UNIVERSITE PAUL SABATIER-TOULOUSE IIIthesesups.ups-tlse.fr/151/1/Leprieur_Fabien.pdfJe remercie vivement Colin Townsend de m’avoir accueilli dans son laboratoire à l’université

Université Toulouse III - Paul Sabatier Ecole doctorale Sciences Ecologiques, UFR Sciences de la Vie et de la Terre Vétérinaires, Agronomiques et Bioingénieries

THESE

pour obtenir le grade de

DOCTEUR DE L’UNIVERSITE TOULOUSE

Délivré par l’Université Toulouse III – Paul Sabatier

Spécialité : Ecologie

Présentée par

Fabien LEPRIEUR

Les introductions d’espèces de poissons d’eau douce : distribution

spatiale, déterminants et impacts sur les espèces natives

Soutenue le 7 décembre 2007

Devant le jury composé de : D. MOUILLOT, Professeur, Université Montpellier 2 Rapporteur D. SIMBERLOFF Professeur, University of Tennessee (USA) Rapporteur E. GARCIA-BERTHOU Professeur, Universitat de Girona (Spain) Examinateur S. LEK Professeur, Université Paul Sabatier (Toulouse) Président T. OBERDORFF Directeur de recherche, IRD (Paris) Examinateur S. BROSSE MCF, Université Paul Sabatier, (Toulouse) Directeur

Université Toulouse III - Paul Sabatier Ecole doctorale Sciences Ecologiques,

UFR Sciences de la Vie et de la Terre Vétérinaires, Agronomiques et Bioingénieries

THESE

pour obtenir le grade de

DOCTEUR DE L’UNIVERSITE TOULOUSE

Délivré par l’Université Toulouse III – Paul Sabatier

Spécialité : Ecologie

Présentée par

Fabien LEPRIEUR

Les introductions d’espèces de poissons d’eau douce : distribution

spatiale, déterminants et impacts sur les espèces natives

Soutenue le 7 décembre 2007

Devant le jury composé de : D. MOUILLOT, Professeur, Université Montpellier 2 Rapporteur D. SIMBERLOFF Professeur, University of Tennessee (USA) Rapporteur E. GARCIA-BERTHOU Professeur, Universitat de Girona (Spain) Examinateur S. LEK Professeur, Université Paul Sabatier (Toulouse) Président T. OBERDORFF Directeur de recherche, IRD (Paris) Examinateur S. BROSSE MCF, Université Paul Sabatier, (Toulouse) Directeur

« Plus je sais, moins je sais. L'essentiel est de ne jamais cesser de chercher. »

(Albert Einstein)

Auteur : Fabien LEPRIEUR Titre : Les introductions d’espèces de poissons d’eau douce : distribution spatiale, déterminants et impacts sur les espèces natives Directeur de thèse : Sébastien BROSSE Lieu et date de soutenance : Toulouse, le 7 décembre 2007 Résumé :

Bien que les espèces non natives de poissons d’eau douce soient bien identifiées, les facteurs déterminant leur distribution spatiale ainsi que leurs impacts sur la biodiversité sont encore peu connus, en particulier à de larges échelles spatiales. Dans ce contexte, cette thèse vise : (i) à une meilleure compréhension de l’impact des espèces non natives de poissons d’eau douce sur les espèces natives ; et (ii) à identifier les facteurs qui contrôlent la distribution spatiale des espèces non natives. Pour cela, différents niveaux de perception du processus d’introduction d’espèces et différentes échelles spatiales ont été considérés.

Les résultats obtenus à l’échelle locale (135 stations au sein d’un bassin

hydrographique de Nouvelle Zélande) ont permis de mettre en évidence que l’impact d’une espèce invasive (la truite, Salmo trutta L.) sur une espèce native (Galaxias anomalus Stockell) peut varier spatialement en fonction des caractéristiques abiotiques locales. En particulier, les perturbations anthropiques, telles que la modification des habitats résultant des variations de débit, ne favorisent pas forcément les espèces invasives. Ainsi, la conservation d’une espèce native menacée nécessite des mesures de gestion adaptées au contexte environnemental local. Enfin, une étude expérimentale souligne le fait qu’une espèce considérée à priori comme invasive et nuisible (le poisson chat, Ameiurus melas Raff.) doit faire l’objet d’études quant à son impact réel sur les espèces natives ; ceci afin de mettre en place des mesures de gestion adaptées aux caractéristiques comportementales et écologiques de l’espèce impactée.

Les résultats obtenus à l’échelle régionale (bassin hydrographique) montrent que les introductions d’espèces de poissons d’eau douce en Europe ont conduit à une augmentation de la diversité alpha des bassins hydrographiques (c'est-à-dire une augmentation du pool régional d’espèces), mais ont provoqué une diminution de la diversité beta (homogénéisation taxonomique). L’augmentation du pool régional de poissons d’eau douce en Europe ne doit pas forcément être interprétée comme bénéfique pour la biodiversité, car les extinctions d’espèces se déroulent généralement à des échelles de temps plus grandes que le phénomène d’introduction d’espèces. Ensuite, il semblerait que la distribution actuelle des poissons d’eau douce exotiques en Europe (c.-à-d. les espèces non européennes) soit le résultat combiné d’une limitation de leur dispersion liée aux activités humaines et d’un contrôle environnemental associé aux contraintes climatiques. Enfin, il est montré que le niveau d’anthropisation d’un bassin hydrographique, et plus particulièrement sa richesse économique, est le principal déterminant de la richesse régional en espèces non natives de poissons d’eau douce. Mots clés : espèces non natives, poissons d’eau douce, assemblages d’espèces, macroécologie, homogénéisation biotique, modèles nuls, filtres environnementaux, hotspots d’invasion.

Discipline : Ecologie Adresse du laboratoire de rattachement : Laboratoire Evolution & Diversité Biologique Bâtiment 4R3 Université Paul Sabatier, 118 route de Narbonne, 31062 Toulouse cedex 4, France

Remerciements

Je tiens tout d’abord à remercier Sébastien Brosse, mon directeur de thèse. Sébastien, tu as toujours été à mon écoute durant cette thèse. Tes conseils pertinents et ta vue globale de l’écologie ont été un atout majeur pour la finalisation de cette thèse. D’un point de vue humain, ta confiance en moi et tes encouragements m’ont été également très précieux, surtout durant ma première année de thèse qui a été un peu mouvementée au niveau personnel.

Je remercie également Sovan Lek qui m’a fait confiance depuis le début. Sovan, il y a quelques années, je suis venu à Toulouse pour te demander des conseils sur le déroulement d’une thèse en écologie. Tes conseils m’ont été très utiles et c’est grâce à toi que je suis parti en Nouvelle Zélande pour commencer ma thèse.

Je remercie vivement Colin Townsend de m’avoir accueilli dans son laboratoire à

l’université d’Otago en Nouvelle Zélande au début de ma thèse. Colin, vous m’avez beaucoup apporté tant d’un point de vue scientifique qu’humain. Mon expérience en Nouvelle Zélande reste un souvenir inoubliable et c’est principalement grâce à vous.

Je remercie David Mouillot et Daniel Simberloff d’avoir accepté d’être rapporteur de cette thèse. En effet, c’est un grand honneur pour moi que vous puissiez juger cette thèse étant donné votre expérience et la qualité de vos travaux en écologie.

Je remercie Emili-García Berthou et Thierry Oberdorff d’avoir accepté d’être membre

de ce jury. Thierry, un grand merci pour ton soutien et tes encouragements. Tu m’as permis de prendre conscience d’une autre facette de la problématique des invasions biologiques. J’espère sincèrement continuer à travailler avec toi.

Je remercie Bernard Hugueny qui m’a donné de nombreux conseils lors de notre

rencontre à Lyon. Bernard, tu m’as éclairé sur la partie « macroécologie » de cette thèse. Nos discussions sur l’intérêt et la philosophie des modèles nuls m’ont été très utiles. Mes neurones ont souffert mais le concept est bien passé.

Je remercie vivement tous mes collègues ami(es) de l’équipe « structure des communautés et macroécologie »: Bobby, Charly, Micky, Ronny, Wendy, Cindy, Sandy Rosy, Domy, Fonzy……. enfin, la bande des « Y ». J’oubliais « Gala » aussi! Votre bonne humeur, gentillesse et votre soutien moral m’ont été indispensable pour réaliser cette thèse. Flamby vous sera toujours fidèle ! Je voudrais particulièrement remercier : Gaël Grenouillet pour sa disponibilité et ses commentaires pertinents sur mes travaux de thèse ; Géraldine Loot pour ses nombreux encouragements ; Simon Blanchet pour ses conseils judicieux (et notre passion commune pour Match Point !) et Laetitia Buisson pour sa bonne humeur et pour m’avoir donné des bonbons nounours.

Je tiens particulièrement à remercier Olivier Beauchard dont le travail impressionnant

(de fourmi) a donné naissance à une base de données mondiale sur la biodiversité des poissons d’eau douce et Karl Kreutzenberger pour l’excellent travail fourni lors de son DESUPS.

Je tiens enfin à remercier toute ma famille (et d’ailleurs l’ensemble de mon arbre généalogique!). En particulier, je remercie mes parents pour leur soutien depuis toujours et leur amour. Un grand merci à mon frère, Frédéric, qui a toujours été un modèle pour moi. Il m’a convaincu de poursuivre mes études… alors que j’étais décidé à suivre des études courtes! Je lui en serais toujours reconnaissant.

Je remercie également mes beaux parents, André et Claudine, et Claude mon « beauf-

ami » pour leur soutien formidable et leur affection. André, vous ne m’avez jamais laissé avoir soif… (sauf à la Gacholle).

Enfin, Magali (la perle), mon épouse, je ne sais pas comment te remercier. Tu m’as tellement soutenu durant cette thèse. Mes remerciements vont bien au-delà du seul cadre de cette thèse. Tu m’as ouvert les yeux…

Partie 1 : Synthèse 1 1. Introduction générale……………………………………………………... 2

1-1. Problématique…………………………………………………………………………………. 2 1-2. Echelles spatiales et niveaux de perception du processus d’introduction d’espèces………….. 4 1-3. Organisation du mémoire……………………………………………………………………… 9

2. Les introductions d’espèces à l’échelle locale……………………………. 15

2-1. Le cas de la truite commune (Salmo trutta L.) introduite en Nouvelle Zélande (P1)………… 15 2-2. Le cas du poisson-chat (Ameiurus melas Raff.) introduit en Europe (P2)……………………. 22 2-3. Conclusion et perspectives……………………………………………………………………. 25

3. Les introductions d’espèces à l’échelle régionale………………………... 27 3-1. Processus d’homogénéisation biotique des assemblages régionaux : le cas des poissons d’eau

douce européens (P3)………………………………………………………………………….. 28 3-2. Rôle des facteurs environnementaux et géographiques dans la structuration des assemblages

régionaux : une comparaison entre les espèces natives et exotiques de poissons d’eau douce européens (P4)………………………………………………………………………………… 31

3-3. Déterminants et répartition géographique mondiale de la richesse en espèces non natives de poisson d’eau douce (P5)………………………………………………………………………. 35

3-4. Conclusion et perspectives……………………………………………………………………... 39 4. Conclusion générale………………………………………………………... 43

Références……………………………………………………………………... 46

Partie 2: Publications 54 P1 Hydrological disturbance benefits a native fish at the expense of an exotic fish

Leprieur F., Hickey M.A., Arbuckle C.J., Closs G.P., Brosse, S. & Townsend

C.R. (2006) Journal of Applied Ecology, 43: 930-939.

