelle <strong>de</strong>s relations<br />

possibles entre les activités humaines et les processus d’invasions <strong>de</strong> certains<br />

types d’écosystèmes par <strong>de</strong>s espèces introduites, l’expansion <strong>de</strong>s roselières étant<br />

la résultante <strong>de</strong> la modification <strong>de</strong>s activités agricoles dans la zone d’étu<strong>de</strong>.<br />

Cependant, le manque <strong>de</strong> données sur la composition <strong>de</strong> la communauté piscicole<br />

durant toute la pério<strong>de</strong> <strong>de</strong> développement <strong>de</strong>s roselières au siècle <strong>de</strong>rnier ne nous<br />

permet pas d’entreprendre une analyse chronologique <strong>de</strong> l’implantation et du<br />

succès d’invasion <strong>de</strong> l’espèce (évolution <strong>de</strong> l’abondance en poissons et <strong>de</strong> la<br />

colonisation <strong>de</strong>s marais). La mise en place d’un tel suivi permettrait à l’avenir <strong>de</strong><br />

mieux appréhen<strong>de</strong>r le déterminisme <strong>de</strong> tels patrons d’invasion, en l’occurrence<br />

pour la présente espèce <strong>de</strong> comprendre les bases <strong>de</strong> sa future dynamique dans les<br />

marais <strong>de</strong> Brière, et aussi <strong>de</strong> savoir dans quelle mesure la restauration <strong>de</strong>s prairies<br />

par <strong>de</strong>struction <strong>de</strong>s roselières peut elle être efficace dans la lutte contre l’invasion<br />

<strong>de</strong> cette espèce. Certaines hypothèses quant à l’avantage <strong>de</strong> l’utilisation <strong>de</strong>s<br />

roselières pour le poisson-chat ont été émises telles que la présence <strong>de</strong><br />

l’écrevisses <strong>de</strong> Louisiane qui pourrait fournir un gain trophique significatif ou<br />

encore les caractéristiques environnementales <strong>de</strong>s roselières (régime thermique ou<br />

structure <strong>de</strong> la végétation) qui faciliteraient la reproduction <strong>de</strong> l’espèce. Ainsi,<br />

une approche comparative <strong>de</strong> certains traits d’histoire <strong>de</strong> vie d’individus <strong>de</strong> cette<br />

espèce sélectionnant <strong>de</strong>s habitats <strong>de</strong> nature différente (e.g. à <strong>de</strong>s <strong>de</strong>grés <strong>de</strong><br />

développement du roseau différents) permettrait <strong>de</strong> tester ce genre d’hypothèses.<br />

Une nouvelle question se pose suite à ce travail, à savoir si la régulation <strong>de</strong>s<br />

effectifs par l’activité <strong>de</strong> pêche actuellement mise en place pour gérer cette espèce<br />

(rappelons que cette espèce est classée ‘espèce nuisible’ au niveau législatif)<br />

influence le niveau d’abondance <strong>de</strong> sa population.<br />

144


Chapitre 9<br />

Effets <strong>de</strong> la <strong>de</strong>struction systématique <strong>de</strong>s individus<br />

capturés par la pêcherie sur la régulation du poisson chat<br />

(Ameiurus melas)<br />

Le poisson-chat, espèce invasive qui domine la communauté piscicole dans les<br />

marais <strong>de</strong> Gran<strong>de</strong> Brière Mottière, pose <strong>de</strong> sérieux problèmes en terme <strong>de</strong> gestion<br />

du peuplement piscicole local. De par son statut d’espèce nuisible, le poisson-chat<br />

est soumis à l’unique législation en vigueur au niveau national, c’est-à-dire une<br />

<strong>de</strong>struction systématique <strong>de</strong>s individus capturés et l’interdiction du transport<br />

d’individus vivants. Les finalités <strong>de</strong> cette gestion sont: (1) <strong>de</strong> limiter le niveau<br />

d’abondance <strong>de</strong> l’espèce dans les zones déjà colonisées, et (2) <strong>de</strong> limiter au<br />

maximum l’expansion <strong>de</strong> l’espèce vers <strong>de</strong>s zones où elle n’est pas encore<br />

présente. Ces mesures <strong>de</strong> gestion <strong>de</strong> l’espèce s’appuient, dans le cas <strong>de</strong> la Brière,<br />

sur l’unique action <strong>de</strong>s pêcheurs amateurs et n’ont cependant jamais été évaluées<br />

auparavant. Nous nous proposons donc ici <strong>de</strong> mesurer l’efficacité <strong>de</strong> la<br />

<strong>de</strong>struction systématique <strong>de</strong>s individus dans les marais <strong>de</strong> Gran<strong>de</strong> Brière Mottière<br />

sur le niveau d’abondance <strong>de</strong> cette espèce.


Chapitre 9<br />

Relève <strong>de</strong> bosselles par un pêcheur amateur <strong>de</strong>s marais <strong>de</strong> Gran<strong>de</strong> Brière Mottière<br />

Individus <strong>de</strong> poisson chat soumis au programme <strong>de</strong> régulation <strong>de</strong> l’espèce<br />

146


Gestion <strong>de</strong> la population <strong>de</strong> poisson chat<br />

Is mass removal an efficient measure to regulate the North<br />

American catfish Ameiurus melas outsi<strong>de</strong> of its native range? 7<br />

Résumé: Le poisson-chat est une espèce originaire d’Amérique du Nord dont <strong>de</strong><br />

nombreuses populations se sont établies avec succès en Europe. La principale mesure <strong>de</strong><br />

régulation <strong>de</strong> l’espèce prise en France consiste à détruire <strong>de</strong> façon systématique tous les<br />

poissons capturés par les pêcheurs. Cependant, l’efficacité <strong>de</strong> cette mesure <strong>de</strong> gestion n’a<br />

jamais été évaluée. Actuellement dans les marais <strong>de</strong> Gran<strong>de</strong> Brière Mottière, le poisson-chat<br />

domine le peuplement piscicole. Dans la présente étu<strong>de</strong>, nous avons cherché à évaluer les<br />

abondances <strong>de</strong> poisson-chat, en ayant recours à la pêche électrique et le piégeage passif. Nous<br />

avons également suivi les pratiques halieutiques <strong>de</strong>s pêcheurs amateurs aux engins. La<br />

distribution <strong>de</strong>s fréquences <strong>de</strong> tailles est significativement différente entre les individus<br />

capturés à l’ai<strong>de</strong> <strong>de</strong>s engins <strong>de</strong>s pêcheurs et ceux échantillonnés par pêche électrique, laissant<br />

penser à une sélectivité <strong>de</strong> capture <strong>de</strong>s poissons soumis à la pêcherie. Par ailleurs, l’abondance<br />

<strong>de</strong> poisson-chat est négativement corrélée à la pression <strong>de</strong> pêche aux engins, mesurée dans<br />

différentes zones du site d’étu<strong>de</strong>. Le rapport d’abondance entre les juvéniles <strong>de</strong> l’année et les<br />

adultes est relativement constant quelle que soit la zone considérée, suggérant ainsi que<br />

l’espèce ne compense pas la mortalité occasionnée par la pêcherie en augmentant le ratio<br />

juvéniles <strong>de</strong> l’année/adultes. Aux vues <strong>de</strong> ces résultats, nous proposons aux gestionnaires <strong>de</strong><br />

maintenir cette mesure <strong>de</strong> gestion, mais aussi d’envisager d’autres alternatives afin d’accroître<br />

l’efficacité <strong>de</strong> la régulation <strong>de</strong> cette espèce en <strong>de</strong>hors <strong>de</strong> son aire <strong>de</strong> distribution d’origine.<br />

7 Ce chapitre est adapté <strong>de</strong> l’article : Cucherousset, J., Paillisson, J.-M., & Carpentier, A. 2006. Is mass removal<br />

an efficient measure to regulate the North American catfish Ameiurus melas outsi<strong>de</strong> of its native range? Journal<br />

of Freshwater Ecology. 21: 699-704.<br />

147


Chapitre 9<br />

Abstract: The black bullhead (Ameiurus melas) is a North American species that has<br />

successfully established populations throughout Europe. The main management policy to<br />

regulate its population in France is systematic mass removal by fishers, but the efficiency of<br />

this measure has not been evaluated. In the Gran<strong>de</strong> Brière Motièrre marsh (northwest France),<br />

this species currently dominates the fish assemblage. We sampled this black bullhead<br />

population with trap nets and by electrofishing. The non-commercial fishery was also<br />

surveyed. Size-frequency distributions were significantly different between trapped<br />

individuals and those sampled by electrofishing, suggesting a size-selectivity of trap nets.<br />

Abundance of black bullhead was negatively correlated with the fishermen activity, measured<br />

at variable locations of the study site. The young-of-the-year/adult ratio was constant at each<br />

site, suggesting that black bullhead might not compensate for fishing mortality with increased<br />

recruitment, and therefore mass removal may not be effective.<br />

148


9.1. Introduction<br />

Gestion <strong>de</strong> la population <strong>de</strong> poisson chat<br />

During the two last centuries, many fish species have been introduced in Europe, notably<br />

from North America, with variable success of naturalization (review in Copp et al. 2005). The<br />

black bullhead (Ameiurus melas), an ictalurid fish native to North America, is an example of a<br />

non-indigenous fish species that has successfully established populations throughout Europe<br />

(Wheeler 1978), and particularly in France where it was introduced in 1871 (Boët 2001).<br />

Whereas this species is well studied in its native range (e.g. Hanchin et al. 2002a, Brown et al.<br />

1999), few studies have been conducted in its non-native range in Europe (but see Boët 1980,<br />

Copp 1993, Poizat & Crivelli 1997).<br />

Despite its classification as a “species liable to cause biological disequilibrium” in<br />

French legislation (article R. 232-3 Co<strong>de</strong> rural, see Guevel 1997), few management measures<br />

have been un<strong>de</strong>rtaken in France to control this invasive species. One such measure is the<br />

obligation for fishers to eliminate all captured individuals. Recently, Louette & Declerck<br />

(2006) experimentally showed that trapping may potentially be a cost-effective tool for the<br />

mass removing of brown bullhead (Ameiurus nebulosus). Nevertheless, the efficiency of this<br />

measure to limit black bullhead population size has never been assessed outsi<strong>de</strong> of its native<br />

range to our knowledge (but see Hanson et al. 1983).<br />

The objective of this study was to investigate the effect of the trap-net fishery on the<br />

black bullhead population in a man-ma<strong>de</strong> wetland (Gran<strong>de</strong> Brière Mottière marsh, northwest<br />

France). First, we compared the size-frequency distributions of black bullhead from the trapnet<br />

fishery with electrofishing to evaluate the size-selectivity of the gears. Then, we<br />

investigated whether black bullhead abundances were related to variable fishermen activity to<br />

assess the efficiency of this management tool.<br />

9.2. Methods and materials<br />

9.2.1. Study area and fishermen activity survey<br />

Based on traditional habits, the study site (Figure 9.1 and see <strong>de</strong>tails in the Chapitre 2) is<br />

divi<strong>de</strong>d into eight zones where fishing is permitted (mean area of land cover = 905 ha ± 366<br />

S.D.). Two protected areas (250 and 700 ha, respectively) where fishing is totally prohibited,<br />

were not inclu<strong>de</strong>d in the study. The study site supports a traditional fishery composed of noncommercial<br />

fishers that target the European eel (Anguilla anguilla) and large piscivorous fish<br />

such as the northern pike (Esox lucius) and the pikeperch (San<strong>de</strong>r lucioperca). The black<br />

bullhead, which was introduced to this area in 1929 (Maillard 1972) and currently dominates<br />

149


Chapitre 9<br />

the marsh fish assemblages, is principally captured by eel fishermen their eel pots (modified<br />

trap nets). In 2005, we questioned 28 fishermen using eel pots (i.e. 58% of the fishermen<br />

using this gear) to evaluate their effort in each of the eight zones of the Gran<strong>de</strong> Brière<br />

Mottière marsh. Fishing effort was calculated as the number of eel pots per kilometer of canal<br />

in each zone.<br />

N<br />

1 Km<br />

Canal network<br />

Fished area<br />

Protected area<br />

Zone <strong>de</strong>limitation<br />

Electrofishing<br />

Trapping<br />

Figure 9.1. Map of the Gran<strong>de</strong> Brière Mottière marsh canal network and location of trapping<br />

surveys and canals sampled by electrofishing in 2004.<br />

9.2.2. Black bullhead sampling<br />

The black bullhead population was sampled in 2004 by electrofishing and using eel pots using<br />

trap nets, which have already proven their efficiency to monitor black bullhead population<br />

(Hanchin et al. 2002b). Eight eel pots (1.5 m long with 1.0 x 0.4 m frames and 10 mm mesh)<br />

were randomly set at three locations in the study area from May to August, representing a<br />

trapping effort of 242 pot-nights.<br />

Fish were sampled by electrofishing (see <strong>de</strong>tails in Chapitre 2). Sixteen canal sections<br />

were sampled in August (i.e. after the spawning period of this species) totalling 401 PAS<br />

(mean = 25.1 ± 3.2 S.D. per section). In total, 5,084 fish were caught, and a minimal subsample<br />

of 100 individuals was measured to the nearest mm in each canal section. From the<br />

150


Gestion <strong>de</strong> la population <strong>de</strong> poisson chat<br />

inspection of size-frequency distributions, we used a 70 mm total length threshold to<br />

distinguish young-of-the-year (YOY) from adults, this threshold being in accordance with<br />

results of Hanchin et al. (2002a). Because YOY are not caught efficiently in eel pots because<br />

of the net mesh-size, YOY were removed from the data set collected by electrofishing to<br />

compare size distributions between the two sampling methods. Black bullhead abundance<br />

(CPUE) was log10(x+1) transformed to conduct linear regression.<br />

9.3. Results and discussion<br />

The size-frequency distributions of black bullhead > 70 mm (Figure 9.2) was different<br />

between trap nets and electrofishing (two-sample Kolmogorov-Smirnov test, KS = 0.170, P <<br />

0.001). This overall difference in size-classes between gears mainly involved the selectivity<br />

by eel pots of individuals between 70 and 100 mm (i.e. certainly age-1 individuals (Hanchin<br />

et al. 2002a)). Black bullhead abundance was negatively related with fishing effort (linear<br />

regression, n = 16, R 2 = 0.609, P < 0.001, Figure 9.3), indicating that eel pot fishing likely<br />

had an effect on bullhead <strong>de</strong>nsity. A major concern in mass removal is whether or not<br />

increased recruitment would negate the benefits of population reduction (Hanson et al. 1983).<br />

In the electrofishing samples, YOY/adult ratio did not vary in relation to the adult abundance<br />

(linear regression, n = 16, R² = 0.03, P = 0.503), suggesting that black bullhead might not<br />

compensate for fishing mortality with increased recruitment, at least un<strong>de</strong>r these metrics. This<br />

result is surprising because mechanical control methods are generally temporary since the<br />

remaining fishes exhibit compensatory survival, increased growth, and increased fecundity,<br />

all of which result in a rapid resi<strong>lien</strong>cy of populations (Wydoski & Wiley 1999). Nevertheless,<br />

Hanson et al. (1983) did not observe any significant increase in black bullhead recruitment. In<br />

the present study, the absence of a significant relation between YOY/adult ratio and adult<br />

abundance might indicate that fishing mortality affects black bullhead <strong>de</strong>nsity but not<br />

sufficiently to activate regulatory mechanisms.<br />

151


Chapitre 9<br />

Relative abundance (%)<br />

12<br />

10<br />

8<br />

6<br />

4<br />

2<br />

0<br />

70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230<br />

Size-classes (mm)<br />

Electrofishing (N = 1,535)<br />

Eel pots (N = 733)<br />

Figure 9.2. Size-frequency distribution for Ameiurus melas sampled by eel pots and electrofishing<br />

in the Gran<strong>de</strong> Brière Mottière marsh in 2004.<br />

Black bullhead abundance<br />

(CPUE, log (x+1) transformed)<br />

4<br />

3.5<br />

3<br />

2.5<br />

2<br />

1.5<br />

1<br />

0.5<br />

R² = 0.60<br />

P < 0.001<br />

0<br />

0 2 4 6 8 10 12 14 16<br />

Fishing effort (eel pots per km)<br />

Figure 9.3. Relationship between total black bullhead abundance and fishermen activity in the<br />

Gran<strong>de</strong> Brière Mottière marsh. Abundance (CPUE) values were log10(x+1) transformed.<br />

152


Gestion <strong>de</strong> la population <strong>de</strong> poisson chat<br />

The systematic mass removal by fishermen seems to be partly successful in limiting<br />

the abundance of black bullhead in the Gran<strong>de</strong> Brière Mottière marsh. Nevertheless, the black<br />

bullhead is still the dominant species in the local fish assemblage. Consequently, the use of<br />

this single management measure currently is not sufficient to regulate the population. Recent<br />

works have shown that alternative measures are efficient in limiting the establishment of nonnative<br />

fishes. In the Gran<strong>de</strong> Brière Mottière, the natural succession of grassland, which is<br />

characterized by reed bed expansion during the last century, has substantially altered<br />

ecosystem function (Eybert et al. 1998). This change is certainly one of major cause for the<br />

dominance of the black bullhead in the fish assemblage. Recently, Scoppettone et al. (2005)<br />

<strong>de</strong>monstrated that habitat rehabilitation could be valuable to control non-native fish species.<br />

Consequently, we propose that managers continue the systematic mass removal but in<br />

conjunction with natural habitat rehabilitation to more efficiently regulate the black bullhead<br />

population in the Gran<strong>de</strong> Brière Mottière marsh.<br />

Acknowledgements<br />

We are grateful to J.-P. Damien, E. Le Mitouard, V. Thoby and the numerous persons who<br />

participated in the field work. We acknowledge D.W. Willis and two anonymous referees for<br />

valuable comments on an earlier draft. We thank the Parc naturel régional <strong>de</strong> Brière for<br />

logistic and the PnrB, FEDER, DIREN, Région Pays <strong>de</strong> la Loire and Agence <strong>de</strong> l’Eau from<br />

Loire-Bretagne for financial support.<br />

153


Chapitre 9<br />

Les résultats obtenus dans ce travail montrent que la <strong>de</strong>struction systématique <strong>de</strong>s<br />

poissons chats capturés par les pêcheurs amateurs permet <strong>de</strong> faire diminuer le<br />

niveau d’abondance <strong>de</strong> cette espèce et ce, proportionnellement à l’effort <strong>de</strong> pêche.<br />

Cependant, même dans les zones où la pression <strong>de</strong> pêche est la plus forte (nombre<br />

<strong>de</strong> pêcheurs et engins utilisés), l’abondance <strong>de</strong> poissons chats reste encore élevée.<br />

Il semble donc que cette mesure <strong>de</strong> gestion, pourtant mise en place <strong>de</strong>puis <strong>de</strong><br />

nombreuses années, ait une efficacité limitée quant à la régulation du niveau<br />

d’abondance <strong>de</strong> cette espèce. Nous avons également regardé, dans cette étu<strong>de</strong>, si<br />

l’espèce présentait une réponse adaptative à une forte pression <strong>de</strong> mortalité sans<br />

que nous puissions la mettre en évi<strong>de</strong>nce. Ce résultat suggère <strong>de</strong>ux hypothèses :<br />

soit le niveau <strong>de</strong> mortalité infligé par la pêcherie est trop faible pour déclencher<br />

une telle réponse, soit le paramètre (ratio juvéniles <strong>de</strong> l’année/adultes) que nous<br />

avons utilisé ne permet pas <strong>de</strong> mesurer une réponse aussi fine. De nouveaux outils<br />

d’investigations doivent donc être mis en place. A l’avenir, il serait donc<br />

souhaitable <strong>de</strong> mesurer certains traits d’histoire <strong>de</strong> vie tels que la fécondité<br />

(nombre et taille <strong>de</strong>s œufs) ou encore la croissance pour vali<strong>de</strong>r cette hypothèse<br />

sur la base d’un échantillonnage stratifié (i.e. pression <strong>de</strong> pêche et niveau<br />

d’envahissement par le roseau). Ces résultats suggèrent également que cette<br />

mesure <strong>de</strong> gestion <strong>de</strong>vrait être couplée à d’autres, telle que la restauration<br />

d’habitat, pour réguler plus efficacement la population <strong>de</strong> poisson chat et<br />

contrecarrer l’évolution paysagère <strong>de</strong>s marais <strong>de</strong> Brière qui favorise cette espèce.<br />

Enfin, une nouvelle question se pose à la suite <strong>de</strong> ce travail : la restauration<br />

d’habitat, et plus particulièrement <strong>de</strong>s MTI, peut-elle être une mesure efficace<br />

dans la gestion d’une autre espèce d’intérêt majeur, non pas nuisible comme celle<br />

que nous venons d’abor<strong>de</strong>r, mais au contraire vulnérable?<br />

154


Chapitre 10<br />

Réponse <strong>de</strong> l’anguille européenne (Anguilla anguilla) à la<br />

mise en place <strong>de</strong> réserves <strong>de</strong> pêche<br />

La gestion d’une espèce uniquement à travers une réglementation <strong>de</strong> la pêche<br />

n’est pas suffisante et a besoin d’être accompagnée d’autres mesures telles que la<br />

gestion <strong>de</strong>s habitats. Dans les années 1980, d’importantes opérations <strong>de</strong><br />

restauration (e.g. travaux curage, fauchage, pâturage) ont été réalisées dans les<br />

marais <strong>de</strong> Brière, notamment sur d’anciennes zones comblées du marais et <strong>de</strong>s<br />

MTI envahis par la roselière, créant ainsi <strong>de</strong> nouveaux habitats aquatiques. Ces<br />

nouveaux territoires ont ensuite bénéficié d’un classement en réserve où accès,<br />

pêche et chasse sont interdits. Principale cible <strong>de</strong> la pêcherie locale, l’anguille<br />

européenne a, elle aussi, colonisé ces habitats comme l’ont prouvé différentes<br />

campagnes <strong>de</strong> pêche scientifiques effectuées dans ces réserves. Alors que <strong>de</strong><br />

nombreuses mesures <strong>de</strong> gestion sont proposées à l’échelle européenne pour<br />

essayer <strong>de</strong> sauvegar<strong>de</strong>r et mieux gérer cette espèce (e.g. réglementation drastique<br />

<strong>de</strong>s pêcheries, soutien <strong>de</strong> population par alevinage et aménagement <strong>de</strong>s ouvrages<br />

hydrauliques), l’alternative <strong>de</strong> mise en réserve n’a pas été évaluée chez l’anguille.<br />

Dans ce chapitre nous allons donc examiner la manière dont les réserves<br />

participent au fonctionnement <strong>de</strong> la population d’anguille <strong>de</strong> Brière et si elles<br />

offrent une solution alternative et valable pour la gestion <strong>de</strong> cette espèce et <strong>de</strong>s<br />

pêcheries traditionnelles.


Chapitre 10<br />

Pêche traditionnelle <strong>de</strong> l’anguille à l’ai<strong>de</strong> d’un carrelet mobile<br />

Entrée <strong>de</strong> la Réserve du Sud <strong>de</strong>s marais <strong>de</strong> Gran<strong>de</strong> Brière Mottière<br />

156


Réserves <strong>de</strong> pêche et anguille<br />

Freshwater protected areas: an effective measure to reconcile<br />

conservation and exploitation of the threatened European eels<br />

(Anguilla anguilla)? 8<br />

Résumé: La population d’anguille européenne a fortement régressé malgré la mise en<br />

place <strong>de</strong> nombreux plans <strong>de</strong> gestion et <strong>de</strong> nouveaux outils dans un but <strong>de</strong> préservation <strong>de</strong><br />

l’espèce sont indispensables. L’utilisation <strong>de</strong> réserves <strong>de</strong> pêche d’eau douce a parfois été<br />

évoquée, mais leur efficacité n’a jamais été évaluée chez cette espèce. Dans cette étu<strong>de</strong>, nous<br />

comparons la structure <strong>de</strong> classe <strong>de</strong> tailles <strong>de</strong> la population et la production d’anguilles<br />

argentées entre les zones pêchées et les réserves <strong>de</strong>s marais <strong>de</strong> Gran<strong>de</strong> Brière Mottière. Pour<br />

cela, nous nous basons sur un suivi intensif <strong>de</strong> la population ainsi que sur une étu<strong>de</strong> <strong>de</strong>s<br />

pratiques halieutiques. Nos résultats montrent que les pêcheurs ciblent les anguilles jaunes <strong>de</strong><br />

plus <strong>de</strong> 320 mm principalement capturées à l’ai<strong>de</strong> <strong>de</strong> bosselles et <strong>de</strong> carrelets. Cependant, la<br />

pression <strong>de</strong> pêche est géographiquement très variable. Nous avons ainsi trouvé <strong>de</strong>s différences<br />

<strong>de</strong> structure <strong>de</strong> tailles et <strong>de</strong> taux <strong>de</strong> mortalité entre les zones pêchées et les réserves, la<br />

mortalité <strong>de</strong>s anguilles supérieures à 320 mm étant positivement corrélée à la pression <strong>de</strong><br />

pêche. De plus, la proportion d’anguilles potentiellement migrantes est plus élevée dans les<br />

réserves que dans les zones pêchées (6.38% vs. 1.42%, respectivement). Nos résultats<br />

montrent également que les réserves produisent 8.4% <strong>de</strong> la production totale <strong>de</strong> géniteurs<br />

potentiels alors qu’elles ne représentent que seulement 2.4% <strong>de</strong> la surface en eau <strong>de</strong>s marais<br />

<strong>de</strong> Gran<strong>de</strong> Brière Mottière. Nous avons estimé qu’un ajustement <strong>de</strong> la taille <strong>de</strong>s réserves à<br />

31.1% <strong>de</strong> la surface <strong>de</strong> marais permettrait la production <strong>de</strong> 50% <strong>de</strong>s géniteurs en l’absence <strong>de</strong><br />

toute source <strong>de</strong> mortalité anthropique. Ainsi, à travers cet exemple, la mise en place <strong>de</strong><br />

réserves <strong>de</strong> pêche d’eau douce apparaît comme une mesure <strong>de</strong> gestion prometteuse, offrant un<br />

consensus constructif qui intègre à la fois les valeurs sociale et patrimoniale <strong>de</strong>s pêcheries<br />

traditionnelles tout en respectant les recommandations <strong>de</strong> gestion <strong>de</strong> l’anguille. De plus, ces<br />

réserves, si elles sont intégrées dans les programmes <strong>de</strong> gestion <strong>de</strong> la biodiversité aquatique,<br />

et si elles s’accompagnent d’autres mesures s’intéressant à tous les sta<strong>de</strong>s <strong>de</strong> vie <strong>de</strong> l’espèce,<br />

peuvent <strong>de</strong>venir une solution locale pour la gestion <strong>de</strong> l’anguille européenne.<br />

8 Ce chapitre est adapté d’un article intitulé « Cucherousset, J., Paillisson, J.-M., Carpentier, A., Thoby, V.,<br />

Damien, J.-P., Eybert, M.-C., Feunteun, E. & Robinet, T. Freshwater protected areas: an effective measure to<br />

reconcile conservation and exploitation of the threatened European eels (Anguilla anguilla)? » soumis et révisé<br />

dans la revue « Ecology of Freshwater Fish ».<br />

157


Chapitre 10<br />

Abstract: For <strong>de</strong>ca<strong>de</strong>s, the European eel Anguilla anguilla (L.) population has been<br />

<strong>de</strong>clining strongly <strong>de</strong>spite several management attempts, so additional experiments need to be<br />

conducted on management measures. The use of freshwater protected areas has been<br />

advocated but their efficiency has never been assessed. In this study, we investigated whether<br />

the population structure and the silver eel (i.e. mature migrating stage) production differ in<br />

fished and protected areas within a marsh wetland (Brière, 7000 ha, Northwest France), using<br />

an intensive biological study (electrofishing and trapping) and a survey of the traditional<br />

fishery. First, we found that fishermen mainly targeted >320 mm yellow eels (i.e. se<strong>de</strong>ntary<br />

stage) using eel pots and square lift nets and that the harvest by fishermen was highly variable<br />

between locations in the study area. Second, we found differences in the size-class structures<br />

and mortality rates of eels between protected and fished areas, mortality rates of eels >340<br />

mm being positively correlated with harvest by fishermen. Furthermore, the proportion of<br />

potentially migrating eels was higher in the protected areas than in fished areas (6.38% vs.<br />

1.42%, respectively). Third, we found that protected areas potentially produce 8.4% of the<br />

total silver eel production whereas they only account for 2.4% of the aquatic habitat area. We<br />

estimated that a size adjustment of protected areas to 31.1% with maintaining the current<br />

fishery would produce 50% of the potential silver eel of a fully protected marsh. Freshwater<br />

protected areas appear to be a promising management measure and a constructive means of<br />

integrating the socioeconomic value of the traditional fishery and the fisheries management<br />

needs to protect European eel stocks at the international scale. Furthermore, freshwater<br />

protective measures can be an effective local solution if they are integrated into the<br />

framework of freshwater biodiversity management and accompanied by other management<br />

measures that focus on all eel life stages.<br />

158


10.1. Introduction<br />

Réserves <strong>de</strong> pêche et anguille<br />

The European eel Anguilla anguilla (L.) population is in a steep <strong>de</strong>cline that began in the<br />

1970s (Moriarty & Dekker 1997, Feunteun 2002, ICES WGEEL 2006). The most frequently<br />

cited causes of the <strong>de</strong>cline are: climate-related changes in marine currents and ocean<br />

productivity, obstructions to migration, over fishing, habitat <strong>de</strong>gradation and non-native<br />

parasite infestations (Feunteun 2002, Robinet & Feunteun 2002, Starkie 2003). Attempts to<br />

manage and restore local eel stocks inclu<strong>de</strong> (see review in Feunteun 2002) (i) regulation of<br />

fisheries at various biological stages (e.g. Rosell et al. 2005), (ii) mitigation of obstacles to<br />

migration in particular the creation of fish passes (e.g. Knights & White 1998), and (iii)<br />

restocking programme (e.g. Moriarty & Dekker 1997). Despite these programs, the general<br />

<strong>de</strong>cline in eel stocks continues and additional management measures need to be <strong>de</strong>veloped.<br />

Since 1999, the ICES Working Group on Eel has recommen<strong>de</strong>d reducing human impacts on<br />

the production and escapement of silver eels (i.e. mature migrating stage) to the lowest<br />

possible levels (ICES WGEEL 2006). Now, the situation is becoming increasingly critical for<br />

the eel fisheries, and ICES experts have expressly requested that “areas producing high<br />

quality silver eels (large sized females, low contaminant and parasite bur<strong>de</strong>ns, unimpacted by<br />

hydropower stations)” be i<strong>de</strong>ntified in or<strong>de</strong>r to prioritize their conservation (Dekker 2003,<br />

ICES WGEEL 2006).<br />

Marine protected areas have been proposed as an easily enforced conservation method<br />

for managers to reduce the impacts of fishing on marine populations and habitats (Apostolaki<br />

2002). Scientists have <strong>de</strong>veloped practical and theoretical approaches for the <strong>de</strong>sign and the<br />

implementation of marine protected areas that have benefits for biodiversity and fisheries<br />

threatened by anthropogenic activities (see review in Leslie 2005). Recent research has shown<br />

that the success of marine protected areas also <strong>de</strong>pends on the integration of social, economic,<br />

political and scientific factors (Lundquist & Granek 2005, Stem et al. 2005). Some attempts<br />

have been recently conducted worldwi<strong>de</strong>, with variable success, to <strong>de</strong>velop freshwater<br />

protected areas (Maitland 1995, Keith 2000, Saun<strong>de</strong>rs et al. 2002). In general, few areas have<br />

been created specifically for freshwater fish, and almost all freshwater protected areas were<br />

inclu<strong>de</strong>d “inci<strong>de</strong>ntally” as part of terrestrial reserves (Eybert et al. 1998, Keith 2000, Self<br />

2005). Although freshwater protected areas have been advocated for management of<br />

American and European eel stocks (Feunteun 2002, Morrison & Secor 2003), their utility for<br />

conservation has not been evaluated.<br />

159


Chapitre 10<br />

Small freshwater coastal marshes are useful for studying this issue because they are<br />

wi<strong>de</strong>ly colonized by eels (Feunteun et al. 1992) and in recent <strong>de</strong>ca<strong>de</strong>s, habitat restoration<br />

programme have been un<strong>de</strong>rtaken and in some cases freshwater protected areas have been<br />

created. These ecosystems also comprise recreational and traditional eel fisheries and their<br />

limited size allows the whole local eel population to be studied. Furthermore, the role played<br />

by many small inland ecosystems in terms of silver eel production remains to be quantified<br />

(Feunteun et al. 2000). The configuration of the Gran<strong>de</strong> Brière Mottière (GBM, western<br />

France) offers good opportunities to test the efficiency of protected areas for eels because this<br />

coastal freshwater marsh has a traditional fishery and two protected areas that were created in<br />

the early 1980s. The objectives of the present study are: (1) to characterize the yellow eel (i.e.,<br />

se<strong>de</strong>ntary stage) size-classes targeted by the local fishery and the spatial distribution of<br />

catches, (2) to compare the eel population structure between fished and protected areas by<br />

analyzing size-class distributions, mortality rates and silver eel production, (3) to measure<br />

how the fisheries impact the eel size-class structure, and finally (4) to estimate the local eel<br />

stock and the differences in silver eel production between fished and protected areas in or<strong>de</strong>r<br />

to evaluate the efficiency of protected areas in terms of quantity and quality of the future<br />

silver eels leaving the GBM.<br />

10.2. Materials and methods<br />

10.2.1. Study area<br />

The GBM is composed of a complex network of permanent canals (144 km representing<br />

206.4 ha) and semi-permanent ponds (392.7 ha, Figure 10.1 and see Chapitre 2 for <strong>de</strong>tails). In<br />

the general framework of restoration programs <strong>de</strong>veloped in the early 1980s to limit the<br />

expansion of reed beds (Bernard & Rolland 1990), two protected areas, where fishing,<br />

hunting and entry are totally prohibited, have been created (Eybert et al. 1998, Figure 10.1).<br />

The southern and northern protected areas were created in 1973 and 1989, and cover 700 ha<br />

and 250 ha of land composed of 8.1 and 6.5 ha of aquatic habitat, respectively. They currently<br />

represent 2.4% of the total aquatic area. Based on traditional habits, the study site is divi<strong>de</strong>d<br />

into eight zones where fishing is permitted plus the two protected areas (Figure 10.1).<br />

160


N<br />

1 km<br />

Habitat:<br />

Fished area<br />

Protected area<br />

Water pond<br />

Canal<br />

8<br />

7<br />

1<br />

10<br />

5<br />

6<br />

9<br />

2<br />

Eel population survey:<br />

1<br />

Fishery survey<br />

Réserves <strong>de</strong> pêche et anguille<br />

4<br />

3<br />

Trapping Electrofishing<br />

Figure 10.1. Map of the Gran<strong>de</strong> Brière Marsh illustrating the canal network, ponds, fished and<br />

protected areas with their co<strong>de</strong>s and the location of eel population survey. Different symbols<br />

represent the location of the fishery surveys () the trapping () and the electrofishing ()<br />

during 2004 and 2005.<br />

161


Chapitre 10<br />

10.2.2. Traditional fishery survey<br />

Since 1784, the GBM marsh had its own property law (“undivi<strong>de</strong>d” and privately owned) and<br />

a specific fishery legislation, the fishery being composed of non-commercial fishermen. In<br />

2005, a fishery survey was conducted to assess the harvest by fishermen (expressed as the<br />

number of eels captured per ha) for each type of gear at different zones of the study area<br />

(Figure 10.1) using three different ways. Firstly, all fishing licences were analysed to count<br />

the number of fishermen and to assess the total number of gear used (product of the number of<br />

fishermen with the mean (± SE) number of each gear per fishermen). Secondly, anonymous<br />

questionnaires were distributed to evaluate the fishing practices since logbooks were rarely<br />

available. During the fishing season, questionnaires were randomly distributed directly to<br />

fishermen on the field or via fishermen associations with follow-up contacts were to improve<br />

the response rate. Fishermen were questioned on the species targeted, their catches, the<br />

number and type of gear and the frequency and location of their trips. The representativeness<br />

of the questioned fishermen was checked to ascertain the wi<strong>de</strong>r application of the data (e.g.<br />

Roth et al. 2001) by comparing these fishermen to the total fishing licenses using a χ² test.<br />

Thirdly, some fishermen were accompanied during their trips (creel survey) to assess the sizeclasses<br />

targeted by comparing the total captures with those fish released. Based on the<br />

fishermen logbooks available (n = 3), we found that eel captures occurred mainly in May<br />

(73% of total captures) as a result of a very limited seasonal efficiency of fishing gears in<br />

respect with the local water regime. Based on questionnaires, we calculated the number of<br />

eels caught during this month by multiplying the number of fishermen with the mean (± SE)<br />

number of each gear per fishermen with the mean (± SE) catch per unit effort (CPUE) for<br />

each gear and then assessed the total number of catches for the whole year.<br />

10.2.3. Eel population surveys<br />

10.2.3.1. Sampling<br />

The eel population was sampled in 2004 and 2005 using trapping and electrofishing (Figure<br />

10.1, Table 10.1 and Chapitre 2). Trapping was randomly performed to assess population<br />

parameters (i.e., size-class profiles, silvering rate and sex ratio) in restricted locations of the<br />

protected and fished areas. It was conducted using fyke nets (two wings 1.2 m high and 3 m<br />

long directing the fish into the 2 m long and 50 cm diameter chamber of 5 mm mesh) and<br />

fishermen eel pots (1.5 m long with 1.0 x 0.4 m frames and 10 mm mesh).<br />

162


Table 10.1. Sampling <strong>de</strong>sign of the eel population (electrofishing and trapping) and traditional fishery survey in protected and fished areas of the Gran<strong>de</strong><br />

Brière Mottière marsh in 2004 and 2005.<br />

Method Year Area Period Habitat<br />

Electrofishing<br />

Trapping<br />

Number<br />

of sites<br />

Sampling effort<br />

Number of<br />

eels<br />

Fished August Canal 15 25.1 PAS/site 244<br />

2004 Protected August Canal 2 31.5 PAS/site 21<br />

2005<br />

Fished June-July<br />

Protected June July<br />

Canal 17 25.0 PAS/site 304<br />

Pond 9 27.8 PAS/site 68<br />

Canal 2 25.0 PAS/site 25<br />

Pond 2 30.0 PAS/site 19<br />

Fished June to July Canal 3 8 gear 125<br />

2004 Protected May to August Canal 1 10 gear 449<br />

2005 Protected March to August Canal 1 10gear 347<br />

Fishery survey 2005 Fished March to June Canal 6 13 fishing trips 266


Chapitre 10<br />

All trapping data (fishermen creel and scientific surveys) were pooled to increase the number<br />

of eels sampled (Table 10.1). Because trapping was not applicable to the whole study area,<br />

electrofishing was randomly conducted at different locations over the entire study area in<br />

2004 and 2005 to assess the geographical variation in eel abundance (Figure 10.1 and Table<br />

10.1). Sampling was conducted using the point abundance sampling method (see <strong>de</strong>tails in<br />

Chapitre 2) and in total, we conducted 1225 PAS in 17 and 30 sites sampled in 2004 and 2005,<br />

respectively (<strong>de</strong>tails in Table 10.1). Abundance was expressed in Catch per Unit Effort<br />

(CPUE), i.e. the number of eels caught per PAS.<br />

For the two sampling methods, eels were anesthetized with eugenol (0.04 mL.L -1 ),<br />

measured (total length TL to the nearest mm), weighed (W, to the nearest g), macroscopic<br />

silvering criteria were assessed (Acou et al. 2005), and then the eels were released into the<br />

water. Given that some differences might occur in the selectivity of trapping gear in relation<br />

to different mesh sizes, only eels longer than the modal body size (i.e. TL = 320 mm) were<br />

used for further analyses (see Naismith & Knights 1990, Knights et al. 1996). Given that<br />

elvers (post larval life stage


Réserves <strong>de</strong> pêche et anguille<br />

where N(t) is the number of individuals of age (t), N (t+1) is the number of individuals of age<br />

(t+1) and Z the total mortality rate per year. The mortality rate calculation was performed at the<br />

study area scale (i.e. protected vs. fished areas), and not for each zone given the insufficient<br />

number of individuals sampled in each zone. At the zone scale, mortality was estimated by<br />

calculating the difference in abundance obtained by electrofishing (expressed in CPUE)<br />

between untargeted and targeted eel size-classes.<br />

In 2005, the proportion of silver eels was <strong>de</strong>termined using a standardized method<br />

based on macroscopic criteria (ocular in<strong>de</strong>x, state of differentiation of the lateral line and<br />

colour contrast) that provi<strong>de</strong>s a quick i<strong>de</strong>ntification in the field of pre-migrant eels in a<br />

standardized way and without sacrificing any individuals (Acou et al. 2005). However, this<br />

method was only <strong>de</strong>veloped on data collected later in the season (i.e. September and October,<br />

Acou et al. (2005)), so we used two earlier criteria of silvering prior to this (i.e. ocular<br />

hypertrophy and differentiation of the lateral line). In<strong>de</strong>ed, the typical pigmentation of silver<br />

eel occurred generally at the end of the silvering process (Acou et al. 2005, Durif et al. 2005).<br />

Because few silver eels were sampled in protected areas by electrofishing, the proportion of<br />

silver eels in the protected areas was calculated using silvering rate in fished area multiplied<br />

by the silvering ratio between protected and fished areas. In addition, we used the Fulton’s<br />

−3<br />

condition factor ( K = W × TL × 100000 ) as a general indicator of pre-migrant quality<br />

(EELREP 2005), and used the 440 mm threshold to assess the sex of silver eels since no<br />

individuals were sacrificed during the study. Individuals longer than 440 mm are known to be<br />

females (e.g. Tesch 2003, Acou et al. in press). The sex ratio was expressed as the proportion<br />

of females among pre-migrant eels.<br />

10.2.4. Stock assessment, fishery mortality and silver eel production<br />

The assessment of eel stocks in extensive areas is particularly difficult. In<strong>de</strong>ed, the most<br />

common technique (<strong>de</strong>pletion sampling associated with electrofishing) consumes manpower,<br />

time and equipment and thus is not easily applicable in extensive areas (e.g. Lobon-Cervia &<br />

Utrilla 1993). Nevertheless, a method has been <strong>de</strong>veloped to point sample the eel abundance<br />

in freshwater areas by establishing the relationship between PAS and <strong>de</strong>pletion samples (see<br />

in Laffaille et al. 2005a). These authors recommen<strong>de</strong>d <strong>de</strong>veloping a single relationship for<br />

each type of equipment and habitat. Furthermore, the authors found that this relationship is<br />

constant at variable eel <strong>de</strong>nsities (expressed as number of eel·100m -2 , Laffaille et al. 2005a).<br />