P2 Impact of the invasive black bullhead (Ameiurus melas Raff.) on the predatory efficiency of northern pike (Esox lucius L.)

Kreutzenberger K., Leprieur F., & Brosse, S.

Journal of Fish Biology

(en révision mineure)

P3 Null model of biotic homogenization: a test with the European freshwater fish fauna

Leprieur F., Beauchard O., Hugueny B., Grenouillet G. & Brosse S. (2007)

Diversity and Distributions

(sous presse)

P4 Patterns and mechanisms of the distance decay of similarity in the European

freshwater fish fauna: contrasting native and exotic species

Leprieur F., Olden, J.D. Lek, S. & Brosse S.

(en préparation)

P5 Fish invasions in the world’s river systems: when natural processes are blurred by human activities

Leprieur F., Beauchard O., Blanchet S., Oberdorff T. & Brosse S.

PLoS Biology

(accepté)

Les introductions d’espèces de poissons d’eau douce 1

Partie 1 : Synthèse


1. Introduction générale 1.1. Problématique

Les introductions d’espèces par l’homme se sont développées dès le néolithique. Le

développement du pastoralisme et de l’agriculture a ainsi entrainé le transport volontaire (p.ex.

les céréales, les animaux domestiques) et/ou involontaire (p.ex. les parasites et les

commensaux des espèces domestiquées) de nombreux organismes animaux et végétaux

(Crosby 1986). Néanmoins, les introductions d’espèces ont considérablement augmenté au

20ème siècle avec le développement du commerce international et des transports humains

(Mack & Londsale 2001 ; Levine & D’Antonio 2003 ; Perrings et al. 2005 ; Meyerson &

Mooney 2007). Cette globalisation du commerce et de l’économie tend ainsi à supprimer

les barrières géographiques limitant la distribution naturelle des espèces et donc à favoriser

leur expansion géographique (Perrings et al. 2005 ; Meyerson & Mooney 2007).

Les espèces non natives en s’établissant dans le milieu d’accueil, peuvent devenir

invasives et par conséquent provoquer des perturbations à différents niveaux d’organisation

écologique : de individu à l’écosystème (Lodge 1993 ; Simberloff 1996 ; Parker et al. 1999 ;

Sakai et al. 2001 ; Simon & Townsend 2003 ; Lockwood et al. 2007). A ces impacts

écologiques s’ajoutent des impacts plus difficiles à quantifier, tels que des changements

évolutifs chez les espèces natives qui peuvent apparaître très rapidement suite à de nouvelles

pressions de sélection imposées par l’introduction d’espèces non natives (Mooney & Cleland

2001 ; Stockwell et al. 2003 ; Strauss et al. 2006). Bien que les invasions biologiques soient

considérées comme la deuxième cause d’extinction d’espèces, après la destruction et la

fragmentation des habitats (Miller et al. 1989 ; Wilcove 1998 ; Woodruff 2001 ; Clavero &

García-Berthou 2005 ; Millenium Ecosystem Assessment 2005), l’impact des introductions

d’espèces est actuellement peu étudié par rapport aux autres types de perturbations

anthropiques (Lawler et al. 2006).


D’un point de vue pratique, les sociétés humaines tirent de nombreux bénéfices du

fonctionnement des écosystèmes (ressources animales et végétales, énergie,…) et la diversité

génétique, taxonomique et écosystémique (c.-à-d. la biodiversité) joue un rôle important dans

le fonctionnement, la résilience et la résistance des écosystèmes (Loreau et al. 2002 ; Hooper

et al. 2005). Les invasions biologiques peuvent ainsi provoquer des pertes importantes de

services écologiques de part leur effet négatif sur la biodiversité (Vitousek et al. 1997 ; Sala

et al. 2000 ; Loreau et al. 2001 ; Hooper et al. 2005), et par conséquent avoir des impacts

économiques non négligeables (Sakai et al. 2001 ; Hooper et al. 2005 ; Pimentel et al. 2005).

Les études portant sur la conservation des écosystèmes aquatiques (marins et d’eau

douce) sont peu nombreuses (Lawler et al. 2006), alors que ces derniers sont les plus

menacées par les activités humaines : surexploitation, pollution, destruction d’habitats,

invasions biologiques (Moyle 1999 ; Millenium Ecosystem Assessment 2005 ; Dudgeon et al.

2005). Les écosystèmes aquatiques continentaux (rivières, lacs et estuaires) font partie des

écosystèmes les plus envahis dans le monde (Moyle 1999 ; Cohen 2002). Les poissons d’eau

douce, plus particulièrement, ont fait l’objet de nombreuses introductions depuis le moyen

âge, de par leur intérêt alimentaire, récréatif ou ornemental (Welcomme 1988 ; Lever 1996).

Comme pour les organismes terrestres, les introductions de poissons d’eau douce ont parfois

été bénéfiques d’un point de vue socio-économique (Lever 1996), mais une grande partie

d’entre elles ont eu des conséquences négatives sur les espèces natives (Ross 1991 ; Miller et

al. 1989 ; Lever 1996 ; Townsend 2003 ; Light & Marchetti 2007). Plusieurs mécanismes sont

à l’origine de l’impact des espèces non natives de poissons d’eau douce. D’abord, la

prédation sur les espèces natives peut entrainer des extinctions locales (Ross 1991 ; Bianco

1995; Fuller 1999 ; Elvira 2001; McDowall 2006 ; Fattini & Petrere 2007), voire des

extinctions globales d’espèces (Barlow et al. 1987 ; Witte et al. 1992 ; Crivelli 1995 ;

McDowall 2006 ; Kirchner & Soubeyran 2007). Ensuite, la compétition interspécifique pour


la ressource trophique ou l’espace peut entrainer des modifications comportementales

(Blanchet et al. 2007 ; Baxter et al. 2007) et altérer la dynamique des populations (p.ex. la

croissance, la survie, la fécondité,…) (Ríncon et al. 2002; Townsend 2003 ; Baxter et al.

2007 ; Blanchet et al. 2007). Enfin, l’hybridation entre espèces natives et non natives (Perry

et al. 2002) et l’introduction de parasites et pathogènes associés aux espèces non natives

(Gozlan 2005) sont susceptibles de mettre en péril les espèces natives. L’ensemble de ces

impacts aux échelles individuelles et populationelles peuvent modifier la structure des

communautés, le fonctionnement des écosystèmes (Vander Zanden et al. 1999 ; Simon &

Townsend 2003 ; Baxter et al. 2005 ; Eby et al. 2006), et avoir des conséquences évolutives

(Strauss et al. 2006).

Bien que les espèces non natives (et invasives) de poissons d’eau douce soient bien

identifiées (Welcomme 1988 ; Elvira, 2001), les facteurs déterminant leur distribution

spatiale ainsi que leurs impacts sur la biodiversité sont encore très peu connus, en

particulier à de larges échelles spatiales. Dans ce contexte, cette thèse vise : (i) à une

meilleure compréhension de l’impact des espèces non natives de poissons d’eau douce

sur les espèces natives ; et (ii) à identifier les facteurs qui contrôlent leur distribution

spatiale ; ceci à différents niveaux de perception du processus d’introduction d’espèces

et à différentes échelles spatiales.

1.2. Echelles spatiales et niveaux de perception du processus d’introduction d’espèces

Un des grands défis de l'écologie des communautés est de comprendre les

interactions entre les échelles d’observations des phénomènes naturels et les mécanismes

écologiques (Levin 1992; O’Neill & King 1998; Willis & Whittaker 2001). Il est aujourd’hui

largement reconnu que les mécanismes écologiques agissant à l’échelle locale sont dépendant


des mécanismes s’exerçant à des échelles spatiales supérieures (Ricklefs 1987; Levin 1992,

Ricklefs & Schulter 1993; Huston 1999). Ceci a été conceptualisé en écologie des

communautés par la notion de « filtres hiérarchiques » conditionnant la composition des

assemblages d’espèces à différentes échelles spatiales (de l’échelle globale à l’échelle locale)

(Simpson 1953; Smith & Powell 1971 ; Tonn 1990 ; Keddy 1992 ; Poff 1997). Ce cadre

conceptuel trouve de nombreuses applications dans l’étude des assemblages de poissons d’eau

douce à différentes échelles spatiales (p.ex. Jackson et al. 2001 ; Quist et al. 2005). Selon

cette approche hiérarchique, chaque filtre serait associé à des processus agissant à des échelles

spatio-temporelles différentes et les assemblages locaux seraient un sous ensemble des

assemblages rencontrés aux échelles supérieures (Figure 1).

•Habitat physique (pente, largeur, profondeur, substrat), température, barrières géographiques

• Intéractions biotiques

• Perturbations naturelles ou anthropiques

A B C D E F G H I

I A B C D E F

I A B C D

Filtre continental

Filtre régional (bassin hydrographique)

Filtre local (tronçon de

rivière)

• Dérive des continents

• Glaciation

• Barrières géographiques

• Conditions climatiques et diversité d’habitats

I A B

Eche

lle s

patia

le (K

m2)

Eche

lle te

mpo

relle

(ann

ée)

10-2 10-2

103 - 107107-108

•Habitat physique (pente, largeur, profondeur, substrat), température, barrières géographiques

• Intéractions biotiques

• Perturbations naturelles ou anthropiques

A B C D E F G H I

I A B C D E F

I A B C D

Filtre continental



rivière)


• Glaciation



I A B

Eche

lle s

patia

le (K

m2)

Eche

lle te

mpo

relle

(ann

ée)

10-2 10-2

103 - 107107-108

A B C D E F G H I

I A B C D E F

I A B C D

Filtre continental



rivière)


• Glaciation



I A B

Eche

lle s

patia

le (K

m2)

Eche

lle te

mpo

relle

(ann

ée)

10-2 10-2

103 - 107107-108

Figure 1 : Cadre conceptuel des filtres hiérarchiques pour l’étude des assemblages de poissons d’eau douce à différentes échelles spatiales et temporelles (d’après Smith & Powell 1991 ; Tonn et al. 1990 ; Poff 1997). Les flèches rouges correspondent au processus d’introduction d’une espèce (I) d’un continent vers un autre. Une fois que l’espèce I a réussi à s’établir localement, elle intègre le pool régional d’espèces ainsi que le pool continental.


Tout d’abord, le pool continental d’espèces (p.ex. en Europe) est un sous-ensemble

du pool mondial d’espèces de poissons d’eau douce, sélectionné suite aux événements

tectoniques (filtre biogéographique, Figure 1). Ensuite, le pool régional d’espèces (c.-à-d.

l’ensemble des espèces d’un bassin hydrographique comme la Garonne) est un sous ensemble

du pool continental d’espèces, dont la composition en espèces est déterminée par des

événements historiques (glaciations, formation de barrières géographiques délimitant les

bassins hydrographiques, transgressions et régressions marines, Figure 1). Ces événements

ont conduit à des taux de spéciation, d’extinction et de colonisation différentiels entre bassins

hydrographiques (voir Tedesco et al. 2006 ; Reyjol et al. 2006). A l’échelle régionale, les

conditions environnementales (c.-à-d. le climat et la diversité d’habitats d’un bassin

hydrographique) limitent également la répartition géographique des espèces (filtre abiotique

régional). Enfin, une succession de filtres réduirait le pool régional d’espèces en un sous-

ensemble d'espèces présentes à l’échelle locale (du micro-habitat au tronçon de rivière d’un

réseau hydrographique, Figure 1). Un assemblage local est ainsi composé d’espèces (i) ayant

eu la capacité de coloniser un habitat local (filtre géographique), (ii) qui sont

physiologiquement adaptées aux conditions abiotiques locales (filtre abiotique) et (iii) dont les

caractéristiques écologiques leur permettent de cohabiter entre elles (filtre biotique).