165


Chapitre 10<br />

Using the habitat differentiation between the shoreline and open water (Broad et al. 2001,<br />

Schulze et al. 2004), we established the following relationship in a typical canal of the GBM<br />

based on 25 PAS distributed in two sites sampled by <strong>de</strong>pletion: Eel <strong>de</strong>nsity (eel·100m -2 ) =<br />

10.59 (± 1.55 S.E.) PAS (in CPUE). No differences in the size-class distribution of eels<br />

between PAS and <strong>de</strong>pletion sampling were found (Kolmogorov-Smirnov two-sample test, KS<br />

= 0.201, P > 0.602, n = 57). Eel stock assessment was <strong>de</strong>rived from this relationship and the<br />

estimation of the area of each type of habitat (shoreline and open water) using a Geographical<br />

Information System (source Parc naturel regional <strong>de</strong> Brière).<br />

Next, we used the mortality rates estimated in protected and fished areas and the<br />

estimated eel stock un<strong>de</strong>r exploitation (eels >340 mm, see below) to evaluate the fishing<br />

mortality based on scientific data. We used the formula:<br />

Z = F + M (2)<br />

where Z is the total mortality, F is the fishing mortality and M is the natural mortality and<br />

making the assumption that recruitment and population parameters were similar in 2004 and<br />

2005 (see before). Thus, in the protected areas, the fishing mortality was assumed to be zero<br />

(F = 0) and thus resulting in M = Z. For the calculation of F at the fished areas, the M value<br />

was subtracted from the Z value in or<strong>de</strong>r to obtain the fishing mortality (F = Z –M). The<br />

number of eels caught by the fishery (NF) was estimated using the following formula and (1)<br />

and (2):<br />

N<br />

F<br />

( −M<br />

)<br />

( −F<br />

)<br />

= N × e × ( 1−<br />

e ) (3)<br />

( t )<br />

where NF is the number of eels that died from fishing mortality. Next, estimated NF was<br />

compared to the results obtained from fishermen questionnaires. Finally, silver eel production<br />

was <strong>de</strong>rived from the estimated eel stock and the proportion of silver eels in both protected<br />

and fished areas. All estimates (number and biomass of eels) and their upper and lower values<br />

(hyphenated) resulted from the products of the lower values (mean – SE), the means and the<br />

upper values (mean + SE) of the population parameters.<br />

166


10.3. Results<br />

10.3.1. The traditional fishery<br />

Réserves <strong>de</strong> pêche et anguille<br />

A total of 521 fishing licences were sold in the GBM marsh in 2005: 34 for the use of eel pots,<br />

66 for square lift nets and 46 for fish spears; the rest of the licences being attributed for gill<br />

nets, rods and multiple gears. Nevertheless, gillnets and rods are principally used to catch<br />

piscivorous fish (northern pike Esox lucius and pikeperch San<strong>de</strong>r lucioperca). The eel fishery<br />

was composed of 48 fishermen using pots, 87 using square lift nets and 60 using spears. In<br />

total, 75 fishermen respon<strong>de</strong>d to the questionnaires and provi<strong>de</strong>d data, including 28 using pots,<br />

43 using square lift nets and 26 using spears, i.e. on average 49.7 % of the total number of<br />

fishermen for each gear. Furthermore, the set of fishermen that respon<strong>de</strong>d to the<br />

questionnaires did not differ significantly from the entire eel fishery (Chi-square, df = 2, χ² =<br />

0.797, P < 0.671). Based on these questionnaires, we estimated that 23 892 eels (18 206 to 29<br />

578 ranging from lowest to highest estimation) were caught in 2005. Given that the mean<br />

weight of eels kept by fishermen was 127.7 g (± 5.7 S.E.), the total biomass of eels fished<br />

reached 3052 kg (2222; 3947).<br />

Based on the spatial distribution of the fishing activity, we found that harvest by<br />

fishermen (in eel·ha -1 ) varied largely between zones and gears (Figure 10.2). The highest<br />

harvest by fishermen was found in the southern part of the GBM (zones 3 and 4) for all gear<br />

types, and the northern part of the GBM (zones 1 and 2) which was mainly fished with pots.<br />

The lowest harvest by fishermen occurred in the eastern part (zones 5, 6, 7 and 8; Figures 10.1<br />

and 10.2). Because fish spears account for a restricted part of the catches (5%), data related to<br />

this gear have been removed from the analysis of targeted eel size-classes per gear type.<br />

Based on creel surveys, fishermen caught eels from 240 to 760 mm total length. Furthermore,<br />

33.5% of the total eel captures (n = 257) were released by fishermen. The size distribution<br />

differed between released and retained eels (Kolmogorov-Smirnov two-sample test, KS =<br />

0.678, P < 0.001) and released individuals were on average smaller than those kept (Mann-<br />

Whitney test, U = 13.950, P < 0.001). Fishermen released on average 79.4% (± 5.1 S.E.) of<br />

smaller eels (from 240 to 320 mm) and kept a high ratio (up to 60 %) of eels measuring more<br />

than 320 mm. From 420 to 620 mm, all eels (100%) were kept. Interestingly, some fishermen<br />

ten<strong>de</strong>d to release some of the larger eels (Figure 10.3). Thus, based on their size, eels were<br />

classified into those untargeted and targeted by the fishery using the 320 mm threshold.<br />

167


Chapitre 10<br />

Harvest by fishermen (eel·ha -1 )<br />

Number of individuals<br />

400<br />

350<br />

300<br />

250<br />

200<br />

150<br />

100<br />

50<br />

0<br />

Fishing gears:<br />

Square lift nets<br />

Eel pots<br />

Fishing spears<br />

1 2 3 4 5 6 7 8<br />

Zone<br />

Figure 10.2. Geographical distribution of the eel harvest by fishermen (eel·ha -1 ) in each zone of<br />

the fished area and for each fishing gear (eel pots, square lift nets and fishing spears).<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

Eel pots<br />

200 240 280 320 360 400 440 480 520 560 600 640 680 720 760 800<br />

Total length (mm)<br />

Square lift nets<br />

Figure 10.3. Size-classes distribution of eels caught by fishermen during the creel survey using<br />

eels pots (n = 184) and square lift nets (n = 82) and percentage of eels kept by fishermen for each<br />

size-class (open circles).<br />

168<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

Percentage of eel kept (%)


10.3.2. Eel population characteristics<br />

Réserves <strong>de</strong> pêche et anguille<br />

In 2004 and 2005, we captured 1868 eels: 681 by electrofishing, 921 by trapping and 266<br />

during the fishermen creel surveys (see Table 10.1 for <strong>de</strong>tails). The following length-weight<br />

relationship was established: W = 13.771 e 0.0045TL (n = 905, r 2 = 0.99). There was no<br />

difference in the size-class distribution between 2004 and 2005 for the eels sampled by<br />

trapping nor by electrofishing (Kolmogorov-Smirnov two-sample test, P > 0.05) or in the<br />

mean abundance per site by electrofishing (Mann-Whitney test, P > 0.05). Eels up to 320 mm<br />

sampled by trapping in fished areas were on average smaller than those from protected areas<br />

(Mann-Whitney test, U = 63.853, P < 0.001, n = 891) and size-class distribution differed<br />

significantly between protected and fished areas (Kolmogorov-Smirnov two-sample test, KS<br />

= 0.164, P < 0.001, n = 891, Figure 10.4).<br />

Number of individuals<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

340 380 420 460 500 540 580 620 660 700 740 780 820 860 900<br />

Total length (mm)<br />

Protected area<br />

Fished area<br />

Figure 10.4. Size-classes distribution of eels >320 mm trapped in fished (n = 251) and protected<br />

(n = 640) areas in 2004 and 2005.<br />

Based on the trapping data and using equation (1), we found a lower mean mortality<br />

rates Z in the protected (0.12) than in the fished areas (0.32). We also found that the<br />

differences in abundance between untargeted (TL < 320 mm) and targeted (>320 mm TL) eels<br />

were positively correlated with the harvest by fishermen from protected to highly fished zones<br />

(Linear regression, r² = 0.51, P = 0.021, n = 10, Figure 10.5).<br />

169


Chapitre 10<br />

Differences in abundance between<br />

untargeted and targeted eels (CPUE)<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

Fished area<br />

Protected area<br />

0 50 100 150 200 250 300 350 400<br />

Harvest by fishermen (eel·ha -1 )<br />

y = 0.0011x + 0.1676<br />

r² = 0.51 p = 0.021<br />

Figure 10.5. Relationship between the difference in abundance between untargeted and targeted<br />

eels and the harvest by fishermen (eel·ha -1 ) in each zone (protected and fished) of the study area<br />

and. See <strong>de</strong>tails in the method section for the calculation of the harvest of fishermen.<br />

Based on the trapping data, we found that the proportion of silver eels was 4.5 times<br />

higher in protected than in the fished areas (Fisher’s exact test, P = 0.003). In<strong>de</strong>ed, 12.83% of<br />

eels >320 mm (n = 265) caught in protected areas presented silvering criteria (i.e., ocular<br />

hypertrophy and differentiation of the lateral line) whereas only 2.86% of eels caught in<br />

fished areas (n = 102) presented these criteria. The proportion of silver eels amongst those<br />

greater > 320 mm TL in the fished areas did not differ between electrofishing (4.87%, n = 86)<br />

and trapping (Fisher’s exact test, P > 0.512). Consequently, the proportion of silver eels was<br />

1.42% (n = 282) when all size-classes from electrofishing were used in the fished areas (Table<br />

10.2). The proportion of silver eels in the protected areas reached 6.38%. The sex ratio of<br />

silver eels was largely biased towards females: 1/36 by trapping and 0/5 by electrofishing,<br />

with no differences between fished and protected areas and sampling methods (Fisher’s exact<br />

test, P > 0.05, Table 10.2). Silver eels weighed 585.4 g (± 46.8 S.E., n = 29) for a mean length<br />

of 675.8 mm (± 17.5 S.E.) on average and their condition factor averaged 0.18 (± 0.01 S.E.).<br />

170


Réserves <strong>de</strong> pêche et anguille<br />

Table 10.2. Number of eels and silver eels (males and females) caught by electrofishing and<br />

trapping in 2005 in the protected and fished areas of Gran<strong>de</strong> Brière Mottière marsh.<br />

Sampling method Parameters Fished area Protected area Total<br />

Total number of eels 282 24 306<br />

Electrofishing<br />

Silver eels 4 1 5<br />

Trapping<br />

Males 0 0 0<br />

Females 4 1 5<br />

Total number of eels 105 265 370<br />

Silver eels 3 34 37<br />

Males 0 1 1<br />

Females 3 33 36<br />

10.3.3. Stock assessment, fishing mortality and silver eel production<br />

The mean estimated eel <strong>de</strong>nsity was highly variable between zones (234.2 ind.ha -1 ± 42.4<br />

S.E.), ranging from 94.1 to 577 ind.ha -1 . The overall stock of eels > 150 mm was estimated to<br />

be 129 076 (74 022 to 200 206) individuals (see Table 10.3 for <strong>de</strong>tails). Based on equation (3),<br />

the total mortality rates in protected and fished areas and the stock assessment of eels un<strong>de</strong>r<br />

full exploitation, we estimated that total harvest by fishermen accounted for 14 123 eels (6295<br />

to 24624, i.e. a mean fishing mortality of 23%). Based on the silvering rates calculated in<br />

fished and protected areas, we estimated the silver eel’s production to be 1961<br />

individuals.year -1 (min = 1431, max = 2000), with a mean production of 11.3 and 3.1 silver<br />

eel·ha -1 in the protected and fished areas, respectively. Thus, the mean production of silver<br />

eels in protected area is 3.6 times higher than in the fished area. The protected areas (i.e. 2.4%<br />

of the total aquatic area), potentially produce 8.4% (± 0.43 S.E.) of the silver eel biomass<br />

produced in the study area.<br />

Were the entire GBM affor<strong>de</strong>d protected status, it would produce 6742 silver eels<br />

(596.6 ha·11.3 eel·ha -1 ). Consi<strong>de</strong>ring that the protected area would always have a mean<br />

production of silver eels 3.6 times higher than those in the fished area, we calculated that only<br />

31.1% (185.5 ha) of the aquatic GBM’s area would need protected status to produce 50% of<br />

the potential eel biomass of a fully protected GBM.<br />

171


Table 10.3. Parameters (number, <strong>de</strong>nsity and biomass) of eel stock and silver eel production assessed at each zone of the Gran<strong>de</strong> Brière Mottière marsh. Values<br />

in parenthesis represent the minimum and maximum estimations. See text for <strong>de</strong>tails on the calculation.<br />

Zone Status<br />

Aquatic<br />

habitat (ha)<br />

1 Fished 69.7<br />

2 Fished 160.7<br />

3 Fished 12.9<br />

4 Fished 25.7<br />

5 Fished 112.0<br />

6 Fished 78.3<br />

7 Fished 39.8<br />

8 Fished 82.9<br />

9 Protected 6.5<br />

10 Protected 8.1<br />

Number of<br />

Individuals<br />

11450<br />

(6201; 20205)<br />

30875<br />

(18266; 35271)<br />

2610<br />

(1205; 6125)<br />

14832<br />

(8825; 30113)<br />

24230<br />

(15317; 36474)<br />

23076<br />

(12816; 40970)<br />

9527<br />

(5282; 14489)<br />

9894<br />

(4574; 13112)<br />

614<br />

(419; 703)<br />

1968<br />

(1117; 2744)<br />

Eel population Silver eel production<br />

Density<br />

(ind·ha -1 Biomass<br />

) (kg·ha -1 Number of Density<br />

) individuals (ind·ha -1 Biomass<br />

) (kg·ha -1 )<br />

164.2<br />

11.84<br />

163 2.3 1.34<br />

(88.9; 289.7) (5.58; 23.58) (88; 287) (1.3; 4.1) (0.75; 2.39)<br />

192.1<br />

8.67<br />

438 2.7 1.57<br />

(113.7; 219.5) (4.86; 10.45) (259; 501) (1.6; 3.1) (0.93; 1.81)<br />

202<br />

10.05<br />

37 2.9 1.69<br />

(93.3; 474.1) (4.21; 25.75) (17; 87) (1.3; 6.7) (0.75; 3.91)<br />

577<br />

37.62<br />

211 8.2 4.79<br />

(343.3; 1171.6) (20.44; 83) (125; 428) (4.9; 16.7) (2.86; 9.75)<br />

216.3<br />

10.42<br />

344 3.1 1.81<br />

(136.7; 325.6) (5.96; 17.17) (218; 518) (1.9; 4.6) (1.11; 2.68)<br />

294.7<br />

24.07<br />

328 4.2 2.45<br />

(163.7; 523.2) (12.34; 46) (182; 582) (2.3; 7.4) (1.34; 4.32)<br />

239.5<br />

13.6<br />

135 3.4 1.98<br />

(132.8; 364.2) (7.13; 21.81) (75; 206) (1.9; 5.2) (1.11; 3.03)<br />

119.4<br />

8.43<br />

140 1.7 0.99<br />

(55.2; 158.2) (3.29; 12.89) (65; 186) (0.8; 2.2) (0.46; 1.28)<br />

94.1<br />

5.85<br />

39<br />

6 3.50<br />

(64.2; 107.8) (3.49; 7.55) (27; 45) (4.1; 6.9) (2.39; 4.03)<br />

242.8<br />

23.24<br />

126 15.5 9.05<br />

(137.8; 338.5) (10.94; 37.91) (71; 175) (8.8; 21.6) (5.14; 12.6)


10.4. Discussion<br />

Réserves <strong>de</strong> pêche et anguille<br />

10.4.1. Freshwater protected areas: a compromise between eel global<br />

management and local fishery activities<br />

The analysis of catch data and scientific surveys in the GBM freshwater marshes during 2004<br />

and 2005 clearly provi<strong>de</strong>d evi<strong>de</strong>nce for the efficiency of a protection policy for guaranteeing a<br />

local production of silver eels and maintaining a traditional fishery activity. In<strong>de</strong>ed, the<br />

protected area showed a mean production of silver eel (ind·ha -1 ) around 3.6 times more than<br />

the fished areas and 2.4% (14.6 ha) of protected area in the GBM produces 8.4% of the<br />

current silver eel production (in biomass). Consequently, a size adjustment of the protected<br />

aquatic areas to 31.1% (185.5 ha of aquatic habitat), whilst maintaining the current fishery<br />

elsewhere, would produce 50% of the potential eel biomass of a fully protected GBM. This<br />

could be a realistic of use to local managers. Nevertheless, the optimal size of protected areas<br />

is difficult to estimate because the consequences of the protected areas extension have never<br />

been thoroughly investigated to establish valid rules for the <strong>de</strong>sign of freshwater protected<br />

areas (size, connectivity, location, land covered, canal <strong>de</strong>nsity, etc…). Another crucial point is<br />

that we do not know the proportion of individuals that escape from the silver eel fishery when<br />

they migrate seaward and the level of eel movements between protected and fished areas.<br />

The conservation of freshwater fish and fisheries is the greatest challenge facing the<br />

sustainable <strong>de</strong>velopment of inland waters (Arlinghaus et al. 2002). Inland fisheries are of high<br />

economic and socio-cultural importance, providing a ‘myriad of benefits to society’<br />

(Arlinghaus et al. 2002, Cowx & Ger<strong>de</strong>aux 2004). By far the most dominant traditional inland<br />

fisheries management in Europe are regulations and stocking practices. To a lesser extent,<br />

inland fisheries management use habitat management techniques to restore <strong>de</strong>gra<strong>de</strong>d habitat<br />

and to increase the potential production of the fishery (see review in Arlinghaus et al. 2002).<br />

The use of freshwater protected areas to manage eel populations is in keeping with the latter<br />

aspect. Because the management of the wi<strong>de</strong> panmictic European eel population is<br />

particularly complex, it faces some highly variable socio-economic and legislation constraints.<br />

Therefore, case-adapted management options with respects to usages, properties and histories,<br />

must be consi<strong>de</strong>red to reach a sufficient magnitu<strong>de</strong> of silver eel production. The use of local<br />

freshwater protected areas appears to be a relevant way to conciliate these aspects and to<br />

respond to both global management constraints and local fisheries subsistence.<br />

173


Chapitre 10<br />

10.4.2. Contribution of small coastal marshes to the European eel population<br />

Small coastal marshes contribute to the overall functioning of the European eel population,<br />

but quantification remains impossible. In the present study, we assessed that a single marsh on<br />

the European Atlantic coast (GBM, 7000 ha total area) supports a sub-population of about<br />

130 000 eels and potentially produces about 1961 silver eels per year, almost exclusively<br />

composed of females. Coastal marshes cover 230 000 ha of land on the western French coast<br />

(Feunteun et al. 1992) and, given our findings in the present study, it can easily be assumed<br />

that these ecosystems produce a significant number of female silver eels. Moreover, eels<br />

produced in coastal marshes are exposed to fewer hazards than those in rivers because such<br />

marshes are not equipped with hydroelectric stations, which damage or kill 20 to 100% of the<br />

silver migrants passing through their turbines (Trava<strong>de</strong> & Larinier 1992, McCleave 2001,<br />

Gibson & Myers 2002). In addition, these marshes are small, only connected to the sea, and<br />

they are part of non-intensive agricultural landscapes. This is probably a good point for the<br />

quality of their silver eel production. In places where the eel abundance is about 100-150<br />

kg.ha -1 (Frémur River, France; Acou et al. in press and Rio Esva, Spain; Lobon-Cervia et al.<br />

1995), silver eels are mainly males (>94%). Whereas where eel abundance is relatively low,<br />

3.5 to 45 kg.ha -1 , (Imsa River, Norway; Vøllestad & Jonsson 1988 and Oir River, France in<br />

Acou et al. (in press)), silver eels are mainly females (≥ 80%). Thus, the observed overdominance<br />

of females found in the present site and the low abundance of yellow eels (Table<br />

10.3) are consistent with observations from other areas with low abundance where females are<br />

the numerically dominant sex. As Vøllestad (1990) recommen<strong>de</strong>d retention of a yellow eel<br />

fishery helps maximize silver eel catches. In the present study, it seems likely that, by limiting<br />

the yellow eel abundance in the GBM to less than 50 kg.ha -1 , the traditional fishery might<br />

contribute to the local production of large silver eels. This might also be influenced by a low<br />

recruitment that leads to a low elver eel abundance, and by the high food availability due to<br />

the recent introduction of the invasive red swamp crayfish (Procambarus clarkii).<br />

Such prospects are crucial from a conservation viewpoint, since one of the main<br />

recommendations of international managers (EELREP 2005, ICES WGEEL 2006) is to<br />

protect aquatic systems with a high proportion of large healthy silver eels. In the present study,<br />

we quantified fishing practices and evaluated their influence on the local eel population. Such<br />

issues are important since the i<strong>de</strong>ntification of mortality causes and their quantification are<br />

difficult in the wild but are keys for international eel management (Feunteun et al. 2000,<br />

Feunteun 2002). The presence of protected areas allowed us to assess that the mean natural<br />

174


Réserves <strong>de</strong> pêche et anguille<br />

mortality was relatively low in the marsh (12%) and comparable to those observed in others<br />

ecosystems for the same life stages (Adam 1997, Feunteun et al. 2000). Nevertheless, our<br />

estimation of fishermen's captures based on biological data was somewhat lower than that<br />

resulting from questionnaires, fishermen having difficulties to evaluate accurately their<br />

catches, un<strong>de</strong>rlining the importance of logbooks to conduct fishery surveys.<br />

10.4.3. Establishment of freshwater protected area<br />

Freshwater protected areas have already been shown to be efficient for conserving bird and<br />

fish diversity (Eybert et al. 1998, Keith 2000, Self 2005) and their adaptation for the local eel<br />

population management could be inclu<strong>de</strong>d in the overall management of freshwater<br />

biodiversity (Noble et al. 2004). Recent research in marine protected areas has <strong>de</strong>monstrated<br />

that fish populations benefit from protected areas not just for overexploited poorly mobile<br />

species, but also (to a lesser extent) for un<strong>de</strong>r-exploited stocks and highly mobile species<br />

(Apostolaki et al. 2002). Thus the creation of freshwater protected areas might also be of<br />

benefit to vulnerable or endangered freshwater fish species, such as has already been<br />

advocated for the northern pike, Esox lucius (Rosell & MacOscar 2002).<br />

The efficiency of protected areas in other inland ecosystems remains to be assessed.<br />

Concerning estuarine and coastal waters, Naismith & Knights (1990) indicated that the<br />

commercial fishery in the Thames estuary was having minimal impacts on the eel stock, and<br />

fishing mortality was masked by natural mortality and migration effects. In the same way, in<br />

the Hudson River estuary, Morrison & Secor (2003) suggested that brackish-water areas could<br />

support a higher fishing mortality than freshwater areas. Such analysis confirms that protected<br />

areas for eels might not be relevantly usable in open habitats like estuaries or coastal areas,<br />

and that this management tool could be preferentially applied in more confined freshwater<br />

areas (Morrison & Secor 2003). Furthermore, the restricted yellow eel home range in several<br />

types of freshwater ecosystems (Baras et al. 1998, Jellyman & Sykes 2003, Laffaille et al.<br />

2005a) is certainly a guarantee for a wi<strong>de</strong>r application of this measure.<br />

175


Chapitre 10<br />

Acknowledgments<br />

We thank the Parc naturel régional <strong>de</strong> Brière for logistic support, the Pnrb, FEDER, DIREN<br />

(MEDD), Région Pays <strong>de</strong> la Loire and Agence <strong>de</strong> l’Eau from Loire-Bretagne for financial<br />

support. We are grateful to the numerous persons that participated to field work and to the<br />

questioned fishermen. We acknowledge two anonymous referees for valuable comments on<br />

an earlier draft. English style was post-edited by Dr R. Britton. We thank A. Acou and E.<br />

Rivot for valuable discussion on earlier drafts.<br />

Les résultats obtenus permettent d’attester que la mise en réserve d’une partie <strong>de</strong>s<br />

marais <strong>de</strong> Brière à la suite d’une restauration <strong>de</strong>s habitats est une mesure <strong>de</strong><br />

gestion qui a permis, avec un relatif succès, <strong>de</strong> participer à la production d’un<br />

nombre non négligeable <strong>de</strong> géniteurs, comparativement aux zones soumises à la<br />

pêcherie. Nos résultats ont démontré que cet outil <strong>de</strong> gestion pouvait être encore<br />

plus favorable à l’espèce s’il était optimisé afin <strong>de</strong> répondre aux recommandations<br />

internationales. Des perspectives très appliquées peuvent ainsi découler <strong>de</strong> ce<br />

travail, telles que l’étu<strong>de</strong> du <strong>de</strong>sign <strong>de</strong>s réserves d’eau douce dans <strong>de</strong> nombreux<br />

types d’écosystèmes aquatiques. En effet, l’élargissement <strong>de</strong>s réserves entraînera<br />

<strong>de</strong>s modifications <strong>de</strong> leurs caractéristiques, et plus particulièrement <strong>de</strong> leur niveau<br />

<strong>de</strong> connectivité avec le reste du réseau hydrographique, et donc modifieront<br />

nécessairement certains paramètres populationnels <strong>de</strong> l’anguille (abondance,<br />

compétition ou sex ratio). Quoi qu’il en soit, les effets <strong>de</strong> telles modifications<br />

<strong>de</strong>vront être suivis afin d’optimiser la mise en place <strong>de</strong> ce type <strong>de</strong> gestion. Le<br />

même type <strong>de</strong> démarche pourrait être entreprise pour évaluer l’efficacité <strong>de</strong> la<br />

mise en place <strong>de</strong> réserve pour d’autres espèces <strong>de</strong> la communauté piscicole <strong>de</strong>s<br />

marais <strong>de</strong> Brière soumises à la pêcherie, telles que les espèces <strong>de</strong> poissons<br />

carnassiers et notamment <strong>de</strong> brochet. Cependant, comme nous l’avons évoqué<br />

précé<strong>de</strong>mment dans ce travail <strong>de</strong> thèse, l’état <strong>de</strong> la population <strong>de</strong> brochet <strong>de</strong>s<br />

marais <strong>de</strong> Brière ne permet pas <strong>de</strong> traiter cet aspect particulier. Ainsi, après avoir<br />

examiné différents aspects <strong>de</strong> gestions <strong>de</strong>s habitats et <strong>de</strong> la pêcherie pour le<br />

poisson chat et l’anguille, nous nous proposons, pour clore cette secon<strong>de</strong> partie,<br />

d’examiner la troisième espèce d’intérêt majeur <strong>de</strong> la communauté piscicole, le<br />

brochet, à travers l’examen <strong>de</strong>s programmes <strong>de</strong> soutien <strong>de</strong> population.<br />

176


Chapitre 11<br />

Estimation <strong>de</strong>s taux <strong>de</strong> survie et i<strong>de</strong>ntification <strong>de</strong>s sources<br />

<strong>de</strong> mortalité <strong>de</strong>s brochets (Esox lucius) issus <strong>de</strong><br />

programmes <strong>de</strong> soutien <strong>de</strong> population<br />

La population <strong>de</strong> brochet, espèce vulnérable et emblématique <strong>de</strong>s MTI, <strong>de</strong>s marais<br />

<strong>de</strong> Brière est à un niveau d’abondance très bas. Ainsi, la population est soutenue<br />

<strong>de</strong>puis plusieurs années par <strong>de</strong>s programmes <strong>de</strong> lâchers dans les MTI <strong>de</strong> juvéniles<br />

<strong>de</strong> l’année nés en pisciculture. Aux dires <strong>de</strong>s différents acteurs locaux et <strong>de</strong>s<br />

résultats obtenus dans la Partie I, le niveau d’abondance <strong>de</strong> la population d’adulte<br />

ne semble pas être en augmentation. Dans les <strong>de</strong>ux prochains chapitres, nous nous<br />

proposons d’examiner l’utilisation faite <strong>de</strong>s MTI par ces individus, à travers une<br />

approche individuelle, pour mieux appréhen<strong>de</strong>r les causes possibles <strong>de</strong><br />

dysfonctionnement <strong>de</strong> la population. La question se pose donc <strong>de</strong> savoir ce que<br />

<strong>de</strong>viennent ces poissons, d’estimer le taux <strong>de</strong> mortalité, d’en i<strong>de</strong>ntifier les causes<br />

possibles et <strong>de</strong> comprendre comment ils utilisent les MTI dans lesquels ils sont<br />

lâchés, ceci afin <strong>de</strong> proposer <strong>de</strong>s solutions concrètes aux gestionnaires afin<br />

d’optimiser la réussite <strong>de</strong> ces actions. Connaître le <strong>de</strong>venir <strong>de</strong>s poissons <strong>de</strong> petites<br />

tailles (juvéniles <strong>de</strong> l’année) lâchés en milieu naturel est un véritable défi. En<br />

effet, peu <strong>de</strong> techniques sont adaptées pour y parvenir. Dans cette étu<strong>de</strong>, nous<br />

avons donc transposé au brochet un outil développé initialement pour suivre les<br />

juvéniles <strong>de</strong> salmonidés (Passive Integrated Transpon<strong>de</strong>r (PIT) tag). A notre<br />

connaissance, cette étu<strong>de</strong> est la première où <strong>de</strong>s PIT tags ont été implantés sur <strong>de</strong>s<br />

juvéniles <strong>de</strong> cette espèce et où leur suivi (PIT télémétrie) est effectué en milieu<br />

lentique et ouvert.


Chapitre 11<br />

Utilisation <strong>de</strong> l’antenne mobile utilisée pour le suivi <strong>de</strong>s juvéniles <strong>de</strong> brochets dans une<br />

prairie temporairement inondée<br />

Un juvénile <strong>de</strong> brochet, marqué individuellement, recapturé lors <strong>de</strong> son émigration <strong>de</strong> la<br />

prairie temporairement inondée<br />

178


Suivi du <strong>de</strong>venir <strong>de</strong>s brochetons<br />

Using PIT technology to study the fate of hatchery-reared YOY<br />

northern pike released into shallow vegetated areas 9<br />

Résumé : Cette étu<strong>de</strong> a pour objectif d’analyser le <strong>de</strong>venir <strong>de</strong> juvéniles (0+) <strong>de</strong><br />

brochet nés en pisciculture et relâchés en milieu naturel, en ayant recours au marquage<br />

individuel (Passive Integrated Transpon<strong>de</strong>r). 192 individus (longueur fourche LF 51.0 mm ±<br />

5.3 S.D.) ont été marqués et relâchés avec 72 individus contrôles témoins (non marqués) dans<br />

une prairie temporairement inondés <strong>de</strong>s marais <strong>de</strong> Brière. De leur date <strong>de</strong> lâcher (20 mai 2005)<br />

jusqu’à l’assèchement <strong>de</strong> la prairie (15 juin 2005), la position et le <strong>de</strong>venir <strong>de</strong> chaque individu<br />

ont été étudiés à l’ai<strong>de</strong> d’une antenne portable et d’un système <strong>de</strong> piégeage passif disposé à<br />

l’unique connexion du site avec le milieu adjacent. Seulement 19.3% et 6.9% <strong>de</strong>s individus<br />

marqués et témoins ont été capturés dans le piège et ont donc quitté avec succès la prairie.<br />

Pour ces individus, aucune perte <strong>de</strong> marque n’a été observée et la croissance mesurée est<br />

conforme aux valeurs reportées dans la littérature. L’antenne portable a détecté 75.8% <strong>de</strong>s<br />

individus marqués. Les plus petits individus (LF < 50 mm marquage) ont souffert du taux <strong>de</strong><br />

mortalité le plus élevé, leurs marques ayant été retrouvées dans la prairie après l’assèchement.<br />

Les effets taille dépendant <strong>de</strong> la procédure <strong>de</strong> marquage, du transport et du cannibalisme sur<br />

le taux <strong>de</strong> survie sont donc suspectés. Plus <strong>de</strong> 30% <strong>de</strong>s individus marqués ont disparu du site<br />

d’étu<strong>de</strong> durant le suivi, et nous suspectons fortement un phénomène <strong>de</strong> prédation exercée par<br />

<strong>de</strong>s oiseaux. Ces résultats confirment qu’une attention toute particulière doit être portée lors<br />

<strong>de</strong> programmes <strong>de</strong> soutien <strong>de</strong> population <strong>de</strong> brochet afin <strong>de</strong> limiter ces différentes sources <strong>de</strong><br />

mortalité i<strong>de</strong>ntifiées dans le cadre <strong>de</strong> ce travail.<br />

9 Ce chapitre est adapté d’un article intitulé « Cucherousset, J., Paillisson, J.-M. & Roussel, J.M. Using PIT<br />

technology to study the fate of hatchery-reared YOY northern pike released into shallow vegetated areas » sous<br />

presse dans la revue « Fisheries Research ».<br />

179


Chapitre 11<br />

Abstract: This study aims to use Passive Integrated Transpon<strong>de</strong>r (PIT) technology to<br />

analyse the fate of 192 PIT-tagged hatchery-reared young-of-the-year northern pike Esox<br />

lucius (fork length FL 51.0 mm ± 5.3 S.D.) released into an experimental area located in the<br />

Brière Marsh (France), together with 72 untagged individuals for control. Individual positions<br />

were studied by trapping and using a portable PIT <strong>de</strong>tector from release (20 May) to complete<br />

drying out of the grassland (15 June 2005). Only 19.3% and 6.9% of the PIT-tagged and<br />

control fish, respectively, successfully emigrated from the experimental area before drying out.<br />

Small individuals (FL < 50 mm at tagging) suffered a higher mortality rate, and size<strong>de</strong>pen<strong>de</strong>nt<br />

effects of cannibalism rather than PIT-tagging procedure on survival were<br />

suspected. No evi<strong>de</strong>nce of tag loss was found for surviving fish, and individual growth<br />

conformed to values reported in literature. The portable PIT antenna <strong>de</strong>tected 71.4% of the<br />

PIT-tagged pike in the floo<strong>de</strong>d grassland. More than 30% of tagged fish disappeared from the<br />

experimental area throughout the study period, and predation by birds is suspected. Results<br />

support the view that more attention should be paid to limiting the <strong>de</strong>trimental effects of<br />

mortality on the stocking success of northern pike.<br />

180


11.1. Introduction<br />

Suivi du <strong>de</strong>venir <strong>de</strong>s brochetons<br />

Northern pike (Esox lucius) is a predatory species that tolerates a broad range of<br />

environmental conditions and is particularly adapted to shallow freshwater environments<br />

(Casselman 1996). For this species, temporary floo<strong>de</strong>d and vegetated areas are critical<br />

spawning and nursery habitats, where larvae and juveniles can find protection against<br />

predators and a<strong>de</strong>quate food resources (Souchon 1983, Wright & Shoesmith 1988). After a<br />

short period of <strong>de</strong>velopment, juvenile pike must emigrate from temporary wetlands to<br />

permanent habitats before total <strong>de</strong>siccation (review in Inskip 1982). Juvenile pike survival<br />

and growth patterns have mainly been investigated un<strong>de</strong>r experimental or semi-natural<br />

conditions (e.g. Wright & Shoesmith 1988, Bry et al. 1995, Le Louarn & Cloarec 1997, Skov<br />

et al. 2003, Skov & Koed 2004). More recently, survival of young-of-the-year (YOY) pike<br />

has been studied in the wild using mass marking (Gronkjaer et al. 2004, Sutela et al. 2004) but<br />

mortality causes were not i<strong>de</strong>ntified. Insights into the early life history of northern pike are<br />

still nee<strong>de</strong>d for <strong>de</strong>veloping proper management policies and for improving the effectiveness<br />

of stocking programs.<br />

In the two last <strong>de</strong>ca<strong>de</strong>s, the Passive Integrated Transpon<strong>de</strong>r (PIT) technology has been<br />

<strong>de</strong>veloped as a novel method for individually tagging fish (Prentice et al. 1990) and has<br />

shown its efficiency in studying fish life-history tactics (e.g. Juanes et al. 2000, Cucherousset<br />

et al. 2005a). Stationary PIT antennas have been increasingly used by ichthyologists to study<br />

fish habitat use and migration processes (see review in Lucas & Baras 2000). More recently,<br />

portable PIT <strong>de</strong>tectors have been <strong>de</strong>veloped to track small fish in shallow streams (Roussel et<br />

al. 2000, Barbin Zydlewski et al. 2001, Cucherousset et al. 2005b, Hill et al. 2006). Such<br />

technological <strong>de</strong>velopments have greatly helped to improve our un<strong>de</strong>rstanding of complex<br />

aspects of the early life ecology and behavior of primarily salmonids (e.g. Martin-Smith &<br />

Armstrong 2002, Roussel et al. 2004, Cunjak et al. 2005), but to our knowledge, PIT<br />

technology has never been used for the study of YOY northern pike in the wild. The objective<br />

of this work was to investigate the usefulness of PIT technology for studying YOY northern<br />

pike in a naturally floo<strong>de</strong>d grassland. We particularly focused on PIT-tag retention, post-<br />

tagging survival, and the efficiency of a portable PIT <strong>de</strong>tector for studying the fate of<br />

hatchery-reared YOY pike after their release in a shallow, still water and vegetated habitat.<br />

Such an investigation is crucial since a quantitative assessment of the fate of hatchery YOY<br />

pike released in the wild is still nee<strong>de</strong>d for improving the effectiveness of stocking programs.<br />

181


Chapitre 11<br />

11.2. Materials and methods<br />

11.2.1. Study area<br />

The survey was conducted in May and June 2005 in a protected area located in the heart of<br />

the Brière Marsh (Figure 11.1 and <strong>de</strong>tails in Chapitre 2). A 0.47 ha floo<strong>de</strong>d grassland was<br />

subdivi<strong>de</strong>d with regularly spaced transect lines in or<strong>de</strong>r to <strong>de</strong>lineate 233 rectangular cells (4 m<br />

wi<strong>de</strong> and 5 m long). Based on measurements of habitat variables performed in each cell, the<br />

substrate consisted exclusively of compact peat with no soft bottom, and the mean water<br />

<strong>de</strong>pth was 15.1 cm (± 5.6 S.D., maximum water <strong>de</strong>pth 36 cm). Emergent (Poaceae and<br />

Phragmites australis) and submerged plants (Ranunculus spp. and Callitriche spp.) covered<br />

47.2% (± 28.5 S.D.) of the total area, offering potential spawning substrate for pike (Bry<br />

1996). A fyke net was set up in a V-shape at this connecting point to trap fish leaving the<br />

grassland (Figure 11.1 and <strong>de</strong>tails in Chapitre 2). During the experiment, the water level in the<br />

grassland fell down continuously until drying out completed (late June).<br />

N<br />

50 m<br />

Floo<strong>de</strong>d<br />

grassland<br />

Pond<br />

Figure 11.1. Map of the study area and <strong>de</strong>scription of the trap setting at the single connecting<br />

point between the temporarily-floo<strong>de</strong>d grassland and the pond.<br />

182<br />

B<br />

a<br />

n<br />

k


11.2.2. Fish tagging and stocking<br />

Suivi du <strong>de</strong>venir <strong>de</strong>s brochetons<br />

Hatchery-reared YOY pike fed with zooplankton were used for this experiment. A total of<br />

197 YOY pike were anaesthetized with eugenol (0.04 mL·L –1 ), measured (fork length, ±1 mm)<br />

and weighed (± 0.1 g). The mean fork length and weight were 51.0 mm (± 5.3 S.D.) and 0.86<br />

g (± 0.3 S.D.), respectively. A PIT tag was inserted into the peritoneal cavity using a sterile<br />

needle mounted on an injector, and the left pelvic fin was clipped to assess tag loss. The PIT<br />

tags used were 11.5 mm long and 2.1 mm in diameter (ID 100; EID Aalten B.V., Aalten, The<br />

Netherlands). In addition, a group of 74 control fish (mean fork length = 54.9 mm ± 4.2 S.D.)<br />

were subjected to the same procedure but not PIT-tagged, the right pelvic fin being clipped to<br />

distinguish them from PIT-tagged individuals. Fish were given 6-8 hours to recover before<br />

being transported in plastic bags by boat to the study area. Mortality, cannibalism and tag loss<br />

were checked at the end of the recovery period and before release in the grassland. Tagged<br />

and control fish were released randomly into the experimental area in groups of 13-15<br />

individuals at 9 p.m. on 20 May, and 25 May, 2005, respectively. The stocking <strong>de</strong>nsity (0.06<br />

ind·m -2 ) was similar to the <strong>de</strong>nsity of wild YOY pike usually observed in the Brière marsh<br />

(unpublished data), but below stocking <strong>de</strong>nsities reported in the literature (e.g. 1 to 3.3 ind·m -2<br />

in Lucchietta (1983) and 0.4 to 3.6 ind·m -2 in Sutela et al. (2004)) to prevent exacerbation of<br />

cannibalism (Bry et al. 1992).<br />

11.2.3. Fyke net and tracking procedure<br />

The fyke net was checked once or twice a day (see <strong>de</strong>tails in Chapitre 2); trapped pike were<br />

anaesthetized with eugenol (0.04 mL·L –1 ), measured, and checked for PIT tags and fin clips.<br />

Then, fish were released into the adjacent pond (Figure 11.1). A portable PIT <strong>de</strong>tector was<br />

also used once a day in search for lost PIT tags insi<strong>de</strong> and un<strong>de</strong>r the fyke net; each<br />

transpon<strong>de</strong>r <strong>de</strong>tected was picked up by the operator. The <strong>de</strong>tector is composed of one RFID<br />

rea<strong>de</strong>r (LID 650; EID Aalten B.V., Aalten, The Netherlands) interfaced with a LCD screen<br />

and powered by a 12-V battery. The rea<strong>de</strong>r is connected to a waterproof antenna (ANT 612;<br />

EID Aalten B.V., Aalten, The Netherlands) which is mounted on a 3-m-long aluminium pole.<br />

When a PIT tag is <strong>de</strong>tected by the antenna, the tag co<strong>de</strong> is displayed and a piezoelectric<br />

buzzer sounds a loud tone to alert the operator. The maximum tag <strong>de</strong>tection distances range<br />

from 30 to 36 cm, <strong>de</strong>pending on the orientation of the PIT tag (see <strong>de</strong>tails in Cucherousset et<br />

al., 2005b). From 21 May to 7 June, the portable PIT <strong>de</strong>tector was used every 3-4 days to<br />

183


Chapitre 11<br />

check for the presence and location of PIT-tagged fish in the experimental area. The operator<br />