Dans le cadre du processus d’introductions d’espèces (Figures 1 et 2), l’homme tend à

supprimer les filtres géographiques (Rahel 2007) en introduisant des espèces d’un continent

vers un autre (espèce exotique) ou bien d’une région à une autre dans la même zone

biogéographique (espèce transloquée). Ces introductions peuvent être accidentelles (p.ex.

suite à la construction de canaux reliant deux bassins hydrographiques, Rahel 2002) ou bien

intentionnelles (p.ex. les introductions espèces ornementales ou d’intérêt halieutique, Lever

1996). Après introduction, une espèce non-native s’établira dans le milieu d’accueil (espèce

établie, Figure 2), si elle arrive à se reproduire avec succès et à se maintenir à long terme


(Richardson, 2000 ; Colautti & MacIsaac 2004 ; Lockwood et al. 2007). Le succès

d’établissement d’une espèce dépend de plusieurs facteurs (Kolar & Lodge 2002 ; Moyle &

Marchetti 2006 ; Lockwood et al. 2007) : (i) la pression de propagules (c.-à-d. le nombre

d’individus introduits) ; (ii) les caractéristiques écologiques de l’espèce introduite (p.ex. la

capacité de dispersion et la tolérance environnementale) ; (iii) les conditions

abiotiques locales (p.ex. l’intensité et la fréquence des perturbations naturelles et

anthropiques) et régionales (p.ex. les conditions climatiques) ; (iiii) les caractéristiques

biotiques locales (p.ex. la présence ou non de prédateurs, l’intensité des interactions biotiques).

Une espèce introduite qui réussit à s’établir avec succès à l’échelle locale, intègre alors le pool

régional d’espèces, voire le pool continental si l’espèce introduite est originaire d’un autre

continent (Figure 1).

Espèce introduite

Espèce établie (non native)

Espèce invasive

Etablissement

Expansion, prolifération

Introduction

Impacts écologiques, économiques et sociaux

• Caractéristiques biotiques et abiotiques du milieu d’accueil

• Caractéristiques écologiques de l’espèce

• Pression de propagule



• Introduction accidentelle ou intentionnelle

Espèce introduite

Espèce établie (non native)

Espèce invasive

Etablissement

Expansion, prolifération

Introduction

Impacts écologiques, économiques et sociaux



• Pression de propagule



• Introduction accidentelle ou intentionnelle

Figure 2 : Les différents niveaux de perception du processus d’introduction d’espèces (introduction, établissement et prolifération) et les facteurs contrôlant le succès ou l’échec de chaque étape du processus d’invasion (d’après Williamson 1996 ; Richardson 2000 ; Lockwood et al. 2007).


Enfin, certaines espèces non natives peuvent dans certaines conditions

environnementales (biotiques et abiotiques) devenir invasives (c.-à-d. proliférer et étendre

leur aire de répartition ; Richardson 2000) et par conséquent avoir des conséquences négatives

sur la biodiversité (Figure 2). D’après Williamson (1996), une espèce introduite sur dix

réussirait à s’établir, et une espèce établie sur dix deviendrait invasive (« la règle des 10% »).

Ceci impliquerait que peu d’espèces introduites deviennent invasives. Cependant, des études

récentes (García-Berthou et al. 2005 ; Jeschke & Strayer 2005) ont montré que cette règle est

loin d’être généralisable. Jeschke & Strayer (2005) ont en effet montré qu’en moyenne, 25%

des vertébrés introduits en Amérique du Nord et en Europe (dont les poissons d’eau douce)

devenaient invasifs.

En résumé, une espèce introduite ne peut s’établir (et devenir invasive) que si ses

caractéristiques écologiques lui permettent de franchir les différents filtres biotiques et

abiotiques du milieu d’accueil (Rahel 2002). Le cadre conceptuel des filtres

hiérarchiques en écologie des communautés est donc tout à fait adapté à l’étude des

introductions d’espèces. Dans le cadre de cette thèse, j’ai tenté d’identifier les filtres

locaux (à l’échelle de la station et du tronçon de rivière) et régionaux (à l’échelle du

bassin hydrographique), qui peuvent limiter la distribution spatiale des espèces non-

natives de poissons d’eau douce. L’étude du processus dynamique des introductions

d’espèces à plusieurs échelles spatiales peut également contribuer à une meilleure

compréhension de ses effets sur la biodiversité et nous aider à identifier des stratégies de

contrôle plus efficaces (Lodge et al. 1998 ; Mack et al. 2000; Pauchard & Shea al. 2006).


1-3. Organisation du mémoire

Cette thèse est composée de cinq publications dont les problématiques s’adressent aux

échelles locales et régionales (Tableau 1). Une échelle spatiale en écologie fait référence à

deux composantes : la résolution spatiale (ou unité spatiale) à laquelle les mesures ou

expérimentations sont effectuées et l’étendue spatiale à laquelle une espèce, un assemblage

ou un écosystème est étudié (Schneider 2001). Dans le cadre de cette thèse, l’échelle spatiale

est synonyme de résolution spatiale.

A l’échelle locale (P1, P2), je me suis intéressé à deux espèces invasives de poissons

d’eau douce. J’ai d’abord exploré la distribution spatiale d’une espèce de salmonidés

invasive (la truite commune Salmo trutta L.) au sein d’un bassin hydrographique de

Nouvelle Zélande, et déterminé son impact sur une espèce endémique menacée

(Galaxias anomalus Stokell) (P1). Je me suis plus particulièrement intéressé au rôle

des modifications locales d’habitats (de type hydrologique) dans le processus

d’invasion. En effet, les modifications d’habitats d’origine anthropique pourraient

favoriser l’établissement des espèces introduites et leur prolifération (voir Moyle &

Light 1996 ; Lockwood et al. 2007). J’ai ensuite tenté de mettre en évidence l’impact

potentiel d’une espèce introduite, le poisson-chat (Ameiurus melas Raff.), sur une

espèce native (Esox lucius L.), en utilisant une approche expérimentale en

microcosmes (P2). Le poisson-chat est originaire d’Amérique du Nord et a été

introduit avec succès en Europe à la fin du 19ème siècle. Bien que cette espèce invasive

soit considérée comme nuisible en Europe (Lever 1996 ; Elvira 2001), aucune étude

n’a jusqu’à présent démontré un impact négatif du poisson-chat sur les espèces

européennes.


Tableau 1 : Problématiques et échelles spatiales abordées dans les différentes publications constituant cette thèse.

Echelle locale

Echelle régionale Problématique

P1

Etendue :

Bassin hydrographique

Résolution : Tronçon de

rivière

Influence des modifications d’habitats sur le potentiel invasif d’une espèce de salmonidés (Salmo trutta L.): conséquences sur la conservation d’une espèce endémique menacée (Galaxias anomalus Stokell).

P2

Etendue : Aquarium

Résolution :

Aquarium

Impact potentiel d’une espèce considérée comme invasive (Ameiurus melas Raff.) sur le succès de prédation d’une espèce native (Esox lucius L.).

P3

Etendue :

Europe

Résolution : Bassin

hydrographique

1) Conséquences des introductions d’espèces sur la diversité alpha et beta des assemblages régionaux. 2) L’homogénéisation biotique est elle un phénomène aléatoire ? 3) Rôle relatif des espèces exotiques vs. transloquées dans le processus d’homogénéisation biotique.

P4

Etendue :

Europe


hydrographique

Importance relative des facteurs environnementaux, humains et géographiques dans la structuration des assemblages régionaux : comparaison entre les espèces natives et exotiques.

P5

Etendue :

Monde


hydrographique

1) Répartition géographique mondiale de la richesse en espèces non natives et identification des hotspots globaux d’invasion de poissons d’eau douce. 2) Identification et quantification des facteurs environnementaux et anthropiques déterminant la richesse mondiale en espèces non natives


A l’échelle régionale (P3, P4, P5), je me suis intéressé à la richesse et à la

composition des assemblages régionaux d’espèces non natives (c.-à-d. le pool

d’espèces non natives d’un bassin hydrographique, lesquelles peuvent potentiellement

s’établir à l’échelle locale). J’ai d’abord évalué les conséquences des introductions

d’espèces de poissons d’eau douce sur la richesse en espèces (diversité alpha) et la

similarité taxonomique (diversité beta) des bassins hydrographiques européens (P3).

En particulier, j’ai testé l’hypothèse que l’homogénéisation biotique (augmentation

de la similarité en espèces entre deux localités suite à des introductions et des

extinctions d’espèces, Olden & Poff 2003, 2004) était le résultat de processus non

aléatoires (Duncan et al. 2001 ; Olden et al. 2004). J’ai ensuite étudié la structuration

spatiale des assemblages régionaux d’espèces exotiques en Europe (c.-à-d. les espèces

de poissons d’eau douce qui ne sont pas natives d’Europe) (P4). Le but in fine était

d’identifier les facteurs régionaux contrôlant la distribution spatiale des espèces

exotiques à large échelle spatiale. Ces résultats ont été comparés à ceux obtenus avec

les espèces natives. J’ai enfin déterminé la répartition géographique de la richesse en

espèces non natives de poissons d’eau douce à l’échelle mondiale (P5). Ce travail se

base sur des informations récoltées sur 1055 bassins hydrographiques recouvrant 80%

de la surface continentale. J’ai ainsi identifié les principaux hotspots globaux

d’invasion de poissons d’eau douce, c'est-à-dire les bassins hydrographiques

comportant une forte proportion d’espèces non natives. Ensuite, j’ai testé, pour la

première fois à l’échelle mondiale, les hypothèses les plus couramment émises dans la

littérature pour expliquer la répartition géographique de la richesse en espèces non

natives (p.ex. Stohlgren et al. 1999 ; Taylor & Irwine 2004 ; Fridley et al. 2007). J’ai

ainsi tenté de caractériser la susceptibilité d’un bassin hydrographique à accueillir un

grand nombre d’espèces non natives.


En macroécologie, qui se définit comme l’étude statistique des patrons d’abondance,

de distribution et de diversité des espèces à des échelles larges (Brown 1989), l'approche

expérimentale n’est pas concevable, ce qui entraîne l’absence d’une situation témoin.

Néanmoins, on peut remédier à ce problème à l'aide de deux types d’approche quantitative. La

première est l’approche comparative que j’ai utilisé dans P1, P4 et P5. Elle implique

l’utilisation de techniques d'analyse de données ou de statistiques traditionnelles pour

identifier des structures ou tester des hypothèses à partir de la variabilité observée entre

situations (Diamond 1983). La deuxième approche est celle des « modèles nuls » (Connor &

Simberloff 1979 ; Gotelli & Graves 1996) que j’ai utilisé dans P3, en simulant des

assemblages d’espèces attendues sous l’hypothèse nulle que le facteur testé n'intervient pas.

Le but est de remplacer par des simulations de « Monte Carlo », le « témoin » d’une approche

expérimentale irréalisable dans la majorité des cas en écologie des communautés et en

biogéographie.

Ce mémoire présente une synthèse des résultats que j’ai obtenus grâce à de

nombreuses collaborations :

Dans P1, P3 et P4, j’ai compilé les données biologiques et environnementales à partir

de nombreuses sources bibliographiques et de bases de données existantes (Tableau 2).