(J. Cucherousset) manoeuvred the antenna above the water surface by wading between two<br />

transects and the entire wetted area was scanned in 8-10 hours. When a PIT tag was <strong>de</strong>tected,<br />

its exact position was plotted in relation to the rectangular cells <strong>de</strong>lineated by the transect<br />

lines. After 7 June, more than 60% of the initial floo<strong>de</strong>d area had completely dried out and the<br />

scanning operations were stopped. On 15 June, the fyke net dried out; two 40-50 m 2 pools<br />

(mean water <strong>de</strong>pth = 4.97 cm ± 4.1 S.D.) persisted but were disconnected from the pond.<br />

These pools were electrofished, but no fish were caught, indicating that all surviving fish had<br />

left the grassland. After total drying out, additional operations were performed on 19 July to<br />

recover PIT tags remaining in the study site. Each rectangular cell <strong>de</strong>lineated by the transect<br />

lines was carefully scanned, the entire experimental area being scanned 3 times consecutively.<br />

Each PIT tag <strong>de</strong>tected was picked up and removed; its position was recor<strong>de</strong>d for comparisons<br />

with the previous locations. During the third and final scanning, no transpon<strong>de</strong>rs were found.<br />

11.2.4. Data analysis<br />

The efficiency of the portable PIT <strong>de</strong>tector was calculated as the proportion of fish<br />

<strong>de</strong>tected by the operator during the first scanning survey (21 May). Only fish trapped in the<br />

fyke net during the subsequent days were consi<strong>de</strong>red to ensure that the <strong>de</strong>tection efficiency<br />

was calculated on living fish.<br />

Individual growth was estimated based on fish trapped in the fyke net. The expected<br />

weight at recapture was calculated based on a length-weight relationship at tagging:<br />

W = 0.<br />

0397e<br />

0.<br />

0589<br />

FL<br />

where W is the weight in g and FL is the fork length in mm (n = 192, R 2 = 0.88). We<br />

used two indicators of growth performance (Bry et al. 1991): the increment in fork length<br />

(IFL, in mm·day –1 ) and the specific growth rate (SGR, in %·day –1 ) using the formulae:<br />

IFL = ( FLf − FLi)<br />

/ t<br />

SGR = [(ln Wf − lnWi)<br />

/ t]<br />

× 100<br />

where FLi and Wi are the fork length and the weight at tagging, FLf the measured fork<br />

length and Wf the back-calculated weight at recapture and t the time spent in the grassland.<br />

184


Suivi du <strong>de</strong>venir <strong>de</strong>s brochetons<br />

The fate of each individually PIT-tagged YOY pike was inferred from recapture<br />

events at the fyke net and PIT tag locations within the experimental area (scanning surveys<br />

with the portable PIT <strong>de</strong>tector). Each individual was assigned to one of the four following<br />

groups. Fish that were trapped in the fyke net belong to Group I. Group II corresponds to<br />

extru<strong>de</strong>d transpon<strong>de</strong>rs found insi<strong>de</strong> or un<strong>de</strong>r the fyke net, whereas extru<strong>de</strong>d transpon<strong>de</strong>rs<br />

collected in the grassland after complete drying out belong to Group III. For Group III fish,<br />

we assume that <strong>de</strong>ath or tag loss occurred when the transpon<strong>de</strong>r started to be consistently<br />

<strong>de</strong>tected in the same rectangular cell until complete drying out (i.e. no movement). Finally,<br />

Group IV corresponds to PIT tags that were <strong>de</strong>tected at least once during the scanning<br />

operations, but the fish were not recaptured in the fyke net, and the transpon<strong>de</strong>rs were not<br />

found after total drying out. PIT tag disappearance was assumed to occur the day after the<br />

transpon<strong>de</strong>r was <strong>de</strong>tected for the last time in the experimental area.<br />

Recapture rates of PIT tagged and control fish in the fyke net were compared using the<br />

Fisher’s exact test. Since <strong>de</strong>viations from normality were <strong>de</strong>tected in the dataset, differences<br />

in fish fork length at tagging between the groups were tested (P ≤ 0.05) using non-parametric<br />

analysis of variance (Kruskal-Wallis tests).<br />

11.3. Results<br />

11.3.1. Mortality at tagging, tag retention and portable PIT <strong>de</strong>tector efficiency<br />

Among the 197 PIT-tagged fish, 5 individuals died during the recovery period (2.5%). Fish<br />

that died were small (FL range 43-52 mm; mean FL = 46.2 mm ± 3.5 S.D.) compared to the<br />

survivors (FL range 42-65 mm, average FL = 51.1 mm ± 5.2 S.D.). No PIT tag was rejected<br />

and no act of cannibalism occurred before stocking. Similarly, two of the control fish (2.7%)<br />

that died were small (FL range 49 and 50 mm; mean FL = 49.5 mm ± 0.7 S.D.) compared to<br />

the survivors (FL range 47-68 mm, average FL = 55.1 mm ± 4.1 S.D.). The mortality rate of<br />

control fish before stocking was not significantly different from that of PIT-tagged fish<br />

(Fisher’s exact test, P = 0.99). On the morning of 21 May, two PIT-tagged fish; released the<br />

day before, were captured in the fyke net. Afterwards, 35 individuals were recaptured in the<br />

fyke net throughout the survey. Among these 35 individuals, 25 have been <strong>de</strong>tected within the<br />

experimental area by the operator during the first scanning survey; the efficiency of the<br />

portable PIT <strong>de</strong>tector was 71.4%.<br />

185


Chapitre 11<br />

11.3.2. Fate of PIT-tagged fish after release<br />

Of the 192 PIT-tagged YOY pike that were released in the grassland, 19.3% of the tagged fish<br />

were captured in the fyke net (Group I, see Table 11.1), 14.5% of the tags were found insi<strong>de</strong><br />

and un<strong>de</strong>r the fyke net (Group II), 35.4% of the tags were collected on the ground after<br />

complete drying out (Group III), and 30.8% of the tags eventually disappeared from the<br />

experimental area (Group IV). All trapped fish with the left pelvic fin clipped had a<br />

transpon<strong>de</strong>r, indicating 100% tag retention for the surviving fish (Group I).<br />

Table 11.1. The fate after complete drying-out of the PIT tags from the 192 YOY pike released<br />

into the experimental floo<strong>de</strong>d grassland.<br />

Fork length at tagging (in mm)<br />

Group Fate of the PIT tag N Mean (± S.D.) [min; max]<br />

I<br />

PIT-tagged fish trapped in the<br />

fyke net<br />

37 52.16 (± 5.9) [44; 65]<br />

I<br />

PIT tags found insi<strong>de</strong> or un<strong>de</strong>r the<br />

fyke net<br />

28 52.32 (± 5.3) [42; 64]<br />

III<br />

PIT tags recovered in the grassland<br />

after complete drying out<br />

68 59.40 (± 4.2) [43; 63]<br />

IV<br />

PIT tags disappeared from the<br />

grassland during the survey<br />

59 51.83 (± 5.4) [43; 63]<br />

The length frequency distributions of tagged fish (Figure 11.2) indicated that higher<br />

survival rates (14.6-33.3%) were observed for fish larger than 50 mm FL. Conversely, fish<br />

that died in the experimental area (Group III) were significantly smaller at tagging (FL 49.40<br />

mm ± 4.2 S.D.) compared with Groups I, II and IV (Table 11.1; Kruskal-Wallis test,<br />

KW=11.79, P = 0.0081). This trend is confirmed by the length frequency distributions (Figure<br />

11.2), which indicate that 44.3% of fish smaller than 50 mm FL at tagging belong to Group III.<br />

The proportion of fish in Group II (11.8-21.4%) was quite constant whatever fish size at<br />

tagging (Figure 11.2), whereas the proportion of PIT-tags that disappeared from the<br />

experimental area tented to be higher (42.9%) for fish larger than 60 mm FL.<br />

186


Proportion of fish (%)<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

n = 3<br />

n = 2<br />

n = 7<br />

n = 5<br />

n = 15<br />

n = 11<br />

n = 40<br />

n = 23<br />

n = 7<br />

n = 9<br />

n = 15<br />

n = 17<br />

Suivi du <strong>de</strong>venir <strong>de</strong>s brochetons<br />

n = 8<br />

n = 3<br />

n = 5<br />

n = 8<br />

42 - 45 46 - 50 51- 55 56 - 60 61 - 65<br />

Fork length at tagging (mm)<br />

n = 4<br />

n = 3<br />

n = 1<br />

n = 6<br />

Figure 11.2. Proportions and numbers by 5-mm size classes at tagging of the 192 YOY PITtagged<br />

pike belonging to Group I, (, PIT-tagged fish trapped in the fyke net), Group II (, PIT<br />

tags found insi<strong>de</strong> or un<strong>de</strong>r the fyke net), Group III (, PIT tag recovered in the grassland after the<br />

complete drying-out) and Group IV (, PIT tag disappeared from the grassland during the survey)<br />

for each 5 mm fork length at-tagging class.<br />

Only five of the 72 control fish (right pelvic fin clipped) survived and were recaptured<br />

in the fyke net. The recapture rate in the fyke net was significantly lower for control fish<br />

(6.9%) compared with PIT-tagged fish (19.3%, Fisher’s exact test, P = 0.014). The small<br />

number of control fish that were recaptured preclu<strong>de</strong>d any comparison of individual growths<br />

between untagged and PIT-tagged fish. For PIT-tagged fish trapped in the fyke net, the mean<br />

SGR was 11.3%·day -1 (± 1.59 S.D.), ranging from 8.18 to 14.7% day -1 ; the mean IFL was<br />

1.92 mm·day -1 (± 0.27 S.D.) and ranged from 1.39 to 2.5 mm·day -1 . Additionally, no<br />

differences in growth performances (SGR and IFL) were found between fish smaller and fish<br />

larger than 50 mm FL at tagging (Kruskal-Wallis test, KW > 186.5, P > 0.637).<br />

The number of PIT-tagged fish captured in the fyke net (Group I) and the number of<br />

transpon<strong>de</strong>rs collected in and un<strong>de</strong>r the fyke net (Group II) did not show explicit trends<br />

throughout the study period (Figure 11.3). For Group III, 50% of the 68 transpon<strong>de</strong>rs were<br />

consistently <strong>de</strong>tected in the same location from the first scanning survey to their final picking<br />

up (19 July), and were assumed to die or to reject their transpon<strong>de</strong>r the day after release. For<br />

Group IV, 42 of the 59 transpon<strong>de</strong>rs (72%) had disappeared from the experimental area<br />

within 5 days after release (Figure 11.3).<br />

187


Chapitre 11<br />

Number of fish by group<br />

35<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

0<br />

20 May<br />

21 May<br />

25 May<br />

29 May<br />

31 May<br />

Tracking date<br />

Figure 11.3. Number of PIT-tagged YOY pike in the grassland throughout the survey (––) and<br />

number of fish belonging to Group I, (, PIT-tagged fish trapped in the fyke net), Group II (,<br />

PIT tags found insi<strong>de</strong> or un<strong>de</strong>r the fyke net), Group III (, PIT tag recovered in the grassland after<br />

the complete drying-out) and Group IV (, PIT tag disappeared from the grassland during the<br />

survey).<br />

11.4. Discussion<br />

11.4.1. Effect of PIT-tagging<br />

Several studies have been conducted to <strong>de</strong>termine the minimum fish size for PIT-tagging,<br />

results being largely variable <strong>de</strong>pending on species and individuals (e.g. Baras et al., 2000;<br />

Sigourney et al., 2005; Mueller et al., 2006). According to investigations that were mostly<br />

conducted on salmonids (Prentice et al., 1990; Ombredane et al., 1998), tag rejection is<br />

usually low if fish body size is not too small. In the present study, no instantaneous tag<br />

rejection was observed after a 6-hour recovery period and no evi<strong>de</strong>nce of tag loss was found<br />

among individuals recaptured in the fyke net while emigrating from the grassland (i.e. all left<br />

pelvic fin-clipped fish captured in the fyke net had a PIT tag). A formal experiment un<strong>de</strong>r<br />

controlled conditions is now nee<strong>de</strong>d to confirm results on tag retention by young-of-the-year<br />

pike. Un<strong>de</strong>r a certain size threshold, PIT-tagging can have negative short-term effects on<br />

survival and growth (Sigourney et al., 2005; Hill et al., 2006). Most of the transpon<strong>de</strong>rs that<br />

were recovered on the ground after complete drying out (Group III) had been inserted into<br />

fish smaller than 50 mm FL and a size-<strong>de</strong>pen<strong>de</strong>nt effect of PIT-tagging procedure on YOY<br />

188<br />

3 June<br />

7 June<br />

15 June<br />

200<br />

175<br />

150<br />

125<br />

100<br />

75<br />

50<br />

25<br />

0<br />

Number of tagged fish still<br />

present in grassland


Suivi du <strong>de</strong>venir <strong>de</strong>s brochetons<br />

survival can be suspected. Unfortunately, the low number of control fish recaptured (n = 5)<br />

preclu<strong>de</strong>s any unambiguous estimation of <strong>de</strong>trimental effect of tagging on fish. However, it is<br />

noteworthy that the growth of PIT-tagged pike in the present study (IFL of 1.92 mm·day –1 and<br />

SGR of 11.3%·day –1 on average) fell within the range of values reported in the literature, i.e.<br />

IFL ranging from 1.2 to 5.6 mm·day –1 un<strong>de</strong>r natural or semi-natural conditions and SGR of<br />

15.2 %·day –1 ranging from 9.8 to 25.0 %·day –1 un<strong>de</strong>r experimental conditions of nearmaximum<br />

growth (see review in Bry et al., 1991). Moreover, we did not <strong>de</strong>tect any<br />

differences in growth performance according to individual body size at tagging. This indicates<br />

that PIT-tagging did not affect the growth of small fish predominantly, as it would have been<br />

expected in our experiment if the effect of tagging was size-<strong>de</strong>pendant. Cannibalism in YOY<br />

pike is known to be size-selective (Bry et al., 1992) and Skov et al. (2003) <strong>de</strong>monstrated that<br />

it can potentially cause a 60% reduction in stocked pike <strong>de</strong>nsity, at least within the first seven<br />

days after release. Despite the low stocking <strong>de</strong>nsities, it is possible that cannibalism occurred<br />

in our experimental area, and the low survival observed for small PIT-tagged and control pike<br />

might be due to predation by larger individual rather than PIT-tagging procedure. The<br />

contribution of each mortality source (cannibalism vs tagging) would <strong>de</strong>serve for more<br />

observations un<strong>de</strong>r controlled conditions, as well as possible changes in fish behaviour after<br />

tagging (Baras et al., 2000).<br />

11.4.2. The use of PIT technology on YOY pike<br />

This work was a first attempt at using PIT technology to study the fate of YOY pike after<br />

stocking in naturally floo<strong>de</strong>d grassland. While tracking YOY pike with the portable PIT<br />

<strong>de</strong>tector, the mean <strong>de</strong>tection efficiency (71.4%) correspon<strong>de</strong>d to values already obtained in<br />

the literature for salmonid species, i.e. 25-85% (Morhardt et al. 2000, Roussel et al. 2000,<br />

Barbin Zydlewski et al. 2001, Cucherousset et al. 2005b, Hill et al. 2006). The high <strong>de</strong>tection<br />

efficiency is not surprising since water <strong>de</strong>pth in the experimental area never excee<strong>de</strong>d the<br />

maximum reading distance of the portable PIT <strong>de</strong>tector. The fish that were not <strong>de</strong>tected were<br />

likely to escape from the antenna while being moved above the water by the operator. This<br />

result indicates that portable PIT <strong>de</strong>tectors can be used to track small-bodied fish such as<br />

YOY pike in shallow, vegetated and open temporary floo<strong>de</strong>d areas. However, some taxa may<br />

be more likely to escape from the antenna compared to the YOY northern pike, and<br />

preliminary tests on <strong>de</strong>tection efficiency are always recommen<strong>de</strong>d.<br />

189


Chapitre 11<br />

11.4.3. The fate of YOY pike after stocking<br />

Hatchery-reared and PIT-tagged YOY pike had a high mortality rate and only 19.3% of<br />

individuals released were recaptured in the fyke net before total drying out. A low survival<br />

rate for YOY hatchery pike has already been reported un<strong>de</strong>r natural or semi-natural conditions<br />

in previous investigations (5-17% in Lucchietta (1983); 2-12% in Gronkjaer et al. (2004);<br />

34% in Bry et al. (1995)), and survival after stocking is known to <strong>de</strong>pend on age at release<br />

and environmental conditions such as temperature, food availability, time of stocking, and<br />

pike <strong>de</strong>nsity (Bry et al. 1995, Skov et al. 2003, Gronkjaer et al. 2004, Sutela et al. 2004).<br />

Among the mortalities, the occurrence of transpon<strong>de</strong>rs that were collected insi<strong>de</strong> and un<strong>de</strong>r<br />

the fyke net (Group II) is unusual. It is unlikely that the lost transpon<strong>de</strong>rs or <strong>de</strong>ad fish could<br />

have been transported to the net by water currents because of the absence of water movement<br />

at the fyke net and the high vegetation <strong>de</strong>nsity in the grassland. Flooding and drying-out in the<br />

Brière marsh is exclusively due to precipitation and evaporation processes. However, the<br />

abundance of the invasive red swamp crayfish (Procambarus clarkii) in the experimental area<br />

was very high, and a total of 4764 juveniles (40-70 mm long) were caught in the fyke net<br />

throughout the experiment. This species is well known to prey upon small fish (Gutierrez-<br />

Yurrita et al. 1998), and we suspect that predation by crayfish on YOY pike in the fyke net<br />

could have been enhanced by the presence of high crayfish numbers in a small water<br />

volume/area (fyke net), particularly at the beginning of the survey. Predation by crayfish in<br />

the fyke net could also explain the very low recovery rate of control fish. The extent to which<br />

crayfish could have killed YOY pike in the grassland and contributed to high number of<br />

transpon<strong>de</strong>rs collected on the ground after completed drying out (Group III) remains unknown.<br />

More than 30.7% of the fish disappeared after release in the grassland (Group IV) and we<br />

were unable to find the transpon<strong>de</strong>rs in the grassland even after complete drying out. The<br />

water <strong>de</strong>pth in the experimental area was lower than the maximum reading distance of the<br />

portable PIT <strong>de</strong>tector at the beginning of the experiment and continuously <strong>de</strong>creased<br />

throughout the study period. Moreover, due to the compactness of the bottom of the<br />

experimental area, it is unlikely that some transpon<strong>de</strong>rs were acci<strong>de</strong>ntally buried by the<br />

operator stepping on lost PIT tags while scanning for fish. Although the fyke nets blocked the<br />

whole water column, the hypothesis that some fish escaped while the operator was checking<br />

the trapping system (1-2 minutes per day) cannot be totally rejected. However, many bird<br />

species breed in the Brière Marsh, and grey herons Ar<strong>de</strong>a cinerea, sacred ibis Threskiornis<br />

aethiopicus and black terns Chlidonias niger were frequently observed in the study site. These<br />

190


Suivi du <strong>de</strong>venir <strong>de</strong>s brochetons<br />

three species are known to prey upon small fish (Del Hoyo et al. 1992, Beintema 1997), and<br />

were occasionally observed fishing in our experimental area. We therefore suspect bird<br />

predation to be the major cause of PIT tag disappearance and YOY pike mortality in our<br />

experiment. One otter Lutra lutra, was also observed in our experimental area, and might<br />

have preyed upon PIT-tagged pike in the grassland.<br />

To conclu<strong>de</strong>, PIT technology provi<strong>de</strong>s an interesting tool for studying YOY northern<br />

pike during their early life period. We found that 11-mm-long transpon<strong>de</strong>rs can be used to tag<br />

YOY northern pike, but complementary investigations un<strong>de</strong>r experimental conditions are<br />

nee<strong>de</strong>d to test the effects of PIT tagging on individual growth, behaviour and mortality.<br />

Portable PIT <strong>de</strong>tectors can efficiently <strong>de</strong>tect the position of PIT-tagged YOY pike provi<strong>de</strong>d<br />

that water <strong>de</strong>pth does not exceed the maximum reading distance (38 cm). Using such<br />

technology, frequent scanning of the experimental area can be used to monitor survival and<br />

movements of PIT-tagged individuals after release. Finally, the temporal pattern of mortality<br />

observed throughout the experiment suggests the existence of a critical period just after<br />

release, when YOY pike are highly vulnerable to predators. This could be a severe drawback<br />

for YOY pike stocking programme.<br />

Acknowledgments<br />

We thank the Parc naturel régional <strong>de</strong> Brière (PnrB) for logistic support, the PnrB, FEDER,<br />

DIREN (MEDD), Région Pays <strong>de</strong> la Loire and Agence <strong>de</strong> l’Eau from Loire-Bretagne for<br />

financial support. The English style was post-edited by Dr R. Britton. We are particularly<br />

grateful to F. Marchand for YOY pike tagging and to J.-P. Damien, M. Buoro, V. Thoby, D.<br />

Huteau, J. Tremblay, E. Le Mitouard and A. Carpentier for assistance during field work.<br />

191


Chapitre 11<br />

Les conclusions <strong>de</strong> cette étu<strong>de</strong> sur le <strong>de</strong>venir <strong>de</strong>s individus issus <strong>de</strong> programmes<br />

<strong>de</strong> soutien <strong>de</strong> populations montrent que les juvéniles ont un faible taux <strong>de</strong> survie<br />

(19.3%) seulement 4 semaines après leur lâcher, limitant ainsi le succès d’une<br />

telle mesure <strong>de</strong> gestion. Depuis ces dix <strong>de</strong>rnières années, d’importants efforts ont<br />

été entrepris par les gestionnaires <strong>de</strong>s marais <strong>de</strong> Brière pour soutenir cette<br />

population locale. Ainsi, pas moins <strong>de</strong> 4340 juvéniles âgés <strong>de</strong> un mois et 5095<br />

juvéniles âgés <strong>de</strong> un été ont été déversés en dix ans dans les marais <strong>de</strong> Brière.<br />

Malgré cela, il apparaît que le niveau d’abondance en brochet reste faible. Cette<br />

étu<strong>de</strong> a permis d’avancer certains éléments explicatifs quant à ce faible taux <strong>de</strong><br />

survie <strong>de</strong>s poissons issus <strong>de</strong> ces alevinages. Par ailleurs, ce travail a permis <strong>de</strong><br />

démontrer que la technologie <strong>de</strong> marquage individuel par PIT tag et <strong>de</strong> télémétrie<br />

était efficace dans ce type d’étu<strong>de</strong>. Nous avons également montré que l’essentiel<br />

<strong>de</strong> la mortalité intervient pendant la première semaine après le lâcher. Il existe<br />

également une relation taille dépendante du taux <strong>de</strong> mortalité <strong>de</strong>s poissons dont le<br />

PIT tag a été récupéré sur le site à l’étiage. Nous pensons que ces individus sont<br />

morts, soit <strong>de</strong>s effets <strong>de</strong> la procédure <strong>de</strong> marquage (anesthésie, pose <strong>de</strong> marque et<br />

transport), soit <strong>de</strong>s effets du cannibalisme, sans pouvoir quantifier l’importance<br />

relative <strong>de</strong> chacune <strong>de</strong> ces <strong>de</strong>ux causes. Evi<strong>de</strong>mment, il est désormais<br />

indispensable d’examiner, en conditions expérimentales, les conséquences fines<br />

du marquage <strong>de</strong> poissons <strong>de</strong> petite taille, notamment d’un point <strong>de</strong> vue<br />

comportemental. Enfin, les individus qui ont survécu ont montré une forte<br />

variabilité en terme <strong>de</strong> temps <strong>de</strong> rési<strong>de</strong>nce et donc d’utilisation <strong>de</strong> la prairie<br />

temporairement inondée. Nous avons souhaité approfondir cet aspect, en<br />

cherchant à i<strong>de</strong>ntifier les mécanismes impliqués dans cette relation<br />

brochetons/prairie temporairement inondée et plus globalement brochetons/MTI.<br />

192


Chapitre 12<br />

Evolution du régime alimentaire et utilisation <strong>de</strong>s milieux<br />

temporairement inondés par les juvéniles <strong>de</strong> brochet (Esox<br />

lucius)<br />

Comme nous l’avons vu, les juvéniles <strong>de</strong> brochet servant aux programmes <strong>de</strong><br />

soutien <strong>de</strong> population optent pour <strong>de</strong>s stratégies individuelles d’utilisation <strong>de</strong> la<br />

prairie inondée différentes, certains individus quittant très rapi<strong>de</strong>ment la prairie<br />

pour rejoindre le plan d’eau, alors que d’autres, au contraire, optent pour une<br />

utilisation maximale jusqu’à l’exondation quasi complète. Il est possible que <strong>de</strong>s<br />

mécanismes ayant trait aux ressources alimentaires conditionnent cette<br />

exploitation différentielle <strong>de</strong> l’habitat. La variabilité individuelle observée en<br />

terme <strong>de</strong> temps <strong>de</strong> rési<strong>de</strong>nce dans la prairie inondée pourrait traduire <strong>de</strong>s<br />

stratégies alimentaires différentes, compte tenu <strong>de</strong>s importants changements qui<br />

s’opèrent pendant cette phase clé du développement <strong>de</strong>s individus. En effet, au<br />

cours du développement juvénile, le brochet voit son régime alimentaire évoluer<br />

drastiquement. Après avoir utilisé leurs réserves vitellines, les alevins<br />

commencent la prise <strong>de</strong> nourriture exogène par du zooplancton. Ensuite, ce<br />

régime change <strong>de</strong> manière transitoire vers les macro- invertébrés pour finir par du<br />

poisson fourrage. Dans le cas <strong>de</strong> notre étu<strong>de</strong>, après leur émigration <strong>de</strong> la prairie,<br />

les individus passent dans le plan d’eau adjacent, utilisé alors comme nourricerie<br />

et doivent ensuite gagner le canal permanent (refuge estival) avant l’étiage. Ici,<br />

nous nous proposons d’examiner les relations entre changements d’habitats et<br />

évolution du régime alimentaire au cours du développement juvénile à l’ai<strong>de</strong> <strong>de</strong><br />

l’analyse <strong>de</strong>s isotopes stables. De plus, ces processus vont aussi être analysés pour<br />

certains individus nés naturellement dans le site d’étu<strong>de</strong>, afin <strong>de</strong> voir s’il existe<br />

une différence avec les individus nés en pisciculture.


Chapitre 12<br />

Prélèvement d’un échantillon <strong>de</strong> nageoire pelvienne lors du marquage individuel<br />

<strong>de</strong>s juvéniles <strong>de</strong> brochet<br />

Prélèvement <strong>de</strong>s proies potentielles <strong>de</strong> juvéniles <strong>de</strong> brochet dans les marais <strong>de</strong> Brière en vue<br />

<strong>de</strong> l’analyse <strong>de</strong>s isotopes stables<br />

194


Régime alimentaire et habitat <strong>de</strong>s brochetons<br />

Individual variation in trophic position and habitat use of juvenile<br />

northern pike (Esox lucius) in temporary floo<strong>de</strong>d habitats 10<br />

Résumé: Des PIT tags et l’analyse <strong>de</strong>s isotopes stables (AIS) ont été utilisé<br />

conjointement pour étudier les changements <strong>de</strong> régimes alimentaires et d’utilisation d’habitats<br />

<strong>de</strong>s juvéniles <strong>de</strong> brochet (Esox lucius) dans <strong>de</strong>s milieux temporairement inondés soumis à un<br />

assèchement progressif. 297 juvéniles <strong>de</strong> brochet (longueur fourche 51.0 ± 5.3 mm),<br />

préalablement élevés en pisciculture et nourris à l’ai<strong>de</strong> <strong>de</strong> zooplancton, ont été marqués<br />

individuellement puis relâchés dans une prairie temporairement inondée (FG, n = 192) et dans<br />

un plan d’eau temporaire adjacent (TP, n = 105). FG s’est asséchée progressivement au<br />

printemps et l’unique connexion entre FG et TP a été équipée d’un verveux afin <strong>de</strong> capturer<br />

les individus quittant FG. De la même manière, TP s’est asséché en été et les poissons ont été<br />

capturés lorsqu’ils rejoignaient le canal permanent adjacent. A chaque fois qu’un brochet a été<br />

manipulé (i.e. marquage et recaptures sur le terrain), un prélèvement <strong>de</strong> nageoire a été réalisé<br />

et <strong>de</strong>s AIS (carbone et azote) ont été réalisées sur ces échantillons ainsi que sur quelques<br />

individus nés en milieu naturel et sur l’ensemble <strong>de</strong>s proies potentielles (zooplancton,<br />

invertébrés, juvéniles d’écrevisse et poissons fourrages). Le taux <strong>de</strong> survie durant l’étu<strong>de</strong> était<br />

19.3% dans FG et 1.9% dans TP, respectivement. Les valeurs <strong>de</strong>s SIA dans la pisciculture ont<br />

révélées l’existence <strong>de</strong> <strong>de</strong>ux groupes distincts d’individus : les zooplanctonivores et les<br />

piscivores (i.e. cannibales) (facteur <strong>de</strong> fractionnement +2.01‰ δ 15 N et +0.82‰ δ 13 C). Nous<br />

avons trouvé que la longueur initiale <strong>de</strong>s poissons était positivement corrélée aux valeurs <strong>de</strong><br />

δ 13 C et <strong>de</strong> δ 15 N. Après leur lâcher, le temps passé par les brochets dans FG était négativement<br />

corrélé aux valeurs <strong>de</strong> δ 13 C et <strong>de</strong> δ 15 N et à la taille <strong>de</strong>s poissons, indiquant que les individus<br />

émigrant précocement étaient les piscivores. La dilution <strong>de</strong>s signatures isotopiques <strong>de</strong> la<br />

pisciculture laisse penser que les brochetons se sont nourris <strong>de</strong> consommateurs primaires dans<br />

FG (i.e. zooplancton et macro invertébrés). Au contraire, les individus se sont enrichis en<br />

azote 15 N (+2.83‰) après leur passage dans TP où ils se sont nourris <strong>de</strong> poissons fourrages.<br />

Des tendances similaires ont été observées pour les individus sauvages. Dans l’ensemble, ces<br />

résultats indiquent que les changements <strong>de</strong> régimes alimentaires et d’habitats se déroulant lors<br />

du développement juvénile <strong>de</strong>s brochets sont fortement associés, et semblent être dépendant<br />

<strong>de</strong> la taille <strong>de</strong>s individus.<br />

10 Ce chapitre est adapté d’un article en préparation écrit par Cucherousset, J., Paillisson, J.-M. & Roussel, J.M<br />

195


Chapitre 12<br />

Abstract: Passive Integrated Transpon<strong>de</strong>r (PIT) techniques and Stable Isotope<br />

Analysis (SIA) were used in conjunction to study early shifts in trophic position and habitat<br />

use of young-of-the-year (YOY) northern pike (Esox lucius) in two contrasted areas of the<br />

Brière marsh (Northwest France) exposed to a progressive drying-out. In total, 297 juvenile<br />

pike (fork length 51.0 mm ± 5.3 S.D.) reared on zooplankton in hatchery tanks were PIT-<br />

tagged and released in a floo<strong>de</strong>d grassland (FG, n = 192) and in the adjacent temporary pond<br />

(TP, n = 105). FG and TP progressively dried out and the only connections with their adjacent<br />

habitat were equipped with fyke nets. Any time individuals were handled (i.e. at tagging and<br />

subsequent recaptures), they were fin clipped. Nitrogen and carbon SIA were run on fin tissue<br />

collection and on potential preys (zooplankton, invertebrate and forage fish taxa). Survival<br />

rate of PIT-tagged fish through the study was 19.3% and 1.9% in FG and TP, respectively.<br />

SIA values at hatchery revealed the existence of two distinct groups i<strong>de</strong>ntified as<br />

zooplanktivorous and piscivorous individuals (fractionation factor +2.01‰ δ 15 N and +0.82‰<br />

δ 13 C). Initial fish length at tagging was positively correlated with δ 13 C and δ 15 N. After release,<br />

time spent in FG was negatively correlated with δ 13 C and δ 15 N and fish size, indicating that<br />

early migrants were likely piscivores. Repeated SIA on recaptured pike clearly suggested a<br />

dilution of initial hatchery signatures and that juvenile pike preyed on primary consumers in<br />

FG (zooplankton and macro invertebrates). Conversely, individuals were significantly 15 Nenriched<br />

(+2.83‰) after their stay in TP where they likely preyed on fish. A similar pattern<br />

was observed for wild juvenile pike. Taken together, the results indicate that shifts in habitat<br />

use by juvenile pike and their trophic position were closely associated, and appeared to be<br />

size <strong>de</strong>pen<strong>de</strong>nt.<br />

196


12.1. Introduction<br />

Régime alimentaire et habitat <strong>de</strong>s brochetons<br />

The diet and habitat use of wild animals are crucial to un<strong>de</strong>rstanding their environmental<br />

biology and as such are prerequisites for effective conservation strategies (Rubenstein &<br />

Hobson 2004). This is particularly important in harsh and spatially-varying environments<br />

(Chesson & Huntly 1997, Chesson 2000) where movements and habitat selection by<br />

individuals directly influence their ability to survive and maximise their fitness. In many<br />

aquatic systems, drought and flood episo<strong>de</strong>s are natural features that dramatically change the<br />

temporal and spatial characteristics of the habitat (Matthews 1998, Humphries & Baldwin<br />

2003, Lytle & Poff 2004) and animals such as fish may display a high level of flexibility in<br />

their response to these varying environments (Lytle & Poff 2004). Inter specific patterns in<br />

the use of temporary waters by fish have already been <strong>de</strong>scribed in the literature (Poizat &<br />

Crivelli 1997, King et al. 2003, Williams 2006, Chapitre 6). Although the existence of life<br />

history variants is central for evolutionary and conservation prospects (see review in Bolnick<br />

et al. 2003), life history variations among individuals of the same population remain poorly<br />

investigated (see for instance Post 2003). Particularly in the case of temporary floo<strong>de</strong>d<br />

habitats, part of literature scarceness on the existence of life history variants within<br />

populations may be that field investigations are technically challenging, especially for smallbodied<br />

fish.<br />

The use of Stable Isotope Analysis (SIA) to trace animal migration and foraging<br />

behaviour is increasingly well documented in literature (e.g. Rubenstein & Hobson 2004,<br />

Dalerum & Angerbjorn 2005, West et al. 2006). The method is based on the existence of<br />

differences between stable isotopic ratios of different habitats and the predictable relationship<br />

between the isotopic composition of animal and their diet (DeNiro & Epstein 1978, Peterson<br />

& Fry 1987, Post 2002, Jardine et al. 2005). However, much of research on animal<br />

movements involving SIA relates to groups of individuals, i.e. with no information on<br />

individually i<strong>de</strong>ntified organisms that could help un<strong>de</strong>rstand the observed behaviour<br />

(Matthews & Mazum<strong>de</strong>r 2004, Cunjak et al. 2005). Consequently, some recent research<br />

programs have been un<strong>de</strong>rtaken to couple SIA information with individual movements data,<br />

for instance by means of satellite telemetry for large-sized organisms (e.g. Hatase et al. 2002,<br />

Webster et al. 2002), or PIT telemetry for small-bodied animals or early life stage (Cunjak et<br />

al. 2005). When animals are individually tagged and successfully recaptured, the potential<br />

advantages of performing repeated SIA on the same individuals remain unexplored, although<br />

197


Chapitre 12<br />

it may provi<strong>de</strong> a mechanistic approach to study behavioural responses to fluctuating<br />

environments at the individual level and the role of life history variants at the population level.<br />

Northern pike (Esox lucius) is a predatory species that is particularly adapted to<br />

shallow freshwater environments (Casselman 1996). Temporary floo<strong>de</strong>d habitats, i.e. aquatic<br />

habitats that cyclically dry out (Williams 2006), are critical spawning and nursery areas of<br />

this species where larvae and juveniles can find protection against predators and large food<br />

supplies (Souchon 1983, Wright & Shoesmith 1988). After the embryonic and larval<br />

<strong>de</strong>velopment stages, juveniles have to reach permanent habitats as the water level<br />

progressively <strong>de</strong>clines (Inskip 1982) and shifts in trophic position occur. Zooplankton is the<br />

first hexogen food source for larvae and juveniles, then juveniles prey upon micro- to macroinvertebrates<br />

and finally become increasingly piscivorous as they grow, these shifts being<br />

variable among individuals (Bry et al. 1995, Elvira et al. 1996, Le Louarn & Cloarec 1997,<br />

Beaudoin et al. 1999). Growth rate during the early juvenile period reaches, on average, 10<br />

mm fork length a week (Casselman 1996). Previous studies have <strong>de</strong>monstrated that several<br />

environmental variables (water and temperature regimes, habitat configuration, food<br />

availability and <strong>de</strong>nsity of conspecific individuals) influence a set of biological <strong>de</strong>scriptors<br />

(survival rate, emigration pattern, growth performance or prey consumption) of the species<br />

(Mann 1976, Inskip 1982, Mann 1982, Bry et al. 1991, Treasurer & Owen 1991, Bry et al.<br />

1995, Casselman 1996, Beaudoin et al. 1999). However, time schedule in habitat and diet<br />

shifts by juveniles pike in the wild and its variability among individuals need to be<br />

documented.<br />

The objective of the present study was to investigate the relationships between<br />

foraging and habitat selection using repeated SIA on individually PIT-tagged young-of-theyear<br />

(YOY) northern pike in temporary floo<strong>de</strong>d habitats. More precisely, we aim to measure<br />

whether habitat use is variable among individuals. If so, then we would expect that it results<br />

from different exploitation in time and space of the food resources available in the different<br />

habitats. Because marked trophic shifts, variable among individuals, occurred during juvenile<br />

<strong>de</strong>velopments, we would expect that individual characteristics, notably their size, could<br />

influence the relationships between foraging and habitat selection. Finally, we investigate the<br />

influence of individual components on growth performance.<br />

198


12.2. Materiel and methods<br />

12.2.1. Study site<br />

Régime alimentaire et habitat <strong>de</strong>s brochetons<br />

The study was carried out from May to August, 2005 in a site located in the heart of the marsh<br />

(see <strong>de</strong>tails in Chapitre 2). This site inclu<strong>de</strong>d a temporary floo<strong>de</strong>d grassland (hereafter called<br />

FG) that covered 0.47 ha at the start of the study that provi<strong>de</strong>s suitable conditions to adult<br />

pike for spawning. The site was also composed of a 2.13 ha temporary pond (TP) used as<br />

nursery and secondary spawning areas, and a permanent canal (PC) offering refuge habitats<br />

during drought (Figure 12.1).<br />

12.2.2. Fish tagging and releasing<br />

Hatchery-reared YOY pike from wild brood stock, solely fed on zooplankton until 3-4 weeks<br />

old, were used. On 20 May 2005, 297 individuals were anaesthetized with eugenol (0.04<br />

mL·L –1 ) and measured for fork length (FL ±1 mm) and weight (± 0.1 g). The mean FL and<br />

weight were 51.3 mm (± 5.5 S.D.) and 0.87 g (± 0.32 S.D.), respectively. A PIT tag was<br />

inserted into the peritoneal cavity using a sterile needle mounted on an injector, and the left<br />

pelvic fin was partly clipped to assess tag loss. The PIT tags used were 11.5 mm long and 2.1<br />

mm in diameter (ID 100; EID Aalten B.V., Aalten, The Netherlands). Fish were given 6-8<br />

hours to recover before being transported in plastic bags by boat to the study area. YOY PITtagged<br />

pike were released into FG (n = 192 individuals) and TP (n = 105) in groups of 13-15<br />

individuals. The two plots of fish had similar size (Mann-Whitney test, U = 10.319, P = 0.734,<br />

n = 297) and weight (U = 10.348, P = 0.702).<br />

12.2.3. Fish displacements and environmental variables<br />

Fish movements between FG and TP and between TP and PC were monitored using pairs of<br />

fyke nets (Figure 12.1 and <strong>de</strong>tails in Chapitre 2). The fyke nets were checked at least once a<br />

day throughout the entire study period. As per French regulations for ethical treatment of wild<br />

animals, trapped YOY pike were anaesthetized in a solution of eugenol (0.04 mg.L -1 ).<br />

Afterwards, they were checked for PIT tags and fin clips, measured for fork length (FL ±1<br />

mm) and weight (± 0.1 g).<br />

199


Chapitre 12<br />

Water level (cm)<br />

120<br />

110<br />

100<br />

90<br />

80<br />

70<br />

60<br />

50<br />

N<br />

50 m<br />

Floo<strong>de</strong>d grassland<br />

(FG)<br />

FL<br />

Temporary pond<br />

(TP)<br />

Permanent canal<br />

(PC)<br />

Fyke net<br />

May, 1 May, 16 May, 31 June, 15 June, 30 July, 15 July, 30 Aug., 14 Aug., 29<br />

Date<br />

Figure 12.1. Localization of the study area composed of a floo<strong>de</strong>d grassland (FG), a temporary<br />

pond (TP) and a permanent canal (PC). Fyke nets were set up at the connection points between<br />

habitats (top). Water level (French General Level, cm, bold line), and water temperature (°C, fine<br />

line) recor<strong>de</strong>d in the permanent canal from May to August 2006 (bottom). The dotted lines<br />

represent the water level nee<strong>de</strong>d to inundate the two temporary floo<strong>de</strong>d habitats.<br />

200<br />

TP<br />

40<br />

35<br />

30<br />

25<br />

20<br />

15<br />

10<br />

Water temperature (°C)


Régime alimentaire et habitat <strong>de</strong>s brochetons<br />

The left pelvic fins were clipped again (see below), and fish were released on the other si<strong>de</strong> of<br />

the fyke net in the direction of movement. Trapping was conducted until the total drying out<br />

of each site, i.e. until 16 June in FG and until 16 August in TP (Figure 12.1). On 16 June, two<br />

isolated pools (mean water <strong>de</strong>pth = 4.97 cm ± 4.1 S.D.) persisted in FG, and those were<br />

electrofished to capture stran<strong>de</strong>d pike, but none were found. The same procedure was also<br />

applied in TP on 16 August: no PIT-tagged pike but four YOY wild pike were captured. For<br />

the fish that survived throughout the survey, individual growth performances were estimated<br />

by calculating FL increment (IFL, in mm·day -1 ) and the Fulton’s condition factor (K); with<br />

the following formulas:<br />

IFL = (FLr – Fli)/t and K = Wr·FLr -3 ·100000<br />

where FLi is FL at tagging, FLr and Wr are FL and weight at recapture, and t is the time spent<br />

in the study area.<br />

To <strong>de</strong>scribe the varying environmental conditions that pike experienced throughout the<br />

progressive drying out, a set of environmental variables were monitored. Water <strong>de</strong>pth and<br />

percentage of vegetation cover (mainly emergent Poaceae and Phragmites australis, and<br />

submerged Ranunculus spp. and Callitriche spp.) were assessed at the start of the survey<br />

based on measurements along transects evenly distributed in FG and TP (631 measurement<br />

points in total). Daily water <strong>de</strong>pth in FG and TP was back-calculated using data of daily<br />

fluctuations of water level measured in the Brière marsh (data courtesy of the Parc naturel<br />

régional <strong>de</strong> Brière). Water temperature was monitored continuously (4 records per hour) in<br />