En particulier, P1 présente les travaux que j’ai menés en Nouvelle Zélande, à

l’université d’Otago, dans le groupe de recherche du Professeur Colin Townsend

(Octobre 2004-Février 2005).

Dans P2, Karl Kreutzenberger (étudiant en DESUPS que j’ai encadré en co-direction

avec Sébastien Brosse) était responsable de la partie expérimentale de l’étude. J’ai

participé à l’échantillonnage des individus (Tableau 2), à la mise en place du plan

expérimental, à l’analyse des données et à la rédaction de la publication.


La base de données que j’ai explorée dans P5 (Tableau 2), à été initiée au début de

cette thèse par l’équipe « structure des communautés & macroécologie » de l’UMR

5174 « Evolution et Diversité Biologique » (EDB, université Paul Sabatier, Toulouse).

L’exploitation de cette base de données fait actuellement l’objet d’une collaboration

entre l’UMR EDB (Toulouse), l’IRD (Paris) et le CEMAGREF (Aix-en-Provence)

dans le cadre d’un projet de recherche financé par l’Agence Nationale pour la

Recherche (ANR biodiversité (2007-2010) « Freshwater Fish Diversity », ANR-06-

BDIV-010). Une grande partie des données a été récoltée par Olivier Beauchard

(ancien membre de l’équipe « structure des communautés & macroécologie » et

actuellement étudiant en thèse à l’université d’Anvers en Belgique). Ces données

proviennent d’une recherche bibliographique intensive (publications, rapports et

documents internet, atlas de poissons d’eau douce nationaux). L’Institut de Recherche

pour le Développement (IRD) a fourni les listes faunistiques des bassins

hydrographiques d’Afrique de l’Ouest et d’Amérique Central et du Sud. J’ai

également participé durant ces 3 années de thèse à la récolte des données. J’ai en

particulier été responsable de vérifier les statuts des espèces (espèces natives vs. non-

natives) des 1055 bassins hydrographiques.


Tableau 2 : Données utilisées dans chaque publication.

Contenu Sources

P1

Données environnementales et occurrences d’espèces pour 135 sites du bassin hydrographique de la rivière Manuherikia (région d’Otago, Nouvelle Zélande).

-Occurrences d’espèces compilées à partir de la « New Zealand Freshwater Fish Database, NIWA » -Données hydrologiques quantifiées à partir des données brutes de l’ « Otago Regional Council ». -Données environnementales compilées à partir du REC (River Environmental Classification) qui est un SIG sur les rivières de Nouvelle Zélande.

P2

Données provenant d’expérimentations menées en microcosmes (aquariums) à l’UPS

Brochets, gardons et poissons-chats collectés en milieu naturel et provenant de vidange d’étang.

P3 P4

-Liste d’espèces natives et non-natives pour les 26 principaux bassins hydrographiques européens. -Données environnementales et humaines pour les 26 principaux bassins hydrographiques européens.

-Liste d’espèces compilée à partir de publications/rapports/ouvrages. -Données environnementales et humaines collectées à partir du World Ressource Institute (2003), de la base de données climatiques de Leemans & Cramer (1991) et de publications/rapports/ouvrages.

P5

-Base de données « Freshwater Fish Diversity » : listes faunistiques pour 1055 bassins hydrographiques. (environ 10 000 espèces et 40 000 occurrences) -Données environnementales et socio-économiques pour 597 bassins hydrographiques.

-Données biologiques collectées à partir de publications/rapports/ouvrages et de bases de données existantes (IRD).

- Données environnementales et socio-économiques collectées à partir du CIESIN, de l’Atlas of Biosphere, d’un atlas géographique et de publications/rapports/ouvrages.


1. Les introductions d’espèces à l’échelle locale Cette thématique à fait l’objet de deux publications : P1 Leprieur F., Hickey M.A., Arbuckle C.J., Closs G.P., Brosse, S. & Townsend C.R.

(2006) Hydrological disturbance benefits a native fish at the expense of an exotic fish. Journal of Applied Ecology, 43, 930-939.

P2 Kreutzenberger K., Leprieur F., & Brosse, S. Impact of the invasive black bullhead

(Ameiurus melas Raff.) on the predatory efficiency of northern pike (Esox lucius L.) Journal of Fish Biology (en révision)

2.1. Le cas de la truite commune (Salmo trutta L.) introduite en Nouvelle Zélande (P1)

Dès lors qu’une espèce introduite s’établit avec succès dans un milieu d’accueil

(espèce non native), son aire géographique peut considérablement augmenter si elle devient

invasive (Lockwood et al. 2007). Néanmoins, plusieurs mécanismes peuvent limiter la

prolifération d’une espèce invasive : (i) la présence de prédateurs et de compétiteurs

(hypothèse de résistance biotique : Elton 1958 ; Levine 2000 ; Kennedy et al. 2002) et (ii)

des conditions abiotiques défavorables pour l’espèce invasive (hypothèse de résistance

abiotique : Elton 1958 ; Simberloff 1986,1989 ; Moyle & Light 1996). Ainsi, l’impact des

espèces invasives sur les espèces natives peut varier localement (c'est-à-dire d’un site à un

autre) au sein de son aire d’introduction (Palmer & Ricciardi 2004).

En Nouvelle Zélande, la truite commune (Salmo trutta L.) a été introduite avec succès

en 1867 pour la pêche sportive (Townsend 1996). Aujourd’hui, la truite a colonisé l’ensemble

des bassins hydrographiques de Nouvelle Zélande (Figure 3). Le succès de son établissement

et de sa prolifération peut s’expliquer par trois principaux facteurs (Townsend 1996): (i) une

très forte pression de propagules par les sociétés de pêche de Nouvelle Zélande (la pêche

sportive est l’une des principales ressources financières en Nouvelle Zélande) ; (ii) des

conditions abiotiques similaires à celles trouvées dans l’aire native de la truite (Europe); (iii)


une très faible résistance biotique. En effet, les rivières de Nouvelle Zélande sont pauvres en

espèces de poissons d’eau douce (comme la plupart des rivières des milieux insulaires). De

plus, les poissons d’eau douce de Nouvelle Zélande ont évolué sans prédateurs naturels,

jusqu’à l’introduction de la truite qui est une espèce piscivore et territoriale, dont le

comportement agressif est largement reconnu.

A BA B

Figure 3 : A) Distribution spatiale de la truite commune (Salmo trutta L.) en Nouvelle Zélande (en rouge foncé) ; B) Distribution spatiale de Galaxias anomalus (Stockell) en Nouvelle Zélande (en rouge foncé). Cette espèce est endémique des rivières Taieri et Clutha (dont l’un des principaux affluents est la rivière Manuherikia).

L’introduction de la truite en Nouvelle Zélande a eu des conséquences négatives sur

les différents niveaux d’organisation écologiques des écosystèmes aquatiques (Townsend

2003). Cette espèce est en particulier responsable du déclin des populations non-migratrices

de galaxidae du genre Galaxias (McDowall 2006). La prédation et la compétition

interspécifique sont les principaux mécanismes responsables de ce déclin (Townsend & Crowl

1991 ; Townsend 2003). Aujourd’hui, la plupart des espèces non-migratrices de galaxidae de

Nouvelle Zélande sont au bord de l’extinction (McDowall 2006).


L’impact de la truite sur les populations de Galaxias en Nouvelle Zélande est connu

depuis l’étude menée par Townsend & Crowl (1991) sur le bassin hydrographique de la

rivière Taieri (région d’Otago, île du sud). Ces auteurs ont montré que les populations de

Galaxias sont fragmentées à l’échelle du bassin. En effet, la distribution des Galaxias est

réduite à l’amont de cascades naturelles (hautes de plus de 3 m), c'est-à-dire aux zones non

accessibles par les truites. En aval des cascades, la truite a provoqué la disparition totale des

Galaxias. La très forte pression de prédation exercée par la truite sur les Galaxias explique

que ces espèces n’occurrent pratiquement jamais ensemble (distribution spatiale disjointe)

dans le bassin de la rivière Taieri (Townsend 2003).

De manière à tester si les résultats obtenus par Townsend & Crowl (1991) étaient

transposable à un bassin hydrographique comportant très peu de cascades naturelles, nous

avons analysé la distribution spatiale de la truite et d’une espèce endémique de

galaxidae (Galaxias anomalus Stockell) au sein du bassin de la rivière Manuherikia

(rivière voisine de la rivière Taieri). De plus, contrairement à la rivière Taieri, la rivière

Manuherikia fait l’objet de nombreux prélèvements d’eau pour l’irrigation. Ces prélèvements

d’eau destinés à l’irrigation des cultures tendent à aggraver les sécheresses naturelles des

rivières de la région d’Otago. Ainsi, certains tronçons de rivière présentent des étiages très

sévères durant une grande partie de l’année. Seules des mouilles (parties profondes des

rivières avec un courant lent, aussi appelées vasques ou pools) persistent, mais présentent des

conditions abiotiques extrêmes en période d’étiage. Des températures de l’eau supérieures à

28°C ont été reportées, ainsi que des taux d’oxygène dissous très faibles.

Dans ce contexte, nous avons voulu déterminer les conséquences de ces

perturbations hydrologiques sur la distribution spatiale de la truite et de G. anomalus au

sein du bassin de la rivière Manuherikia. En effet, d’après de nombreuses études (Moyle

Light 1996 ; Byers 2002 ; Stromberg et al. 2007 ; Lockwood et al. 2007), les modifications


d’habitats d’origine anthropique entraineraient le déclin des populations natives et

favoriseraient l’établissement et la prolifération d’espèces non-natives. Ces dernières seraient

de part leur forte tolérance environnementale mieux adaptées aux nouvelles conditions

abiotiques crées par les perturbations anthropiques (Stromberg et al. 2007 ; Lockwood et al.

2007).

Pour analyser la distribution spatiale de la truite et de G. anomalus, nous avons

compilé des données d’occurrence de ces espèces dans 135 sites distribués dans l’ensemble

du bassin de la rivière Manuherikia. Pour chaque site, nous avons collecté des données

relatives à différents descripteurs environnementaux (à l’échelle du site et du tronçon de

rivière). Par exemple, nous avons quantifié, pour chaque site, le risque d’étiage du aux

prélèvements d’eau pour l’irrigation en amont de chaque site (c.-à-d. le risque que la rivière

ait un débit très faible voir quasiment nul). Nous avons d’abord cherché à déterminer si les

sites comportant des truites et G. anomalus en sympatrie et en allopatrie étaient différents

d’un point de vue de leurs caractéristiques environnementales. Ensuite, nous avons cherché à

prédire l‘occurrence de la truite et de G. anomalus dans les 135 sites. Enfin, la quantification

de la contribution de chaque variable environnementale dans le modèle prédictif de type

réseau de neurones artificiels (ANN) a permis d’identifier les variables responsables de la

distribution spatiale de la truite et de G. anomalus dans la rivière Manuherikia.

D’abord, nous montrons que la truite et G. anomalus ont une distribution disjointe,

confirmant ainsi les études menées sur d’autres rivières de Nouvelle Zélande (dont la rivière

Taieri), d’Australie et de Tasmanie (Crowl & Townsend 1992; McIntosh 2000 ; McDowall

2006). Dans le bassin de la Manuherikia, 75% des sites étudiés sont habités uniquement par la

truite, confirmant ainsi son caractère invasif. De plus, l’absence de G. anomalus dans ces sites

a conforté l’idée que la truite est le principal responsable du déclin des populations de G.

anomalus.


Ensuite, les sites où la truite est absente (15%) sont ceux où l’on observe uniquement

G. anomalus. Ces sites sont caractérisés par un risque d’étiage maximal. De plus, les faciès

d’écoulement de l’eau (radier, plat, rapide), caractéristiques des milieux lotiques, sont très peu

représentés dans ces sites, qui sont composés essentiellement de mouilles (signe d’un étiage

sévère). Ainsi, il semblerait que les conditions environnementales crées par les étiages

(aggravés dans l’espace et dans le temps par les prélèvements d’eau) aient à la fois limité

l’invasion de la truite et protégé les populations de G. anomalus. En effet, les espèces non

migratrices du genre Galaxias sont capables de s’enfouir dans le substrat lors de conditions de

très faibles débits (Dunn 2003 ; Davey et al. 2006). Ces poissons sont également capables de

supporter les températures élevées rencontrées dans les mouilles durant les périodes d’étiage

(voir par exemple Closs & Lake 1996). Au contraire, il est largement reconnu que la truite

(espèce d’eau froide) ne supporte pas les températures d’eau élevées (et en règle générale les

conditions environnementales crées par les étiages, Matthews & Berg 1997). En réponse à ces

perturbations, la truite a tendance à migrer vers des zones non impactées ayant des eaux plus

fraîches et plus oxygénées (Gowan & Fausch 1996).

Ces résultats ont deux implications principales, l’une fondamentale et l’autre en

relation avec la conservation de G. anomalus :

Les perturbations hydrologiques d’origine anthropique (c.-à-d. des étiages aggravés par

les prélèvements d’eau pour l’irrigation) limitent la prolifération de la truite et permettent

aux populations de G. anomalus de se maintenir (Figure 4). Ce résultat va à l’encontre de

beaucoup d’études montrant l’influence positive des perturbations anthropiques sur le

succès d’invasion (voir Lockwood et al. 2007 pour une synthèse). En fait, une

perturbation, qu’elle soit naturelle ou anthropique, peut favoriser une espèce

invasive (ou native) si les nouvelles conditions environnementales (filtre abiotique)

sont en adéquation avec les caractéristiques écologiques et l’histoire évolutive de


l’espèce considérée (Townsend 2003). Dans le cas présent, les espèces non migratrices

du genre Galaxias sont adaptées aux conditions climatiques de la région d’Otago

(tolérance aux fortes températures, capacité à s’enfouir dans le substrat et de survivre dans

des mouilles résiduelles en cas d’assèchement temporaire du cours d’eau). Au contraire, la

truite est incapable de faire face à de telles conditions environnementales. Lors des

périodes d’étiages sévères, le maintien de la truite dans le bassin de la rivière Manuherikia

est probablement du à sa forte capacité de dispersion, qui lui permet de migrer dans des

zones plus stables d’un point de vue hydrologique. Récemment, le mécanisme de

résistance environnementale à l’invasion de la truite a également été reporté dans des

rivières de la côté ouest de l’île du sud de Nouvelle Zélande (Olsson et al. 2006). Dans

cette étude, il est montré que l’acidité naturelle de certaines rivières empêche l’invasion de

la truite et permet ainsi à des populations de Galaxias de persister.

Ce travail a confirmé le mécanisme initialement détecté sur la rivière Taieri (Townsend &

Crowl 1991), c'est-à-dire un contrôle biotique et abiotique de la distribution spatiale

des espèces non-migratrices de galaxidae dans la région d’Otago. La truite aurait

limité tout d’abord la distribution des Galaxias (contrôle biotique : prédation et

compétition) et les facteurs abiotiques de certains sites auraient permis ensuite aux

Galaxias d’y survivre en empêchant la colonisation de la truite (c.-à-d. les cascades de

plus de 3 m pour la rivière Taieri et les sites impactés par les prélèvements d’eau pour la

rivière Manuherikia). Ainsi, au regard de la conservation de G. anomalus (qui est l’une

des espèces non-migratrices de galaxidae les plus menacées d’extinction en Nouvelle

Zélande), nous avons préconisé (i) la mise en place de cascades artificielles

supérieures à 3 mètres dans certaines rivières tests et (iii) l’éradication de la truite en

amont de ces cascades. Nous avons également préconisé une gestion contrôlée de la

ressource en eau, car bien que les sites impactés par les prélèvements d’eau servent de


zones refuge pour G. anomalus, leurs conditions abiotiques peuvent avoir des effets

négatifs à long terme sur la reproduction des Galaxias (Allibone 2000). Au vu de nos

résultats, il est très important que la restauration des débits ne soit pas conduite avant la

création de cascades permettant aux Galaxias de maintenir une population viable en

amont. En effet, un retour à la normale des débits entrainerait la disparition des sites

refuges pour G. anomalus. La colonisation de ces sites par la truite mettrait alors en péril

les dernières populations de G. anomalus en Nouvelle Zélande.

Figure 4 : Schéma synthétique montrant l’impact négatif de la truite sur les Galaxias dans des sites non impactés par les prélèvements d’eau (flèche rouge). Au contraire, dans les sites impactés par les prélèvements d’eau (flèche orange), la truite ne peut pas survivre et donc n’interagit pas avec les Galaxias.

Impact de la truite

+++

Prédation/compétition

Truite

Galaxias

Perturbations hydrologiques dues aux prélèvements

d’eau : étiage prolongé

Impact de la truite

- - -

Impact de la truite

+++

Prédation/compétition

Truite

Galaxias



Impact de la truite

- - -

Truite

Galaxias



Impact de la truite

- - -


2.2. Le cas du poisson-chat (Ameiurus melas Raff.) introduit en Europe (P2)

Bien que les espèces invasives de poissons d’eau douce soient bien identifiées à

travers le monde (Welcomme 1988 ; Lever 1996 ; Elvira 2001), beaucoup d’entre elles n’ont

jamais fait l’objet d’études quant à leur impact sur la biodiversité native. Pourtant, la

connaissance de l’impact réel d’une espèce invasive est un pré-requis indispensable pour la

gestion de ces espèces ainsi que pour la mise en place de mesures de conservation des espèces

impactées (Lodge et al. 1998).

Le poisson-chat (Ameiurus melas Raff.) est originaire d’Amérique du Nord et a été

introduit en Europe (et plus particulièrement en France) à la fin du 19ième siècle. Cette espèce

a ensuite rapidement colonisé une grande partie des bassins hydrographiques européens, de la

péninsule ibérique jusqu’au sud du Danemark (Elvira 2001). Bien que le poisson-chat soit

considéré comme nuisible et invasif (Elvira 2001; Keith & Allardi 2001 ; Cucherousset et al.

2006), aucune étude n’a jusqu’à présent mis en évidence l’impact du poisson-chat sur les

espèces natives européennes.

Nous avons ici tenté de quantifier l’impact du poisson-chat sur le brochet (Esox

lucius L.) une espèce prédatrice européenne, considérée comme vulnérable (Keith &

Allardi 2001). Nous avons en particulier mis en place une approche expérimentale

permettant de quantifier l’impact du poisson chat sur l’efficacité de prédation du

brochet (P2). Le poisson-chat et le brochet cohabitent dans différents types de milieux

aquatiques : réservoirs, étangs, marais, partie basse des rivières et fleuves (Cucherousset et al.

2006), et le poisson chat est susceptible d’affecter la prédation du brochet car son régime

alimentaire comprend une proportion non négligeable de poissons (Boët 1980). De plus, de

part sa forte densité dans les milieux qu’il envahit (Cucherousset et al. 2006), le poisson-chat

peut perturber le comportement de prédation du brochet et/ou le comportement anti-prédateur

des proies. Enfin, le poisson-chat peut générer une importante turbidité de l’eau (Braig &


Johnson 2003), susceptible d’affecter l’efficacité de prédation du brochet qui est un prédateur

visuel. Nous avons donc ici testé (i) un effet direct du poisson-chat sur l’efficacité de

prédation du brochet (par compétition pour la ressource trophique et/ou par interférences

comportementales avec la proie et/ou le prédateur), et (ii) un effet de la turbidité générée par

le poisson-chat sur l’efficacité de prédation du brochet.

Nous avons donc utilisé un protocole expérimental de type « multi-prédateur »

(Tableau 3), qui consiste à comparer la prédation de chaque espèce séparément à la prédation

générée par les deux espèces (brochet + poisson-chat) (Griffen 2006). Les expérimentations

ont été conduites dans des aquariums de 200 litres à deux niveaux de turbidité (eau turbide

(TW) et eau non turbide (CW). En fin d’expérience (après 72 h) le nombre de poissons

restants ainsi que le nombre de poissons ingérés par chaque prédateur a été compté.

Tableau 3 : protocole expérimental. Chaque traitement a été répliqué 8 fois. Entre 2 répliques, les poissons et l’eau des aquariums ont été renouvelés de manière à éviter des biais potentiels du à des substances chimiques émises par les poissons ou à une pseudoreplication.

Nombre de poissons introduits

Bac expérimental

TW1

CW1

1 -

TW2

TW3

CW2

CW3

Brochet Poisson-chat Gardon

3

10

10

10

10

10

3-

10

1

1 -

- 3

1 3

TW, eau turbide ; CW, eau non turbide

Turbiditémoyenne (NTU)

70.1±2.5

72.5±2.5

73.2±3.0

1.5±0.3

1.5±0.3

1.6±0.3

Nombre de poissons introduitsBac expérimental

TW1

CW1

1 -

TW2

TW3

CW2

CW3

Brochet Poisson-chat Gardon

3

10

10

10

10

10

3-

10

1

1 -

- 3

1 3

TW, eau turbide ; CW, eau non turbide

Turbiditémoyenne (NTU)

70.1±2.5

72.5±2.5

73.2±3.0

1.5±0.3

1.5±0.3

1.6±0.3


Les résultats montrent qu’une turbidité élevée (70 NTU) n’influence pas l’efficacité de

prédation du brochet, ce qui est inattendu étant donné l’abondance des études montrant un

effet négatif de la turbidité sur l’efficacité de prédation des prédateurs visuels (p.ex. Reid et al.

1999 ; Pekcan-Hekim & Lappalainen 2006). Ceci peut s’expliquer par le fait que la turbidité

peut également influencer le comportement de la proie (en particulier son comportement anti-

prédateur). En effet, une proie peut moins bien discerner un prédateur en milieu turbide et

ainsi être plus vulnérable à la prédation (Gregory 1993), annulant donc l’impact négatif de la

turbidité sur le prédateur.

Indépendamment du niveau de turbidité, nos résultats révèlent une diminution

significative du nombre de proies ingérées par le brochet en présence du poisson-chat (Figure

5). Ceci semble dû à une forte interférence comportementale entre le brochet et les poissons-

chats, puisque ceux-ci ont eu un comportement agressif envers le brochet (traitement CW3,

poisson-chat et brochet en sympatrie). Ces agressions ont sûrement affecté le bon déroulement

des séquences comportementales de prédation du brochet (Eklöv & Diehl 1994), entraînant

ainsi une diminution significative de l’efficacité de sa prédation.

Brochet Poisson-chat

Brochet+

Poisson-chat

Controle0

2

4

6

8

10 CWTW

Figure 5: Nombre de proies restantes (± SE, n = 8) dans chaque traitement : eau non turbide (CW) et eau turbide (TW). Le contrôle correspond à un traitement où il y 10 gardons et aucun prédateur.