FG and TP using two automatic sensors (Sensor StowAway TidBit, Onset Computer<br />

Corporation) set up in the middle of each habitat.<br />

The food available to YOY pike was estimated in the two habitats. First, other fish<br />

species captured in the fyke nets (young-of-the-year Cyprinidae and mosquitofish Gambusia<br />

holbrooki) were counted to assess the abundance of forage fish (individuals.m -2 ) in FG and<br />

TP. Second, zooplankton (mainly Cladocera and Copepoda) abundance was estimated by<br />

filtering (0.1-mm mesh) 15 L of water sampled in the water column. This procedure was<br />

repeated weekly in four different locations in FG and TP until the total drying out. Samples<br />

were preserved in 70% ethanol and zooplankton biomass was estimated in the laboratory by<br />

counting the total number of items using a Dollfus counting chamber and a binocular<br />

microscope. A fraction of the zooplankton samples (100 items) was dried at 60°C for 48 h and<br />

weighted to express zooplankton abundance in biomass (mg.L -1 ).<br />

201


Chapitre 12<br />

12.2.4. Sample collection and stable isotope analysis (SIA)<br />

The left pelvic fin of each individual was partly clipped at tagging in or<strong>de</strong>r to get a small<br />

amount of fresh tissue, corresponding to approximately 200 μg of dry tissue. Because body<br />

growth is very high in YOY pike, the left pelvic fin re<strong>de</strong>veloped very rapidly and it was<br />

possible to clip it again at each recapture throughout the experiment (time interval up to five<br />

days). SIA was performed on pike that survived and were recaptured at least once throughout<br />

the study, i.e. 35 individuals in total. The different causes of mortality of YOY pike, among<br />

which the avian predation is a major cause, are thoroughly discussed elsewhere (Chapitre 11).<br />

A sample of zooplankton was collected at the hatchery; it represented the unique food<br />

resource used to rear YOY before PIT-tagging. Potential preys were sampled in FG and TP on<br />

13 May, just before the release of PIT-tagged pike: zooplankton (mainly Cladocera and<br />

Copepoda) with a net (0.1-mm mesh), macro invertebrates (amphipoda: Gammarus sp. and<br />

juvenile red swamp crayfish Procambarus clarkii) with a net (2-mm mesh) and available<br />

forage fish (mosquitofish) by electrofishing. YOY cyprinids hatched later in the season (June-<br />

July), and complementary samples were collected in the fyke nets when YOY common carp<br />

(Cyprinus carpio), rudd (Scardinius erythrophtalmus) and bream (Blicca bjoerkna) emigrated<br />

from the two habitats.<br />

Zooplankton and amphipoda were kept alive during 24h for gut clearance whereas<br />

digestive tracts were discar<strong>de</strong>d from crayfish and prey fish before freezing. Zooplankton,<br />

invertebrates and prey fish samples consisted of several individuals from each taxon or<br />

species in each site. All samples (zooplankton, amphipoda, crayfish, forage fish taxa, and pike<br />

fin clips) were oven dried (60°C for 48h) and ground to a homogeneous pow<strong>de</strong>r using a mixer<br />

mill (Retsch MM 200). Approximately 0.2 mg of sample were weighed (Sartorius M2P) into<br />

a tin cup and combusted in a Carlo Erba NC2500 elemental analyzer at high temperature for<br />

C and N isotope analysis. Resultant gases were <strong>de</strong>livered via continuous flow to a Finnigan<br />

Mat Delta XP isotope ratio mass spectrometer (Thermo Finnigan, Bremen, Germany). Isotope<br />

ratios are reported in <strong>de</strong>lta notation relative to international standards: Pee<strong>de</strong>e Belemnite<br />

Carbonate (C) and atmospheric nitrogen (N). Data were corrected using working standards<br />

(bass muscle, bovine liver, nicotinami<strong>de</strong>) that were previously calibrated against International<br />

Atomic Energy Agency (IAEA) standards.<br />

202


12.2.5. Statistical analysis<br />

Régime alimentaire et habitat <strong>de</strong>s brochetons<br />

Analysis of the environmental variables dataset revealed <strong>de</strong>viations from normality; therefore<br />

the Mann-Whitney U test was used to test differences in water level and vegetation cover<br />

between habitats (FG and TP). Wilcoxon signed rank test was used to test for differences in<br />

zooplankton abundance between habitats at each common sampling date (n = 10). The<br />

relationships between fish size at tagging, the time spent in each site and stable isotope values<br />

(δ 15 N and δ 13 C) were analysed using linear regression analyses. After the i<strong>de</strong>ntification of<br />

pike diet types in the hatchery, analyses of variance (ANOVA) and non-parametric equivalent<br />

were performed as appropriate to test for differences in FL at tagging, in time spent in FG and<br />

TP, and in growth performances between zooplanktivorous and piscivorous pike. Values are<br />

mean ± standard <strong>de</strong>viation (S.D.). A rejection level of 0.05 was used in all tests.<br />

12.3. Results<br />

12.3.1. Environmental variables: site characteristics and temporal changes<br />

On 15 May, water level was significantly higher (Mann-Whitney U test, P < 0.001, n = 631)<br />

in TP (63.9 cm ± 16.6 S.D) than in FG (15.3 cm ± 5.6) and TP was significantly (P < 0.001)<br />

less vegetated (6.5% ± 22.1) than FG (50.3% ± 65.8). Water level continuously <strong>de</strong>creased as<br />

the season progressed and drying out was completed in FG on 16 June, whereas TP<br />

disconnected with the permanent canal on 16 August, before complete drying out. In the<br />

meantime, mean water temperature increased from 17.3°C (± 2.1) to 19.6°C (± 1.3) in FG and<br />

from 17.0°C (± 0.9) to 20.4°C (± 4.4) in TP. The abundance of zooplankton was relatively<br />

constant from 15 May to 16 June in the two habitats, and significantly higher (Wilcoxon<br />

signed rank test, Z = -1.988, p = 0.047, n = 10) in FG (185.5 mg.L -1 ± 94.6) than in TP (104.0<br />

mg.L -1 ± 77.9). Afterwards, it strongly increased in the TP (up to 1946.9 mg.L -1 ± 481.1)<br />

when FG dried out. In total, 444 and 4338 forage fish emigrated respectively from FG (G.<br />

holbrooki and C. carpio) and TP (C. carpio, B. bjoerkna, S. erythrophthalmus and Rutilus<br />

rutilus). Abundance of forage fish (all species pooled) available to YOY pike was twofold in<br />

TP (0.20 ind.m -2 ) compared with FG (0.09 ind.m -2 ).<br />

203


Chapitre 12<br />

12.3.2. YOY pike displacements<br />

Of the 192 YOY PIT-tagged pike initially released in FG, 37 were captured in the fyke net<br />

from 21 May to 12 June while moving to TP (survival rate of 19.3%). Among them, eight<br />

individuals emigrated within five days post release, and were not fin-clipped for SIA. In TP,<br />

seven PIT-tagged pike were caught in the fyke net from 19 July to 29 July among which only<br />

two individuals (of 105) were fish released in TP at the beginning of the experiment (survival<br />

rate of 1.9%). Few wild YOY were captured in the fyke net in FG (n=2) and in TP (n=7). PITtagged<br />

and wild YOY pike emigrated from the two habitats at similar time, but wild<br />

individuals ten<strong>de</strong>d to be bigger than PIT-tagged individuals (Table 12.1). After release, the<br />

smaller YOY PIT-tagged pike tented to stay for a longer period in FG than larger individuals<br />

(Figure 12.2, linear regression, r 2 = 0.68, P < 0.001, n = 29). No significant relationship was<br />

found between body size and time spent in TP (r 2 = 0.32, P = 0.183, n = 7).<br />

Time spent in the temporary floo<strong>de</strong>d habitat (in days)<br />

80<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

Temporary pond<br />

Floo<strong>de</strong>d grassland<br />

0<br />

44 48 52 56 60 64<br />

Fish size at tagging (in mm)<br />

Figure 12.2. Time spent (in days) by young-of-the-year (YOY) PIT tagged pike in the temporary<br />

pond (black diamonds) and the floo<strong>de</strong>d grassland (white diamonds) as a function of their body size<br />

at tagging (fork length in mm). Linear regressions are displayed (solid lines).<br />

204


Table 12.1. Summary statistics (n: sample number; SD: standard <strong>de</strong>viation; Min: minimum and Max: maximum) of fork length (FL) and C and N stable<br />

isotopes (δ 13 C and δ 15 N, ‰) in fin tissue sampled from wild and PIT-tagged young-of-the-year pike captured in the hatchery, the floo<strong>de</strong>d grassland (FG) and<br />

the temporary pond (TP) different habitats throughout the survey.<br />

Fork length (FL, mm) δ 13 C (‰) δ 15 N (‰)<br />

Type Location Date n Mean<br />

Mean (S.D.) Min Max<br />

(S.D.)<br />

Min Max Mean<br />

(S.D.)<br />

Min Max<br />

PIT- Hatchery 20 May 35 51.7 (5.6) 44 64 -27.2 (0.5) -28.2 -26.4 10.6 (1.1) 8.7 12.4<br />

tagged FG 26 May – 11 June 29 79.5 (6.8) 68 94 -29.6 (0.7) -31.5 -28.0 9.2 (0.7) 7.9 10.3<br />

TP<br />

(electrofishing)<br />

30 June 2 128.5 (2.1) 127 130 -27.8 (0.1) -27.8 -<br />

.27.7<br />

11.4 (0.1) 11.3 11.4<br />

TP<br />

(fyke net)<br />

19 July – 27 July 7 183.9 (11.2) 167 196 -27.5 (0.2) -27.8 -27.2 12.2 (0.3) 11.9 12.6<br />

Wild FG 18 May - 26 May 2 101.0 (4.2) 98 104 -28.5 (1.4) -29.6 -27.5 9.5 (0.1) 9.4 9.5<br />

TP 22 July – 16 Aug. 7 254.1 (20.0) 227 277 -26.9 (0.7) -27.6 -26.0 11.7 (0.5) 11.0 12.2


Chapitre 12<br />

12.3.3. Stable Isotope Analyses on pike before release<br />

Zooplankton (δ 13 C = -28.66‰ and δ 15 N = 7.38‰, Table 12.2) was the unique source of food<br />

<strong>de</strong>livered to pike at the hatchery. Frequency distribution of the number of pike according to<br />

δ 15 N and δ 13 C values and the combined δ 15 N-δ 13 C scatter plot showed that YOY pike at the<br />

hatchery are split into two distinct groups (Figure 12.3). The first group (δ 13 C = -27.73‰ ±<br />

0.21 and δ 15 N = 9.21‰ ± 0.36) was i<strong>de</strong>ntified as zooplanktivorous YOY. The second group<br />

(δ 13 C = -26.89‰ ± 0.26 and δ 15 N = 11.24‰ ± 0.58) was i<strong>de</strong>ntified as piscivorous (i.e.<br />

cannibalistic) YOY. On average, trophic enrichment was +0.82‰ δ 13 C (± 0.13) and +2.01‰<br />

δ 15 N (± 0.10) at the hatchery. Zooplanktivores (mean size = 46.5 mm ± 1.9) were significantly<br />

smaller (ANOVA, F = 22.35, P < 0.001, n = 35) than piscivores (mean FL = 54.0 mm ± 5.1).<br />

The size of fish at tagging was positively correlated with δ 15 N (linear regression, r 2 = 0.69, P<br />

< 0.001, n = 29) and δ 13 C values (r² = 0.65, P < 0.001) in fin tissue. Moreover, the time spent<br />

in the grassland (in days) was negatively correlated with δ 15 N (r 2 = 0.47, P < 0.001) and δ 13 C<br />

(R 2 = 0.47, P < 0.001). The mean time spent by pike in FG was significantly longer (ANOVA,<br />

F = 10.53, P = 0.003, n = 29) for zooplanktivores (18.1 days ± 2.3) than for piscivores (12.5<br />

days ± 5.2). Regarding the time spent in TP, no significant differences were found between<br />

zooplanktivores and piscivores (Mann-Whitney U test, P = 0.845, n = 7).<br />

Table 12.2. Summary values of the C- and N-isotopic compositions (δ 13 C and δ 15 N, ‰) of the<br />

food resources for YOY pike in the hatchery, the floo<strong>de</strong>d grassland (FG) and the temporary pond<br />

(TP).<br />

Site Organisms Genus or species n Date δ 13 C (‰) δ 15 N (‰)<br />

Hatchery Zooplankton Cladocera & Copepoda ≈ 50 May, 20 -28.7 7.4<br />

FG<br />

TP<br />

Zooplankton Cladocera & Copepoda ≈ 50 May, 13 -33.2 6.9<br />

Macro invertebrates Amphipoda 8 May, 13 -29.4 6.8<br />

P. clarkii 3 May, 13 -29.5 6.4<br />

Forage fish G. holbrooki 3 May, 13 -29.0 10.4<br />

C. carpio 3 June, 3 -29.2 9.3<br />

Zooplankton Cladocera & Copepoda ≈ 50 May, 13 -33.9 8.9<br />

Macro invertebrates Amphipoda 8 May, 13 -28.8 8.9<br />

P. clarkii 3 May, 13 -31.6 7.5<br />

Forage fish G. holbrooki 3 May, 13 -28.0 10.2<br />

S. erythrophtalmus 3 Aug., 16 -26.5 9.2<br />

B. bloerkna 3 July, 5 -30.4 12.6<br />

206


δ 15 N (‰)<br />

13<br />

12<br />

11<br />

10<br />

9<br />

8<br />

Number of individuals<br />

7<br />

6<br />

5<br />

4<br />

3<br />

2<br />

1<br />

0<br />

Zooplankton<br />

7.2<br />

7.6<br />

8<br />

8.4<br />

8.8<br />

9.2<br />

9.6<br />

10<br />

10.4<br />

10.8 10.8<br />

11.2<br />

11.6<br />

12<br />

12.4<br />

12.8<br />

13.2<br />

δ15 δ N (‰)<br />

15N (‰)<br />

Régime alimentaire et habitat <strong>de</strong>s brochetons<br />

Zooplankton<br />

Zooplantivorous<br />

Piscivorous<br />

7<br />

-29 -28 -27 -26<br />

δ13C (‰)<br />

Figure 12.3. Combined δ 13 C and δ 15 N (‰) of young-of-the-year (YOY) PIT tagged pike at the<br />

hatchery when tagged. Black circle represents the zooplankton, black diamonds the<br />

zooplanktivorous pike and white diamonds the piscivorous pike. The triangles represent mean<br />

values of each group and the associated error bars are standard <strong>de</strong>viation. The solid line<br />

represents the trophic enrichment line. The histogram represents the <strong>de</strong>nsity plot of δ 15 N values<br />

measured on pike.<br />

12.3.4. Repeated Stable Isotope Analyses on recaptured pike<br />

Assuming a trophic enrichment of +2.01‰ δ 15 N (± 0.10), results <strong>de</strong>rived from pike recapture<br />

suggested that juvenile pike preyed upon zooplankton and macro-invertebrates in FG whereas<br />

they likely preyed upon forage fish in TP (Figure 12.4). Wild YOY pike ten<strong>de</strong>d to be bigger<br />

(FG) than hatchery-born individuals in both FG and TP but similar stable N signatures<br />

suggested no differences in resources used (Table 12.1 and Figure 12.4). Repeated SIA on<br />

individually PIT tagged pike showed variations in isotopic dilution from hatchery to FG and<br />

TP among individuals (Figure 12.5 and Figure 12.6). Despite the small number of fish<br />

recaptured in both FG and TP (n = 3), trophic shift from zooplankton – benthic invertebrates<br />

to forage fish associated with displacement from one habitat to the other is confirmed at the<br />

individual level. Variation in isotopic trajectories among YOY pike highlighted that timing of<br />

habitat and trophic shift after release is related to fish initial isotopic signatures, i.e. feeding<br />

behaviour (zooplanktivorous or piscivorous) at the hatchery (Figure 12.5 and Figure 12.6).<br />

207


Chapitre 12<br />

δ 15 N (‰)<br />

13<br />

12<br />

11<br />

10<br />

9<br />

8<br />

7<br />

Hatchery Floo<strong>de</strong>d Grassland Temporary Pond<br />

δ 15 N (‰)<br />

13<br />

12<br />

11<br />

10<br />

9<br />

8<br />

7<br />

6<br />

6<br />

6<br />

0 60 120 180 240 300 0 60 120 180 240 300 0 60 120 180 240 300<br />

Fish size (mm) Fish size (mm) Fish size (mm)<br />

Figure 12.4. δ 15 N (‰) values as a function of fork length (mm) of young-of-the-year (YOY) pike<br />

at the hatchery (open diamonds) when PIT-tagged (left), when emigrating from the floo<strong>de</strong>d<br />

grassland (centre) and when emigrating from the temporary pond (right) with wild YOY pike<br />

(filled diamonds) in Brière marsh, May-August 2005. The horizontal lines represent δ 15 N (‰) of<br />

food resources in each habitat: zooplankton (small dotted line), macro-invertebrates (large dotted<br />

line) and forage fish (full line).<br />

12.3.5. Growth of survival PIT-tagged pike<br />

Among the seven PIT tagged individuals recaptured at the end of the survey, four were<br />

zooplanktivorous and three were piscivorous at tagging (Table 12.3). The mean growth (FL<br />

increment, mm.day -1 ) of zooplantivorous pike was not statistically different (Mann-Whitney<br />

U test, P = 0.724, n = 7) than for piscivorous pike. Similar results were observed for the<br />

Fulton’s condition factor (K) with no statistical differences (Mann-Whitney U test, U = 7.0, p<br />

= 0.724, n = 7) between zooplantivores and piscivores (Table 12.3).<br />

208<br />

δ 15 N (‰)<br />

13<br />

12<br />

11<br />

10<br />

9<br />

8<br />

7


Zooplankton<br />

in the hatchery<br />

Zooplankton<br />

in the grassland<br />

Zooplankton<br />

in the pond<br />

Macro invertebrates<br />

in the grassland<br />

Macro invertebrates<br />

in the pond<br />

Régime alimentaire et habitat <strong>de</strong>s brochetons<br />

Zooplantivorous pike<br />

in the hatchery<br />

Forage fish<br />

in the grassland<br />

Forage fish<br />

in the pond<br />

PIT-tagged pike<br />

in the hatchery<br />

PIT-tagged pike<br />

in the grassland<br />

PIT-tagged pike<br />

in the pond<br />

-34 -33 -32 -31 -30<br />

δ<br />

-29 -28 -27<br />

5<br />

-26<br />

13C (‰)<br />

Figure 12.5. δ 13 C and δ 15 N (‰) of young-of-the-year (YOY) PIT-tagged pike (diamonds) and<br />

their food resources at the hatchery, in the floo<strong>de</strong>d grassland and in the temporary pond<br />

(theoretical food sources, see text for <strong>de</strong>tails on fractionation values).<br />

PIT-tagged pike<br />

in the hatchery<br />

PIT-tagged pike<br />

in the grassland<br />

PIT-tagged pike<br />

in the pond<br />

-32 -31 -30 -29 -28 -27 -26<br />

δ13 δ C (‰)<br />

13C (‰)<br />

Figure 12.6. δ 13 C and δ 15 N (‰) of young-of-the-year (YOY) PIT-tagged pike (diamonds) and<br />

individual isotopic trajectory of YOY PIT-tagged pike recaptured in the floo<strong>de</strong>d grassland (fine<br />

lines) and in the temporary pond (thick lines) in the observed food web.<br />

209<br />

13<br />

12<br />

11<br />

10<br />

9<br />

8<br />

7<br />

6<br />

δ 15 N (‰)<br />

13<br />

12<br />

11<br />

10<br />

9<br />

8<br />

7<br />

δ15 δ N (‰)<br />

15N (‰)


Chapitre 12<br />

PIT-tag ID<br />

Table 12.3. Characteristics of the seven PIT-tagged YOY pike recaptured at the end of the study<br />

when emigrating from the temporary pond to the permanent canal. Growth parameters are the<br />

increment in fork length (IFL, in mm.day -1 ) and the Fulton’s condition factor (K). See text for<br />

<strong>de</strong>tails on calculation.<br />

Characteristics at tagging Characteristics at recapture Growth parameters<br />

Diet<br />

FL<br />

(mm)<br />

Date<br />

Size<br />

(mm)<br />

Weight<br />

(g)<br />

IFL K<br />

674680B Zooplanktivores 49 July, 23 193 49 2.25 0.68<br />

6747314 Zooplanktivores 45 July, 23 173 32 2.00 0.62<br />

6702EAF Zooplanktivores 46 July, 23 167 30 1.89 0.64<br />

6704BB4 Zooplanktivores 47 July, 25 184 37 2.08 0.59<br />

Means (± S.D.) for zooplanktivores 179.3 (11.6) 37.0 (8.5) 2.06 (0.15) 0.63 (0.04)<br />

6701B89 Piscivores 48 July, 27 180 35 1.94 0.6<br />

6749336 Piscivores 61 July, 23 196 52 2.11 0.69<br />

67480A2 Piscivores 60 July, 19 194 47 2.23 0.64<br />

12.4. Discussion<br />

Means (± S.D.) for piscivores 190.0 (8.7) 44.7 (8.7) 2.09 (0.15) 0.64 (0.05)<br />

The grassland in which adult pike preferentially spawn had fewer forage fish and higher<br />

zooplankton abundance than the temporary pond. As such, the pond was a much better<br />

nursery area for YOY pike and this was apparent in their habitat shifts. In<strong>de</strong>ed, when released<br />

in a habitat with low abundance of forage fish, piscivorous individuals (i<strong>de</strong>ntified thanks to<br />

their stable N and C isotope signatures) left the site earlier than zooplanktivorous individuals.<br />

Because they had already accomplished their trophic shift towards piscivory when reared at<br />

the hatchery, it is likely that these individuals emigrated to the nursery habitat in search for<br />

forage fish, instead of remaining in FG and shifting their diet to the main available food<br />

resources (zooplankton). Cannibalism in YOY pike is usually observed during the early<br />

<strong>de</strong>velopment stage (FL = 21-60 mm), and it is known to be size-<strong>de</strong>pen<strong>de</strong>nt (as measured here)<br />

and size-selective, i.e. the smallest individuals having a higher probability of being eaten by<br />

larger ones (Bry et al. 1992, Skov & Koed 2004), and exacerbated un<strong>de</strong>r high rearing fish<br />

<strong>de</strong>nsities at hatchery (Giles et al. 1986, Bry et al. 1992, Juanes 2003). Releasing PIT-tagged<br />

hatchery-born juvenile pike in the study site likely loosened <strong>de</strong>nsity constraint and diminish<br />

opportunities for piscivorous to prey upon congeners, forcing them to emigrate rapidly to<br />

<strong>de</strong>eper habitats in search for appropriate preys. This suggests that shift in habitat use by YOY<br />

pike during the early juvenile period would be mainly driven by energetic constraints. The<br />

energetic constraint hypothesis is in line with recent laboratory experiments by Galarowicz &<br />

210


Régime alimentaire et habitat <strong>de</strong>s brochetons<br />

Wahl (2005), who <strong>de</strong>monstrated that energy return (J.min -1 ) and growth (g.day -1 ) of large<br />

YOY walleye (San<strong>de</strong>r vitreus) are high when feeding on fish, intermediate when feeding on<br />

benthic invertebrates and low when feeding on zooplankton.<br />

Conversely to the previous group of individuals released, zooplanktivorous pike<br />

remained in the floo<strong>de</strong>d and shallow grassland for a longer period, some individuals leaving<br />

the site just before total drying out. As water level <strong>de</strong>creases in temporary floo<strong>de</strong>d habitats,<br />

food availability and water quality (high temperature, low dissolved oxygen) concurrently<br />

<strong>de</strong>clined (Capone & Kushlan 1991, Magoulick & Kobza 2003), and the risk of being trapped<br />

in <strong>de</strong>siccating pools is real (Lowe-McConnell 1964, Kushlan 1976, Capone & Kushlan 1991,<br />

Chapman & Kramer 1991, Poizat & Crivelli 1997). Since zooplankton abundance was higher<br />

in the shallow spawning habitat than in the <strong>de</strong>ep nursery habitat, such habitat use tactic might<br />

be the result of a tra<strong>de</strong>-off between the risks of being stran<strong>de</strong>d and suffering poor water<br />

conditions in shallow spawning habitat and the benefits from higher food abundance and so,<br />

higher trophic returns. In<strong>de</strong>ed, Galarowicz & Wahl (2005) <strong>de</strong>monstrated that small YOY<br />

walleye had similar energy return and growth when feeding on zooplankton, benthic<br />

invertebrates and fish. In the present study, SIA results showed that, when small YOY pike<br />

chose to emigrate to the <strong>de</strong>eper habitat, they have most exclusively preyed upon zooplankton<br />

or macro-invertebrates. These individuals optimized their foraging behaviour by staying in the<br />

spawning habitat and preying upon primary consumers until they need more energy than<br />

offered in this habitat, and then moved towards the <strong>de</strong>ep nursery habitat. Whatever the tactic<br />

they displayed (rapid shift to <strong>de</strong>eper habitat to feed on forage fish, or feeding on the spawning<br />

habitat to feed on zooplankton until drying out), our results suggest that selection of the<br />

habitat, including the timing of habitat shift, varies in conjunction with their body-size. Gape<br />

size, which is a linear function of pike body length, limits eaten in pike (Nilsson & Bronmark<br />

2000) and this feature certainly influenced the patterns we measured. Nevertheless, these<br />

results, obtained here in harsh and spatially varying environments, need further investigations<br />

to un<strong>de</strong>rstand if similar habitat shifts would occur in more stabilized environments.<br />

Although early piscivory is known to affect life history in fish (e.g. Jonsson et al. 1999,<br />

Post 2003, Galarowicz & Wahl 2005), we did not find any differences in growth parameters<br />

between zooplanktivorous and piscivorous individuals, suggesting that their was no energetic<br />

advantage of early piscivory associated with an early migration. However, earlier migration<br />

might be indirectly beneficial to fish survival, since mortality risk likely increases as water<br />

level <strong>de</strong>clines. Furthermore, the earlier size advantage of piscivorous individuals might<br />

increase their probability of surviving the winter (Post 2003). The high mortality rates<br />

211


Chapitre 12<br />

observed throughout the study are similar to data reported in the literature for this species (e.g.<br />

Lucchietta 1983, Gronkjaer et al. 2004 and <strong>de</strong>tails in Chapitre 11). It is therefore possible that<br />

the low number of pike recaptures in TP preclu<strong>de</strong>s any <strong>de</strong>monstration of differences in<br />

individual growth performance between zooplanktivores and piscivores. Further<br />

investigations are nee<strong>de</strong>d to un<strong>de</strong>rstand the role played by early specialization of juvenile pike<br />

for piscivory on individual survival and fecundity and its substantial fitness advantage upon<br />

individuals able to make the early transition to piscivory (Post 2003).<br />

In conclusion, pike movements during the early juvenile period appear to be driven by<br />

three main forces: temporal changes in environmental characteristics (e.g. water level and<br />

temperature), early dietary shift to piscivory, and food availability. Shifts in habitat use by<br />

juvenile pike and their trophic position in the food web were closely associated, and appeared<br />

to be size <strong>de</strong>pen<strong>de</strong>nt. It was observed that the timing of habitat and diet shift was highly<br />

variable between individuals and that individual specialization cannot be explained if the<br />

initial fish characteristics (i.e. size and diet) are ignored. To that end, combining markrecapture<br />

and repeated SIA on individually tagged fish provi<strong>de</strong>s a key method to <strong>de</strong>scribe<br />

individual habitat use and diet trajectories, to study temporal processes un<strong>de</strong>rlying individual<br />

specialization, and to un<strong>de</strong>rstand the mechanisms leading to the existence of life history<br />

variants within the same cohort. Further investigations are nee<strong>de</strong>d to discriminate the role of<br />

genetic, maternal or <strong>de</strong>velopmental effects that might cause these variations.<br />

Acknowledgements<br />

We thank the Parc naturel régional <strong>de</strong> Brière (PnrB) for logistic support, the PnrB,<br />

FEDER, DIREN (MEDD), Région Pays <strong>de</strong> la Loire, Agence <strong>de</strong> l’Eau from Loire-Bretagne,<br />

for financial support. We are especially grateful to D. Huteau, F. Marchand, J.-P. Damien, M.<br />

Buoro, V. Thoby, J. Tremblay, E. Le Mitouard and A. Carpentier for assistance during field<br />

and laboratory work.<br />

212


Régime alimentaire et habitat <strong>de</strong>s brochetons<br />

Juvéniles <strong>de</strong> brochet marqué individuellement recapturé lors <strong>de</strong> son émigration du plan d’eau<br />

213


Chapitre 12<br />

Les résultats <strong>de</strong> ce chapitre montrent l’intérêt <strong>de</strong> coupler le marquage individuel à<br />

l’analyse <strong>de</strong>s isotopes stables pour comprendre le déterminisme <strong>de</strong> processus<br />

écologiques, en l’occurrence ici les relations entre changements <strong>de</strong> régimes<br />

alimentaires et utilisation <strong>de</strong>s MTI par les brochetons ainsi que pour en mesurer le<br />

niveau <strong>de</strong> variabilité inter-individuelle. Plus particulièrement, il a été montré que<br />

le statut trophique <strong>de</strong>s individus au moment <strong>de</strong> leur lâcher (i.e. piscivore vs.<br />

zooplanctonivore) lié à leur taille initiale va conditionner le temps qu’ils passeront<br />

dans les MTI, et ce en réponse aux caractéristiques environnementales du site, et<br />

notamment les ressources alimentaires disponibles. Sans qu’il soit possible <strong>de</strong><br />

mesurer les conséquences <strong>de</strong> cette variabilité <strong>de</strong> stratégies alimentaires en terme<br />

<strong>de</strong> croissance ou <strong>de</strong> survie du fait du nombre limité <strong>de</strong> survivants, il apparaît que<br />

la stratégie alimentaire vont conditionner l’utilisation spatio-temporelle <strong>de</strong>s MTI<br />

et donc le fonctionnement <strong>de</strong> la population. L’origine <strong>de</strong> la variabilité liée à la<br />

taille <strong>de</strong>s individus et à leur régime trophique peut être liée à <strong>de</strong>s facteurs<br />

génétiques, maternels ou du développement et doivent être étudiée plus en détails.<br />

Prochainement, <strong>de</strong>s investigations complémentaires seront engagées pour<br />

répondre à ces hypothèses. Elles allieront un lot <strong>de</strong> variables morphologiques<br />

mesurées individuellement (collaboration en cours) à une caractérisation <strong>de</strong> la<br />

mobilité <strong>de</strong>s individus (suivie à l’ai<strong>de</strong> <strong>de</strong> l’antenne portable, cf. Chapitre 11). In<br />

fine, cette approche pourrait permettre éventuellement <strong>de</strong> comprendre les<br />

interactions (cannibalisme et utilisation <strong>de</strong> l’habitat) existantes entre les <strong>de</strong>ux<br />

types d’individus. Les résultats issus <strong>de</strong> ce chapitre montrent aussi que les<br />

processus observés chez les individus nés en pisciculture semblent semblables à<br />

ceux observés chez les individus nés en milieu naturel, malgré une légère<br />

différence <strong>de</strong> taille tout au long du suivi. Ainsi, nous sommes en mesure <strong>de</strong><br />

fournir <strong>de</strong>s recommandations lors <strong>de</strong>s programmes <strong>de</strong> soutien <strong>de</strong> population <strong>de</strong><br />

brochet, même si la question générale du soutien <strong>de</strong> population doit être abordée<br />

dans un contexte plus global en intégrant notamment une expertise du<br />

fonctionnement <strong>de</strong> l’écosystème et <strong>de</strong> la communauté piscicole.<br />

214


Discussion générale et perspectives<br />

Chapitre13<br />

Un brochet dans une prairie temporairement inondée <strong>de</strong>s marais <strong>de</strong> Brière


Chapitre 13<br />

13.1. Synthèse <strong>de</strong>s principaux résultats obtenus<br />

Au cours <strong>de</strong> ce travail <strong>de</strong> thèse, nous avons vu que l’utilisation <strong>de</strong>s MTI dans les marais <strong>de</strong><br />

Brière était un phénomène fondamental pour le fonctionnement <strong>de</strong> la communauté piscicole,<br />

très variable et relativement restreint dans le temps et dans l’espace. La figure 13.1 permet<br />

d’avoir une vision synthétique <strong>de</strong> cette utilisation au travers <strong>de</strong>s résultats obtenus au cours <strong>de</strong><br />

ce travail grâce à la mise en place <strong>de</strong> protocoles à différentes échelles spatiales et temporelles,<br />

et ce pour différentes espèces et dans différents types <strong>de</strong> MTI.<br />

Abondance<br />

Abondance<br />

(a) (b)<br />

(c)<br />

Distance<br />

Distance<br />

Prairie<br />

Phase d’inondation<br />

Prairie<br />

Phase d’inondation<br />

Abondance<br />

Abondance<br />

(d)<br />

Distance<br />

Distance<br />

Roselière<br />

Phase d’inondation<br />

Roselière<br />

Phase d’inondation<br />

Figure 13.1. Synthèse <strong>de</strong>s résultats obtenus sur l’utilisation spatiale et temporelle <strong>de</strong>s milieux<br />

temporairement inondés dans les marais <strong>de</strong> Brière. La ligne pleine représentante l’abondance dans<br />

les milieux temporairement inondés et la ligne pointillée représente l’abondance dans les milieux<br />

permanents à travers les exemples d’une espèce « obligatoire » (a & b) et d’une espèce<br />

« facultative » (c & d).<br />

Il existe <strong>de</strong>s espèces plus ou moins abondantes dans la communauté piscicole occupant<br />

les milieux permanents à l’étiage (Chapitres 3, 4 & 5) et qui présentent une forte variabilité<br />

en terme d’utilisation <strong>de</strong>s MTI. Prenons le cas <strong>de</strong> <strong>de</strong>ux exemples très contrastés représentaifs<br />

<strong>de</strong>s grands patrons observés : une espèce native qui se reproduit dans les MTI (espèce<br />

« obligatoire », Figure 13.1.a.b) et une espèce introduitve dont la reproduction se déroule<br />

surtout dans les milieux permanents (espèce « facultative » ; Figure 13.1.c.d.). Lors <strong>de</strong> la<br />

phase d’inondation, les différentes espèces colonisent les MTI relativement tardivement<br />

(Chapitres 5 & 7). Comparativement à son niveau d’abondance dans les milieux permanents,<br />

plutôt bas (Chapitres 3 & 4), l’espèce native va, en terme d’abondance, fortement coloniser<br />

216


Discussion générale et perspectives<br />

les MTI, se disperser à <strong>de</strong>s distances élevées (Chapitres 4 & 5) et émigrer relativement tôt<br />

lors <strong>de</strong> la phase d’exondation (Chapitres 6 & 7). Ces processus vont se dérouler <strong>de</strong> la même<br />

manière dans les prairies et dans les roselières (Chapitre 4). A l’inverse, comparativement à<br />

son niveau d’abondance dans les milieux permanents, plutôt élevé (Chapitres 3 & 4), l’espèce<br />

introduite va, en terme d’abondance, faiblement coloniser les MTI, se disperser à <strong>de</strong>s<br />

distances limitées (Chapitres 4 & 5) et émigrer relativement tardivement lors <strong>de</strong> la phase<br />

d’exondation (Chapitres 6 & 7). De plus, ces processus vont se dérouler différemment dans<br />

les prairies et dans les roselières, avec un niveau d’abondance plus fort, une dispersion à <strong>de</strong>s<br />

distances plus éloignées et une émigration plus tardive dans les roselières que dans les praires<br />

(Chapitres 4 & 6).<br />

En plus <strong>de</strong> ces espèces échantillonnées <strong>de</strong> manière suffisante dans les MTI pour<br />

pouvoir en étudier l’utilisation (cf. précé<strong>de</strong>mment), nous avons aussi observé que d’autres<br />

espèces étaient relativement rares dans ces milieux. Soit il s’agit d’espèces présentes à <strong>de</strong>s<br />

niveaux d’abondances normaux dans les milieux permanents mais qui n’utilisent pas les MTI<br />

au cours <strong>de</strong> leur cycle biologique (par exemple le sandre, la perche franche ou l’épinoche).<br />

Soit il s’agit d’espèces qui doivent utiliser les MTI au cours <strong>de</strong> leur cycle biologique mais<br />

dont le niveau global d’abondance dans le site d’étu<strong>de</strong> est très faible, entraînant <strong>de</strong> très faibles<br />

abondances dans les MTI (par exemple le brochet ou la carpe).<br />

Nous avons vu que l’utilisation <strong>de</strong>s MTI était contrôlée par <strong>de</strong>ux paramètres<br />

principaux. D’abord, par <strong>de</strong> nombreux facteurs environnementaux abiotiques et biotiques et<br />

plus particulièrement <strong>de</strong> leurs évolutions synergiques au cours du cycle hydrologique. Il s’agit<br />

plus particulièrement du niveau d’eau, <strong>de</strong> la température <strong>de</strong> l’eau et <strong>de</strong> l’abondance <strong>de</strong><br />

nourriture (Chapitres 5 & 7). Ensuite, par les caractéristiques intrinsèques <strong>de</strong>s individus,<br />

qui changent en fonction du niveau d’organisation auquel l’analyse est réalisée. Il s’agit soit<br />

<strong>de</strong>s caractéristiques <strong>de</strong> l’espèce (tolérance physiologique, cycle <strong>de</strong> vie, origine, Chapitres 4, 5<br />

& 6), soit <strong>de</strong>s caractéristiques du sta<strong>de</strong> <strong>de</strong> vie (adulte, juvénile ou juvénile <strong>de</strong> l’année,<br />

Chapitre 7) ou soit <strong>de</strong>s caractéristiques propres <strong>de</strong> l’individus (taille, régime alimentaire,<br />

Chapitre 12).<br />

L’ensemble <strong>de</strong> ces résultats convergents, concernant l’utilisation <strong>de</strong>s MTI par la<br />

communauté piscicole, nous permet <strong>de</strong> suggérer que les modifications subies par les MTI,<br />

notamment en <strong>lien</strong> avec le changement <strong>de</strong>s activités anthropiques, ont certainement<br />

contribué à l’évolution <strong>de</strong> la composition <strong>de</strong> la communauté piscicole (i.e. implantations<br />

massives d’espèces introduites et raréfaction <strong>de</strong> certaines espèces natives) lors <strong>de</strong>s <strong>de</strong>rnières<br />

décennies (Chapitre 3 & 8).<br />

217


Chapitre 13<br />

La partie I <strong>de</strong> cette thèse a aussi permis d’i<strong>de</strong>ntifier certaines espèces d’intérêt<br />

majeur pour la communauté et son fonctionnement. Pour ces espèces, nous avons plus<br />

particulièrement analysé les interférences directes avec l’homme en examinant les réponses<br />

fonctionnelles aux caractéristiques <strong>de</strong> l’habitat et aux programmes <strong>de</strong> gestion dans la partie II.<br />

Le tableau 13.1 permet d’avoir une vision synthétique <strong>de</strong> ces réponses fonctionnelles, ainsi<br />

que <strong>de</strong> mesurer l’efficacité <strong>de</strong>s mesures <strong>de</strong> gestion actuellement mise en place pour les trois<br />

espèces d’intérêt majeur, à savoir le poisson chat, l’anguille et le brochet.<br />

Tableau 13.1. Synthèse <strong>de</strong>s principaux résultats obtenus sur les réponses fonctionnelles <strong>de</strong>s trois<br />

espèces d’intérêt majeur aux caractéristiques <strong>de</strong> l’habitat et aux différentes mesures <strong>de</strong> gestion.<br />

Espèce Poisson-chat Anguille Brochet<br />

Habitat sélectivité <strong>de</strong>s habitats restauration habitat utilisation MTI<br />

Mesure <strong>de</strong> gestion <strong>de</strong>struction par la pêcherie mise en réserve soutien population<br />

Objectif ↓ <strong>de</strong> l’abondance ↑ production géniteurs ↑ population<br />

Efficacité limitée intéressante très limitée<br />

Causes<br />

abondance globale production insuffisante forte mortalité<br />

d’inefficacité<br />

trop forte<br />

<strong>de</strong> géniteurs post-lâcher<br />

Concernant le poisson chat, espèce introduite qui domine actuellement la<br />

communauté piscicole, nous avons mis en évi<strong>de</strong>nce que les juvéniles et les adultes<br />

présentaient une forte affinité pour les roselières et l’ensemble <strong>de</strong>s modifications du milieu<br />

entraîné par leur expansion (modification <strong>de</strong>s MTI, atterrissement <strong>de</strong>s canaux, production <strong>de</strong><br />

litière, Chapitre 8). La <strong>de</strong>struction systématique <strong>de</strong>s individus capturés par la pêcherie<br />

entraîne une diminution <strong>de</strong>s abondances <strong>de</strong> l’espèce. Cependant, même dans les zones où la<br />

pression <strong>de</strong> pêche est la plus forte, l’abondance <strong>de</strong> poisson chat reste très élevée, démontrant<br />

les limites <strong>de</strong> cette mesure <strong>de</strong> gestion malgré plusieurs années <strong>de</strong> mise en place (Chapitre 9).<br />