Ces résultats obtenus en milieu expérimental montrent pour la première fois que

le poisson-chat peut avoir un effet négatif sur une espèce native européenne. En effet, la

diminution de l’efficacité de prédation du brochet peut entraîner une diminution de sa

croissance ou/et affecter la survie individuelle. Elle peut également influencer la

sélection des proies disponibles, et par conséquent modifier la structure locale de la

communauté (Eklöv & Hamrin 1989). Cependant, la variabilité environnementale

(disponibilité des ressources, structures de l’habitat, …) étant difficile à reproduire en

laboratoire, nos résultats doivent être confirmés en milieu naturel ou semi-naturel.

2.3. Conclusion et perspectives Les résultats obtenus dans P1 et P2 à l’échelle locale ont permis de mettre évidence que :

L’impact d’une espèce invasive sur une espèce native peut varier spatialement en

fonction des caractéristiques abiotiques locales (P1)

Les perturbations anthropiques telles que la modification des habitats ne favorisent

pas forcément les espèces invasives (P1)

La conservation d’une espèce native menacée nécessite des mesures de gestion

adaptées au contexte environnemental local (P1)

Une espèce considérée à priori comme invasive et nuisible doit faire l’objet

d’études quant à son impact réel sur les espèces natives ; ceci afin de mettre en place

des mesures de gestion adaptées aux caractéristiques comportementales et écologiques

de l’espèce impactée. (P2)

Les perspectives associées à P1 et P2 sont nombreuses :

Il serait intéressant de comparer le taux de croissance et de survie (« fitness ») de

populations de G. anomalus dans des sites non perturbés d’un point de vue


hydrologique et non colonisés par la truite (sites situés en amont de cascades) à ceux

observés dans les sites impactés d’un point de vue hydrologique. En effet, bien que les

sites impactés par les prélèvements d’eau servent de zone refuge pour G. anomalus,

les conditions hydrologiques rencontrées dans ces sites sont susceptibles de perturber à

long terme la dynamique des populations de Galaxias en termes de succès

reproducteur, croissance et survie individuelle (Allibone 2000). Il serait également

utile de comparer la diversité génétique (p.ex. le niveau de consanguinité) des

populations de G. anomalus occupant des milieux impactés ou non impactés par les

prélèvements d’eau. Ceci permettrait d’évaluer si les conditions abiotiques provoquées

par les prélèvements d’eau accentuent le risque à long terme d’extinction des

populations de G. anomalus.

Il est prévu d’étendre l’étude menée sur le bassin hydrographique de la rivière

Manuherikia à l’ensemble des bassins de la région d’Otago qui sont également

impactées par des prélèvements d’eau et qui comportent pour certains des cascades

naturelles (projet en cours avec Chris Arbuckle « Southland Regional Council » et

Colin Townsend « University of Otago, Dunedin »). Ceci permettrait de cartographier

les zones refuges pour les populations de différentes espèces de Galaxias et ainsi

d’identifier les zones prioritaires de conservation.

Enfin, il est nécessaire de mettre en place des études en milieux semi-naturels, tels que

des enclos expérimentaux installés dans des étangs, afin de vérifier et d’étendre les

résultats obtenus en laboratoire sur l’impact du poisson chat (P2). Ce type

d’expérimentations permettrait de tester différentes conditions environnementales afin

d’obtenir des résultats directement transposables au milieu naturel (voir Lodge et al.

1998 ; White & Harvey 2001). Par exemple, il serait utile de manipuler la densité des

proies et prédateurs natifs (gardon et brochet) ainsi que la densité de l’espèce invasive


(le poisson-chat). De plus, le poisson chat colonisant des milieux fréquemment

affectés par des perturbations physiques, la prise en compte des modifications

d’habitats dans les expérimentations semble nécessaire (cf. P1) à la prédiction de

l’impact réel du poisson chat sur les assemblages de poissons Européens.

3. Les introductions d’espèces à l’échelle régionale Cette thématique à fait l’objet de trois publications : P3 Leprieur F., Beauchard O., Hugueny B., Grenouillet G. & Brosse S. (2007) Null

model of biotic homogenization: a test with the European freshwater fish fauna. Diversity and Distributions (sous presse).

P4 Leprieur F., Olden, J.D. Lek, S. & Brosse S. Patterns and mechanisms of the

distance decay of similarity in the European freshwater fish fauna: contrasting native and exotic species. (en préparation).

P5 Leprieur F., Beauchard O., Blanchet S., Oberdorff T. & Brosse S. Fish invasions in

the world’s river systems: when natural processes are blurred by human activities. (en révision dans PLoS Biology)

Les publications P3, P4 et P5 portent sur les introductions d’espèces de poissons d’eau

douce à l’échelle régionale, c’est à dire à l’échelle du basin hydrographique (unité spatiale).

L’ensemble des espèces de poissons d’eau douce d’un bassin hydrographique constitue un

pool régional d’espèces distinct de celui des autres bassins. En effet, les zones terrestres et les

océans constituent des barrières géographiques infranchissables pour les poissons d’eau douce,

et chaque bassin hydrographique peut par conséquent être considéré comme une île

biogéographique (Hugueny 1989 ; Oberdorff 1995 ; Reyjol et al. 2006). En replaçant l’étude

des introductions d’espèces dans le cadre de la théorie de la biogéographique insulaire

(McArthur & Wilson 1967 ; Whittaker & Palacios 2007), il est alors possible d’évaluer leur

impact sur la richesse et la composition du pool régional d’espèces (P3). De plus,

l’identification des déterminants de la composition (P4) et de la richesse (P5) en espèces non

natives à l’échelle régionale semble être un pré-requis indispensable à une meilleure


prédiction des assemblages d’espèces non natives à l’échelle locale. En effet, il est

aujourd’hui largement reconnu que la richesse et la composition des espèces à l’échelle locale

sont fortement dépendantes des facteurs agissant à l’échelle régionale (Ricklefs 1987

Hugueny & Paugy 1995 ; Angermeir & Winston 1998).

3.1 . Processus d’homogénéisation biotique des assemblages régionaux : le cas des poissons d’eau douce européens (P3)

Ces dernières années, de nombreuses études se sont intéressées à l’impact des

introductions d'espèces sur la diversité alpha (richesse en espèces) et beta (différence de

composition d’espèces entre localités) (voir la revue de Sax & Gaines 2003). En effet,

certaines régions faunistiques ou floristiques sont devenues de plus en plus similaires d’un

point de vue taxonomique (diminution de la diversité beta) suite à des introductions et/ou des

extinctions d’espèces, un processus appelé « homogénéisation biotique » (Rahel 2002 ;

Olden & Poff 2002; Qian & Ricklefs 2006 ; Cassey et al. 2007).

Dans P3, un modèle nul d’homogénéisation biotique a été développé et appliqué aux

assemblages régionaux de poissons d’eau douce européens. En utilisant un modèle nul, nous

avons voulu tester si l’homogénéisation biotique est un phénomène non aléatoire (c'est-à-dire

le résultat d’une distribution géographique non aléatoire des espèces non natives, Duncan et al.

2001 ; Olden et al. 2004). Ce modèle nul consiste à simuler un grand nombre de fois une

distribution aléatoire des espèces non natives entre les principaux bassins hydrographiques

européens (Figure 6), puis à calculer les changements de similarité taxonomique entre bassins

causés par les introductions d’espèces exotiques. Les résultats observés en milieu naturel sont

ensuite comparés aux résultats obtenus avec le modèle nul (10 000 simulations).

Nous avons ensuite détaillé cette approche en considérant séparément les espèces

exotiques (espèces non natives d’origine non européenne) et les espèces transloquées (espèces


non natives d’origine européenne). Nous avons testé l’hypothèse que contrairement aux

espèces exotiques, les espèces transloquées tendent à augmenter la similarité taxonomique

entre localités (McKinney 2005 ; La Sorte & McKinney 2006).

B

1000Km

N

D

1

23

45

67

8 9

10

1112 13 14

15

16

17 18

19

20

21

2223

24

25

A [ -6 ; -4 [

[ -4 ; -2 [

[ -2 ; 0 [

[ 0 ; 2 [

[ 2 ; 4 [

[ 4 ; 6 [

[ 6 ; 8 [

[ 8 ; 10 [

[ 10 ; 13 [C

Figure 6: Distribution géographique du taux d’homogénéisation/différentiation par bassin hydrographique (Basin ΔCS %) (a) Carte des 25 principaux basins hydrographiques européens ; 1: Guadalquivir; 2: Tagus; 3: Douro; 4: Ebro; 5: Garonne; 6: Loire; 7: Seine; 8: Rhône; 9: Pô; 10: Rhine; 11: Weser; 12: Elbe; 13: Oder; 14: Wisla; 15: Danube; 16: Dniestr; 17: Dniepr; 18: Don; 19: Volga; 20: Ural; 21: Petchora; 22: Dniva; 23: Neva; 24: Kemijoki; 25: Glomma). (b) Basin ΔCS (%) calculé en prenant en compte à la fois les espèces exotiques et transloquées. (c) Basin ΔCS (%) calculé en prenant en compte que les espèces exotiques. (d) Basin ΔCS (%) calculé en prenant en compte que les espèces transloquées. Le dégradé de vert indique une différentiation taxonomique et le dégradé de rouge indique une homogénéisation taxonomique. Dans P3, nous montrons que :

L’homogénéisation biotique n’est pas un phénomène écologique induit par des

processus aléatoires, confirmant ainsi pour la première fois les hypothèses émises

dans de nombreuses études. Ceci peut s’expliquer par de nombreux facteurs tels

que la sélection par l’homme des espèces non natives (Blackburn & Duncan 2001),


les caractéristiques environnementales des milieux d’accueil (Kennard et al. 2005),

ainsi que les caractéristiques écologiques des espèces non natives (Moyle & Marchetti

2006).

Les bassins hydrographiques qui sont similaires quant à leur faune native, et par

conséquent proches géographiquement (Nekola & White 2001), ont plus de

chance d’avoir des assemblages similaires d’espèces non natives qu’attendu par

le hasard. Deux mécanismes non mutuellement exclusifs peuvent expliquer ce

résultat : (i) les espèces natives et non natives seraient sélectionnées par les mêmes

filtres environnementaux; (ii) les transferts accidentels (p.ex. par la construction de

canaux reliant les bassins) ou intentionnels (p.ex. pour la pêche sportive) d’espèces

seraient structurés géographiquement. Ces deux hypothèses ont été testées dans P4.

Les espèces exotiques tendent à diminuer la similarité taxonomique entre bassins

(différentiation) alors que les espèces transloquées tendent à augmenter leur

similarité taxonomique (homogénéisation) (Figure 6). En effet, les espèces

exotiques ont une distribution localisée en Europe (une espèce exotique est observée

en moyenne dans moins de bassins que les espèces transloquées). Au contraire, une

grande partie des espèces transloquées (originaire d’Europe de l’Est) ont vu leur aire

de distribution augmenter suite à leur introduction dans les bassins d’Europe de

l’Ouest qui sont moins riches en espèces (Reyjol et al. 2006). Au final, les espèces

exotiques et transloquées ont entraîné conjointement une augmentation moyenne de la

similarité taxonomique de 2%.

Les introductions d’espèces de poissons d’eau douce en Europe n’ont pas causé

d’extinction d’espèces natives à l’échelle régionale. Ainsi, la richesse régionale en

poissons d’eau douce a augmenté en Europe suite aux introductions d’espèces. Ceci a

également été reporté aux Etats-Unis (Gido & Brown 1999). Néanmoins, des


extinctions locales causées par certaines espèces non natives ont été observées en

Europe, et en particulier dans plusieurs bassins hydrographiques de la région

méditerranéenne (p.ex. Bianco 1995 ; Elvira 2001). De plus, les extinctions d’espèces

se déroulent généralement à des échelles de temps plus grandes que le phénomène

d’introduction d’espèces (Sax et al. 2002). Cela implique que notre perception de

l’impact des espèces non natives est dépendante de l’échelle temporelle à laquelle

les observations sont réalisées. Ainsi, l’augmentation actuelle du pool régional de

poissons d’eau douce en Europe ne doit pas être interprétée comme forcément

bénéfique pour la biodiversité.

3.2. Rôle des facteurs environnementaux et géographiques dans la

structuration des assemblages régionaux : une comparaison entre les espèces natives et exotiques de poissons d’eau douce européens (P4)

Les géographes ont depuis longtemps observé que la similarité entre observations

diminuait en fonction de la distance géographique les séparant (Tobler 1970). En

biogéographie, il est également très courant d’observer une relation négative entre la

similarité taxonomique et la distance géographique (Nekola & White 2001 ; Soininen et al.

2007). Selon Soininen et al. (2007), ce phénomène serait du : (i) aux différences de tolérance

physiologique des espèces le long de gradients environnementaux qui sont très souvent

spatialement structurés (hypothèse des filtres environnementaux) ; et/ou (ii) à la

configuration du paysage (p.ex. la présence de barrières géographiques limitant le mouvement

des organismes) et aux capacités différentielles de dispersion des espèces (hypothèse de

limitation de la dispersion).

Récemment, des études ont reporté que la similarité des plantes exotiques entre

régions (c.-à-d. les espèces originaires d’une autre zone biogéographique) diminuait en

fonction de la distance géographique les séparant (La Sorte & McKinney 2006 ; Qian &


Ricklefs 2006). Néanmoins, les mécanismes responsables de ce patron spatial sont encore peu

documentés chez les espèces exotiques (Qian & Ricklefs 2006) et plus particulièrement chez

les vertébrés exotiques. Les espèces exotiques étant moins limitées dans leur dispersion

que les espèces natives (p.ex. Rahel 2007 ; Lockwood et al. 2007), on devrait observer un

taux de remplacement des espèces en relation avec la distance géographique (c.-à-d. le

taux de turnover) plus faible chez les espèces exotiques.

Dans P4, nous avons d’abord exploré la structure géographique des assemblages

régionaux de poissons d’eau douce exotiques en Europe (c.-à-d. les espèces de poissons d’eau

douce qui ne sont pas natives d’Europe ; même base de données que dans P3). Ensuite, nous

avons testé si les différences de composition spécifique entre bassins étaient le résultat d’un

contrôle environnemental ou bien d’une limitation de la dispersion des espèces. Selon

l’hypothèse de limitation de la dispersion, seule la distance géographique entre bassins

hydrographiques expliquerait les différences de similarité taxonomique à travers

l’Europe. Enfin, nous avons comparé les résultats obtenus avec les espèces exotiques à ceux

observés chez les espèces natives.

Nos résultats montrent que :

La similarité taxonomique entre bassins diminue en fonction de la distance

géographique les séparant (Figure 7), aussi bien pour les espèces natives et que

pour les exotiques. Néanmoins, le taux de turnover des espèces exotiques est

supérieur à celui des espèces natives, bien que la différence soit très faible. Cela

confirme les résultats de P3, c’est à dire une faible diminution de la similarité

taxonomique entre bassins causée par l’introduction d’espèces exotiques.

La similarité taxonomique entre bassins est également corrélée négativement à la

différence de conditions climatiques entre bassins (c.-à-d. la température moyenne


annuelle et le nombre de jour de pluie), aussi bien pour les espèces natives et que pour

les exotiques. Autrement dit, les bassins hydrographiques ayant des conditions

climatiques similaires ont plus d’espèces exotiques et natives en commun que des

bassins ayant des conditions climatiques contrastées. Ceci semble indiquer que les

conditions climatiques sélectionnent les espèces exotiques et natives adaptées à ces

conditions. Nos résultats sont donc à première vue concordant avec l’hypothèse des

filtres environnementaux. Néanmoins, comme les bassins ayant des conditions

climatiques similaires sont proches géographiquement (c.-à-d. une forte auto-

corrélation spatiale du climat), il est difficile de distinguer l’effet de la distance

géographique (correspondant à l’hypothèse de limitation de la dispersion) de celui du

climat (correspondant à l’hypothèse de filtres environnementaux). En effet, d’un point

de vue statistique, la plus grande part de la variation de similarité taxonomique est

expliquée conjointement par la distance géographique et les différences de conditions

climatiques. De plus, de nombreuses études ont montré que les poissons d’eau douce

natifs en Europe ont été limités dans leur dispersion lors d’événements historiques tels

que les glaciations (voir Reyjol et al. 2006). Les poissons d’eau douce exotiques sont

probablement eux aussi limités dans leur dispersion, non pas par des facteurs

historiques, mais plutôt par l’homme qui sélectionnent les espèces à introduire. En

effet, une grande majorité des introductions de poissons d’eau douce en Europe ont été

intentionnelles (García-Berthou et al. 2005 ; García-Berthou 2007). Ainsi, on peut

conclure que l’hypothèse des filtres environnementaux et celle de la limitation de

la dispersion ne sont pas mutuellement exclusives pour expliquer la distribution

des poissons d’eau douce natifs et exotiques en Europe.


Log

sim

ilar

ité

taxo

no

miq

ue

0

0.2

0.4

0.6

0.8

0 1 2 3 4 5

0 1 2 3 4 50

0.2

0.4

0.6

0.8

Log distance géographique

Espèces natives (A)

Espèces exotiques (B)

Figure 7: Relation entre la similarité en espèce native (A) et exotique (B) par paire de bassins hydrographiques (indice de similarité de Jaccard) et la distance géographique les séparant. La distance géographique (euclidienne) et quantifiée à partir de la latitude et longitude moyenne de chaque bassin.

Nos résultats ne corroborent pas nos prédictions initiales, ni les résultats observés

chez les plantes aux Etats-Unis par La Sorte & McKinney (2006) et Qian & Ricklefs

(2006) ; c'est-à-dire un plus faible taux de remplacement des espèces en relation avec la

distance géographique chez les espèces exotiques. Plusieurs facteurs peuvent expliquer la


différence observée entre nos résultats et ceux obtenus avec les plantes exotiques aux Etats-

Unis. Le plus important d’entre eux semble être la différence de capacité de dispersion entre

les plantes et les poissons d’eau douce. En effet, un poisson d’eau douce introduit dans un

bassin hydrographique ne peut pas franchir sans l’aide de l’homme les barrières

géographiques séparant les bassins voisins. Par conséquent, les poissons d’eau douce sont

beaucoup plus limités au niveau de leur expansion géographique que les plantes. Ces

dernières peuvent, en effet, se disperser naturellement via différents vecteurs tels que le vent

(anémochorie) ou les animaux (zoochorie).

Enfin, le processus d’introduction d’espèces étant dynamique dans le temps (voir

Clavero & García-Berthou 2006), il ne serait pas surprenant que les poissons exotiques en

Europe voient leur aire de distribution augmenter suite à des introductions intentionnelles par

l’homme. Cette augmentation de l’aire de distribution des espèces exotiques ne sera possible

que si les caractéristiques environnementales des bassins récepteurs sont en adéquation avec

les exigences physiologiques des espèces introduites. Dans ce contexte, on devrait observer au

cours du temps une diminution du taux de remplacement des espèces exotiques en relation

avec la distance géographique (c.-à-d. une plus faible différence moyenne de composition

d’espèces exotiques entre les différents bassins hydrographiques européens).

3.3. Déterminants et répartition géographique mondiale de la richesse en

espèces non natives de poisson d’eau douce (P5)

De nombreux travaux ont cherché à expliquer les gradients globaux de richesse en

espèces natives chez différents groupes taxonomiques (p.ex. Pianka 1966 ; Oberdorff et al.

1995 ; Hawkins et al. 2003; Kreft & Jetz 2007). Concernant les poissons d’eau douce, la

distribution mondiale de la richesse régionale en espèces natives (c.-à-d. à l’échelle du bassin

hydrographique) est principalement expliquée par l’hétérogénéité de l’habitat (exprimée par

l’aire du bassin et le débit à l’embouchure ; hypothèse aire-espèces) et l’énergie disponible


dans le système (exprimée par la production primaire nette ; hypothèse énergie-espèces ;

Guéguan et al. 1998). Concernant les espèces non natives, les seules études ayant tenté

d’expliquer les gradients de richesse régionale en espèces non natives, ont été réalisées aux

Etats-Unis (voir Gido & Brown 1999 ; Stohlgren et al. 2003 ; Taylor & Irwine 2004 ;

McKinney 2006). Dans ce contexte, nous avons tenté dans P5 d’identifier pour la

première fois les principaux déterminants de la richesse régionale en espèces non-natives

en analysant une base de données mondiale sur la biodiversité des poissons d’eau douce

(c.-à-d. environ 40 000 occurrences de 10 000 espèces dans 1055 bassins

hydrographiques). Nous avons également identifié les « hotspots » d’invasion de poissons

d’eau douce, c'est-à-dire les bassins comportant une forte proportion d’espèces non

natives.

Trois principales hypothèses, non mutuellement exclusives, ont été émises pour

expliquer le nombre d’espèces non natives présentes dans un écosystème. L’hypothèse de

résistance biotique (Levine 2000) prédit que les communautés riches en espèces sont une

barrière à l’établissement des espèces introduites. Ainsi, selon cette hypothèse, on devrait

observer une relation négative entre les richesses en espèces natives et non natives.

L’hypothèse d’acceptation biotique prédit que les conditions abiotiques (p.ex. l’énergie

disponible, la diversité des ressources et des habitats) qui permettent la mise en place de

communautés riches en espèces natives, facilitent également l’établissement d’un grand

nombre d’espèces non natives (Fridley 2004). Selon cette hypothèse, on devrait observer une

relation positive entre les richesses en espèces natives et non natives et les variables

abiotiques. Enfin, l’hypothèse anthropique (Taylor & Irwine 2004 ; Meyerson & Mooney

2007) prédit que les activités humaines facilitent l’établissement des espèces introduites : (i)

en étant la principale source de propagules d’espèces (en particulier dans le cadre

d’introductions intentionnelles); (ii) en créant des milieux dégradés et/ou artificialisés qui


peuvent favoriser des espèces introduites ayant une forte tolérance environnementale. Selon

cette hypothèse, on devrait observer une relation positive entre des descripteurs des activités

humaines (p.ex. le PIB, la densité de population, le taux d’urbanisation) et la richesse en

espèces non natives.

Figure 8 : A) Distribution du pourcentage d’espèces non natives de poissons d’eau douce dans 1055 bassins hydrographiques ; B) Distribution géographique du nombre d’espèces non natives de poissons d’eau douce dans 1055 bassins hydrographiques. Les bassins indiqués en rouge dans la figure A sont considérés comme des hotspots d’invasion car ils comportent plus d’un quart d’espèces non natives.