Concernant l’anguille, espèce amphihaline et classée vulnérable, nous avons montré<br />

que la mise en place <strong>de</strong> réserves <strong>de</strong> pêche à la suite <strong>de</strong> la restauration d’habitat envahis par la<br />

roselière permettait une production intéressante <strong>de</strong> géniteurs. Cependant, en l’état actuel, les<br />

mises en réserves ne permettent pas <strong>de</strong> produire suffisamment <strong>de</strong> géniteurs au regard <strong>de</strong>s<br />

recommandations internationales concernant la gestion <strong>de</strong> l’espèce. A terme, cette pratique <strong>de</strong><br />

gestion pourrait être un moyen pertinent pour concilier la gestion durable <strong>de</strong> l’espèce et le<br />

maintien <strong>de</strong>s pêcheries traditionnelles (Chapitre 10).<br />

218


Discussion générale et perspectives<br />

Enfin, concernant le brochet, espèce native, localement vulnérable et exploitée, nous<br />

avons vu que le soutien <strong>de</strong> population réalisé à l’ai<strong>de</strong> <strong>de</strong> juvéniles <strong>de</strong> l’année présentait une<br />

efficacité très limitée du fait <strong>de</strong> la forte mortalité que subissent ces individus dans les MTI,<br />

notamment lors <strong>de</strong> la première semaine suivant leur lâcher (Chapitre 11). En s’intéressant <strong>de</strong><br />

plus près aux stratégies d’utilisation <strong>de</strong>s MTI et alimentaires <strong>de</strong>s individus ayant survécu,<br />

nous avons montré l’existence d’une forte variabilité interindividuelle provenant<br />

essentiellement <strong>de</strong>s caractéristiques <strong>de</strong>s individus au moment <strong>de</strong> leur lâcher (i.e. taille et<br />

régime alimentaire ; Chapitre 12), suggérant une adaptation <strong>de</strong> programme <strong>de</strong> soutien <strong>de</strong><br />

population vis-à-vis <strong>de</strong> ces résultats.<br />

Des carpes en train <strong>de</strong> se reproduire dans un milieu temporairement inondé <strong>de</strong>s<br />

marais <strong>de</strong> Brière<br />

219


Chapitre 13<br />

13.2. Quelle utilisation fonctionnelle <strong>de</strong>s milieux temporairement inondés<br />

par l’ichtyofaune ?<br />

L’utilisation <strong>de</strong>s MTI par l’ichtyofaune est apparue comme étant un phénomène fondamental<br />

du fonctionnement <strong>de</strong> la communauté piscicole et <strong>de</strong> l’accomplissement du cycle biologique<br />

<strong>de</strong> nombreuses espèces qui la composent, que ce soit les espèces « obligatoires » qui s’y<br />

reproduisent ou, à un moindre <strong>de</strong>gré, les espèces « facultatives » qui y croissent. Au cours<br />

<strong>de</strong>s différents chapitres <strong>de</strong> ce travail <strong>de</strong> thèse et quelles qu’ont être les approches utilisées<br />

(synchronique ou diachronique, communautaires, populationnelle ou individuelle, large ou<br />

fine échelle spatiale), il est apparu que cette utilisation était limitée et variable, à la fois dans<br />

le temps et dans l’espace et ce, en réponse aux caractéristiques propres <strong>de</strong>s individus<br />

(caractéristiques <strong>de</strong>s espèces, <strong>de</strong>s sta<strong>de</strong>s biologiques ou <strong>de</strong>s individus eux-mêmes) et <strong>de</strong>s<br />

conditions environnementales très variables entre MTI et au cours du cycle hydrologique.<br />

Elle peut néanmoins être aussi très profitable pour certaines espèces, sans certainement être<br />

optimale.<br />

13.2.1. Quelle importance pour l’ichtyofaune ?<br />

Au cours <strong>de</strong> ce travail <strong>de</strong> thèse, nous avons démontré la transposition possible <strong>de</strong>s gran<strong>de</strong>s<br />

caractéristiques <strong>de</strong>s MTI (voir la synthèse <strong>de</strong> Williams (2006)) au niveau du site d’étu<strong>de</strong>.<br />

Comme cela a déjà été montré, nous avons trouvé que les milieux permanents présentent <strong>de</strong>s<br />

conditions environnementales relativement stables comparées aux MTI qui sont <strong>de</strong>s milieux<br />

pulsés, inondés <strong>de</strong> manière saisonnière et durant une pério<strong>de</strong> relativement courte. Nous avons<br />

notamment vu que les MTI présentaient <strong>de</strong>s ressources trophiques abondantes, une bonne<br />

qualité physico-chimique <strong>de</strong> l’eau comparée à celle <strong>de</strong>s milieux permanents ainsi qu’une<br />

végétation <strong>de</strong>nse qui limite la prédation et fournit un support <strong>de</strong> ponte pour <strong>de</strong> nombreuses<br />

espèces <strong>de</strong> poissons, comme cela a déjà été démontré par <strong>de</strong> nombreux auteurs (e.g. Capone<br />

& Kushlan 1991, Snodgrass et al. 1996, Corti et al. 1999).<br />

Certaines étu<strong>de</strong>s ont démontré que les MTI d’un écosystème peuvent être<br />

partiellement occupés par l’ichytofaune, dans le sens où le taux d’occupation varie entre 21 et<br />

71% <strong>de</strong>s sites échantillonnés (Snodgrass et al. 1996, Baber et al. 2002). Dans les marais <strong>de</strong><br />

Brière, tous les MTI échantillonnés étaient colonisés par l’ichtyofaune. Ainsi, ces résultats<br />

confirment leur importance fonctionnelle dans ce site d’étu<strong>de</strong> et leur utilisation peut paraître<br />

bonne si l’on se limite au seul critère d’occurrence. Cependant, ces résultats sont à relativiser<br />

220


Discussion générale et perspectives<br />

lorsque l’on s’intéresse plus finement à l’utilisation temporelle mais aussi spatiale <strong>de</strong> ces<br />

milieux par l’ichtyofaune. Sur ce <strong>de</strong>rnier point, ce travail <strong>de</strong> thèse a confirmé l’existence d’un<br />

gradient d’abondance <strong>de</strong> poissons au sein <strong>de</strong>s MTI à partir <strong>de</strong>s milieux sources en individus,<br />

les abondances maximales étant observées pour les distances les plus proches <strong>de</strong>s milieux<br />

permanents. Ce résultat rejoint les quelques précé<strong>de</strong>nts travaux ayant abordé cette notion<br />

(Carpentier et al. 2004, Rehage & Trexler 2006).<br />

Cette caractéristique en terme d’exploitation spatiale <strong>de</strong>s MTI par les poissons limite<br />

probablement quantitativement et <strong>de</strong> façon significative le rôle fonctionnel <strong>de</strong> ces milieux<br />

dans les marais <strong>de</strong> Brière, puisqu’ils ne sont pas colonisés sur toute leur surface. Ce gradient<br />

spatial n’est pas lié aux paramètres environnementaux puisque nous n’avons pas observé <strong>de</strong><br />

changements significatifs <strong>de</strong>s paramètres mesurés (profon<strong>de</strong>ur, ressources trophiques ou<br />

couvert végétal) selon la distance aux milieux permanents sur les MTI <strong>de</strong>s marais <strong>de</strong> Brière.<br />

On peut émettre l’hypothèse que la distance <strong>de</strong> colonisation est dépendante d’une notion <strong>de</strong><br />

risque pris par les poissons présents sur les MTI. En effet, plus les poissons colonisent <strong>de</strong><br />

gran<strong>de</strong> distances dans les MTI plus le temps nécessaire pour rejoindre les milieux permanents<br />

sera long, la difficulté à regagner le milieu permanent via la connexion élevée et les risques <strong>de</strong><br />

mortalité accrus en cas, par exemple, d’exondation brutale. Ces résultats démontrent que toute<br />

la surface <strong>de</strong>s MTI <strong>de</strong>s marais <strong>de</strong> Brière n’est pas utilisée par l’ichtyofaune, et ce en réponse à<br />

<strong>de</strong>s distances trop élevées avec les milieux permanents, alors que les conditions<br />

environnementales ne sont pas plus adverses. A terme, il doit être envisagé <strong>de</strong> réussir à définir<br />

finement <strong>de</strong>s « surfaces utiles », c’est-à-dire <strong>de</strong>s surfaces minimales qui doivent être<br />

colonisées afin que les fonctions <strong>de</strong>s MTI <strong>de</strong>s marais <strong>de</strong> Brière pour la communauté piscicole<br />

soient améliorées. Ainsi, les résultats obtenus dans ce travail <strong>de</strong> thèse vont à l’encontre <strong>de</strong><br />

l’idée reçue qui consiste à penser que les MTI <strong>de</strong>s marais <strong>de</strong> Brière, comme c’est<br />

certainement le cas dans d’autres sites présentant cette particularité d’habitats (gran<strong>de</strong>s<br />

surfaces inondables), sont fortement utilisés par l’ichtyofaune au regard <strong>de</strong> la durée<br />

d’inondation et <strong>de</strong> la surface totale inondée. Les résultats <strong>de</strong> ce travail <strong>de</strong> thèse démontrent<br />

que l’exploitation <strong>de</strong>s MTI par l’ichtyofaune est fortement limitée, même si ces habitats sont<br />

tout <strong>de</strong> même attractifs et fondamentaux.<br />

Ce travail <strong>de</strong> thèse a permis <strong>de</strong> préciser le rôle <strong>de</strong>s MTI pour la reproduction et la<br />

croissance <strong>de</strong> nombreuses espèces <strong>de</strong> poissons. Cependant, il n’a pas été possible <strong>de</strong><br />

clairement quantifier le nombre d’individus d’une espèce ou plus précisément d’un sta<strong>de</strong> <strong>de</strong><br />

vie donné qui rejoignent ces milieux lors du cycle d’inondation. Certes, la mesure <strong>de</strong> la<br />

sélectivité d’habitat (indice d’Ivlev (1961)) et les données <strong>de</strong> capture par pêche électrique ou<br />

221


Chapitre 13<br />

par piégeage passif ont permis <strong>de</strong> se faire une idée à ce sujet et <strong>de</strong> démontrer que ce processus<br />

est variable entre les espèces. L’utilisation d’un protocole <strong>de</strong> capture/marquage/recapture<br />

(voir par exemple Hohausova 2000), pourrait permettre d’y parvenir. Cependant, à l’échelle<br />

<strong>de</strong> la communauté piscicole et <strong>de</strong> l’ensemble <strong>de</strong>s sta<strong>de</strong>s <strong>de</strong> vie, une telle approche semble peu<br />

envisageable car très difficilement applicable.<br />

En revanche, l’utilisation <strong>de</strong> l’outil isotopique pourrait être un moyen pertinent pour<br />

répondre à cette question. En effet, Wantzen et al. (2002) et Oliveira et al. (2006) ont<br />

récemment démontré que l’analyse <strong>de</strong>s isotopes stables permet <strong>de</strong> mettre en évi<strong>de</strong>nce la forte<br />

variabilité isotopique entre milieux permanents et MTI dans <strong>de</strong>s écosystèmes amazoniens en<br />

terme <strong>de</strong> sources <strong>de</strong> carbone (plantes en C4 et C3). Ces différences ont ainsi permis<br />

d’i<strong>de</strong>ntifier clairement la zone (milieu permanent ou MTI) dans laquelle se trouve la source<br />

(niveau trophique) sur laquelle les individus appartenant à différentes espèces <strong>de</strong> poissons<br />

s’alimentent. En effet, Wantzen et al. (2002) ont synthétisé l’ensemble <strong>de</strong>s facteurs qui<br />

influence les valeurs <strong>de</strong>s isotopes stables <strong>de</strong>s poissons utilisant les MTI, ce qui permet <strong>de</strong><br />

comprendre pourquoi cet outil est intégrateur et performant dans l’étu<strong>de</strong> <strong>de</strong>s relations<br />

ichtyofaune / MTI. (Figure 13.2). De plus, certains <strong>de</strong> nos résultats préliminaires montrent<br />

que l’existence <strong>de</strong> signatures différentes sur l’isotope du carbone le long du gradient<br />

d’inondation entre milieu (canal, plan d’eau et prairie) est aussi observable dans les marais <strong>de</strong><br />

Brière (Cucherousset, données non publiées). On peut aisément imaginer qu’une analyse <strong>de</strong>s<br />

isotopes stables réalisée sur <strong>de</strong>s individus collectés dans les milieux permanents au début <strong>de</strong><br />

l’étiage permettrait certainement <strong>de</strong> quantifier la proportion d’individus (d’une espèce ou d’un<br />

sta<strong>de</strong> <strong>de</strong> vie) ayant utilisé les MTI lors <strong>de</strong> la pério<strong>de</strong> d’inondation précé<strong>de</strong>nte. Enfin,<br />

l’utilisation <strong>de</strong> l’analyse <strong>de</strong>s isotopes stables dans <strong>de</strong> telles problématiques a aussi permis <strong>de</strong><br />

montrer qu’une partie <strong>de</strong> l’ichtyofaune jeûne pendant l’étiage et que l’intégralité <strong>de</strong>s prises<br />

alimentaires se déroule dans les MTI pendant l’inondation dans un écosystème soumis à un<br />

rythme saisonnier d’inondation (Wantzen et al. 2002). D’une manière plus globale, <strong>de</strong> telles<br />

analyses dans les marais <strong>de</strong> Brière permettraient d’approfondir les connaissances sur le rôle<br />

fonctionnel <strong>de</strong>s MTI pour l’ichtyofaune dans les écosystèmes <strong>de</strong>s régions tempérées.<br />

222


Etat saisonnier<br />

<strong>de</strong>s MTI<br />

Items alimentaires<br />

disponibles<br />

Historique local<br />

Biochimie, géochimie et<br />

connectivité du site d’étu<strong>de</strong><br />

Préférences alimentaires<br />

individuelles<br />

Valeur <strong>de</strong>s isotopes stables<br />

<strong>de</strong> l’ichtyofaune utilisant les MTI<br />

Régime alimentaire<br />

<strong>de</strong> l’espèce<br />

Discussion générale et perspectives<br />

Etat saisonnier<br />

<strong>de</strong>s poissons<br />

Stockage <strong>de</strong> graisses<br />

Taille <strong>de</strong>s ressources<br />

trophiques<br />

Sta<strong>de</strong>s <strong>de</strong> développement<br />

<strong>de</strong>s poissons<br />

Figure 13.2. Représentation schématique <strong>de</strong>s facteurs qui influencent, à différentes échelles<br />

temporelles, les valeurs isotopiques <strong>de</strong>s poissons utilisant les milieux temporairement inondés<br />

(d’après Wantzen et al. 2002).<br />

13.2.2. Quelles conséquences <strong>de</strong> la variabilité <strong>de</strong>s conditions<br />

environnementales ?<br />

Au cours <strong>de</strong> ce travail <strong>de</strong> thèse, nous avons vu que les caractéristiques environnementales<br />

dans les MTI ne sont pas stables au cours du cycle d’inondation et nous avons quantifié les<br />

échanges d’individus entre milieux permanents et MTI (immigration et d’émigration) en<br />

fonction <strong>de</strong> l’évolution <strong>de</strong> ces paramètres environnementaux.<br />

Lors <strong>de</strong> la phase d’inondation, la montée du niveau d’eau conduit à une augmentation<br />

<strong>de</strong> la surface inondée et donc à un accroissement <strong>de</strong> l’accessibilité <strong>de</strong>s MTI pour les poissons<br />

aux cours <strong>de</strong> l’hiver (janvier/février). Parallèlement, la température <strong>de</strong> l’eau monte et<br />

s’accompagne d’une augmentation <strong>de</strong> l’abondance <strong>de</strong> zooplancton (convergence <strong>de</strong>s <strong>de</strong>ux<br />

patrons au cours du temps). A ce sta<strong>de</strong>, qui peut avoir lieu plusieurs mois après le début <strong>de</strong><br />

l’inondation (avril), l’immigration <strong>de</strong> poissons dans les MTI est observée, principalement en<br />

réponse aux changements <strong>de</strong> température, avec <strong>de</strong> fortes variations individuelles dont les<br />

sources sont essentiellement interspécifiques. Ces températures ont une valeur différente<br />

selon que les espèces se reproduisent ou non dans les MTI. Dans le premier cas, elles peuvent<br />

être nécessaires pour déclencher la maturation et la ponte et sont très variables selon les<br />

espèces (voir revue dans Keith & Allardi 2001). Dans le second cas, elles représentent <strong>de</strong>s<br />

températures seuils nécessaires à la reprise d’activité alimentaire après la phase <strong>de</strong> léthargie<br />

223


Chapitre 13<br />

hivernale. Une fois ces températures atteintes, les différentes espèces colonisent les MTI au<br />

fur et à mesure <strong>de</strong> l’augmentation <strong>de</strong> la température moyenne et se dispersent <strong>de</strong> plus en plus<br />

loin en réponse à l’augmentation du niveau d’eau et <strong>de</strong> la surface inondée dans ces milieux.<br />

Une fois le pic d’inondation atteint (fin avril), les conditions environnementales se dégra<strong>de</strong>nt<br />

en réponse à la diminution du niveau d’eau. Ainsi, la disponibilité en ressources alimentaires<br />

diminue (e.g. Capone & Kushlan 1991), la pression <strong>de</strong> prédation augmente (e.g. Kushlan<br />

1976), et la qualité <strong>de</strong> physico-chimique <strong>de</strong> l’eau se détériore (augmentation <strong>de</strong> la température<br />

et diminution <strong>de</strong> la concentration en dioxygène, e.g. Magoulick & Kobza 2003). De plus, les<br />

poissons risquent d’être piégés dans <strong>de</strong>s mares déconnectées, ceci pouvant conduire à une<br />

forte prédation <strong>de</strong>s poissons (e.g. Lowe-McConnell 1964, Kushlan 1976, Capone & Kushlan<br />

1991, Chapman & Kramer 1991, Poizat & Crivelli 1997).<br />

Dans le site d’étu<strong>de</strong>, nous n’avons pas observé <strong>de</strong> phénomènes <strong>de</strong> piégeages massifs<br />

<strong>de</strong> poissons dans <strong>de</strong>s mares isolées. On peut donc émettre l’hypothèse que l’ichtyofaune<br />

locale a certainement s’adapter aux conditions locales, notamment en terme d’utilisation <strong>de</strong><br />

l’habitat, du fait <strong>de</strong> la prédictibilité saisonnière relativement forte du cycle hydrologique et<br />

d’une évolution relativement progressive du niveau d’eau que ce soit lors <strong>de</strong> l’inondation ou<br />

<strong>de</strong> l’exondation (cf. Figure 2.3). L’une d’entre elles est la capacité à émigrer en réponse à la<br />

dégradation <strong>de</strong>s conditions environnementales, entre les mois <strong>de</strong> mai et juin. Ce processus<br />

d’émigration est apparu comme étant sous l’influence <strong>de</strong> nombreux facteurs et présentant une<br />

forte variabilité entre les espèces, entre les sta<strong>de</strong>s <strong>de</strong> vie et entre les individus.<br />

Au niveau interspécifique, ce travail <strong>de</strong> thèse a permis d’i<strong>de</strong>ntifier la tolérance<br />

physiologique comme étant un <strong>de</strong>s facteurs intrinsèques aux espèces permettant <strong>de</strong><br />

comprendre la dynamique d’émigration <strong>de</strong>s MTI. Plus particulièrement, nous avons montré<br />

qu’une tolérance physiologique élevée, caractéristique <strong>de</strong>s espèces invasives (Marchetti et al.<br />

2004), conduit à une émigration tardive et donc une exploitation maximale <strong>de</strong>s ressources<br />

disponibles dans les MTI par ces espèces. De plus, les espèces introduites <strong>de</strong>s marais <strong>de</strong><br />

Brière (principalement le poisson chat, la perche soleil et la gambusie) utilisent les MTI<br />

principalement comme zone <strong>de</strong> croissance (espèces dites « facultatives »). En effet, ces<br />

espèces présentent une reproduction tardive et leur cycle biologique n’est donc pas fortement<br />

inféodé aux MTI (voir la synthèse <strong>de</strong>s différents guil<strong>de</strong>s reproducteurs dans Keith & Allardi<br />

(2001)). L’ensemble <strong>de</strong> ces caractéristiques écologiques laisse à penser que les individus <strong>de</strong>s<br />

espèces introduites <strong>de</strong>s marais <strong>de</strong> Brière ont une utilisation <strong>de</strong> l’habitat très opportuniste et<br />

tirent <strong>de</strong>s gains supplémentaires en terme <strong>de</strong> croissance grâce à une exploitation prolongée <strong>de</strong>s<br />

MTI, gains qui pourront ensuite être investis dans l’effort <strong>de</strong> reproduction au moment <strong>de</strong><br />

224


Discussion générale et perspectives<br />

l’étiage et ainsi accroître les fortes capacités d’invasion <strong>de</strong> ces espèces. De plus, le niveau<br />

d’eau <strong>de</strong> Brière sont contrôlés et stabilisés par <strong>de</strong>s ouvrages et Vila-Gispert et al. (2005) ont<br />

démontré que la régulation et la stabilisation <strong>de</strong>s régimes hydrologiques (réduction <strong>de</strong><br />

l’amplitu<strong>de</strong> et <strong>de</strong> la durée <strong>de</strong> l’inondation) par la mise en place d’ouvrages hydrauliques<br />

pourraient favoriser les espèces <strong>de</strong> poissons introduites. En effet, la plupart d’entre elles sont<br />

adaptées à ces conditions car originaires d’écosystèmes dits stables. Cependant, une analyse<br />

plus fine <strong>de</strong> certains traits d’histoire <strong>de</strong> vie, tels que la croissance et la fécondité, est<br />

nécessaire afin <strong>de</strong> vali<strong>de</strong>r plus précisément ces hypothèses et comprendre <strong>de</strong> manière plus<br />

précise les changements <strong>de</strong> composition <strong>de</strong> la communauté piscicole et le succès d’invasion<br />

<strong>de</strong> certaines espèces.<br />

Au niveau intraspécifique, ce travail <strong>de</strong> thèse a montré le rôle <strong>de</strong> la taille <strong>de</strong>s<br />

individus dans le processus d’émigration, soit entre les différents sta<strong>de</strong>s <strong>de</strong> vie d’une même<br />

population, soit entre les individus d’une même cohorte. Dans un premier temps, nous avons<br />

vu que les individus les plus petits (juvéniles <strong>de</strong> l’année) émigrent le plus tardivement alors<br />

que individus les plus gros (adultes puis juvéniles >0+) émigrent le plus précocement. Dans<br />

un second temps, nous avons démontré, grâce à l’utilisation du marquage individuel par PIT<br />

tag que les juvéniles <strong>de</strong> brochets issus d’une même cohorte émigrent <strong>de</strong>s MTI, et ce <strong>de</strong><br />

manière très prédictive, en fonction <strong>de</strong> la taille initiale au marquage, la date d’émigration étant<br />

négativement corrélée avec la taille initiale.<br />

Plusieurs interprétations peuvent être avancées pour expliquer le patron d’émigration<br />

<strong>de</strong>s individus en fonction <strong>de</strong> leur taille. D’abord, la taille d’un individu conditionne son<br />

régime alimentaire pour la majorité <strong>de</strong>s espèces qui se caractérisent par une évolution du<br />

régime au cours du développement ontogénique (voir revue complète dans Keith & Allardi<br />

(2001)). Dans ce cas, les poissons <strong>de</strong> petite taille sont souvent zooplanctivores et trouvent<br />

d’importantes ressources trophiques dans les MTI. Ils optimisent donc leur croissance en<br />

restant dans ces milieux tant que les conditions environnementales sont supportables et que le<br />

zooplancton reste plus abondant que dans les milieux permanents. A l’inverse les plus grands<br />

poissons qui possè<strong>de</strong>nt d’autres régimes alimentaires (consommateurs <strong>de</strong> macro invertébrés<br />

ou piscivores) ne trouvent peut être pas ces conditions dans les MTI et quittent donc plus<br />

précocement ces milieux. Ensuite, la taille d’un individu conditionne sa capacité à tolérer<br />

<strong>de</strong>s concentrations en dioxygène très faibles. En effet, <strong>de</strong> récentes étu<strong>de</strong>s ont permis <strong>de</strong><br />

confirmer que les petits individus d’une espèce sont moins sensibles aux hypoxies (e.g. Robb<br />

& Abrahams 2003) du fait <strong>de</strong> la relation allométrique négative entre la taille du corps et la<br />

surface branchiale, les petits individus ayant une meilleur capacité d’échange <strong>de</strong> gazeux avec<br />

225


Chapitre 13<br />

leur environnement (Hugues 1984). De plus, les petits individus sont souvent davantage<br />

capables <strong>de</strong> capter l’oxygène au niveau <strong>de</strong> l’interface eau-air que les plus gros poissons (voir<br />

dans Sargent & Galat 2002). Enfin, la taille d’un individu influence sa capacité à se disperser<br />

dans <strong>de</strong>s milieux peu profonds et à quitter <strong>de</strong>s milieux présentant <strong>de</strong>s niveaux d’eau très bas.<br />

Les sources <strong>de</strong> variabilité <strong>de</strong> croissance <strong>de</strong>s juvéniles <strong>de</strong> poissons sont nombreuses, et on peut<br />

citer les causes génétiques, les effets maternels et l’influence <strong>de</strong>s conditions<br />

environnementales lors du développement (voir revue dans Matthews 1998). Toutes ces<br />

sources <strong>de</strong> variabilités individuelles <strong>de</strong> croissance ont <strong>de</strong>s conséquences sur l’utilisation <strong>de</strong>s<br />

MTI, et il sera nécessaire <strong>de</strong> comprendre en quoi cela influence les stratégies d’histoire <strong>de</strong> vie<br />

<strong>de</strong>s différents individus, notamment à travers l’analyse fine <strong>de</strong> certains traits.<br />

13.2.3. Prairies, roselières et piar<strong>de</strong>s : quelles différences pour l’ichtyofaune ?<br />

Au cours <strong>de</strong> ce travail <strong>de</strong> thèse, nous avons vu que l’utilisation <strong>de</strong>s MTI par les poissons<br />

(composition <strong>de</strong> la communauté et dispersion <strong>de</strong>s individus sur les MTI à partir <strong>de</strong>s milieux<br />

sources) était fortement influencée par leur nature. Dans les marais <strong>de</strong> Brière, les MTI,<br />

distribués le long d’un gradient <strong>de</strong> durée et d’amplitu<strong>de</strong> d’inondation, sont principalement<br />

composés <strong>de</strong> prairies, <strong>de</strong> roselières et <strong>de</strong> piar<strong>de</strong>s qui se différencient par leurs caractéristiques<br />

environnementales (notamment la profon<strong>de</strong>ur, le niveau <strong>de</strong> recouvrement végétal ou la<br />

composition du substrat) causées par <strong>de</strong>s patrons d’inondation très variables (cf. Figure 2.4).<br />

Cependant, aucune différence n’a été démontrée entre les différents MTI en terme<br />

d’abondance générale et <strong>de</strong> richesse spécifique. Mais, nous avons mis en évi<strong>de</strong>nce une forte<br />

variation en terme <strong>de</strong> composition spécifique.<br />

L’approche prédictive développée par Williams (2006) portant sur les espèces<br />

d’invertébrés présentes dans les MTI dans une gamme théorique allant d’une inondation nulle<br />

à une inondation permanente peut être reprise par analogie dans le cas <strong>de</strong> la communauté<br />

piscicole <strong>de</strong>s marais <strong>de</strong> Brière. Celle-ci prédit en effet que plus le milieu est inondé (durée et<br />

amplitu<strong>de</strong>), plus la proportion d’espèces dites « facultatives » (c’est-à-dire <strong>de</strong>s espèces dont<br />

le cycle <strong>de</strong> vie est peu inféodé aux MTI) présentes dans ce milieu sera forte. Inversement, plus<br />

le milieu est asséché longtemps, plus la proportion d’espèces dites « obligatoires » (c’est-àdire<br />

<strong>de</strong>s espèces dont le cycle <strong>de</strong> vie est fortement inféodé aux MTI) sera importante dans ce<br />

milieu (Figure 13.3).<br />

226


Maximum<br />

Proportion d’espèces facultatives<br />

dans le milieu temporairement inondé<br />

Minimum<br />

Minimum<br />

Durée <strong>de</strong> la phase<br />

d’assèchement<br />

Discussion générale et perspectives<br />

Maximum<br />

Maximum<br />

Proportion d’espèces obligatoires<br />

dans le milieu temporairement inondé<br />

Minimum<br />

Figure 13.3. Représentation schématique <strong>de</strong>s prédictions <strong>de</strong> changements dans la composition <strong>de</strong><br />

la communauté le long du gradient <strong>de</strong> conditions environnementales (durée d’inondation) entre<br />

milieux temporairement inondés et milieux permanents (d’après Williams 2006).<br />

Cette prédiction s’est confirmée dans ce travail <strong>de</strong> thèse, notamment au regard <strong>de</strong>s<br />

guil<strong>de</strong>s reproducteurs <strong>de</strong>s différentes espèces <strong>de</strong> la communauté piscicole présentes dans les<br />

prairies, les roselières et les piar<strong>de</strong>s. Par exemple, nous avons observé que les espèces, dont la<br />

reproduction ne peut se dérouler que dans <strong>de</strong>s milieux faiblement inondés (espèces dites<br />

« obligatoires » telle que le rotengle) colonisent majoritairement les prairies, c’est-à-dire les<br />

MTI <strong>de</strong>s marais <strong>de</strong> Brière avec la durée d’assèchement la plus longue. Inversement, nous<br />

avons démontré que la communauté piscicole <strong>de</strong>s plans d’eau <strong>de</strong>s marais <strong>de</strong> Brière est<br />

principalement composé d’espèces dites « facultatives » (telles que la brème ou la poisson<br />

chat), c’est-à-dire dont la reproduction se déroule dans les milieux permanents ou semipermanents.<br />

Une telle approche montre d’ailleurs l’intérêt <strong>de</strong> bien caractériser le cycle d’inondation,<br />

notamment en terme <strong>de</strong> durée et d’amplitu<strong>de</strong>. En effet, ce travail <strong>de</strong> thèse s’est aussi attaché à<br />

analyser la distribution spatiale <strong>de</strong>s individus dans les différents types <strong>de</strong> MTI. Nous avons<br />

ainsi démontré l’existence d’un fort gradient d’abondance <strong>de</strong> poissons concordant avec le<br />

227


Chapitre 13<br />

gradient <strong>de</strong> distance dans les prairies et les roselières <strong>de</strong>puis les milieux permanents.<br />

Cependant, une analyse préliminaire <strong>de</strong> cette distribution dans les plans d’eau a démontré que<br />

ce gradient en était absent. De plus, lors <strong>de</strong> crues exceptionnellement fortes comme en 2001<br />

(cf. Figure 2.3 & Carpentier, données non publiées), le gradient d’abondance en fonction <strong>de</strong> la<br />

distance tend à disparaître et la répartition spatiale <strong>de</strong>s poissons tend à augmenter à la fois<br />

dans les prairies et les roselières. Dans les Evergla<strong>de</strong>s (Flori<strong>de</strong>) où les inondations sont<br />

intermédiaires à celles <strong>de</strong>s prairies/roselières et <strong>de</strong>s piar<strong>de</strong>s <strong>de</strong>s marais <strong>de</strong> Brière, Rehage &<br />

Trexler (2006) ont aussi observé l’existence d’un tel patron <strong>de</strong> réponse lié au niveau <strong>de</strong><br />

stabilité <strong>de</strong>s MTI. Il sera donc nécessaire <strong>de</strong> comprendre à terme comment <strong>de</strong>s conditions<br />

hydrologiques différentes peuvent modifier cette distribution spatiale et donc l’utilisation<br />

faite <strong>de</strong>s MTI par l’ichtyofaune, que ce soit en comparant différents types <strong>de</strong> MTI ou en<br />

comparant le même type <strong>de</strong> MTI avec <strong>de</strong>s conditions hydrologiques très variables (sur un pas<br />

<strong>de</strong> temps supérieur à 3 ans).<br />

De plus, nous avons démontré que la distribution spatiale <strong>de</strong>s poissons dans les MTI<br />

varie dans les prairies et les roselières, et ce en fonction <strong>de</strong>s espèces. Des espèces telles que<br />

les brèmes, la gambusie ou la perche soleil ne se dispersent qu’à <strong>de</strong>s distances réduites dans<br />

les MTI (<strong>de</strong> l’ordre <strong>de</strong> 50 mètres) alors que le rotengle, le poisson chat ou l’anguille se<br />

dispersent sur <strong>de</strong>s distances plus élevées (<strong>de</strong> l’ordre <strong>de</strong> 200 mètres). De plus, pour les espèces<br />

natives, il n’existe pas <strong>de</strong> différence en terme <strong>de</strong> distance <strong>de</strong> colonisation à partir du milieu<br />

source entre les prairies et roselières, alors que pour les espèces introduites, la distance <strong>de</strong><br />

colonisation par les espèces introduites est plus élevée dans les roselières que dans les prairies.<br />

Les perturbations environnementales facilitent les phénomènes d’invasion (Moyle &<br />

Light 1996, Vila-Gispert et al. 2005). Ainsi, la modification et la fragmentation <strong>de</strong> l’habitat<br />

pénalisent l’ichtyofaune native au profit <strong>de</strong>s espèces introduites souvent plus tolérantes à <strong>de</strong>s<br />

conditions environnementales plus dégradées (Marchetti et al. 2004). Dans les marais <strong>de</strong><br />

Brière, les résultats montrent que l’expansion <strong>de</strong> la roselière (c’est-à-dire modification<br />

naturelle mais brutale <strong>de</strong>s MTI <strong>de</strong> type prairie, production forte <strong>de</strong> litière, atterrissement <strong>de</strong>s<br />

canaux) a certainement eu <strong>de</strong>s conséquences sur la composition <strong>de</strong> la communauté piscicole<br />

en facilitant les espèces introduites, sans forcément influencer directement négativement les<br />

espèces natives (en tout cas celles étudiées en détail). Ces résultats soutiennent l’hypothèse <strong>de</strong><br />

Simberloff & Von Holle (1999), qui consiste à postuler que la présence d’une espèce invasive<br />

(ici le roseau) facilite l’établissement d’autres espèces indigènes (ici les espèces <strong>de</strong> poissons<br />

introduites), en modifiant les caractéristiques <strong>de</strong> l’habitat qui s’avèrent favorables aux espèces<br />

introduites.<br />

228


Discussion générale et perspectives<br />

L’ensemble <strong>de</strong> ces résultats concernant les espèces introduites <strong>de</strong> poissons conduit à la<br />

convergence <strong>de</strong> <strong>de</strong>ux phénomènes. Ces espèces utilisent « facultativement » les MTI et<br />

semblent profiter <strong>de</strong>s modifications que ceux-ci ont subies dans les marais <strong>de</strong> Brière. Dans le<br />

cas présent, trois hypothèses principales peuvent être avancées pour expliquer les possibles<br />

modifications <strong>de</strong> l’habitat qui auraient favorisé le succès d’invasion <strong>de</strong>s espèces introduites.<br />

La première est d’ordre trophique, car l’expansion <strong>de</strong> la roselière a pour conséquence une<br />

modification <strong>de</strong> la composition <strong>de</strong> la communauté d’invertébrés associée qui est<br />

caractéristique d’habitats riches en litière (Mason & Bryant 1975, Bedford & Powell 2005).<br />

Cette communauté d’invertébrés correspond peut-être mieux aux régimes alimentaires <strong>de</strong>s<br />

espèces introduites ou est peut-être mieux exploitée par ces <strong>de</strong>rnières que par les espèces<br />

natives. La secon<strong>de</strong> hypothèse fait référence à la tolérance physiologique <strong>de</strong>s espèces<br />

introduites. En effet, il a déjà été évoqué que les espèces introduites sont souvent<br />

particulièrement tolérantes (Marchetti et al. 2004), soit à <strong>de</strong>s températures élevées soit à <strong>de</strong>s<br />

concentrations en dioxygène très basses (Otto 1973, McKinsey & Chapman 1998, Sargent &<br />

Galat 2002). La décomposition <strong>de</strong> la litière modifie fortement les conditions<br />

environnementales (élévation <strong>de</strong> la température et diminution <strong>de</strong> la concentration en<br />

dioxygène) et la tolérance physiologique <strong>de</strong>s espèces introduites leur permet peut être <strong>de</strong><br />

maximiser l’utilisation <strong>de</strong>s ressources trophiques présentes dans ces milieux. Enfin, la<br />

troisième hypothèse est liée à <strong>de</strong>s aspects <strong>de</strong> reproduction. La perche soleil et le poisson<br />

chat sont <strong>de</strong>s espèces dites « gardiens <strong>de</strong> nids » qui nécessitent l’utilisation <strong>de</strong> débris végétaux<br />

pour la construction <strong>de</strong> nids (Keith & Allardi 2001). Or ces débris sont produits en fortes<br />

quantités dans les roselières. Ces <strong>de</strong>ux espèces, qui se reproduisent tardivement dans la saison<br />

et principalement dans les milieux permanents, bénéficient certainement <strong>de</strong> la présence d’un<br />

tel substrat dans les milieux bordés <strong>de</strong> roselières.<br />

Les trois hypothèses précé<strong>de</strong>mment avancées pourraient être étudiées pour<br />

comprendre les processus écologiques facilitant l’invasion <strong>de</strong>s espèces introduites. Il est<br />

proposé pour cela d’engager une démarche basée sur un échantillonnage stratifié (individus<br />

d’espèces natives et introduites utilisant soit les prairies soit les roselières), couplant à la fois<br />

l’étu<strong>de</strong> <strong>de</strong>s traits d’histoire <strong>de</strong> vie (croissance, fécondité et survie) et analyse <strong>de</strong>s isotopes<br />

stables.<br />

229


Chapitre 13<br />

13.2.4. Quels compromis pour l’ichtyofaune ?<br />

Selon Wellborn et al. (1996), les mécanismes d’utilisation <strong>de</strong>s différents habitats le long d’un<br />

gradient <strong>de</strong> durée d’inondation constituent <strong>de</strong>s facteurs forts <strong>de</strong> structuration et <strong>de</strong><br />

fonctionnement d’une communauté d’invertébrés. De la même façon, le fonctionnement <strong>de</strong> la<br />

communauté piscicole <strong>de</strong>s marais <strong>de</strong> Brière peut être analysé en prenant en compte les rôles<br />

fonctionnels <strong>de</strong>s MTI pour la croissance et la reproduction <strong>de</strong> l’ichtyofaune.<br />

En effet, tous nos résultats suggèrent que les processus liés à l’utilisation <strong>de</strong>s MTI par<br />

l’ichtyofaune sont assez prédictibles (c’est-à-dire non aléatoires) et on peut donc dire que le<br />

comportement <strong>de</strong> sélection volontaire <strong>de</strong> l’habitat est impliqué. En effet, l’interprétation <strong>de</strong> la<br />

sélection d’habitat en terme <strong>de</strong> stratégies au sens évolutionniste du terme (Morris 2003), est<br />

un moyen utile <strong>de</strong> percevoir <strong>de</strong> manière synthétique l’utilisation <strong>de</strong>s MTI par l’ichtyofaune<br />

dans un écosystème spatialement très variable. Cette interprétation fait référence à <strong>de</strong>s<br />

déplacements non aléatoires qui résultent <strong>de</strong> mouvements volontaires <strong>de</strong>s organismes (Kramer<br />

et al. 1997). Une manière simple <strong>de</strong> percevoir ces mécanismes <strong>de</strong> sélection <strong>de</strong> l’habitat est<br />

donc <strong>de</strong> les analyser comme étant la résultante <strong>de</strong> compromis (« tra<strong>de</strong>-off ») entre les<br />

bénéfices que fournissent les MTI et les coûts qu’entraînent leur utilisation, conduisant à une<br />

optimisation <strong>de</strong> la fitness individuelle (Morris 2003). De plus, une telle interprétation permet<br />

<strong>de</strong> comprendre l’ensemble <strong>de</strong>s processus écologiques observés et leurs conséquences dans le<br />

fonctionnement d’une communauté (Kneitel & Chase 2004), et dans notre cas une<br />

communauté piscicole.<br />

De tels compromis ont lieu lorsqu’un trait physiologique, comportemental ou<br />

écologique confère <strong>de</strong>s avantages pour une fonction en même temps qu’il confère <strong>de</strong>s<br />

désavantages à une autre fonction (Chase & Leibold 2003). Dans le cadre <strong>de</strong> ce travail <strong>de</strong><br />

thèse, l’utilisation <strong>de</strong>s MTI par l’ichtyofaune confère, par exemple, <strong>de</strong>s avantages aux<br />

fonctions <strong>de</strong> reproduction <strong>de</strong>s adultes et <strong>de</strong> croissance <strong>de</strong>s juvéniles (comme cela a pu être<br />

démontré par Neckles et al. (1990) et Schlosser (1987)) mais aussi <strong>de</strong>s désavantages en<br />

terme <strong>de</strong> mortalité (risque <strong>de</strong> prédation) ou <strong>de</strong> maintien physiologique lors <strong>de</strong> la décrue<br />

progressive (comme cela a pu être démontré par Magoulick & Kobza (2003) et Matthews &<br />

Marsh-Matthews (2003)). Ainsi, la synthèse <strong>de</strong>s coûts et bénéfices, lors <strong>de</strong> la phase<br />

d’inondation, peut expliquer l’utilisation <strong>de</strong>s MTI et les processus d’émigration et<br />

d’immigration (Figures 13.4). Ces schémas se basent sur une différence <strong>de</strong> compromis entre<br />

les MTI et les milieux permanents qui évolue en fonction <strong>de</strong>s changements <strong>de</strong> conditions<br />

environnementales dans les MTI au cours du cycle d’inondation.<br />

230


a)<br />

b)<br />

c)<br />

Coûts<br />

Bénéfices<br />

Compromis<br />

(bénéfices – coûts)<br />

-<br />

MTI<br />

Milieu<br />

permanent<br />

Cycle d’inondation<br />

Milieu<br />

permanent<br />

MTI<br />

Cycle d’inondation<br />

+ Milieu<br />

permanent<br />

Immigration<br />

Emigration<br />

-<br />

MTI<br />

Cycle d’inondation<br />

Discussion générale et perspectives<br />

Lors <strong>de</strong> la pério<strong>de</strong> d‘inondation,<br />

les caractéristiques du MP sont considérées<br />

comme relativement stables (pointillés), alors<br />

qu’elles évoluent drastiquement dans le MTI<br />

(courbe pleine). Par exemple, au pic <strong>de</strong><br />

l’inondation, les risques <strong>de</strong> mortalité sont<br />

moindres dus à une profon<strong>de</strong>ur maximale<br />

(diminution <strong>de</strong>s risques <strong>de</strong> prédation) et à une<br />

qualité physico-chimique <strong>de</strong> l’eau maximale<br />

(maintien physiologique).<br />

En terme <strong>de</strong> bénéfices, les tendances sont<br />

inverses. Alors qu’ils restent relativement<br />

stables dans le milieu permanent (pointillés), ils<br />

évoluent drastiquement dans le MTI (courbe<br />

pleine). Au pic <strong>de</strong> l’inondation, les surfaces<br />

colonisables sont maximales et les ressources<br />

trophiques disponibles abondantes. Ces<br />

conditions se détériorent ensuite au cours <strong>de</strong><br />

l’exondation progressive <strong>de</strong>s MTI.<br />

Le compromis (vision simplifiée) conduit à une<br />

immigration dans les MTI par les poissons<br />

lorsque la synthèse <strong>de</strong>s coûts et <strong>de</strong>s bénéfices<br />

(courbe pleine) est supérieure à celle dans les<br />

milieux permanents (pointillés).<br />

Lors <strong>de</strong> l’exondation progressive <strong>de</strong>s MTI, cette<br />

tendance s’inverse et conduit à l’émigration <strong>de</strong>s<br />

MTI et à l’utilisation <strong>de</strong>s milieux permanents<br />

comme refuges estivaux.<br />

Figure 13.4. Représentation schématique <strong>de</strong> l’évolution <strong>de</strong>s coûts (a) et <strong>de</strong>s bénéfices (b) dans les<br />

MTI (courbe pleine) et les milieux permanents (pointillés) au cours <strong>de</strong> la pério<strong>de</strong> d’inondation.<br />

Cette évolution conduit à une sélection d’habitat variable au cours <strong>de</strong> ce cycle hydrologique (c),<br />

présentant le compromis où les bénéfices sont plus important que les coûts (+), puis l’inverse (-).<br />

Cependant, il a été précé<strong>de</strong>mment démontré qu’il existait <strong>de</strong> nombreuses sources <strong>de</strong><br />

variabilité relatives à l’utilisation <strong>de</strong>s MTI qui vont conduire à une modification <strong>de</strong> ce<br />

compromis en fonction <strong>de</strong> certains paramètres. Nous allons donc voir à travers <strong>de</strong>ux exemples<br />

comment ce compromis, défini pour <strong>de</strong>ux espèces, peut évoluer et quelles en sont les<br />

conséquences sur l’utilisation <strong>de</strong>s MTI par l’ichtyofaune.<br />

Prenons d’abord le cas le cas <strong>de</strong> <strong>de</strong>ux espèces, l’une native (par exemple l’anguille) et<br />

l’autre introduites (par exemple le poisson chat), qui présentent une tolérance physiologique<br />

différente, i.e. plus élevée pour l’introduite. Ces <strong>de</strong>ux espèces utilisent les MTI comme zone<br />

<strong>de</strong> croissance et se nourrissent <strong>de</strong>s mêmes items alimentaires (Keith & Allardi 2001). La<br />

courbe <strong>de</strong>s bénéfices peut donc être considérée comme semblable pour ces <strong>de</strong>ux espèces<br />

(Figure 13.5.a). Lors <strong>de</strong> l’inondation, les individus <strong>de</strong> ces espèces colonisent en même temps<br />

le MTI car elles possè<strong>de</strong>nt la même sensibilité aux températures basses (Aston & Brown 1978,<br />

231


Chapitre 13<br />

Arrignon 1998). Ensuite, lors <strong>de</strong> la phase d’exondation, la courbe <strong>de</strong> coûts <strong>de</strong> l’espèce native<br />

augmente plus fortement que l’espèce introduite, en raison <strong>de</strong> sa plus faible tolérance à la<br />

dégradation <strong>de</strong>s conditions environnementales (Figure 13.5.b). En terme d’utilisation <strong>de</strong>s MTI,<br />

les <strong>de</strong>ux espèces immigrent en même temps mais l’espèce introduite profitera plus longtemps<br />

que l’espèce native <strong>de</strong> la ressource trophique plus élevée dans le MTI que dans le milieu<br />

permanent car elle émigrera plus tardivement (Figure 13.5.c). In fine, les individus<br />

appartenant à l’espèce introduite bénéficieront certainement <strong>de</strong> performances <strong>de</strong> croissance<br />

supérieures à celles <strong>de</strong> l’espèce native. Ils pourraient ainsi optimiser leur fitness individuelle<br />

ce qui pourrait avoir <strong>de</strong>s conséquences sur le fonctionnement <strong>de</strong> la communauté.<br />

a)<br />

Bénéfices<br />

b)<br />

Coûts<br />

c)<br />

Compromis<br />

(bénéfices – coûts)<br />

a = b<br />

a > b<br />

a<br />

Cycle d’inondation<br />

Cycle d’inondation<br />

b<br />

Cycle d’inondation<br />

L’espèce a (native, trait fin) et l’espèce b<br />

(introduite, trait gras) utilisent toutes les <strong>de</strong>ux<br />

les MTI comme zone <strong>de</strong> croissance. De plus,<br />

ces <strong>de</strong>ux espèces se nourrissent <strong>de</strong>s mêmes<br />

items alimentaires. On peut émettre<br />

l’hypothèse que l’abondance <strong>de</strong> ces proies suit<br />

le même profil que l’abondance <strong>de</strong> zooplancton<br />

dans les MTI. Alors les profils <strong>de</strong> bénéfices<br />

obtenus dans un type donné <strong>de</strong> MTI sont<br />

semblables pour les <strong>de</strong>ux espèces. Les<br />

ressources trophiques dans les milieux<br />

permanents sont constantes pendant<br />

l’inondation (pointillés).<br />

La courbe <strong>de</strong>s coûts suit le profil du niveau<br />

d’eau, qui est semblable à la qualité générale<br />

du milieu aquatique incluant la profon<strong>de</strong>ur, la<br />

température ou la concentration en dioxygène.<br />

Au cours <strong>de</strong> l’inondation, les courbes <strong>de</strong> coûts<br />

<strong>de</strong>s <strong>de</strong>ux espèces sont similaires. Mais<br />

pendant l’exondation progressive, l’espèce b<br />

(trait gras), plus tolérante, voit sa courbe <strong>de</strong><br />

coûts augmenter plus lentement que celle <strong>de</strong><br />

l’espèce a (trait fin).<br />

Les <strong>de</strong>ux espèces colonisent le MTI en même<br />

temps mais l’espèce b (trait gras) émigre plus<br />

tardivement que l’espèce a (trait fin). Elle<br />

profite ainsi plus longtemps <strong>de</strong>s conditions<br />

trophiques supérieures dans le MTI que dans<br />

le milieu permanent.<br />

Figure 13.5. Représentation schématique <strong>de</strong> l’évolution <strong>de</strong>s bénéfices (a) et <strong>de</strong>s coûts (b) dans les<br />

MTI (courbe pleine) et les milieux permanents (pointillés) au cours <strong>de</strong> la pério<strong>de</strong> d’inondation<br />

pour une espèce introduite tolérante (trait gras) et une espèce native moins tolérante (trait fin).<br />

Cette évolution conduit à une sélection d’habitat variable entre les <strong>de</strong>ux espèces au cours <strong>de</strong> ce<br />

cycle hydrologique (c).<br />

232


Discussion générale et perspectives<br />

Analysons désormais la distribution spatiale <strong>de</strong>s individus <strong>de</strong>s <strong>de</strong>ux mêmes espèces<br />

dans un MTI. Partant <strong>de</strong> l’hypothèse que les proies (ici <strong>de</strong>s macro-invertébrés) sont<br />

abondantes et distribuées équitablement le long du gradient <strong>de</strong> distance au milieu permanent<br />

dans le MTI, une distance plus élevée <strong>de</strong> dispersion entraînerait une légère augmentation <strong>de</strong>s<br />

gains trophiques. En effet, la ressource est abondante, mais moins exploitée dans les zones<br />

éloignées du milieu permanent car les abondances <strong>de</strong> poissons sont plus faibles. (Figure<br />

13.6.a). Cependant, la dispersion <strong>de</strong> distances éloignées conduit à <strong>de</strong>s coûts croissants. Par<br />

exemple, le simple fait <strong>de</strong> se déplacer dans <strong>de</strong>s milieux peu profonds et très végétalisés, le<br />

risque <strong>de</strong> rester piégé dans <strong>de</strong>s mares isolées qui s’assècheront ensuite, ou <strong>de</strong> s’éloigner <strong>de</strong> la<br />

connexion avec les milieux permanents augmenteront les coûts. Mais l’évolution <strong>de</strong> ces coûts<br />

avec la distance est variable entre les <strong>de</strong>ux espèces, notamment parce qu’elles présentent <strong>de</strong>s<br />

tolérances physiologiques différentes. En effet, lors <strong>de</strong> l’exondation progressive, plus un<br />

individu est loin dans le MTI plus il sera soumis à un stress physiologique élevé et long<br />

(Figure 13.6.b). Ainsi, l’espèce introduite colonisera <strong>de</strong>s distances plus élevées que l’espèce<br />

native (Figure 13.6.c). Elle pourra ainsi bénéficier <strong>de</strong> ressources peu exploitées par le reste <strong>de</strong><br />

la communauté et donc maximiser la fitness <strong>de</strong> ses individus.<br />

a)<br />

Bénéfices<br />

b)<br />

c)<br />

Coûts Coûts<br />

Compromis<br />

(bénéfices – coûts)<br />

a<br />

a = b<br />

a > b<br />

Distance<br />

b<br />

Distance<br />

Distance<br />

L’espèce a (native, trait fin) et l’espèce b<br />

(invasive, trait gras) voient leur bénéfices<br />

augmenter légèrement au fur et à mesure<br />

qu’elles colonisent <strong>de</strong>s zones éloignées du<br />

milieu permanent. En effet, elles accè<strong>de</strong>nt à<br />

<strong>de</strong>s habitats <strong>de</strong> moins en moins exploités et<br />

dans lesquels les gains trophiques<br />

augmentent.<br />

Au fur et à mesure que la distance<br />

augmente, les coûts s’accroissent mais <strong>de</strong><br />

manière différente entre les <strong>de</strong>ux espèces.<br />

Le stress physiologique lors <strong>de</strong> l’exondation<br />

et les coûts seront d’autant plus élevés et<br />

longs que l’individu est loin du milieu<br />

permanent. L’espèce b (trait gras) tolérera<br />

mieux ce stress que l’espèce a (trait fin), et<br />

verra donc ces coûts plus limités.<br />

Par conséquence, l’espèce invasive b (trait<br />

gras), plus tolérante au stress physiologique<br />

associé à l’exondation, colonisera <strong>de</strong>s<br />

distances plus élevées que l’espèce native<br />

a (trait fin).<br />

Figure 13.6. Représentation schématique <strong>de</strong> l’évolution <strong>de</strong>s bénéfices (a) et <strong>de</strong>s coûts (b) en<br />

fonction <strong>de</strong> la distance colonisée par les poissons dans les MTI pour une espèce introduite<br />

tolérante (trait gras) et une espèce native moins tolérante (trait fin). Cette évolution conduit à<br />

distribution spatiale variable entre les <strong>de</strong>ux espèces au cours <strong>de</strong> ce cycle hydrologique (c).<br />

233


Chapitre 13<br />

Bien entendu, d’autres facteurs influencent les évolutions <strong>de</strong>s coûts et <strong>de</strong>s bénéfices<br />

mais les principaux résultats liés à l’utilisation spatiale et temporelle et issus <strong>de</strong> ce travail ont<br />

été présentés ici. On peut notamment citer le succès reproducteur lié à la qualité et à la<br />

quantité <strong>de</strong> substrat <strong>de</strong> ponte, la survie <strong>de</strong>s juvéniles liée à la disponibilité trophique, le risque<br />

<strong>de</strong> mortalité lié notamment à la prédation aviaire et bien d’autres qui influenceront ces<br />

compromis. Des étu<strong>de</strong>s plus fines sont donc désormais nécessaires pour en mesurer<br />

l’évolution dans le temps et dans l’espace afin <strong>de</strong> mieux quantifier les différents coûts et<br />

bénéfices et affiner ces courbes théoriques pour in fine améliorer la perception du<br />

fonctionnement d’une communauté piscicole dans les MTI.<br />

Une telle approche d’analyse <strong>de</strong>s compromis, issue <strong>de</strong> l’écologie évolutive et<br />

appliquée à l’étu<strong>de</strong> <strong>de</strong> la sélection <strong>de</strong> l’habitat (Morris 2003), nous permet <strong>de</strong> mieux<br />

interpréter l’utilisation spatiale et temporelle <strong>de</strong>s MTI par les différentes espèces <strong>de</strong> la<br />

communauté piscicole et donc <strong>de</strong> mieux en appréhen<strong>de</strong>r son fonctionnement. Cette approche<br />

s’appuie sur <strong>de</strong>s concepts issus notamment <strong>de</strong> la théorie <strong>de</strong>s jeux (voir synthèse dans<br />

Dugaktin & Reeve 1998) mais n’abor<strong>de</strong> pas les processus <strong>de</strong> <strong>de</strong>nsité-dépendance. Or ces<br />

processus influencent certainement l’ensemble <strong>de</strong>s paramètres évalués précé<strong>de</strong>mment (par<br />

exemple la disponibilité trophique en fonction <strong>de</strong> l’abondance <strong>de</strong> poissons, la qualité du<br />

substrat <strong>de</strong> ponte en fonction <strong>de</strong> son utilisation, etc...). Une prise en compte <strong>de</strong> ces<br />

phénomènes <strong>de</strong>nsité-dépendant permettrait d’affiner, à terme, la compréhension du<br />

fonctionnement <strong>de</strong> la communauté. Enfin, cette approche en terme <strong>de</strong> compromis est aussi un<br />

outil qui peut être utilisé <strong>de</strong> manière prédictive, afin d’envisager les changements <strong>de</strong> ce<br />

fonctionnement par rapport aux modifications <strong>de</strong>s paramètres qui le contrôlent. Nous<br />

reviendrons sur cet aspect à la fin du présent chapitre.<br />

234


Discussion générale et perspectives<br />

13.3. Fonctionnement <strong>de</strong> la communauté dans un écosystème sous fortes<br />

contraintes anthropiques : vers quelles mesures <strong>de</strong> gestion ?<br />

Cette partie <strong>de</strong> la discussion s’intéresse aux aspects <strong>de</strong> fonctionnement <strong>de</strong> la<br />

communauté piscicole dans le contexte d’un écosystème soumis à <strong>de</strong> fortes contraintes<br />

anthropiques pour analyser quelles propositions <strong>de</strong> gestion peuvent être réalisées afin<br />

d’optimiser le fonctionnement <strong>de</strong> cette communauté dans ce type d’écosystème.<br />

13.3.1. Efficacité <strong>de</strong>s mesures actuelles ?<br />

Comme <strong>de</strong> nombreux autres écosystèmes aquatiques soumis à <strong>de</strong> fortes contraintes<br />

anthropiques, les marais <strong>de</strong> Brière ont subi l’introduction <strong>de</strong> nombreuses espèces invasives<br />

animales et végétales, la construction <strong>de</strong> nombreux barrages et vannages conduisant à une<br />

forte altération du régime hydrologique naturel et une modification importante <strong>de</strong>s activités<br />

agricoles conduisant à une profon<strong>de</strong> modification paysagère. Les conséquences néfastes <strong>de</strong> ce<br />

type <strong>de</strong> modifications sur le fonctionnement <strong>de</strong>s écosystèmes aquatiques sont largement<br />

reconnues (e.g. Bunn & Arthington 2002, Byers 2002, Copp et al. 2005, García-Berthou et al.<br />

2005, Kolar & Lodge 2001, Marchetti et al. 2004, Vitousek et al. 1997, Wolter 2001). Nos<br />

résultats laissent penser que ces modifications ont eu <strong>de</strong>s conséquences notoires sur le<br />

fonctionnement <strong>de</strong> la communauté piscicole. Afin d’enrayer ces effets, <strong>de</strong>s mesures <strong>de</strong><br />

gestion <strong>de</strong> certaines populations d’intérêt majeur ont été mises en place localement.<br />

Nos résultats ont montré que la <strong>de</strong>struction systématique d’une espèce introduite<br />

(poisson chat) ne permet pas d’éradiquer les populations, même après plusieurs années d’une<br />

telle gestion. De plus, cette pratique connaît <strong>de</strong>s limites car son efficacité est souvent<br />

temporaire. Les poissons qui survivent à cette régulation présentent <strong>de</strong>s réponses adaptatives<br />

pour augmenter leur croissance, leur fécondité et leur capacité <strong>de</strong> rési<strong>lien</strong>ce (Wydoski &<br />

Wiley 1999). Nos résultats montrent <strong>de</strong> plus que les différentes espèces introduites tirent <strong>de</strong>s<br />

profits importants dans l’utilisation <strong>de</strong>s MTI, et plus particulièrement <strong>de</strong>s roselières, et une<br />

réflexion doit être engagée vis-à-vis <strong>de</strong> la réhabilitation <strong>de</strong>s habitats naturels, qui a prouvé<br />

son efficacité pour lutter contre certaines espèces introduites (Scoppettone et al. 2005).<br />

Une telle réflexion a été engagée, à petite échelle, dans les marais <strong>de</strong> Brière, il y a une<br />

trentaine d’années, principalement dans le but d’optimiser l’avifaune (Eybert et al. 1998).<br />

Ainsi, <strong>de</strong>s réserves ont été créées et nous avons pu démontré leur efficacité potentielle sur<br />

l’anguille européenne, espèce vulnérable et exploitée. La création <strong>de</strong> réserves serait une<br />

235


Chapitre 13<br />

mesure <strong>de</strong> gestion novatrice à l’échelle <strong>de</strong> l’aire distribution <strong>de</strong> cette espèce et pourrait venir<br />

compléter <strong>de</strong> manière intéressante l’ensemble <strong>de</strong>s mesures actuellement mises en place (voir<br />

revue dans Feunteun (2002)). En l’état actuel, cette mesure n’est cependant pas suffisante<br />

vis-à-vis <strong>de</strong>s consignes internationales et mériterait une adaptation. De plus, cela pourrait<br />

aussi bénéficier à d’autres espèces à forte intérêt écologique et qui sont peu exploitées<br />

(Apostolaki et al. 2002) mais aussi à d’autres espèces vulnérables, telles que le brochet<br />

(Rosell & MacOscar 2002).<br />

Nos résultats concernant le brochet, ont montré que les mesures <strong>de</strong> soutien <strong>de</strong><br />

population ont une efficacité limitée du fait <strong>de</strong> la forte mortalité subie par les juvéniles après<br />

leur lâcher. Ces résultats, obtenus sur une seule année, correspon<strong>de</strong>nt cependant aux valeurs<br />

généralement décrites dans la littérature (Lucchietta 1983, Gronkjaer et al. 2004, Bry et al.<br />

1995). Les très bas niveaux d’abondance <strong>de</strong> cette espèce dans les marais <strong>de</strong> Brière confirment<br />

l’efficacité très limitée <strong>de</strong> cette mesure <strong>de</strong> renforcement <strong>de</strong> population. De plus, les<br />

abondances d’individus sauvages capturés dans les MTI, comparées nettement plus faibles<br />

que celles <strong>de</strong> Carpentier et al. (2001), laissent penser que la population <strong>de</strong> brochet <strong>de</strong>s marais<br />

<strong>de</strong> Brière est actuellement faible, et que le soutien <strong>de</strong> population réalisé actuellement ne<br />

permet pas d’augmenter significativement ses effectifs.<br />

L’ensemble <strong>de</strong>s résultats obtenus ici souligne le fait que ces opérations <strong>de</strong> gestion sont,<br />

en l’état, insuffisamment efficaces. On peut donc conclure que le type <strong>de</strong> pratiques qui<br />

consiste à intervenir directement voire ponctuellement sur certaines espèces, ne permet pas<br />

<strong>de</strong> répondre <strong>de</strong> manière globale à la problématique <strong>de</strong> gestion d’une communauté piscicole<br />

dans ce type d’écosystème. Ainsi une gestion intégrative doit être mise en place en réponse<br />

aux facteurs qui génèrent le dysfonctionnement global <strong>de</strong> cette communauté et qui ont été<br />

i<strong>de</strong>ntifiés au cours <strong>de</strong> ce travail <strong>de</strong> thèse.<br />

13.3.2. Quels facteurs globaux <strong>de</strong> dysfonctionnement <strong>de</strong> l’écosystème ?<br />

Ce travail <strong>de</strong> thèse a aussi permis <strong>de</strong> mettre en évi<strong>de</strong>nce certains dysfonctionnements <strong>de</strong><br />

l’écosystème ayant <strong>de</strong>s sources différentes mais influençant <strong>de</strong> manière synergique le<br />

fonctionnement <strong>de</strong> la communauté piscicole. Ainsi, la fermeture du milieux par le roseau est<br />

un facteur facilitant l’établissement <strong>de</strong> plusieurs espèces introduites dont les effets négatifs sur<br />

l’ichtyofaune native sont nombreux et en partie i<strong>de</strong>ntifiés (voir les revues <strong>de</strong> Copp et al.<br />

(2005), García-Berthou et al. (2005)). En l’état actuel <strong>de</strong> la composition <strong>de</strong> la communauté<br />

piscicole, nous n’avons pas mis en évi<strong>de</strong>nce d’effets négatifs <strong>de</strong> l’envahissement par la<br />

236


Discussion générale et perspectives<br />

roselière sur l’ichtyofaune native. Cependant, on peut probablement penser que cela a pu se<br />

produire par le passé et avoir déjà eu un impact sur les espèces natives. Par exemple, il est<br />

vraisemblable que les caractéristiques hypoxiques et hyperthermiques <strong>de</strong>s roselières posent<br />

<strong>de</strong>s problèmes pour la reproduction <strong>de</strong> certaines espèces natives. Outre cet aspect, les effets<br />

du développement <strong>de</strong>s roselières, qui sont notamment <strong>de</strong>s milieux très favorables pour le<br />

poisson-chat, ne peuvent être contrebalancés par les tentatives <strong>de</strong> régulation <strong>de</strong> l’espèce par la<br />

pêcherie.<br />

A cet envahissement par la roselière est venu se greffer l’invasion par l’écrevisse <strong>de</strong><br />

Louisiane, dont les effets ne sont pas clairement mesurables. Cependant, les effets<br />

dévastateurs <strong>de</strong> cette espèce dans <strong>de</strong>s écosystèmes semblables aux marais <strong>de</strong> Brière ont déjà<br />

été nettement démontrés (voir notamment dans Rodríguez et al. 2005). Ainsi, cette espèce<br />

détruit notamment les végétaux aquatiques dans les milieux permanents (Rodríguez et al.<br />

2005) et exerce une prédation sur les juvéniles <strong>de</strong> poissons (Gutierrez-Yurrita et al. 1998).<br />

Dans les marais <strong>de</strong> Brière, on peut penser que cela rend les espèces phytolitophiles encore<br />

plus dépendantes <strong>de</strong>s MTI pour leur reproduction car ces zones sont les seules encore<br />

végétalisées. Cela augmente aussi la mortalité <strong>de</strong> certaines espèces vulnérables car cette<br />

prédation est certainement surexprimée par certaines pratiques halieutiques (par exemple, la<br />

concentration d’un grand nombre d’écrevisses et <strong>de</strong> juvéniles <strong>de</strong> brochet dans les bosselles).<br />

On peut donc émettre l’hypothèse que l’écrevisse <strong>de</strong> Louisiane peut accroître les effets<br />

négatifs (mortalité, déficit <strong>de</strong> croissance) <strong>de</strong> l’étiage sur l’ichtyofaune. En effet, la <strong>de</strong>struction<br />

<strong>de</strong> la végétation qui sert <strong>de</strong> refuge vis-à-vis <strong>de</strong>s prédateurs et fournit <strong>de</strong> la nourriture à <strong>de</strong><br />

nombreuses espèces (e.g. Rozas & Odum 1988, Gozlan et al. 1998, Grenouillet & Pont 2001)<br />

a certainement <strong>de</strong>s répercussions pour les poissons en terme <strong>de</strong> croissance et <strong>de</strong> survie estivale<br />

et on peut penser qu’il en a été ainsi en Brière. A cela, il convient <strong>de</strong> prendre en compte aussi<br />

d’autres facteurs (sévérité <strong>de</strong>s étiages ou sélectivité <strong>de</strong> la pêcherie) qui peuvent agir <strong>de</strong><br />

manière synergique en affectant principalement les espèces natives les moins tolérantes.<br />

Au cours <strong>de</strong> ce travail, nous avons vu qu’il existait aussi <strong>de</strong>s problèmes au niveau en<br />

terme <strong>de</strong> connectivité <strong>de</strong> l’écosystème. D’abord au niveau <strong>de</strong> la connectivité longitudinale,<br />

la présence <strong>de</strong> barrages et <strong>de</strong> vannages entraîne une rupture <strong>de</strong> la continuité hydrologique et<br />

biologique avec l’estuaire <strong>de</strong> la Loire qui limite drastiquement tous les types d’échanges. La<br />

communauté se trouve ainsi modifiée en terme <strong>de</strong> composition (notamment vis-à-vis <strong>de</strong>s<br />

espèces amphihalines) mais aussi en terme <strong>de</strong> fonctionnement (notamment vis-à-vis <strong>de</strong>s flux<br />

<strong>de</strong> matières et <strong>de</strong> nutriments). De plus, le cycle hydrologique, intégralement contrôlé par<br />

l’homme à l’ai<strong>de</strong> <strong>de</strong>s différents ouvrages hydrauliques, ne semble pas forcément être couplé<br />

237


Chapitre 13<br />

avec les différents cycles biologiques <strong>de</strong>s espèces. Cela est particulièrement vrai pour les<br />

espèces natives qui se reproduisent au printemps dans les MTI et entraîne certainement <strong>de</strong>s<br />

problèmes <strong>de</strong> recrutements pour <strong>de</strong> nombreuses espèces (King et al. 2003). Ensuite au niveau<br />

<strong>de</strong> la connectivité latérale, nous avons vu que la faible connectivité entre milieux permanents<br />

et MTI était un facteur limitant qui affecte fortement la colonisation et la distribution spatiale<br />

<strong>de</strong>s poissons dans les MTI, car les conditions environnementales dans ces milieux restent<br />

constantes en fonction du gradient <strong>de</strong> distance au milieu permanent. De plus, ce facteur<br />

limitant est surexprimé dans les marais <strong>de</strong> Brière par la présence <strong>de</strong> bourrelets <strong>de</strong> curage (cf.<br />

Figure 2.5) et d’un nombre très limité <strong>de</strong> points <strong>de</strong> connexion entre les milieux permanents et<br />

les MTI par rapport au linéaire total <strong>de</strong> canaux. Ainsi, <strong>de</strong>s points <strong>de</strong> connexion très éloignés<br />

‘stérilisent’ pour l’ichtyofaune <strong>de</strong> gran<strong>de</strong>s surfaces <strong>de</strong> MTI, qui ne peuvent être exploitées par<br />

l’ichtyofaune car elles sont trop distantes <strong>de</strong>s milieux permanents, alors qu’elles sont<br />

potentiellement très attractives.<br />

13.3.3. Quelles voies <strong>de</strong> gestion à promouvoir ?<br />

Compte tenu <strong>de</strong>s différents éléments qui ont été précé<strong>de</strong>mment évoqués, il est donc nécessaire<br />

<strong>de</strong> promouvoir une gestion intégrative <strong>de</strong> ce type d’écosystème et <strong>de</strong> sa communauté<br />

piscicole en agissant directement sur le milieu. Une telle gestion doit être développée au<br />

regard du cadre conceptuel qui régit le fonctionnement <strong>de</strong> tels écosystèmes (cf. Chapitre 1).<br />

Clairement, l’une <strong>de</strong>s alternatives à proposer est d’avoir recours à <strong>de</strong>s opérations <strong>de</strong><br />

réhabilitation <strong>de</strong>s caractéristiques structurelles (i.e. systèmes prairiaux) et fonctionnelles<br />

(i.e. connectivité et accessibilité) initiales <strong>de</strong>s MTI, du régime hydrologique naturel (i.e.<br />

hydropério<strong>de</strong> et couplage aux régimes biologiques) et <strong>de</strong> la continuité <strong>de</strong> l’écosystème (i.e.<br />

connectivité hydrologique et biologique avec l’amont et l’aval), tout en prenant en compte les<br />

différentes activités anthropiques et plus particulièrement le rôle <strong>de</strong> la pêcherie dans le<br />

fonctionnement <strong>de</strong> la communauté. La restauration <strong>de</strong>s caractéristiques structurelles et<br />

fonctionnelles <strong>de</strong>s habitats, et notamment <strong>de</strong>s MTI, a déjà montré son efficacité pour <strong>de</strong><br />

nombreux compartiments biologiques (Eybert et al. 1998, Carpentier et al. 2001) mais aussi<br />

pour lutter efficacement contre les espèces introduites (Scoppettone et al. 2005). Cette<br />

restauration passe aussi par une prise en compte <strong>de</strong> certains principes issus <strong>de</strong> l’écologie du<br />

paysage, et notamment par l'amélioration <strong>de</strong> la connectivité physique et biologique entre ces<br />

MTI et les milieux permanents. Les matériaux <strong>de</strong> curage qui sont actuellement stockés le long<br />

238


Discussion générale et perspectives<br />

<strong>de</strong>s canaux et qui constituent une barrière physique importante à la diffusion <strong>de</strong>s poissons<br />

dans les MTI lors du cycle d’inondation doivent être gérés différemment.<br />

La prise en compte du « Flood Pulse Concept » est certainement pertinente et<br />

intéressante à développer lors <strong>de</strong> programmes <strong>de</strong> restauration (Middleton 2000, Middleton<br />

2002, Juradja et al. 2004). En effet, cela permet <strong>de</strong> comprendre le rôle fonctionnel <strong>de</strong>s<br />

échanges latéraux et d’optimiser les programmes <strong>de</strong> restauration au regard <strong>de</strong>s différentes<br />

activités anthropiques (Tockner et al. 2000, Angeler & Alvarez-Cobelas 2005). La prise en<br />

compte <strong>de</strong> ce concept conduit par conséquent à rétablir la dynamique hydrologique<br />

naturelle et ses <strong>de</strong>grés d’imprédictibilités <strong>de</strong> l’évolution saisonnière <strong>de</strong>s niveaux d’eau<br />

(fréquence, amplitu<strong>de</strong> et durée d’inondation) pour maintenir un certain niveau d’intégrité<br />

écologique pour ces écosystèmes (Tockner & Stanford 2002). Trexler et al. (2005) ont ainsi<br />

démontré que, dans les Evergla<strong>de</strong>s, une telle gestion a permis la modification <strong>de</strong> la structure<br />

<strong>de</strong> la communauté piscicole et le rétablissement <strong>de</strong> sa dynamique naturelle. De manière<br />

encore plus probante, Marchetti & Moyle (2001) ont démontré qu’une telle mesure, appliquée<br />

dans une rivière régulée, permet <strong>de</strong> mieux gérer les espèces natives et <strong>de</strong> contrôler les espèces<br />

introduites. La prise en compte du « River Continuum Concept » développé par Vannote et al.<br />

(1980) permet <strong>de</strong> comprendre en quoi les différents secteurs d’un écosystème aquatique sont<br />

dépendant <strong>de</strong>s flux biologiques et <strong>de</strong> nutriments. Cette notion fondamentale <strong>de</strong> continuité<br />

graduelle peut aussi être appliquée aux écosystèmes lentiques (e.g. Euliss et al. 2004). La<br />

restauration d’une telle continuité est fondamentale car elle permet <strong>de</strong> rétablir les échanges<br />

longitudinaux fondamentaux pour son fonctionnement et plus particulièrement ceux avec<br />

l’écosystème estuarien adjacent.<br />

Enfin, toutes ces mesures doivent nécessairement être accompagnées par une gestion<br />

pérenne <strong>de</strong> la pêcherie. En effet, la conservation <strong>de</strong>s pêcheries et <strong>de</strong> l’ichtyofaune<br />

continentales est l’un <strong>de</strong>s grands challenges <strong>de</strong> la gestion durable <strong>de</strong>s écosystèmes aquatiques<br />

continentaux (Arlinghaus et al. 2002). Les pêcheries ont en effet <strong>de</strong> fortes valeurs<br />

économiques et socioculturelles et engendrent <strong>de</strong> multiples bénéfices à la société mais aussi à<br />

<strong>de</strong> nombreux écosystèmes (Arlinghaus et al. 2002, Cowx & Ger<strong>de</strong>aux 2004). Ainsi, la gestion<br />

<strong>de</strong> la pêcherie doit être intégrée à la gestion globale <strong>de</strong> l’écosystème et <strong>de</strong> la communauté<br />

piscicole, pouvant ainsi servir ‘d’outil d’appoint’, notamment pour la régulation <strong>de</strong>s espèces<br />

introduites et la gestion <strong>de</strong>s MTI, mais qui, <strong>de</strong> toute façon, ne <strong>de</strong>vra agir qu’en synergie avec<br />

les autres mesures proposées.<br />

Sur la base <strong>de</strong> cette réflexion intégrative, il est possible <strong>de</strong> dresser <strong>de</strong>s prédictions sur<br />

le fonctionnement <strong>de</strong> la communauté piscicole et <strong>de</strong>s espèces d’intérêt majeur (Tableau 13.2).<br />

239


Chapitre 13<br />

L’ensemble <strong>de</strong> ces prédictions convergent vers une amélioration du fonctionnement <strong>de</strong> la<br />

communauté d’espèces natives dulçaquicoles et amphihalines, à l’inverse <strong>de</strong>s espèces<br />

introduites. In fine, cela pourrait permettre <strong>de</strong> rétablir la dynamique naturelle <strong>de</strong> cette<br />

communauté. Cependant, certains problèmes liés par exemple à l’écrevisse restent difficiles à<br />

gérer et <strong>de</strong>s travaux plus précis doivent être entrepris. De plus, Harding et al. (1998) ont<br />

démontré que les effets <strong>de</strong>s pratiques anthropiques anciennes ont une longue rési<strong>lien</strong>ce pour le<br />

fonctionnement <strong>de</strong> l’écosystème.<br />

Tableau 13.2. Synthèse <strong>de</strong>s possibles effets <strong>de</strong>s mesures <strong>de</strong> gestion <strong>de</strong> l’habitat proposées sur les<br />

différents groupes d’espèces <strong>de</strong> la communauté piscicole. + : effet positif, - : effet négatif, = : pas<br />

d’effet, ?: effet inconnu, ↑ : augmentation et ↓ : diminution.<br />

Mesures <strong>de</strong> gestion Effets Native<br />

(brochet)<br />

Groupe d’espèces<br />

Introduite<br />

(poisson chat)<br />

Amphihaline<br />

(anguille)<br />

Réhabilitation<br />

MTI<br />

↑ <strong>de</strong>s surfaces prairiales<br />

↑ connectivité<br />

<strong>de</strong>s MTI<br />

+<br />

+<br />

-<br />

=<br />

?<br />

+<br />

Restauration ↓ <strong>de</strong> la prédictibilité + - ?<br />

du cycle<br />

hydrologique<br />

↑ <strong>de</strong> durée et amplitu<strong>de</strong><br />

<strong>de</strong>s inondations<br />

+ - ?<br />

Restauration<br />

<strong>de</strong> la continuité<br />

hydrologique<br />

↑ <strong>de</strong>s échanges<br />

biologiques<br />

↑ <strong>de</strong>s échanges<br />

<strong>de</strong> nutriments<br />

+<br />

+<br />

?<br />

+<br />

+<br />

+<br />

Régulation <strong>de</strong> ↑ sélectivité + - +<br />

la pêcherie ↑ réserves + ? +<br />

Le barrage <strong>de</strong> Méan, exutoire <strong>de</strong>s marais <strong>de</strong> Brière<br />

240


Discussion générale et perspectives<br />

13.4. Avenir <strong>de</strong> cette ichtyofaune dans le contexte <strong>de</strong>s changements globaux ?<br />

Au-<strong>de</strong>là d’une gestion intégrative <strong>de</strong> ce type d’écosystème sous fortes contraintes<br />

anthropiques, il est indispensable <strong>de</strong> raisonner aussi en fonction <strong>de</strong>s changements globaux<br />

dus aux activités humaines et <strong>de</strong>s conséquences présentes et surtout futures qu’ils auront sur<br />

le fonctionnement <strong>de</strong> ce type d’écosystème et <strong>de</strong> la communauté piscicole associée. Dans<br />

l’ensemble, ces changements pourraient conduire à différents scénarii, notamment la<br />

diminution d’aires <strong>de</strong> distribution d’espèces natives voire même l’extinction <strong>de</strong> certaines <strong>de</strong><br />

ces espèces ou, à l’inverse, l’augmentation <strong>de</strong> l’aire <strong>de</strong> distribution d’espèces introduites (e.g.<br />

Johnson & Evans 1990, Matthews & Zimmerman 1990, Poff 1991, Carpenter et al. 1992,<br />

Dukes & Mooney 1999, Jager et al. 1999, King et al. 1999, Morgan et al. 2001, Xenopoulos<br />

et al. 2005).<br />

De part leurs caractéristiques <strong>de</strong> zones transitionnelles entre écosystèmes terrestres et<br />

aquatiques, les MTI sont particulièrement sensibles et vulnérables aux changements globaux<br />

(Poff et al. 2002, Williams 2006). Les conséquences <strong>de</strong>s changements globaux sur les<br />

écosystèmes aquatiques sont nombreuses. Citons la modification <strong>de</strong>s flux <strong>de</strong> nutriments, la<br />

fragmentation <strong>de</strong> l’habitat, la diminution <strong>de</strong> la surface <strong>de</strong>s écosystèmes aquatiques,<br />

l’augmentation <strong>de</strong> la salinité liée à une plus forte évaporation, l’augmentation <strong>de</strong>s intrants, <strong>de</strong>s<br />

polluants, etc... (Williams 2006). Plus particulièrement, Poff et al. (2002) ont isolé trois<br />

paramètres qui risquent d’avoir <strong>de</strong>s conséquences très importantes pour ces écosystèmes, à<br />

savoir la modification du régime thermique, l’altération du régime hydrologique et<br />

l’introduction <strong>de</strong> nouvelles espèces. L’écologie a trop souvent été considérée comme une<br />

science qui possè<strong>de</strong> <strong>de</strong> faibles capacités prédictives mais le niveau actuel <strong>de</strong> connaissance <strong>de</strong>s<br />

processus écologiques permet <strong>de</strong> fournir aux gestionnaires certaines prédictions qualitatives<br />

(Poff 2002). C’est ce qui va être entrepris dans cette <strong>de</strong>rnière partie.<br />

Une <strong>de</strong>s premières conséquences fortes <strong>de</strong>s changements globaux est la modification<br />

du régime thermique et une augmentation générale <strong>de</strong> la température moyenne. Même si les<br />

prédictions concernant les valeurs <strong>de</strong> cette augmentation varient entre un et plusieurs <strong>de</strong>grés<br />

Celsius lors <strong>de</strong>s cent prochaines années, quasiment toutes s’orientent vers une augmentation<br />

significative <strong>de</strong> la température (voir par exemple la synthèse dans Morgan et al (2001)). A<br />

large échelle, une telle trajectoire risque <strong>de</strong> conduire à une modification <strong>de</strong>s aires <strong>de</strong><br />

distribution <strong>de</strong>s espèces et donc à un bouleversement <strong>de</strong> la structure <strong>de</strong>s communautés<br />

animales et végétales (Poff et al. 2002). A l’échelle locale, cette évolution du régime<br />

241


Chapitre 13<br />

thermique s’accompagne <strong>de</strong> l’arrivée <strong>de</strong> nouvelles espèces et la disparition d’autres. De même,<br />

une élévation <strong>de</strong> la température moyenne <strong>de</strong> l’eau peut affecter la durée du cycle <strong>de</strong> vie <strong>de</strong><br />

nombreuses espèces (croissance juvénile plus forte et maturité plus précoce) et donc<br />

influencer le fonctionnement <strong>de</strong> nombreuses populations piscicoles (Dembski et al. 2006,<br />

Gillet & Quetin 2006). D’après les résultats issus <strong>de</strong> ce travail <strong>de</strong> thèse, il semblerait que les<br />

espèces les plus sensibles à ces changements globaux soient les espèces natives, car les moins<br />

tolérantes à une température élevée et à l’ensemble <strong>de</strong>s changements physico-chimiques<br />

associés. En effet, nous avons montré ici que la tolérance physiologique est un <strong>de</strong>s facteurs<br />

forts conditionnant notamment la durée d’utilisation <strong>de</strong>s MTI lors <strong>de</strong> la phase d’exondation.<br />

Une augmentation <strong>de</strong> la température <strong>de</strong> l’eau conduit à une diminution <strong>de</strong> la<br />

production <strong>de</strong> phytoplancton qui risque d’entraîner une diminution <strong>de</strong> la ressource en<br />

zooplancton (Williams 2006) et donc une diminution <strong>de</strong>s bénéfices pour les poissons<br />

consommateurs <strong>de</strong> cette ressource. De même, <strong>de</strong>s températures <strong>de</strong> l’eau plus élevées peuvent<br />

encore réduire les surfaces <strong>de</strong> MTI utilisables par les poissons car les coûts liés à cette<br />

utilisation risquent d’augmenter en liaison avec l’accroissement du stress physiologique<br />

(Figure 13.7.a). Dans ces conditions, les coûts liés à l’exploitation <strong>de</strong> ces milieux pourraient<br />

augmenter <strong>de</strong> manière plus nette pour les espèces natives, moins tolérantes, que pour les<br />

espèces introduites, plus tolérantes (Figure 13.7.b). Ces modifications <strong>de</strong>s profils <strong>de</strong> coûts et<br />

<strong>de</strong> bénéfices ten<strong>de</strong>nt à prévoir une diminution <strong>de</strong> la durée <strong>de</strong> l’utilisation <strong>de</strong>s MTI par toutes<br />

les espèces. Cependant, cette diminution est là encore plus prononcée pour les espèces natives<br />

que pour les introduites (Figure 13.7.c). A terme, si les températures continuent à augmenter,<br />

le temps d’utilisation <strong>de</strong>s MTI diminuera <strong>de</strong> plus en plus. In fine, les espèces natives, qui<br />

utilisent principalement les MTI pour s’y reproduire, seront les plus fortement pénalisées.<br />

Bien entendu, ces prédictions supposent que les autres paramètres environnementaux<br />

conditionnant les valeurs <strong>de</strong>s coûts et <strong>de</strong>s bénéfices dans les MTI ainsi que dans les milieux<br />

permanents n’évoluent pas. On pourrait ainsi penser qu’une colonisation plus précoce <strong>de</strong>s<br />

MTI par l’ichtyofaune pourrait compenser les effets <strong>de</strong> cette augmentation <strong>de</strong> température.<br />

Cependant, cette phase <strong>de</strong> colonisation est aussi influencée par le régime hydrologique, et il<br />

n’est pas du tout sûr que les MTI seront inondés plus tôt dans la saison et pourront donc être<br />

exploités plus tôt par l’ichtyofaune.<br />

242


a)<br />

Bénéfices<br />

b)<br />

Coûts<br />

c)<br />

Compromis<br />

↑ T°C<br />

Phase d’exondation<br />

↑ T°C<br />

Phase d’exondation<br />

↑ T°C<br />

Phase d’exondation<br />

Discussion générale et perspectives<br />

Au cours <strong>de</strong> l’exondation, les bénéfices<br />

décroissent pour l’espèce native (trait fin)<br />

et introduite (trait gras) en réponse à la<br />

diminution <strong>de</strong> la surface inondée et <strong>de</strong>s<br />

ressources trophiques disponibles<br />

(situation actuelle, trait gris). Une<br />

augmentation <strong>de</strong> la température peut<br />

conduire à une diminution <strong>de</strong> cette<br />

productivité trophique et <strong>de</strong>s surfaces<br />

utilisables par l’ichtyofaune (simulation,<br />

trait noir) par rapport aux milieux<br />

permanents (pointillés).<br />

Au cours <strong>de</strong> l’exondation, les coûts<br />

augmentent, mais plus fortement pour<br />

une espèce native, peu tolérante (trait<br />

fin), que pour une espèce introduite, plus<br />

tolérante (trait gras). Une augmentation<br />

<strong>de</strong>s températures entraînera une<br />

augmentation <strong>de</strong> ces coûts (trait noir<br />

gras) et a priori plus fortement pour<br />

l’espèce la moins tolérante (trait gris fin).<br />

Cette augmentation <strong>de</strong> température<br />

conduira à une diminution <strong>de</strong> la durée <strong>de</strong><br />

l’utilisation <strong>de</strong>s MTI pour les <strong>de</strong>ux types<br />

d’espèces, et cette diminution risque<br />

d’être plus prononcée pour les espèces<br />

natives (trait fin) que pour les espèces<br />

introduites (trait gras).<br />

Figure 13.7. Prédiction <strong>de</strong>s conséquences d’une élévation <strong>de</strong>s températures sur l’utilisation <strong>de</strong>s<br />

MTI par une espèce native, peu tolérante (trait fin) et une espèce introduite, plus tolérante (trait<br />

gras), lors <strong>de</strong> la phase d’exondation progressive du milieu (approche exprimée en coûts, bénéfices<br />

et compromis). La situation actuelle est figurée en gris et la prédiction future en noir.<br />

En effet, les changements globaux risquent d’entraîner une augmentation <strong>de</strong> la fréquence<br />

<strong>de</strong>s évènements extrêmes (fortes inondations et étiages sévères), et ainsi influencer le rythme<br />

<strong>de</strong> la variabilité naturelle <strong>de</strong> nombreux écosystèmes aquatiques (Carpenter et al. 1992). A<br />

cette modification <strong>de</strong> la fréquence <strong>de</strong>s événements extrêmes, l’ensemble <strong>de</strong> la communauté<br />

scientifique s’accor<strong>de</strong> pour y ajouter une raréfaction globale <strong>de</strong> la ressource en eau liée à une<br />

diminution <strong>de</strong>s précipitations (Poff et al. 2002). Les écosystèmes qui souffriront le plus <strong>de</strong><br />

cette raréfaction sont les marais dont l’approvisionnement en eau se réalise surtout par<br />

précipitation (Poff et al. 2002). Là encore, d’après les résultats issus <strong>de</strong> ce travail <strong>de</strong> thèse, on<br />

peut prédire que cette modification du régime hydrologique va, <strong>de</strong> manière synergique, agir<br />

dans le même sens que l’élévation <strong>de</strong>s températures. En effet, les surfaces inondées vont être<br />

<strong>de</strong> moins en moins importantes, les étiages très forts et les inondations très brutales risquent<br />

243


Chapitre 13<br />

<strong>de</strong> ne pas permettre une exploitation correcte <strong>de</strong>s MTI par l’ichtyofaune. Là encore, les<br />

espèces natives qui utilisent surtout ces milieux pour la reproduction risquent <strong>de</strong> subir plus<br />

fortement, notamment pour l’accomplissement <strong>de</strong> certaines phases clés du cycle biologique,<br />

ces effets négatifs que les espèces introduites.<br />

Enfin, les changements globaux liés aux activités humaines auront comme conséquence<br />

<strong>de</strong> faciliter l’introduction <strong>de</strong> nouvelles espèces (Vitousek et al. 1997, Dukes & Mooney 1999)<br />

et l’expansion d’autres espèces introduites déjà installées (Simberloff & Von Holle 1999). Les<br />

conséquences <strong>de</strong> ces nouvelles introductions sont très difficiles à quantifier car <strong>de</strong> nouveaux<br />

assemblages d’espèces natives et introduites risquent fort <strong>de</strong> bouleverser le fonctionnement<br />

actuel <strong>de</strong> la communauté piscicole (Trexler et al. 2000, Poff et al. 2002).<br />

Un poisson chat (Ameiurus melas), une écrevisse <strong>de</strong> Louisiane (Procambarus clarkii) et un<br />

plant <strong>de</strong> jussie (Ludwigia uruguayensis) échantillonnés dans les marais <strong>de</strong> Brière<br />

244


Bibliographie générale<br />

Un canal <strong>de</strong>s marais <strong>de</strong> Brière jalonné par une roselière<br />

Chapitre 14


Chapitre 14<br />

Bibliographie générale<br />

Acou A., Boury P., Laffaille P., Crivelli A.J. & Feunteun E. 2005. Towards a standardized<br />

characterization of the potentially migrating silver European eel (Anguilla anguilla L.). Archiv für<br />

Hydrobiologie. 164: 237-255.<br />

Acou A., Gabriel G. & Feunteun E. In press. Differential production and condition indices of<br />

premigrant eels (Anguilla anguilla) in two small Atlantic coastal catchments of France. American<br />

Fisheries Society Symposium.<br />

Adam G. 1997. L’anguille européenne (Anguilla anguilla L. 1758): dynamique <strong>de</strong> la sous-population<br />

du lac <strong>de</strong> Grand-Lieu en relation avec les facteurs environnementaux et anthropiques. Thèse,<br />

<strong>Université</strong> <strong>de</strong> Toulouse III, France. 353 pp.<br />

Ailstock M.S., Norman C.M. & Bushmann P.J. 2001. Common reed Phragmites australis: control and<br />

effects upon biodiversity in freshwater nontidal wetlands. Restoration Ecology. 9: 49-59.<br />

Alcaraz C., Vila-Gispert A. & García-Berthou E. 2005. Profiling invasive fish species: the importance<br />

of phylogeny and human use. Diversity and Distributions. 11: 289-298.<br />

Amoros C. & Bornette G. 2002. Connectivity and biocomplexity in waterbodies of riverine<br />

floodplains. Freshwater Biology. 47: 761-776.<br />

Angeler D.G. & Alvarez-Cobelas M. 2005. Island biogeography and landscape structure: integrating<br />

ecological concepts in a landscape perspective of anthropogenic impacts in temporary wetlands.<br />

Environmental Pollution. 138: 420-424.<br />

Apostolaki P., Milner-Gulland E.J., McAllister M.K. & Kirkwood G.P. 2002. Mo<strong>de</strong>ling the effects of<br />

establishing a marine reserve for mobile fish species. Canadian Journal of Fisheries and Aquatic<br />

Sciences. 59: 405-415.<br />

Arlinghaus R., Mehner T. & Cowx I.G. 2002. Reconciling traditional inland fisheries management and<br />

sustainability in industrialized countries, with emphasis on Europe. Fish and Fisheries. 3: 261-316.<br />

Arrignon J. 1998. Aménagement Piscicole <strong>de</strong>s Eaux Douces. Lavoisier, Paris, France.<br />

Aston R.J. & Brown D.J.A. 1978. Fish farming in heated effluents. In: Pastakia C.M.R. (ed.)<br />

Proceeding of the conference on fish farming and wastes. Institute of Fisheries Management,<br />

London, U.K. pp. 39-62.<br />

Baber M.J., Chil<strong>de</strong>rs D.L., Babbitt K.J. & An<strong>de</strong>rson D.H. 2002. Controls on fish distribution and<br />

abundance in temporary wetlands. Canadian Journal of Fisheries and Aquatic Sciences. 59: 1441-<br />

1450.<br />

Balon E.K. 1975. Reproductive guilds of fishes: a proposal and <strong>de</strong>finition. Journal of the Fisheries<br />

Research Board of Canada. 32: 821-864.<br />

Baras E., Jeandrain D., Serouge B. & Philippart J.C. 1998. Seasonal variations in time and space<br />

utilization by radio-tagged yellow eels Anguilla anguilla (L.) in a small stream. Hydrobiologia.<br />

371/372: 187-198.<br />

246


Bibliographie générale<br />

Baras E., Malbrouck C., Houbart M., Kestemont P. & Mélard C. 2000. The effects of PIT tags on<br />

growth and physiology of age-0 cultured Eurasian perch Perca fluviatilis of variable size.<br />

Aquaculture. 185: 159-173.<br />

Barbin Zydlewski G.B., Haro A., Whalen G. & McCormick S.D. 2001. Performance of stationary and<br />

portable passive transpon<strong>de</strong>r <strong>de</strong>tection systems for monitoring of fish movements. Journal of Fish<br />

Biology. 58: 1471-1475.<br />

Beaudoin C.P., Tonn W.M., Prepas E.E. & Wassenaar L.I. 1999. Individual specialization and trophic<br />

adaptability of northern pike (Esox lucius): an isotope and dietary analysis. Oecologia. 120: 386-<br />

396.<br />

Bedford A.P. & Powell I. 2005. Long-term changes in the invertebrates associated with the litter of<br />

Phragmites australis in a managed reedbed. Hydrobiologia. 549: 267-285.<br />

Beintema A.J. 1997. European Black Tern (Childonias niger) in trouble: examples of dietary problems.<br />

Colonial Waterbirds. 20: 558-565.<br />

Beltz J.R., Johnson J.E., Cohen D.L. & Pratt F.B. 1974. An annotated bibliography of the effects of<br />

temperature on fish with special reference to the freshwater and anadromous species of New<br />

England. Research Bulletin Massachusetts Agricultural Experimental Station. University of<br />

Massachusetts, Amherst, U.S. 97 pp.<br />

Bernard J.Y. & Rolland R. 1990. Restoration of the “Gran<strong>de</strong> Brière Mottière” marsh overhauling’s<br />

project of the stretches of water. Bulletin d’Ecologie. 21: 15-19.<br />

Boët P. 1980. L’alimentation du poisson-chat (Ictalurus melas Raf.) dans le lac <strong>de</strong> Créteil. Annales <strong>de</strong><br />

Limnologie. 16: 255-270.<br />

Boët P. 2001. Le Poisson-chat Ameiurus melas (Rafinesque, 1820). In: Keith, P. & Allardi, J. (eds.).<br />

Atlas <strong>de</strong>s Poisons d’Eau Douce <strong>de</strong> France. Patrimoines naturels, MNHN, Paris, France. pp. 222-<br />

223.<br />

Bolnick D.I., Svanback R., Fordyce J.A., Yang L.H., Davis J.M., Hulsey C.D. & Foristeer M.L. 2003.<br />

The ecology of individuals: inci<strong>de</strong>nce and implications of individual specialization. The American<br />

Naturalist. 161: 1-28.<br />

Borcherding J., Bauerfeld M., Hintzen D. & Neumann D. 2002. Lateral migrations of fishes between<br />

floodplain lakes and their drainage channels at the Lower Rhine: diel and seasonal aspects. Journal<br />

of Fish Biology. 61: 1154-1170.<br />

Braig E.C. & Johnson D.L. 2003. Impact of black bullhead (Ameiurus melas) on turbidity in a diked<br />

wetland. Hydrobiologia. 490: 11-21.<br />

Broad T.L., Townsend C.R., Closs G.P. & Jellyman D.J. 2001. Microhabitat use by longfin eels in<br />

New Zealand streams with contrasting riparian vegetation. Journal of Fish Biology. 59: 1385-1400.<br />

Brocks M.A., Nielsen D.L., Shiel R.J., Green J.D. & Langley J.D. 2003. Drought and aquatic<br />

community resi<strong>lien</strong>ce: the role of eggs and seeds in sediments of temporary wetlands. Freshwater<br />

Biology. 48: 1207-1218.<br />

247


Chapitre 14<br />

Brown M.L., Willis D.W. & Blackwell B.G. 1999. Physiochemical and biological influences on black<br />

bullhead populations in Eastern South Dakota glacial lakes. Journal of Freshwater Ecology. 14:<br />

47-60.<br />

Bry, C. 1996. Role of vegetation in the life cycle of pike. In: Craig, J.F. (ed.). Pike: biology and<br />

exploitation. Chapman & Hall, London, U.K. pp. 45-67.<br />

Bry C., Basset E., Rognon X. & Bonamy, F. 1992. Analysis of sibling cannibalism among pike, Esox<br />

lucius, juveniles reared un<strong>de</strong>r semi-natural conditions. Environmental Biology of Fishes. 35: 75-84.<br />

Bry C., Bonamy F., Manelphe J. & Duranthon B. 1995. Early life characteristics of pike, Esox lucius,<br />

in rearing ponds: temporal survival pattern and ontogenetic diet shifts. Journal of Fish Biology. 46:<br />

99-113.<br />

Bry C., Hollebecq M.G., Ginot V., Israel G. & Manelphe J. 1991. Growth patterns of pike (Esox lucius<br />

L.) larvae and juveniles in small ponds un<strong>de</strong>r natural temperature regimes. Aquaculture. 97: 155-<br />

168.<br />

Bunn S.E. & Arthington A.H. 2002. Basic principles and ecological consequences of altered flow<br />

regime for aquatic biodiversity. Environmental Management. 30: 492-507.<br />

Byers J.E. 2002. Impact of non-indigenous species on natives enhanced by anthropogenic alteration of<br />

selection regimes. Oikos. 97: 449-458.<br />

Byers J.E. 2002. Physical habitat attribute mediates biotic resistance to non-indigenous species<br />

invasion. Oecologia. 130: 146-156.<br />

Capone T.A. & Kushlan J.A. 1991. Fish in dry-season stream pools. Ecology. 72: 983-992.<br />

Carpenter S.R., Fisher S.G., Grimm N.B. & Kitchell J.F. 2002. Global change and freshwater<br />

ecosystems. Annual Review of Ecology and Systematics. 23: 119-139.<br />

Carpentier A., Paillisson J.-M. & Damien J.-P. 2002. Evolution <strong>de</strong> l’intérêt écologique et fonctionnel<br />

<strong>de</strong>s cariçaies gyrobroyées pour la faune piscicole: comparaison avec d’autres zones inondables en<br />

Gran<strong>de</strong> Brière Mottière. Rapport <strong>Université</strong> <strong>de</strong> <strong>Rennes</strong> 1 – Parc naturel régional <strong>de</strong> Brière. 52 pp.<br />

Carpentier A., Paillisson J.-M., Damien J.-P. & Feunteun E. 2001. Gyrobroyage <strong>de</strong> cariçaies en<br />

Gran<strong>de</strong> Brière Mottière : nouvelles zones d’accueil pour la faune piscicole? Rapport <strong>Université</strong> <strong>de</strong><br />

<strong>Rennes</strong> 1 – Parc naturel régional <strong>de</strong> Brière. 22 pp.<br />

Carpentier A., Paillisson J.-M., Feunteun E. & Marion L. 2004. Fish community structure in<br />

temporary lowland floo<strong>de</strong>d grasslands. Bulletin Français <strong>de</strong> la Pêche et <strong>de</strong> la Pisciculture. 375: 1-<br />

14.<br />

Casanova M. & Brock M.A. 2000. How do <strong>de</strong>pth, duration and frequency of flooding influence the<br />

establishment of wetland plant communities. Plant Ecology. 147: 237-250.<br />

Casselman J.M. 1978. Effects of environmental factors on growth, survival, activity, and exploitation<br />

of northern pike. American Fisheries Society Special Publications. 11: 114-128.<br />

Casselman J.M. 1996. Age, growth and environmental requirements of pike. In: Craig, J. F. (ed.). Pike:<br />

biology and exploitation. Chapman & Hall, London, U.K. pp. 69-101.<br />

248


Bibliographie générale<br />

Chapman L.J. & Kramer D.L. 1991. The consequences of flooding for the dispersal and fate of<br />

poeciliid fish in an intermittent tropical stream. Oecologia. 87: 299-306.<br />

Chase J.M. & Leibold M.A. 2003. Ecological niches. Linking classical and contemporary approaches.<br />

University of Chicago Press, Chicago, U.S.<br />

Cherry D.S., Rodgers J.H., Cairns J., Dickson K.L. & Guthrie R.K. 1976. Responses of mosquitofish<br />

(Gambusia affinis) to ash effluent and thermal stress. Transaction of the American Fisheries<br />

Society. 105: 686-694.<br />

Chesson P. & Huntly N. 1997. The roles of harsh and fluctuating conditions in the dynamics of<br />

ecological communities. American Naturalist. 150: 519-553.<br />

Chesson P. 2000. General theory of competitive coexistence in spatially-varying environments.<br />

Theoretical Population Biology. 58: 211-237.<br />

Chick J.H., Ruetz C.R. & Trexler J.C. 2004. Spatial scale and abundance patterns of large fish<br />

communities in freshwater marshes of the Florida evergla<strong>de</strong>s. Wetlands. 3: 652-654.<br />

Conseil Supérieur <strong>de</strong> la Pêche. 1985. Etu<strong>de</strong> <strong>de</strong>s populations piscicoles <strong>de</strong>s marais du Brivet. Rapport<br />

du Ministère <strong>de</strong> l’Environnement – Conseil Supérieur <strong>de</strong> la Pêche. 14 p. + annexes.<br />

Constant P. 1972. L’avifaune <strong>de</strong> la Gran<strong>de</strong> Brière. Penn Ar Bed. 69: 296-303.<br />

Constant P., Bonnet P., Eybert M.-C. & Hédin J. 1987. Importance <strong>de</strong>s zones humi<strong>de</strong>s <strong>de</strong> la dépression<br />

brièronne pour l’avifaune aquatique : reproduction et estivage. Bulletin d’Ecologie. 18: 169-182.<br />

Copp G.H. 1989a. The habitat diversity and fish reproductive function of floodplain ecosystems.<br />

Environmental Biology of Fish. 26: 1-26.<br />

Copp G.H. 1989b. Electrofishing for fish larvae and juveniles: equipment modifications for increased<br />

efficiency with short fishes. Aquaculture and Fisheries Management. 20: 453-462.<br />

Copp G.H. 1993. The upper River Rhône revisited: an empirical mo<strong>de</strong>l of micohabitat use by 0+<br />

juvenile fishes. Folia Zoologica. 42: 329-340.<br />

Copp G.H., Bianco P.G., Bogutskaya N.G., Erős T., Falka I., Ferreira M.T., Fox M.G., Freyhof J.,<br />

Gozlan R.E., Grabowska J., Kováč V., Moreno-Amich R., Naseka A.M., Peňáz M., Povž M.,<br />

Przybylski M., Robillard M., Russell I.C., Stakėnas S., Šumer S., Vila-Gispert A. & Wiesner C.<br />

2005. To be, or not to be, a non-native freshwater fish? Journal of Applied Ichthyology. 21: 242-<br />

262.<br />

Copp G. & Penaz M. 1988. Ecology of spawning and nursery zones in the flood plain, using a new<br />

sampling approach. Hydrobiologia. 169: 209-224.<br />

Copp G.H. & Garner P. 1995. Evaluating microhabitat use of fish larvae and juveniles with Point<br />

Abundance Sampling. Folia Zoologica. 44: 145-158.<br />

Corti D.L., Kohler S.L. & Sparks R.E. 1999. Effects of hydroperiod and predation on a Mississippi<br />

River floodplain invertebrate community. Oecologia. 109: 154-165.<br />

249


Chapitre 14<br />

Courtenay W.R. Jr. & Meffe G.K. 1989. Small fishes in strange places: A review of introduced<br />

poeciliids. In: Meffe G.K. & Snelson F.F. Jr. (eds.). Ecology and Evolution of Livebearing fishes<br />

(Poeciliidae). Prentice Hall, Englewood Cliffs, New Jersey, U.S. pp. 319-331.<br />

Coutant C.C. 1977. Compilation of temperature preference data. Journal of the Fisheries Research<br />

Board of Canada. 34: 739-745.<br />

Cowardin L.M., Carter V., Golet F.C. & LaRoe E.T. 1979. Classification of wetlands and <strong>de</strong>epwater<br />

habitats of the United States. Fish and Wildlife Service, US Department of the Interior ed.,<br />

FWS/OBS-79/31, 103 p.<br />

Cowx I.G. 1996. Stock Assessment in Inland Fisheries. Fishing with electricity, applications in<br />

freshwater fisheries management. Fishing News Book. Blackwell Science, Oxford, U.K.<br />

Cowx I.G. & Ger<strong>de</strong>aux D. 2004. The effects of fisheries management practises on freshwater<br />

ecosystems. Fisheries Management and Ecology. 11: 145-151.<br />

Cowx I.G. & Lamarque P. 1990. Fishing with electricity, applications in freshwater fisheries<br />

management. Fishing News Book. Blackwell Science, Oxford, U.K.<br />

Crain P.K., Whitener K. & Moyle P.B. 2004. Use of restored central California floodplain by larvae of<br />

native and a<strong>lien</strong> fishes. American Fisheries Society Symposium. 39: 125-140.<br />

Cucherousset J., Eybert M.-C. & Feunteun E. In press. La faune piscicole, les milieux temporairement<br />

inondés et les marais <strong>de</strong> Brière. Æstuaria, sciences humaines et environnement.<br />

Cucherousset J., Ombredane D., Charles K., Marchand F. & Baglinière J.-L. 2005a. A continuum of<br />

life history tactics in a brown trout (Salmo trutta) population. Canadian Journal of Fisheries and<br />

Aquatic Sciences. 62: 1600-1610.<br />

Cucherousset J., Roussel J.-M., Keeler R., Cunjak R.A. & Stump R. 2005b. The use of two new<br />

portable 12-mm PIT tag <strong>de</strong>tectors to track small fish in shallow streams. North American Journal<br />

of Fisheries Management. 25: 270-274.<br />

Cunjak R. A., Roussel J.-M., Gray M.A., Dietrich J.P., Cartwright D.F., Munkittrick K.R. & Jardine<br />

T.D. 2005. Using stable isotope analysis with telemetry or mark-recapture data to i<strong>de</strong>ntify fish<br />

movement and foraging. Oecologia. 144: 636-646.<br />

Dalerum F. & Angerbjorn A. 2005. Resolving temporal variation in vertebrate diets using naturally<br />

occurring stable isotope. Oecologia. 144: 647-658.<br />

DeAngelis D.L., Trexler J.C. & Loftus W.F. 2005. Life history tra<strong>de</strong>-offs and community dynamics of<br />

small fishes in a seasonally pulsed wetland. Canadian Journal of Fisheries and Aquatic Sciences.<br />

62: 781-790.<br />

Dekker W. 2003. Did lack of spawners cause the collapse of the European eel, Anguilla anguilla?<br />

Fisheries Management and Ecology. 10: 365-376.<br />

Del Hoyo J., Elliott A. & Sargatal J. 1992. Handbook of the Birds of the World, volume 1, Lynx<br />

Edicions, Barcelona.<br />

250


Bibliographie générale<br />

Dembski S., Masson G., Monnier D., Wagner P. & Pihan J.C. 2006. Consequences of elevated<br />

temperatures on life-history traits of an introduced fish, pumpkinseed Lepomis gibbosus. Journal of<br />

Fish Biology. 69: 331-346.<br />

DeNiro M.J. & Epstein S. 1978. Influence of diet on the distribution of carbon isotopes in animals.<br />

Geochimica et Cosmochimica Acta. 42 : 495-506.<br />

Doadrio I. 2001. Atlas y Libro Rojo <strong>de</strong> los Peces Continentales <strong>de</strong> España. Dirección General <strong>de</strong><br />

Conservación <strong>de</strong> la Naturaleza, Madrid, Espana. 364 pp.<br />

Dufour E. & Ger<strong>de</strong>aux D. 2001. Apports <strong>de</strong>s isotopes stables ( 13 C/ 12 C, 15 N/ 14 N, 18 O/ 16 O, 36 S/ 34 S,<br />

87 Sr/ 86 Sr) aux étu<strong>de</strong>s écologiques sur les poissons. Cybium. 25: 369-382.<br />

Dugatkin L.A. & Reeve H.K. 1998. Game theory and animal behaviour. Oxford University Press,<br />

New York, U.S.<br />

Dukes J.S. & Mooney H.A. 1999. Does global change increase the success of biological inva<strong>de</strong>rs?<br />

Trends in Ecology and Evolution. 14: 135-139.<br />

Durif C., Dufour S. & Elie P. 2005. The silvering process of Anguilla anguilla: a new classification<br />

from the yellow resi<strong>de</strong>nt to the silver migrating stage. Journal of Fish Biology 66: 1025-1043.<br />

Dynesius M. & Nilsson C. 1994. Fragmentation and flow regulation of river systems in the northern<br />

3 rd of the world. Science. 266: 753-762.<br />

EELREP. 2005. Estimation of the reproduction capacity of European eel. Final report, EU contract<br />

Q5RS-2001-01836, 272 p.<br />

Elliott J.M. 1981. Some aspects of thermal stress on freshwater teleosts. In: Pickering, A.D. (ed.).<br />

Stress and Fish. Aca<strong>de</strong>mic Press, London, U.K. pp. 209-245.<br />

Elvira B. 1984. First records of the North American catfish Ictalurus melas (Rafinesque, 1920) (Pisces,<br />

Ictaluridae) in Spanish waters. Cybium. 8: 96-98.<br />

Elvira B., Nicola G.G. & Almodovar A. 1996. Pike and red swamp crayfish: a new case on predatory-<br />

prey relationship between a<strong>lien</strong>s in central Spain. Journal of Fish Biology. 48: 437-446.<br />

Euliss N.H., LaBaugh J.W., Fredrickson L.H., Mushet D.M., Laubhan M.K., Swanson G.A., Winter<br />

T.C., Rosenberry D.O. & Nelson R.D. 2004. The wetland continuum: a conceptual framework for<br />

interpreting biological studies. Wetlands. 24: 448-458.<br />

Evans D.O. 1977. Seasonal changes in standard metabolism, upper and lower thermal tolerance and<br />

thermoregulatory behaviour of the pumpkinseed, Lepomis gibbosus, Linnaeus. Doctoral<br />

Dissertation, Department of Zoology, University of Toronto, Toronto, Canada. 429 p.<br />

Eybert M.C., Bernard J.-Y., Constant P., Feunteun E., Hédin J. & Questiau S. 1998. Réhabilitation <strong>de</strong>s<br />

prairies inondables dans les marais brièrons: évolution <strong>de</strong> la flore, <strong>de</strong>s poissons et <strong>de</strong>s oiseaux.<br />

Gibier Faune Sauvage, Game Wildlife. 15: 999-1016.<br />

Eybert M.-C., Hédin J. & Feunteun E. 1997. Réserves <strong>de</strong> Gran<strong>de</strong> Brière Mottière (réserve du Nord et<br />

réserve <strong>de</strong>s Grands Charreaux): peuplements ornithologiques et ichtyologiques. Rapport <strong>Université</strong><br />

<strong>de</strong> <strong>Rennes</strong> 1 – Parc naturel régional <strong>de</strong> Brière. 90 p.<br />

251


Chapitre 14<br />

Feunteun, E. 2002. Management and restoration of European eel population (Anguilla anguilla): an<br />

impossible bargain. Ecological Engineering. 18: 575-591.<br />

Feunteun E. & Constant P. 1992. Le peuplement piscicole <strong>de</strong> la Gran<strong>de</strong> Brière Mottière. Inventaire<br />

dans la Réserve Nord. Rapport <strong>Université</strong> <strong>de</strong> <strong>Rennes</strong> 1 – Parc naturel régional <strong>de</strong> Brière. 13 p.<br />

Feunteun E., Acou A., Laffaille P. & Legault A. 2000. European eel (Anguilla anguilla): prediction of<br />

spawner escapement from continental population parameters. Canadian Journal of Fisheries and<br />

Aquatic Sciences. 57: 1627-1635.<br />

Feunteun E., Rigaud C., Elie P. & Lefeuvre J.C. 1992. Les marais doux endigués <strong>de</strong> Bourgneuf-<br />

Machecoul (Pays <strong>de</strong> Loire). Premiers éléments <strong>de</strong> connaissance du peuplement piscicole. Relation<br />

ichtyofaune-habitat et problèmes majeurs <strong>de</strong> gestion. Revue Française <strong>de</strong>s Sciences <strong>de</strong> l’Eau. 5:<br />

509-528.<br />

Fiers V., Gauvrit B.,Gavazzi P., Haffner H. & Maurin H. 1997. Statut <strong>de</strong> la faune <strong>de</strong> France<br />

métropolitaine. Statuts <strong>de</strong> protection, <strong>de</strong>grés <strong>de</strong> menace, statuts biologiques. Col. Patrimoines<br />

naturels. Service du Patrimoine / IEGB / MNHN, Réserves Naturelles <strong>de</strong> France, Ministère <strong>de</strong><br />

l’Environnement. vol.24. Paris, France.<br />

Fuller P.L., Nico L.G. & Williams J.D. 1999. Nonindigenous Fishes Introduced into Inland Waters of<br />

the United States. Special Publication 27. American Fisheries Society, Bethesda, Maryland, U.S.<br />

613 pp.<br />

Fustec E. & Lefeuvre J.-C. 2000. Fonctions et valeurs <strong>de</strong>s zones humi<strong>de</strong>s. Dunod, Paris, France. 426<br />

pp.<br />

Galarowicz T.L. & Wahl D.H. 2005. Foraging by a young-of-the-year piscivore: the role of predator<br />

size, prey type, and <strong>de</strong>nsity. Canadian Journal of Fisheries and Aquatic Sciences. 62: 2330-2342.<br />

Gante H.F. & Santos C.D. 2002. First records of the North American catfish Ameiurus melas in<br />

Portugal. Journal of Fish Biology. 61: 1643-1646.<br />

García-Berthou E., Alcaraz C., Pou-Rovira Q., Zamora L., Coen<strong>de</strong>rs G. & Feo C. 2005. Introduction<br />

pathways and establishments rates of invasive aquatic species in Europe. Canadian Journal of<br />

Fisheries and Aquatic Sciences. 62: 453-463.<br />

Garner P. 1997. Sample sizes for length and <strong>de</strong>nsity estimation of 0+ fish when using point sampling<br />

by electrofishing. Journal of Fish Biology. 50: 107-119.<br />

Gozlan R.E., St Hilaire S., Feist S.W., Martin P. & Kent M.L. 2005. Disease threats on European fish.<br />

Nature. 435: 1003-1136.<br />

Giattinna J.D. & Garton R.R. 1982. Graphical mo<strong>de</strong>l of thermoregulatory behaviour by fishes with a<br />

new measure of eurythermality. Canadian Journal of Fisheries and Aquatic Sciences. 39: 524-528.<br />

Gibson A.J. & Myers R.A. 2002. A logistic regression mo<strong>de</strong>l for estimating turbine mortality at<br />

hydroelectric generating stations. Transactions of the American Fisheries Society. 131: 623-633.<br />

Giles N., Wright R.M. & Nord M.E. 1986. Cannibalism in pike fry, Esox lucius L.: some experiments<br />

with fry <strong>de</strong>nsities. Journal of Fish Biology. 29: 107-113.<br />

252


Bibliographie générale<br />

Gillet C. & Quetin P. 2006. Effect of temperature changes on the reproductive cycle of roach in Lake<br />

Geneva from 1983 to 2001. Journal of Fish Biology. 69: 518-534.<br />

Gozlan R.E., Mastrorillo S., Dauba F., Tourenq J.-N. & Copp G.H. 1998. Multi-scale analysis of<br />

habitat use during late summer for 0+ fishes in the River Garonne (France). Aquatic Sciences. 60:<br />

99-117.<br />

Grenouillet G. & Pont D. 2001. Juvenile fishes in macrophyte beds: influence of food resources,<br />

habitat structure and body size. Journal of Fish Biology. 59: 939-959.<br />

Gronkjaer P., Skov C. & Berg S. 2004. Otolith-based analysis of survival and size-selective mortality<br />

of stocked 0+ year pike related to time of stocking. Journal of Fish Biology. 64: 1625-1637.<br />

Guevel B. 1997. La loi pêche (Co<strong>de</strong> Rural) et l’introduction <strong>de</strong>s espèces piscicoles. Bulletin Français<br />

<strong>de</strong> la Pêche et <strong>de</strong> la Pisciculture. 344/345: 43-51.<br />

Gutierrez-Yurrita P. J., Sancho G., Bravo M.A. & Montes C. 1998. Diet of the red swamp crayfish<br />

Procambarus clarkii in natural ecosystems of the Doñana National Park temporary fresh-water<br />

marsh (Spain). Journal of Crustacean Biology. 18: 120-127.<br />

Halliwell D.B., Langdon R.W., Daniels R.A., Kurtenbach J.P. & Jacobson R.A. 1999. Classification<br />

of freshwater fish species of the Northeastern United States for use in the <strong>de</strong>velopment of indices of<br />

biological integrity with regional applications. In: Simon T.P. (ed.). Assessing the Sustainability<br />

Biological Integrity of Water Resources using Fish Communities. CRC Press, New York, NY, U.S.<br />

pp. 301-338.<br />

Hanchin P.A., Willis D.W. & Hubers M.J. 2002a. Black bullhead growth in South Dakota waters:<br />

limnological and community influences. Journal of Freshwater Ecology. 17: 65-73.<br />

Hanchin P.A., Willis D.W. & St. Sauver T.R. 2002b. Comparison of concurrent trap-net and gill-net<br />

samples for black bullheads. Journal of Freshwater Ecology. 17: 233-237.<br />

Hanson D.A., Belonger B.J. & Schoenike D.L. 1983. Evaluation of a mechanical population reduction<br />

of black crappie and black bullheads in a small Wisconsin lake. North American Journal of<br />

Fisheries Management. 3: 41-47.<br />

Hanson M.A., Zimmer K.D., Butler M.G., Tangen B.A., Herwig B.R. & Euliss N.H. 2005. Biotic<br />

interactions as <strong>de</strong>terminants of ecosystem structure in prairie wetlands : an example using fish.<br />

Wetlands. 25: 764-775.<br />

Harding J.S., Benfield E.F., Bolstad P.V., Helfman G.S. & Jones E.B.D. 1998. Stream biodiversity:<br />

the ghost of land use past. Proceedings of the national Aca<strong>de</strong>my of Sciences of the United States of<br />

America. 95: 14843-14847.<br />

Hatase H., Takai N., Matsuzawa Y., Sakamoto W., Omuta K., Goto K., Arai N. & Fujiwara T. 2002.<br />

Size-related differences in feeding habitat use of adult female loggerhead turtles Caretta caretta<br />

around Japan <strong>de</strong>termined by stable isotope analyses and satellite telemetry. Marine Ecology<br />

Progress Series. 233: 273-281.<br />

253


Chapitre 14<br />

Hatzenbeler G.R., Bozek M.A., Jennings M.J. & Emmons E.E. 2000. Seasonal variation in fish<br />

assemblage structure and habitat structure in the nearshore littoral zone of Wisconsin lakes. North<br />

American Journal of Fisheries Management. 20: 360-368.<br />

Heermann L. & Borcherding J. 2006. Winter short-distance migration of juvenile fish between two<br />

floodplain water bodies of the Lower River Rhine. Ecology of Freshwater Fish. 15: 161-168.<br />

Hill M.S., Zydlewski G.B., Zydlewski J.D. & Gasvoda J.M. 2006. Development and evaluation of<br />

portable PIT tag <strong>de</strong>tection units: PITpacks. Fisheries Research. 77: 102-109.<br />

Hladik M. & Kubecka J. 2003. Fish migration between a temperate reservoir and its main tributary.<br />

Hydrobiologia. 504: 251-266.<br />

Hoeinghaus D.J., Layman C.A., Arrington D.A. & Winemiller K.O. 2003. Spatiotemporal variation in<br />

fish assemblage structure in tropical floodplain creeks. Environmental Biology of Fishes. 67: 379-<br />

387.<br />

Hohausova E. 2000. Exchange rate and small scale movements of fish between a river and its<br />

backwater. Archiv für Hydrobiologie. 4: 485-504.<br />

Hohausova E., Copp G.H. & Jankovsky P. 2003. Movement of fish between a river and its backwater:<br />

diel activity and relation to environmental gradients. Ecology of Freshwater Fish. 12: 107-117.<br />

Hugues G.M. 1984. Scaling of respiratory areas in relation to oxygen consumption of vertebrates.<br />

Experientia. 40: 519-652.<br />

Humphries P. & Baldwin D.S. 2003. Drought and aquatic ecosystems: an introduction. Freshwater<br />

Biology. 48: 1141-1146.<br />

ICES WGEEL. 2006. Report of the ICES/EIFAC Working Group on Eels. 23-27 January 2006, Rome,<br />

Italy.<br />

Inskip P.D. 1982. Habitat suitability in<strong>de</strong>x mo<strong>de</strong>ls: northern pike. U.S. Department of the Interior,<br />

Fish and Wildlife Service, FWS/OBS-82/10.10.17, Washington, D.C.<br />

Ivlev V.S. 1961. Experimental Ecology of the Feeding of Fishes. New Haven, CT: Yale University<br />

Press, U.S.<br />

Jackson D.A., Peres-Neto P.R. & Ol<strong>de</strong>n J.D. 2001. What controls who is where in freshwater<br />

communities: The roles of biotic, abiotic, and spatial factors. Canadian Journal of Fisheries and<br />

Aquatic Sciences. 58: 157-170.<br />

Jager H.I., Van Winkle W. & Holcomb B.D. 1999. Would hydrologic climate changes in Sierra<br />

Nevada streams influence trout persistence? Transactions of the American Fisheries Society. 128:<br />

222-240.<br />

Jardine T.D., Gray M.A., McWilliam S.M. & Cunjak R.A. 2005. Stable isotope variability in tissues of<br />

stream fishes. Transactions of the American Fisheries Society. 134: 1103-1110.<br />

Jardine T.D., McGeachy S.A., Paton C.M., Savoie M. & Cunjak R.A. 2003. Stable isotope in aquatic<br />

systems: sample preparation, analysis and interpretation. Canadian Manuscript Report of Fisheries<br />

and Aquatic Sciences. No. 2656: 39 p.<br />

254


Bibliographie générale<br />

Jellyman D.J. & Sykes J.R.E. 2003. Diel and seasonal movements of radio-tagged freshwater eels<br />

Anguilla spp., in two New Zealand streams. Environmental Biology of Fishes. 66: 143-154.<br />

Johnson T.B. & Evans D.O. 1990. Size-<strong>de</strong>pen<strong>de</strong>nt winter mortality of young-of-the-year white perch:<br />

climate warming and invasion of the Laurentian Great Lakes. Transactions of the American<br />

Fisheries Society. 119: 301-313.<br />

Johnson L.B. & Gage S.H. 1997. Landscape approaches to the analyses of aquatic ecosystems.<br />

Freshwater Biology. 37: 113-132.<br />

Jonsson N., Naesje T.F., Jonsson B., Saksgard R. & Sandlund O.T. 1999. The influence of piscivory<br />

on life history traits of brown trout. Journal of Fish Biology. 55: 1129-1141.<br />

Juanes F. 2003. The allometry of cannibalism in piscivorous fish. Canadian Journal of Fisheries and<br />

Aquatic Sciences. 60: 594-602.<br />

Juanes F., Letcher B. H. & Gries G. 2000. Ecology of stream fish: insights gained from an individual-<br />

based approach to juvenile Atlantic salmon. Ecology of Freshwater Fish. 9: 65-73.<br />

Junk W.J. & Wantzen K.M. 2004. The flood pulse concept: new aspects, approaches and applications<br />

– an update. In: Welcomme R.L. & Petr T. (eds.). Proceedings of the second international<br />

symposium on the management of large rivers for fisheries, Volume 2. Food and Agriculture<br />

Organization & Mekong River Commission. FAO Regional Office for Asia and the Pacific,<br />

Bangkok. RAP Publication 2004/16, pp. 117-149.<br />

Junk WJ., Bayley P.B. & Sparks R.E. 1989. The flood pulse concept in riverfloodplain systems. In:<br />

Didge D.P. (ed.). Proceedings of the International Large River Symposium (Lars). Canadian<br />

Special Publication of Fisheries and Aquatic Sciences, Ottawa, Canada. pp. 110-127.<br />

Juradja P., Ondrackova M. & Reichard M. 2004. Managed flooding as a tool for supporting the natural<br />

fish reproduction in man-ma<strong>de</strong> lentic water bodies. Fisheries Management and Ecology. 11: 237-<br />

242.<br />

Keith P. 2000. The part played by protected areas in conservation of threatened French freshwater fish.<br />

Biological Conservation. 92: 265-273.<br />

Keith P. & Allardi J. 2001. Atlas <strong>de</strong>s poissons d’eau douce <strong>de</strong> France. Patrimoines Naturels. 47: 387 p.<br />

King A.J., Humphries P. & Lake P.S. 2003. Fish recruitment on floodplains: the roles of patterns of<br />

flooding and life history characteristics. Canadian Journal of Fisheries and Aquatic Sciences. 60:<br />

773-786.<br />

King J.R., Shuter B.J. & Zimmerman A.P. 1999. Empirical links between thermal habitat, fish growth<br />

and climate change. Transactions of the American Fisheries Society. 128: 656-665.<br />

Klaar M., Copp G.H. & Horsfield R. 2004. Autumnal habitat use of non-native pumpkinseed Lepomis<br />

gibbosus and associations with native fish species in small English streams. Folia Zoologica. 53:<br />

189-202.<br />

Kneithel J.M. & Chase J.M. 2004. Tra<strong>de</strong>-offs in community ecology: linking spatial scales and species<br />

coexistence. Ecology Letters. 7: 69-80.<br />

255


Chapitre 14<br />

Knights B. & White E. 1998. Enhancing immigration and recruitment of eels: the use of passes and<br />

associated trapping systems. Fisheries Management and Ecology. 5: 459-471.<br />

Knights B., White E. & Naismith I.A. 1996. Stock assessment of European eel, Anguilla anguilla L. In<br />

Cowx I.G. (ed.). Stock Assessment in Inland Fisheries. Fishing News Books, Oxford, U.K. pp. 431-<br />

447.<br />

Kobza R.M., Trexler J.C., Loftus W.F. & Perry S.A. 2004. Community structure of fishes inhabiting<br />

aquatic refuges in a threatened Karst wetland and its implications for ecosystem management.<br />

Biological Conservation. 116: 153-165.<br />

Kolar C.S. & Lodge D.M. 2001. Progress in invasion biology: predicting inva<strong>de</strong>rs. Trends in Ecology<br />

and Evolution. 16: 199-204.<br />

Kramer D.L., Rangeley R.W. & Chapman L.J. 1997. Habitat selection: Patterns of spatial distribution<br />

from behavioural <strong>de</strong>cisions. In: Godin J-G. J. & Fitzgerald, G.J. (eds.). Behavioural Ecology of<br />

Fishes. Oxford University Press, Oxford, U.K. pp. 37-80.<br />

Kushlan J.A. 1974. Effects of a natural fish kill on the water quality, plankton, and fish population of a<br />

pond in the Big Cypress Swamp, Florida. Transactions of the American Fisheries Society. 103:<br />

235-243.<br />

Kushlan J.A. 1976. Environmental stability and fish community diversity. Ecology. 57: 821-825.<br />

Laffaille P., Acou A. & Guillouët J. 2005. The yellow European eel (Anguilla anguilla L.) may adopt<br />

a se<strong>de</strong>ntary lifestyle in inland freshwaters. Ecology of Freshwater Fish. 14: 191-196.<br />

Laffaille P., Briand C., Fatin D., Lafage D. & Lasne, E. 2005. Point sampling the abundance of<br />

European eel (Anguilla anguilla) in freshwater areas. Archiv für Hydrobiologie. 162: 91-98.<br />

Laffaille P., Pétillon J., Parlier E., Valéry L., Ysnel F., Radureau A., Feunteun E. & Lefeuvre, J.C.<br />

2005. Does the invasive plant Elymus athericus modify fish diet in tidal salt marshes? Estuarine,<br />

Coastal and Shelf Science. 65: 739-746.<br />

Lake P.S. 2003. Ecological effects of perturbation by drought in flowing waters. Freshwater Biology.<br />

48: 1161-1172.<br />

Le Louarn H. & Cloarec A. 1997. Insect predation on pike fry. Journal of Fish Biology. 50: 366-370.<br />

Lee C.E. 2002. Evolutionary genetics of invasive species. Trends in Ecology and Evolution. 17: 386-<br />

391.<br />

Legendre P. & Legendre L. 1998. Numerical Ecology. 2 nd English Edition. Elsevier, Amsterdam, The<br />

Netherlands.<br />

Leslie H. 2005. A synthesis of marine conservation planning approaches. Conservation Biology. 19:<br />

1701-1713.<br />

Lewis W.M. Jr. 1970. Morphological adaptations of cyprinodontoids for inhabiting oxygen <strong>de</strong>ficient<br />

waters. Copeia. 70: 319-326.<br />

Lobon-Cervìa J. & Utrilla C. 1993. A simple method to <strong>de</strong>termine stream trout (Salmo trutta L.)<br />

<strong>de</strong>nsities based on one removal fish electrofishing. Fisheries Research. 15: 369-373.<br />

256


Bibliographie générale<br />

Lobon-Cervìa J., Utrilla C.G. & P.A. Rincon. 1995. Variations in the population dynamics of the<br />

European eel Anguilla anguilla (L.) along the course of a Cantabrian river. Ecology of Freswater<br />

Fish. 4: 17-27.<br />

Louette G. & Declerck S. 2006. Assessment and control of the non-indigenous brown bullhead<br />

Ameiurus nebulosus populations using fyke nets in shallow ponds. Journal of Fish Biology. 68:<br />

522-531.<br />

Lowe-McConnell R.H. 1964. The fishes of the Rupununi savana district of British Guiana, South<br />

America. Part I. Ecological groupings of fish species and effects of the seasonal cycle on the fish.<br />

Journal of the Linnean Society 45: 103-144.<br />

Lowe-McConnell R.H. 1975. Fish Communities in Tropical Freshwaters. Longman, London, U.K.<br />

Lucas M.C. & Baras E. 2000. Methods for studying spatial behaviour of freshwater fishes in the<br />

natural environments. Fish and Fisheries. 1: 283-316.<br />

Lucas M.C. & Baras E. 2001. Migration of Freshwater Fish. Blackwell Science Ltd., Oxford, U.K.<br />

Lucchetta J.C. 1983. Utilisation <strong>de</strong> sites naturels pour le grossissement d’alevins <strong>de</strong> brochet produit en<br />

écloserie. In: Billard, R. (ed.). Le Brochet: Gestion dans le Milieu Naturel et Elevage. INRA<br />

Publication, Paris, France. pp. 199-208.<br />

Lundquist C.J. & Granek E.F. 2005. Strategies for successful marine conservation: integrating<br />

socioeconomic, political and scientific factors. Conservation Biology. 19: 1771-1778.<br />

Lusk S., Halacka K., Luskova V. & Horak V. 2001. Annual dynamics of the fish stock in a backwater<br />

of the River Dyje. Regulated Rivers: Research and Management. 17: 571-581.<br />

Lytle D.A. 2001. Disturbance regimes and life-history evolution. The American Naturalist. 157: 525-<br />

536.<br />

Lytle D.A. & Poff N.L. 2004. Adaptation to the natural flow regime. Trends in Ecology and Evolution.<br />

19: 94-100.<br />

Magoulick D.D. & Kobza R.M. 2003. The role of refugia for fish during drought: a review and<br />

synthesis. Freshwater Biology. 48: 1186-1198.<br />

Magoulick D.D. 2000. Spatial and temporal variation in fish assemblages of drying stream pools: the<br />

role of abiotic and biotic factors. Aquatic Ecology. 34: 29-41.<br />

Maillard Y. 1972. Poissons et problèmes piscicoles du bassin du Brivet. Penn ar Bed. 71: 386-399.<br />

Maitland P.S. 1995. The conservation of freshwater fish: past and present experience. Biological<br />

Conservation. 72: 259-270.<br />

Mann R.H.K. 1976. Observations on the age, growth, reproduction and food of the pike Esox lucius<br />

(L.) in two rivers in southern England. Journal of Fish Biology. 8: 179-197.<br />

Mann R.H.K. 1982. The annual food consumption and prey preferences of pike (Esox lucius) in the<br />

River Frome, Dorset. Journal of Animal Ecology. 51: 81-95.<br />

Marchetti M.P. & Moyle P.B. 2001. Effects of flow regime on fish assemblages in a regulated<br />

California stream. Ecological Applications. 11: 530-539.<br />

257


Chapitre 14<br />

Marchetti M.P., Moyle P.B. & Levine R. 2004. A<strong>lien</strong> fishes in California watersheds: characteristics<br />

of successful and failed inva<strong>de</strong>rs. Ecological Applications. 14: 587-596.<br />

Martin-Smith K.M. & Armstrong J. D. 2002. Growth rates of wild stream-dwelling Atlantic salmon<br />

correlate with activity and sex but not dominance. Journal of Animal Ecology. 71: 413-423.<br />

Masson C.F. & Bryant R.J. 1975. Production, nutrient content and <strong>de</strong>composition of Phragmites<br />

communis Trin. and Typha latifolia L. Journal of Ecology. 63: 71-95.<br />

Matthews B. & Mazum<strong>de</strong>r A. 2004. A critical evaluation of intrapopulation variation of and isotopic<br />

evi<strong>de</strong>nce of individual specialization. Oecologia. 140: 361-371.<br />

Matthews W.J. 1998. Patterns in Freshwater Fish Ecology. Chapman and Hall, New-York, U.S.<br />

Matthews W.J. & Marsh-Matthews E. 2003. Effects of drought on fish across axes of spaces, time and<br />

ecological complexity. Freshwater Biology. 48: 1232-1253.<br />

Matthews W.J. & Zimmerman E.G. 1990. Potential effects of global warming on native fishes of the<br />

Southern Great Plains and the Southwest. Fisheries. 15: 26-32.<br />

McCleave J.D. 2001. Simulation of the impact of dams and fishing weirs on reproductive potential of<br />

silver-phase American eels in the Kennebec River Basin, Maine. North American Journal of<br />

Fisheries Management. 21: 592-605.<br />

McKinsey D. & Chapman L.J. 1998. Dissolved oxygen and fish distribution in a Florida spring.<br />

Environmental Biology of Fishes. 53: 211-223.<br />

Middleton B. 2000. Wetland Restoration: Flood Pulsing and Disturbance Dynamics. John Wiley &<br />

Sons Inc., Chichester, U.K.<br />

Middleton B.A. 2002. The flood pulse concept in wetland restoration. In: Middleton B.A. (ed.). Flood<br />

Pulsing in Wetlands: Restoring the Natural Hydrological Balance. John Wiley & Sons Inc.,<br />

Chichester, U.K. pp. 1-10.<br />

Mills M.D., Ra<strong>de</strong>r R.B. & Belk M.C. 2004. Complex interactions between native and invasive fish:<br />

the simultaneous effects of multiple negative interactions. Oecologia. 141: 713-721.<br />

Morgan I.J., McDonald D.G. & Wood C.M. 2001. The cost of living for freshwater fish in a warmer,<br />

more polluted world. Global Change Biology. 7: 345-355.<br />

Morhardt J.E., Bishir D., Handlin C.I. & Mul<strong>de</strong>r S.D. 2000. A portable system for reading large<br />

passive integrated transpon<strong>de</strong>r tags from wild trout. North American Journal of Fisheries<br />

Management. 20: 276-283.<br />

Moriarty C. & Dekker W. 1997. Management of European eel fisheries. Second report of EC<br />

concerted action AIR A94-1939. Fisheries Bulletin Marine Institute, Vol. 15, Dublin, Ireland.<br />

Morris D.W. 2003. Toward an ecological synthesis: a case for habitat selection. Oecologia. 136: 1-13.<br />

Morrison W.E. & Secor D.H. 2003. Demographic attributes of yellow-phase American eels (Anguilla<br />

rostrata) in the Hudson River estuary. Canadian Journal of Fisheries and Aquatic Sciences. 60:<br />

1487-1501.<br />

258


Bibliographie générale<br />

Moyle P.B. & Light T. 1996. Biological invasions of fresh water: empirical rules and assembly theory.<br />

Biological Conservation. 78: 149-161.<br />

Mueller R.P., Moursund R.A. & Bleich M.D. 2006. Tagging juvenile lamprey with passive integrated<br />

transpon<strong>de</strong>rs: methodology, short-term mortality, and influence on swimming performance. North<br />

American Journal of Fisheries Management. 26: 361-366.<br />

Naismith I.A. & Knights B. 1990. Mo<strong>de</strong>lling of unexploited and exploited populations of eels,<br />

Anguilla anguilla (L.), in the Thames estuary. Journal of Fish Biology. 37: 975-986.<br />

Naismith I.A. & Knights B. 1990. Studies of sampling methods and of techniques for estimating<br />

populations of eels, Anguilla anguilla L. Aquaculture and Fisheries Management. 21: 357-367.<br />

Neckles H.A., Murkin H.R. & Cooper J.H. 1990. Influences of seasonal flooding on<br />

macroinvertebrates abundance in wetland habitats. Freshwater Biology. 23: 311-322.<br />

Nelva A., Persat H. & Chessel D. 1979. Une nouvelle métho<strong>de</strong> d’étu<strong>de</strong> <strong>de</strong>s peuplements<br />

ichthyologiques dans les grands cours d’eau par échantillonnage ponctuel d’abondance. Comptes<br />

Rendus <strong>de</strong> l’Académie <strong>de</strong>s Sciences. Série III, Sciences <strong>de</strong> la Vie. 289: 1295-1298.<br />

Nicolas Y. & Pont D. 1995. Importance d’annexes latérales artificielles pour le recrutement en<br />

juvéniles <strong>de</strong> poissons dans un fleuve aménagé, le Bas-Rhône. Bulletin Français <strong>de</strong> la Pêche et <strong>de</strong> la<br />

Pisciculture. 337/338/339: 249-257.<br />

Nilsson C., Reidy C.A., Dynesius M. & Revenga C. 2005. Fragmentation and flow regulation of the<br />

world’s large river systems. Science. 308: 405-408.<br />

Nilsson P.A. & Bronmark C. 2000. Prey vulnerability to a gape-sized limited predator: behavioural<br />

and morphological impacts on northern pike piscivory. Oikos. 88: 539-546.<br />

Noble R.A.A., Harvey J.P. & Cowx I.G. 2004. Can management of freshwater fish populations be<br />

used to protect and enhance the conservation status of a rare, fish-eating bird, the bittern, Botaurus<br />

stellaris, in the UK? Fisheries Management and Ecology. 11: 291-302.<br />

Oberdorff T., Pont D., Hugueny R. & Porcher J.-P. 2002. Development and validation of a fish-based<br />

in<strong>de</strong>x for the assessment of ‘river health’ in France. Freshwater Biology. 47: 1720-1734.<br />

Ol<strong>de</strong>n J.D., Jackson D.A. & Peres-Neto P.R. 2001. Spatial isolation and fish communities in drainage<br />

lakes. Oecologia. 127: 572-585.<br />

Ol<strong>de</strong>n J.D., Poff N.L. & Bestgen K.R. 2006. Life-history strategies predict fish invasions and<br />

extirpations in the Colorado River Basin. Ecological Monographs. 76: 25-40.<br />

Oliveira A.C.B., Soares M.G., Martinelli L.A. & Moreira M.Z. 2006. Carbon sources of fish in an<br />

Amazonian floodplain lake. Aquatic Sciences. 68: 229-238.<br />

Ombredane D., Baglinière J.-L. & Marchand F., 1998. The effects of passive integrated transpon<strong>de</strong>r<br />

tags on survival and growth of juvenile brown trout (Salmo trutta L.) and their use for studying<br />

movement in a small river. Hydrobiologia. 371/372: 99-106.<br />

259


Chapitre 14<br />

Ostrand K.G. & Wil<strong>de</strong> G.R. 2001. Temperature, dissolved oxygen, and salinity tolerances of five<br />

prairie stream fish and their role in explaining fish assemblage patterns. Transactions of the<br />

American Fisheries Society. 130: 742-749.<br />

Ostrand K.G. & Wil<strong>de</strong> G.R. 2002. Seasonal and spatial variation in a prairie stream-fish assemblage.<br />

Ecology of Freshwater Fish. 11: 137-149.<br />

Otto R.G. 1973. Temperature tolerance of the mosquitofish, Gambusia affinis (Baird and Girard).<br />

Journal of Fish Biology. 5: 575-585.<br />

Parc naturel régional <strong>de</strong> Brière (2000). Charte du Parc naturel régional <strong>de</strong> Brière, Saint-Joachim,<br />

France.<br />

Perrow M.R., Jowitt A.J.D. & Gonzalez L.Z. 1996. Sampling fish communities in shallow lowland<br />

lakes: point-sample electric fishing vs electric fishing within stop-nets. Fisheries Management and<br />

Ecology. 3: 303-313.<br />

Persat H. & Copp G.H. 1989. Electric fishing and point abundance sampling for the ichthyology of<br />

large rivers. In: Cowx, I. (ed.). Development in Electric Fishing. Blackwell Scientific Publications,<br />

Oxford, U.K. pp. 197-209.<br />

Peterson B.J. & Fry B. 1987. Stable isotopes in ecosystem studies. Annual Review of Ecology and<br />

Systematics. 18: 293-320.<br />

Pimentel D. 2005. Aquatic nuisance species in the New York State canal and Hudson River systems<br />

and the Great Lakes Basin: an economic and environmental assessment. Environmental<br />

Management. 35: 692-701.<br />

Poff N.L. 1991. Regional hydrologic response to climate change: an ecological perspective. In: Firth P.<br />

& Fisher S.G. (eds.). Global Climate Change and Freshwater Ecosystems. Springer-Verlag, New-<br />

York, U.S. pp. 88-115.<br />

Poff N.L. 2002. Ecological response to and management of increased flooding caused by climate<br />

change. Philosophical Transitions of the Royal Society A: Mathematical, Physical and Engineering<br />

Sciences. 360: 1497-1510.<br />

Poff N.L., Allan J.D., Bain M.B., Karr J.R., Prestegaard K.L., Richter B.D., Sparks R.E. & Stromberg<br />

J.C. 1997. The natural flow regime, a paradigm for river conservation and restoration. BioScience.<br />

47: 769-784.<br />

Poff N.L., Brinson M.M. & Day J.W. 2002. Freshwater coastal ecosystems and global change: a<br />

review of projected impacts for the United States. Pew Center on Global Change, Arlington, U.S.<br />

Poizat G. & Crivelli A.J. 1997. Use of seasonally floo<strong>de</strong>d marshes by fish in a Mediterranean wetland:<br />

timing and <strong>de</strong>mographic consequences. Journal of Fish Biology. 51: 106-119.<br />

Poizat G. & Pont D. 1996. Multi-scale approach to species-habitat relationships: juvenile fish in a<br />

large river section. Freshwater Biology. 36: 611-622.<br />

260


Bibliographie générale<br />

Pont D., Hugueny B. & Oberdorff T. 2005. Mo<strong>de</strong>lling habitat requirement of European fishes: do<br />

species have similar responses to local and regional environmental constraints? Canadian Journal<br />

of Fisheries and Aquatic Sciences. 62: 163-173.<br />

Pont D., Hugueny B., Beier U., Goffaux D., Melcher A., Noble R., Rogers C., Roset N. & Schmutz S.<br />

2006. Assessing river biotic condition at a continental scale: a European approach using functional<br />

metrics and fish assemblages. Journal of Applied Ecology. 43: 70-80.<br />

Post D.M. 2002. Using stable isotopes to estimate trophic position: mo<strong>de</strong>ls, methods, and assumptions.<br />

Ecology. 83: 703-718.<br />

Post D.M. 2003. Individual variation in the timing of ontogenetic niche shifts in largemouth bass.<br />

Ecology. 84: 1298-1310.<br />

Prentice E.F., Flagg T.A. & McCutcheon C.S. 1990. Feasibility of using implantable Passive<br />

Integrated Transpon<strong>de</strong>r (PIT) tags in salmonids. American Fisheries Society Symposium. 7: 317-<br />

322.<br />

Rehage J.S. & Trexler J.C. 2006. Assessing the net effect of anthropogenic disturbance on aquatic<br />

communities in wetlands: community structure relative to the distance from canals. Hydrobiologia.<br />

569: 359-373.<br />

Robb T. & Abrahams M.V. 2003. Variation in tolerance to hypoxia in a predator and prey species: an<br />

ecological advantage of being small. Journal of Fish Biology. 62: 1067-1081.<br />

Robinet T. & Feunteun E. 2002. Sublethal effects of exposure to chemical compounds: a cause for the<br />

<strong>de</strong>cline in Atlantic eels? Ecotoxicology. 11: 265-277.<br />

Rodríguez C.F., Bécares E., Fernán<strong>de</strong>z-Aláez M. & Fernán<strong>de</strong>z-Aláez C. 2005. Loss of diversity and<br />

<strong>de</strong>gradation of wetlands as a result of introducing exotic crayfish. Biological Invasions. 7: 75-85.<br />

Rosell R., Evans D. & Allen M. 2005. The eel fishery in Lough Neagh, Northern Ireland – an example<br />

of sustainable management? Fisheries Management and Ecology. 12: 377-385.<br />

Rosell R.S. & MacOscar K.C. 2002. Movements of pike, Esox lucius, in Lower Lough Erne,<br />

<strong>de</strong>termined by mark-recapture between 1994 and 2000. Fisheries Management and Ecology. 9:<br />

189-196.<br />

Roth E., Toivonen A.L., Navrud S., Bengtsson B., Godbergsson G., Tuunainen P., Appelblad H. &<br />

Weissglas G. 2001. Methodological, conceptual and sampling practices in surveying recreational<br />

fisheries in Nordic countries – experiences from a valuation survey. Fisheries Management and<br />

Ecology. 8: 355-367.<br />

Roussel J.-M., Cunjak R.A., Newbury R., Caissie D. & Haro A. 2004. Movements and habitat use by<br />

PIT-tagged Atlantic salmon parr in early winter: the influence of anchor ice. Freshwater Biology.<br />

49: 1026-1035.<br />

Roussel J.-M., Haro A. & Cunjak R.A. 2000. Field test of a new method for tracking small fishes in<br />

shallow rivers using passive integrated transpon<strong>de</strong>r (PIT) technology. Canadian Journal of<br />

Fisheries and Aquatic Sciences. 57: 1326-1329.<br />

261


Chapitre 14<br />

Rozas G. & Odum W.E. 1988. Occupation of submerged aquatic vegetation by fishes: testing the roles<br />

of food and refuge. Oecologia. 77: 101-106.<br />

Rubenstein D.R. & Hobson K.A. 2004. From birds to butterflies: animal movement patterns and stable<br />

isotope. Trends in Ecology and Evolution. 19: 256-263.<br />

Sadler K. 1979. Effects of temperature on the growth and survival of the European eel, Anguilla<br />

anguilla L. Journal of Fish Biology. 15: 499-507.<br />

Sakaris P.C., Jesien R. & Pinkney A.E. 2005. Brown bullhead as an indicator species: seasonal<br />

movement patterns and home ranges within the Anacostia River, Washington, D.C. Transactions of<br />

the American Fisheries Society. 134: 1262-1270.<br />

Santoul F., Cayrou J., Mastrorillo S. & Céréghino R. 2005. Spatial patterns of the biological traits of<br />

freshwater fish communities in south-west France. Journal of Fish Biology. 66: 301-314.<br />

Sargent J.C. & Galat D.L. 2002. Fish mortality and physicochemistry in a managed floodplain wetland.<br />

Wetland Ecology and Management. 10: 115-121.<br />

Saun<strong>de</strong>rs D.L., Meeuwig J.J. & Vincent A.C.J. 2002. Freshwater protected areas: strategies for<br />

conservation. Conservation Biology. 16: 30-41.<br />

Sayer M.D.J. 2005. Adaptations of amphibious fish for surviving life out of water. Fish and Fisheries.<br />

6: 186-211.<br />

Schlosser I.J. 1987. The role of predation in age- and size-related habitat use by stream fishes. Ecology.<br />

68: 651-659.<br />

Schulze T., Kahl U., Radke R.J. & Benndorf J. 2004. Consumption, abundance and habitat use of<br />

Anguilla anguilla in a mesotrophic reservoir. Journal of Fish Biology. 65: 1543-1562.<br />

Scoppettone G.G., Rissler P.H., Gourley C. & Martinez C. 2005. Habitat restoration as a means of<br />

controlling non-native fish in a Mojave <strong>de</strong>sert oasis. Restoration Ecology. 13: 247-256.<br />

Self M. 2005. A review of management for fish and bitterns, Botaurus stellaris, in wetland reserves.<br />

Fisheries Management and Ecology. 12: 387-394.<br />

Shea K. & Chesson P. 2002. Community ecology theory as a framework for biological invasions.<br />

Trends in Ecology and Evolution. 17: 170-176.<br />

Sigourney D.B., Horton G.E., Dubreuil T.L., Varaday A.M. & Letcher B.H. 2005. Electroshocking<br />

and PIT tagging of juvenile Atlantic salmon: are there interactive effects on growth and survival?<br />

North American Journal of Fisheries Management. 25: 1016-1021.<br />

Sih A. 1997. To hi<strong>de</strong> or not to hi<strong>de</strong>: refuge use in a fluctuating environment. Trends in Ecology and<br />

Evolution. 12: 375-376.<br />

Simberloff D. & Von Holle B. 1999. Positive interactions of nonindigenous species: invasional<br />

meltdown? Biological Invasions. 1: 21-32.<br />

Skov C. & Koed A. 2004. Habitat use of 0+ year pike in experimental ponds in relation to cannibalism,<br />

zooplankton, water transparency and habitat complexity. Journal of Fish Biology. 64: 448-459.<br />

262


Bibliographie générale<br />

Skov C., Jacobsen L. & Berg S. 2003. Post-stocking survival of 0+ year pike in ponds as a function of<br />

water transparency, habitat complexity, prey availability and size heterogeneity. Journal of Fish<br />

Biology. 62: 311-322.<br />

Smale M.A. & Rabeni C.F. 1995. Influences of hypoxia and hyperthermia on fish species composition<br />

in headwater streams. Transactions of the American Fisheries Society. 124: 711-725.<br />

Snodgrass J.W., Bryan A.L., Li<strong>de</strong> R.F. & Smith G.M. 1996. Factors affecting the occurrence and<br />

structure of fish assemblage in isolated wetlands of the upper coastal plain, U.S.A. Canadian<br />

Journal of Fisheries and Aquatic Sciences. 53: 443-454.<br />

Souchon Y. 1983. La reproduction du brochet (Esox lucius L., 1758) dans le milieu naturel. In: Billard,<br />

R. (ed.). Le Brochet: Gestion dans le Milieu Naturel et Elevage. INRA Publication, Paris, France.<br />

pp. 21-37.<br />

Sparre P. & Venema S.C. 1998. Introduction to tropical fish stock assessment. Part 1. Manual. FAO<br />

Fisheries Technical Paper. No. 306.1, Rev. 2. FAO, Rome, Italy.<br />

Starkie A. 2003. Management issues relating to the European eel, Anguilla anguilla. Fisheries<br />

Management and Ecology. 10: 361-364.<br />

Stem C., Margoluis R., Salafsky N. & Brown M. 2005. Monitoring and evaluation in conservation: a<br />

review of trends approaches. Conservation Biology. 19: 295-309.<br />

Sutela T., Korhonen P. & Nyberg K. 2004. Stocking success of newly hatched pike evaluated by<br />

radioactive strontium ( 85 Sr) marking. Journal of Fish Biology. 64: 653-664.<br />

Taylor C.M. 1997. Fish species richness and inci<strong>de</strong>nce patterns in isolated and connected stream pools:<br />

effect of pool volume and spatial position. Oecologia. 110: 560-566.<br />

Ter Braak C.J.F. 1986. Canonical correspon<strong>de</strong>nce analysis: a new eigenvector technique for<br />

multivariate direct gradient analysis. Ecology. 67: 1167-1179.<br />

Ter Braak C.J.F. 1996. Unimodal mo<strong>de</strong>ls to relate species to environment. DLO-Agricultural<br />

Mathematics Group, Wageningen, The Netherlands.<br />

Ter Braak C.J.F. & Prentice I.C. 1988. A theory of Gradient Analysis. Advances in Ecological<br />

Research. 18: 272-317.<br />

Ter Braak C.J.F. & Smilauer P. 2003. CANOCO for Windows 4.51. Biometris - Plant Research<br />

International, Wageningen, The Netherlands.<br />

Tesch F.W. 2003. The eel. Fifth edition. Blackwell Publishing, Oxford, U.K..<br />

Thioulouse J., Chessel D., Dolé<strong>de</strong>c S. & Olivier J-M. 1997. ADE-4: a multivariate analysis and<br />

graphical display software. Statistics and Computing. 7: 75-83.<br />

Thoby V. 2005. Etu<strong>de</strong> <strong>de</strong> l’exploitation halieutique dans les marais du Brivet. Parc Naturel régional <strong>de</strong><br />

Brière - <strong>Université</strong> <strong>de</strong> <strong>Rennes</strong> 1, <strong>Rennes</strong>, France.<br />

Tockner K. & Stanford J.A. 2002. Riverine flood plains: present state and future trends.<br />

Environmental Conservation. 29: 308-330.<br />

263


Chapitre 14<br />

Tockner K., Malar F. & Ward J.V. 2000. An extension of the flood pulse concept. Hydrological<br />

Processes. 14: 2861-2883.<br />

Townsend C.R., Dolé<strong>de</strong>c S., Norris R., Peacock K. & Arbuckle C. 2003. The influence of scale and<br />

geography on relationships between stream community composition and landscape variables:<br />

<strong>de</strong>scription and prediction. Freshwater Biology. 48: 768-785.<br />

Trava<strong>de</strong> F. & Larinier M. 1992. La migration <strong>de</strong> dévalaison: problèmes et dispositifs. Bulletin<br />

Français <strong>de</strong> la Pêche et <strong>de</strong> la Pisciculture. 326: 165-176.<br />

Treasurer, J. & Owen R. 1991. Food and growth of pike, Esox lucius, in simple fish communities in<br />

lakes of different trophic status. Aquatic Living Resources. 4: 289-292.<br />

Trexler J.C. & Travis J. 1993. Nontradiotional regression analyses. Ecology. 74: 1629-1637.<br />

Trexler J.C., Loftus W.F. & Perry S. 2005. Disturbance frequency and community structure in a<br />

twenty-five year intervention study. Oecologia. 145: 140-152.<br />

Trexler J. C., Loftus W. F., Jordan C. F., Chick J., Kandl K. L., McElroy T. C. & Bass O. L. 2001.<br />

Ecological scale and its implications for freshwater fishes in the Florida Evergla<strong>de</strong>s. In: Porter J. W.<br />

& Porter K.G. (eds.). The Evergla<strong>de</strong>s, Florida Bay, and Coral Reefs of the Florida Keys: An<br />

Ecosystem Sourcebook. CRC, Boca Raton, U.S. pp. 153-181.<br />

Trexler J.C., Loftus W.F., Jordan F., Lorenz J.J., Chick J.H. & Kobza R.M. 2000. Empirical<br />

assessment of fish introductions in a subtropical wetland: an evaluation of contrasting views.<br />

Biological Invasions. 2: 265-277.<br />

Vannote R.L., Minshall G.W., Cummins K.W., Se<strong>de</strong>ll J.R. & Cushing C.E. 1980. The river continuum<br />

concept. Canadian Journal of Fisheries and Aquatic Sciences. 37: 130-137.<br />

Verneaux J. 1981. Les poisons et la qualité <strong>de</strong>s cours d’eau. Annales Scientifiques <strong>de</strong> l’<strong>Université</strong> <strong>de</strong><br />

Franche-Comté. 2: 33-41.<br />

Vila-Gispert A., Alcaraz C. & Garcia-Berthou E. 2005. Life-history of invasive fish in small<br />

Mediterranean streams. Biological Invasions. 7: 107-116.<br />

Vitousek P.M., D’Antonio C.M., Loope L.L., Rejmanek M. & Westbrooks R. 1997. Introduced<br />

species: a significant component of human-caused global change. New Zealand Journal of Ecology.<br />

21: 1-16.<br />

Vitousek P.M., Mooney H.A., Lubchenco J. & Melillo J.M. 1997. Human domination of Earth’s<br />

ecosystems. Science. 277: 494-499.<br />

Vøllestad L.A. 1990. Optimal management of European eel in the Imsa River, Norway. Journal du<br />

Conseil international d’Exploration <strong>de</strong> la Mer. 46: 115-120.<br />

Vøllestad L.A. & B. Johnson 1988. A 13-year study of the population dynamics and growth of the<br />

European eel Anguilla anguilla in a Norwegian river: evi<strong>de</strong>nce for <strong>de</strong>nsity-<strong>de</strong>pen<strong>de</strong>nt mortality,<br />

and <strong>de</strong>velopment of a mo<strong>de</strong>l for predicting yield. Journal of Animal Ecology. 57: 983-997.<br />

Wantzen K.M., Machado F.d.A., Voss M., Boriss H. & Junk W. 2002. Seasonal isotopic shifts of the<br />

Pantanal wetland, Brazil. Aquatic Sciences. 64: 239-251.<br />

264


Bibliographie générale<br />

Ward J.V., Tockner K., Uehlinger U. & Malard F. 2001. Un<strong>de</strong>rstanding natural patterns and processes<br />

in river corridors as the basis for effective river restoration. Regulated Rivers: Research and<br />

Management. 17: 311-323.<br />

Webster M.S., Marra P.P., Haig S.M., Bensch S. & Holmes R.T. 2002. Links between worlds:<br />

unravelling migratory connectivity. Trends in Ecology and Evolution. 17: 76-83.<br />

Welcomme R.L. 1979. Fisheries Ecology of Floodplain Rivers. Longmann, London, U.K.<br />

Wellborn G.A., Skelly D.K. & Werner E.E. 1996. Mechanisms creating community structure across a<br />

freshwater habitat gradient. Annual Review of Ecology and Systematics. 27: 337-363.<br />

West J.B., Bowen G.J., Cerling T.E. & Ehleringer J.R. 2006. Stable isotope as one of nature’s<br />

ecological recor<strong>de</strong>rs. Trends in Ecology and Evolution. 21: 408-414.<br />

Wheeler A. 1978. Ictalurus melas (Rafinesque, 1820) and I. nebulosus (Lesueur, 1819): the North<br />

American catfishes in Europe. Journal of Fish Biology. 12: 435-439.<br />

Wiens J.A. 2002. Riverine landscapes: taking landscape ecology onto the water. Freshwater Biology.<br />

47: 501-515.<br />

Williams D.D. 2006. The biology of temporary waters. Oxford University Press Inc., New York, U.S.<br />

Winemiller K.O. 2005. Life history strategies, population regulation, and implications for fisheries<br />

management. Canadian Journal of Fisheries and Aquatic Sciences. 62: 872-885.<br />

Wolter C. 2001. Conservation of fish species diversity in navigable waterways. Landscape and Urban<br />

Planning. 53: 135-144.<br />

Wright R.M. & Shoesmith E.A. 1988. The reproductive success of Pike, Esox lucius: aspects of<br />

fecundity, egg <strong>de</strong>nsity and survival. Journal of Fish Biology. 33: 623-636.<br />

Wydoski R.S. & R.W. Wiley. 1999. Management of un<strong>de</strong>sirable fish species. In: Kohler C.C. &<br />

Hubert W.A. (eds.). Inland fisheries management in North America. American Fisheries Society,<br />

Bethesda, Maryland, U.S. pp. 403-430.<br />

Xenopoulos M.A., Lodge D.M., Alcamo J., Marker M., Schulze K. & Van Vuuren D.P. 2005.<br />

Scenarios of freshwater fish extinctions from climate change and water withdrawal. Global Change<br />

Biology. 11: 1557-1564.<br />

Zimmer K.D., Hanson M.A. & Butler M.G. 2000. Factors influencing invertebrate communities in<br />

prairie wetlands: a multivariate approach. Canadian Journal of Fisheries and Aquatic Sciences. 57:<br />

76-85.<br />

Zitek A., Schmutz S., Unfer G. & Ploner A. 2004. Fish drift in a Danube si<strong>de</strong>arm-system: I. Site-,<br />

inter- and intraspecific patterns. Journal of Fish Biology. 65: 1319-1338.<br />

265


Coordonnées <strong>de</strong>s co-auteurs<br />

A. Carpentier UMR 6553 <strong>Ecobio</strong> CNRS <strong>Université</strong> <strong>de</strong> <strong>Rennes</strong> 1. Equipe<br />

Biologie <strong>de</strong>s Populations et <strong>de</strong> la Conservation. Campus <strong>de</strong><br />

Beaulieu. Avenue <strong>de</strong> Général Leclerc. 35042 <strong>Rennes</strong> Ce<strong>de</strong>x,<br />

France.<br />

L.J. Chapman (1) McGill University, Department of Biology, 1205 Av.<br />

Docteur Penfield, Montreal, PQ, Canada H3A 1B1 and<br />

(2) Wildlife Conservation Society, 2300 Southern Boulevard,<br />

Bronx, NY 10460, USA.<br />

J.-P. Damien Parc naturel régional <strong>de</strong> Brière. 177, Ile <strong>de</strong> Fédrun. 44720 Saint-<br />

Joachim, France.<br />

M.-C. Eybert UMR 6553 <strong>Ecobio</strong> CNRS <strong>Université</strong> <strong>de</strong> <strong>Rennes</strong> 1. Equipe<br />

E. Feunteun<br />

J.D. Ol<strong>de</strong>n<br />

Biologie <strong>de</strong>s Populations et <strong>de</strong> la Conservation. Campus <strong>de</strong><br />

Beaulieu. Avenue <strong>de</strong> Général Leclerc. 35042 <strong>Rennes</strong> Ce<strong>de</strong>x,<br />

France.<br />

(1) Centre <strong>de</strong> Recherche sur les Ecosystèmes Littoraux<br />

Anthropisés (CRELA), UMR6217 CNRS-IFREMER-<strong>Université</strong><br />

<strong>de</strong> La Rochelle, La Rochelle, France.<br />

(2) Muséum National d’Histoire Naturelle, Laboratoire maritime<br />

<strong>de</strong> Dinard, Dinard, France.<br />

School of Aquatic and Fishery Sciences, University of<br />

Washington, Box 355020, Seattle Washington 98195, USA<br />

J.-M. Paillisson UMR 6553 <strong>Ecobio</strong> CNRS <strong>Université</strong> <strong>de</strong> <strong>Rennes</strong> 1. Equipe<br />

Biologie <strong>de</strong>s Populations et <strong>de</strong> la Conservation. Campus <strong>de</strong><br />

Beaulieu. Avenue <strong>de</strong> Général Leclerc. 35042 <strong>Rennes</strong> Ce<strong>de</strong>x,<br />

France.<br />

266


T. Robinet<br />

Centre <strong>de</strong> Recherche sur les Ecosystèmes Littoraux Anthropisés<br />

(CRELA), UMR6217 CNRS-IFREMER-<strong>Université</strong> <strong>de</strong> La<br />

Rochelle, La Rochelle, France.<br />

J.-M. Roussel Institut National <strong>de</strong> la Recherche Agronomique, UMR 985,<br />

V. Thoby<br />

<strong>Ecobio</strong>logie et Qualité <strong>de</strong>s Hydrosystèmes Continentaux, 65 rue<br />

<strong>de</strong> Saint-Brieuc, 35042 <strong>Rennes</strong> Ce<strong>de</strong>x, France.<br />

(1) Parc naturel régional <strong>de</strong> Brière, Saint Joachim, France.<br />

(2) Bureau d’étu<strong>de</strong> RIVE Sarl, 37500 Chinon, France.<br />

Levé <strong>de</strong> soleil sur la Gran<strong>de</strong> Brière Mottière <strong>de</strong>puis le port <strong>de</strong> Bréca<br />

267


Résumé<br />

Les milieux temporairement inondés, i.e. les habitats qui s’assèchent <strong>de</strong> manière cyclique, ont<br />

un rôle fondamental pour <strong>de</strong> nombreux organismes et ce, malgré les conditions<br />

environnementales très fluctuantes qui les caractérisent. De plus, en tant que zones <strong>de</strong><br />

transition entre écosystèmes aquatiques et terrestres, les milieux temporairement inondés sont<br />

soumis à <strong>de</strong> fortes contraintes anthropiques. Ce travail <strong>de</strong> thèse s’est attaché à caractériser<br />

l’utilisation fonctionnelle <strong>de</strong>s milieux temporairement inondés d’un écosystème sous forte<br />

emprise humaine (les marais <strong>de</strong> Brière, Nord Ouest <strong>de</strong> la France) par l’ichtyofaune, et <strong>de</strong><br />

déterminer les réponses <strong>de</strong>s poissons aux contraintes dynamiques imposées par ces habitats.<br />

Cette démarche a été entreprise à différentes échelles spatiales (micro- et macro-habitat) et<br />

temporelles (à un moment précis du cycle d’inondation ou <strong>de</strong> manière continue au cours d’un<br />

cycle complet) ainsi qu’à différents niveaux d’organisation (communauté, population et<br />

individu). Une analyse fine a porté sur les réponses fonctionnelles aux contraintes d’habitat et<br />

<strong>de</strong> gestion <strong>de</strong> trois espèces i<strong>de</strong>ntifiées comme d’intérêt majeur <strong>de</strong> la communauté, à savoir le<br />

poisson chat (Ameiurus melas), l’anguille (Anguilla anguilla) et le brochet (Esox lucius). Il<br />

ressort <strong>de</strong> ce travail que l’utilisation <strong>de</strong>s milieux temporairement inondés par la communauté<br />

piscicole est très restreinte dans le temps (phase particulière <strong>de</strong>s régimes hydrologique et<br />

thermique) et dans l’espace (diffusion limitée <strong>de</strong>s poissons par rapport au milieu source).<br />

Cette utilisation est dépendante à la fois <strong>de</strong>s caractéristiques spécifiques (tolérance<br />

physiologique, cycle <strong>de</strong> vie), <strong>de</strong>s sta<strong>de</strong>s <strong>de</strong> vie particuliers (juvéniles <strong>de</strong> l’année, juvéniles >0+<br />

et adultes) et individuelles (régime alimentaire, taille) <strong>de</strong>s poissons et <strong>de</strong> l’évolution <strong>de</strong>s<br />

conditions environnementales au cours <strong>de</strong> la phase d’inondation. De plus, l’utilisation <strong>de</strong>s<br />

milieux temporairement inondés semble avoir <strong>de</strong>s conséquences importantes dans le<br />

fonctionnement <strong>de</strong> la communauté piscicole. En effet, certaines espèces introduites présentent<br />

<strong>de</strong>s stratégies d’utilisation <strong>de</strong> ces habitats très opportunistes et il semble que les importantes<br />

modifications subies par les marais <strong>de</strong> Brière au cours <strong>de</strong>s <strong>de</strong>rnières décennies, notamment<br />

une forte expansion du roseau (Phragmites australis), aient profité à ces espèces introduites<br />

tolérantes qui utilisent les milieux temporairement inondés uniquement à <strong>de</strong>s fins <strong>de</strong><br />

croissance (« espèces facultatives »). A l’opposé, les espèces natives plus sensibles, qui<br />

utilisent ces milieux pour s’y reproduire (« espèces obligatoires »), semblent avoir été<br />

affectées par à cette dynamique paysagère. Face à cette évolution du milieu, certaines mesures<br />

<strong>de</strong> gestion ont été entreprises pour gérer les populations <strong>de</strong>s trois espèces d’intérêt majeur<br />

(<strong>de</strong>struction <strong>de</strong>s poissons chats, réserves <strong>de</strong> pêche pour l’anguille et programmes <strong>de</strong> soutien<br />

<strong>de</strong> la population <strong>de</strong> brochet). Il ressort <strong>de</strong> cette thèse que les mesures actuellement mises en<br />

place pour ces trois espèces ont une efficacité variable, mais globalement limitée, et ne<br />

semblent pas être en mesure <strong>de</strong> contrecarrer les dysfonctionnements plus globaux observés à<br />

l’échelle <strong>de</strong> la communauté piscicole et <strong>de</strong> l’écosystème. Une gestion intégrative, prenant en<br />

compte les processus fonctionnels contrôlant l’utilisation <strong>de</strong>s milieux temporairement inondés,<br />

notamment à travers une approche basée sur les compromis coûts/bénéfices, est proposée dans<br />

ce travail. Elle <strong>de</strong>vra s’orienter vers une prise en compte du fonctionnement global <strong>de</strong> ce type<br />

d’écosystème et <strong>de</strong> son icthyofaune et ce notamment à travers <strong>de</strong>s programmes <strong>de</strong> restauration<br />

<strong>de</strong>s milieux temporairement inondés et <strong>de</strong> la connectivité latérale et longitudinale <strong>de</strong><br />

l’écosystème. Une telle mesure est d’autant plus urgente que les dysfonctionnements liés à<br />

l’utilisation <strong>de</strong>s milieux temporairement inondés par l’ichtyofaune, i<strong>de</strong>ntifiés au cours du<br />

présent travail, risquent d’être amplifiés par les changements globaux.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!