Nos résultats montrent que :

Les bassins comportant une forte proportion d’espèces non natives (c.-à-d. des

hotspots d’invasion, Figure 8A) sont localisés sur la côte pacifique d’Amérique du

Nord et d’Amérique Central, en Europe de l’Ouest, en Afrique et du Sud et à

Madagascar et enfin dans le sud de l’Australie, de la Nouvelle Zélande et de

l’Amérique du Sud. L’ensemble de ces bassins sont également caractérisés par la plus

forte proportion d’espèces menacées d’extinction selon les critères de l’IUCN (2006).

Parmi les 3 hypothèses proposées pour expliquer la richesse en espèces non natives,

seule l’hypothèse anthropique est vérifiée. En effet, les variables associées aux

activités humaines, indépendamment des variables environnementales (aire du bassin,

range d’altitude, productivité primaire nette, richesse en espèces natives), expliquent la

majeure partie de la variation de richesse en espèces non natives entre bassins

hydrographiques. Parmi ces variables anthropiques, le Produit Intérieur Brut

explique le plus la variation mondiale de richesse en espèces non natives entre

bassins hydrographiques.

Les résultats obtenus dans P5 ont deux implications majeures pour la conservation de la

biodiversité. D’abord, contrairement à de nombreuses études réalisées à des étendues plus

fines (continent et région) (p.ex. Stohlgren et al. 2003 ; Taylor & Irwine 2004 ; Evans et al.

2005 ; Chown et al. 2005 ; Fridley et al. 2007), les conditions abiotiques et la richesse en

espèces natives influencent très peu la richesse en espèces non natives. Au contraire, ce sont

les activités humaines et plus particulièrement la richesse économique d’un bassin

hydrographique qui détermine sa susceptibilité à accueillir un grand nombre d’espèces

non natives.


Ainsi, ces résultats suggèrent que le développement économique prévu dans les

pays en voie de développement devrait s’accompagner d’un accroissement du nombre

d’espèces non natives de poissons d’eau douce. Un tel scénario serait préjudiciable au

maintien de la biodiversité aquatique de ces régions du monde qui comportent pour la plupart

un grand nombre d’espèces endémiques (Moyle & Cech 2004). Des mesures de prévention

sont donc nécessaires dans les pays en voie de développement car une fois établie, une espèce

non native est très difficile à éradiquer et cela engendre des coûts économiques très

importants (Pimentel et al. 2005). Ensuite, des mesures efficaces de contrôle de l’expansion

des espèces non natives doivent être mises en place dans l’ensemble des hotspots d’invasion,

lesquels comportent une forte proportion d’espèces menacées d’extinction. En effet, les

espèces invasives de poissons d’eau douce sont directement responsables du déclin de 20%

des poissons d’eau douce listés par l’IUCN (Olden et al. 2007).

3.4. Conclusions et perspectives

Les résultats obtenus à l’échelle régionale dans P3, P4 et P5 ont permis de mettre en

évidence que :

Les introductions d’espèces non natives de poissons d’eau douce ont conduit à une

augmentation de la diversité alpha des bassins hydrographiques européens (c.-à-d.

une augmentation du pool régional d’espèces), mais ont provoqué une diminution de

la diversité beta (homogénéisation taxonomique).

Les bassins hydrographiques européens ayant des assemblages similaires d’espèces

natives tendent aussi à avoir des assemblages similaires d’espèces exotiques.

La distribution actuelle des poissons d’eau douce exotiques en Europe semble être le

résultat combiné d’une limitation de la dispersion des espèces associée aux activités

humaines et d’un contrôle environnemental associé aux contraintes climatiques.


Le niveau d’anthropisation d’un bassin hydrographique et plus particulièrement sa

richesse économique est le principal déterminant de la richesse régionale en espèces

non natives de poisson d’eau douce.

Les perspectives associées à P3, P4 et P5 sont nombreuses :

Les travaux menés dans P3 et P4 sur les assemblages régionaux d’espèces non natives

sont uniquement basés sur la composition taxonomique des assemblages. Une

approche incorporant les traits biologiques des espèces permettrait de considérer

le rôle fonctionnel de chacune d’entre elles (Mason et al. 2007 ; Mouillot et al.

2007), tout en s’affranchissant le plus possible de contraintes biogéographiques et

historiques qui sont en partie responsables des différences de composition

taxonomique entre bassins (p.ex. Poff et al. 2006). D’abord, l’utilisation des traits

biologiques des espèces permettrait de mieux rendre compte de l’influence des filtres

environnementaux (Keddy 1992 ; Stazner et al. 2004 ; Mason et al. 2007) sur la

composition des espèces à l’échelle régionale et ainsi de dégager des règles

d’assemblages des espèces exotiques. Cela permettrait, par exemple, d’identifier

les traits biologiques qui favorisent la présence d’espèces exotiques dans certains

types de bassins hydrographiques (par exemple ceux de la région

méditerranéenne ayant des conditions hydrologiques variables). Ensuite, il serait

possible de tester l’hypothèse d’une convergence fonctionnelle entre assemblages

exotiques et natifs en termes de diversité et de distribution des traits biologiques (une

hypothèse émise suite aux résultats de P4). Enfin, la comparaison des résultats obtenus

dans P3 à ceux obtenus en utilisant uniquement les traits biologiques des espèces

exotiques et transloquées permettrait de déterminer si ces espèces ont provoqué

respectivement une différentiation et une homogénéisation fonctionnelle des


assemblages de poissons d’eau douce européens (voir Olden et al. 2004). Des travaux

portant sur ces trois thématiques sont actuellement initiés par la construction

d’une base de données sur les traits biologiques d’environ 300 espèces de poissons

d’eau douce européens (équipe « structure des communautés et macroécologie »,

EDB). Cette base de données sera couplée à une base de données existante sur la

composition en espèces d’environ 160 bassins hydrographiques européens dont les

caractéristiques environnementales et anthropiques sont connues.

A l’échelle mondiale, il serait intéressant de mettre en place des modèles

prédictifs du risque de colonisation des bassins hydrographiques par des espèces

connues pour être invasives (c.-à-d. le risque qu’une espèce introduite s’établisse

avec succès dans le pool régional d’espèces). Les risques de colonisation d’une

espèce seront quantifiés grâce à des méthodes statistiques prédictives (p.ex. Zambrano

et al. 2006 ; Ficetola et al. 2007) qui se basent sur la niche écologique réalisée de

l’espèce ; celle-ci correspondant à l’ensemble des conditions abiotiques et biotiques

permettant la survie de l’espèce (Hutchinson 1957). Une étude préliminaire sur les

risques globaux de colonisation de la Gambusie, Gambusia sp. (Leprieur et al. en

préparation) montre qu’il est possible d’identifier les bassins hydrographiques à fort

risque de colonisation (Figure 9). Dans ces bassins, des travaux pourront ensuite se

focaliser sur l’impact potentiel de l’espèce considérée à une échelle plus fine, c'est-à-

dire à l’échelle locale. Ces travaux visent également à aider les gestionnaires à diriger

leurs efforts vers le contrôle et la prévention des introductions d’espèces ayant un fort

risque de colonisation (p.ex. Zambrano et al. 2006 ; Mercado–Silva et al. 2006). Enfin,

ces modèles peuvent être utilisés pour étudier l’évolution de la distribution des espèces

invasives de poissons d’eau douce en vue des changements climatiques à venir (voir

Thuiller et al. 2007).


Figure 9 : Risques de colonisation de la gambusie (Gambusia affinis & Holbrooki) dans 616 bassins hydrographiques (PR : Présence actuelle correspondant à l’aire de distribution native et exotique ; les classes de couleur de R0 à R5 indiquent un risque de colonisation croissant. L’utilisation combinée de plusieurs modèles prédictifs (« Ensemble forecasting models ») pour déterminer les risques de colonisation d’une espèce, suggérée par Araújo & New (2006), repose sur le fait que l’exactitude des prédictions augmente avec la concordance des modèles. Ici, cinq méthodes de modélisation sont utilisées : les Modèles Linéaires Généralisés (GLMs), les Modèles Additifs Généralisés (GAMs), les Analyses Factorielles Discriminantes (AFD), les arbres de classification (CART) et les réseaux d’arbres (Boosted Trees). Les modèles consistent à prédire la présence et l’absence de la gambusie dans 616 bassins hydrographiques à partir 11 variables environnementales (p.ex. température moyenne annuelle ; nombre de jour de pluie, pluviométrie annuelle moyenne). Chaque modèle est construit selon la méthode du k-fold 10 consistant à utiliser 90% des données comme jeu d’apprentissage et les 10% restantes comme jeu de test indépendant. Pour les cinq méthodes, 1000 modèles sont générés par choix aléatoire de 62 bassins test permettant d’obtenir une probabilité d’occurrence de l’espèce (valeur entre 0 et 1). La moyenne de ces probabilités pour chacun des 616 bassins est alors calculée et transformée en présence/absence après détermination du seuil par la méthode des Receiver Operating Cuves (ROC). Une échelle de risque de colonisation (variant de 0 à 5) est établie en comptant le nombre de méthodes prédisant le bassin comme colonisable. (D’après Leprieur et al. actuellement en préparation).


4. Conclusion générale

Les résultats présentés dans ce mémoire, obtenus par l’utilisation d’approches

comparatives et expérimentales, ont mis en évidence (i) que les introductions d’espèces

peuvent avoir des impacts sur la biodiversité à plusieurs échelles spatiales (P1, P2 et P3 ;

Figure 11) et (ii) que des filtres abiotiques et/ou anthropiques associés à différentes échelles

spatiales conditionnent la richesse et la composition locale et régionale en espèces non natives

de poissons d’eau douce (P1, P4 et P5 ; Figure 12). De plus, les résultats confirment que

l’étude, à différentes échelles spatiales, du processus dynamique que constituent les

introductions d’espèces, peut contribuer à une meilleure compréhension de ses effets sur la

biodiversité (P1, P2 et P3) et nous aider à identifier des stratégies les plus adaptées (P1, P2 et

P5).

Echelle régionale


Impact des espèces non natives (P3):

Augmentation de la diversitéalpha du pool régional

d’espèces mais diminution de la diversité beta (c.-à-d.

une homogénéisation taxonomique)

EtendueRésolution

Echelle

Echelle locale :

Station ou tronçon de rivière

Impact des espèces non natives (P1 et P2):

P1: Déclin d’une espèce endémique menacée d’extinction (mécanisme : prédation/compétition)

P2 : Perturbation de l’efficacité de prédation d’une espèce native (mécanisme : interférence comportementale)

Echelle régionale


Impact des espèces non natives (P3):

Augmentation de la diversitéalpha du pool régional

d’espèces mais diminution de la diversité beta (c.-à-d.

une homogénéisation taxonomique)

EtendueRésolution

Echelle

Echelle locale :


Impact des espèces non natives (P1 et P2):

P1: Déclin d’une espèce endémique menacée d’extinction (mécanisme : prédation/compétition)

P2 : Perturbation de l’efficacité de prédation d’une espèce native (mécanisme : interférence comportementale)

Figure 11 : Schéma synthétique de l’impact des espèces non natives de poissons d’eau douce étudiés durant cette thèse à différentes échelles spatiales


Echelle régionale


Richesse en espèces non natives (P5) :

Filtre anthropique(richesse économique)

Composition en espèces non natives (P4):

Filtre abiotique (climat)Filtre anthropique

EtendueRésolution

Echelle

Echelle locale :



Filtre abiotique : perturbations hydrologiques

Echelle régionale


Richesse en espèces non natives (P5) :

Filtre anthropique(richesse économique)


Filtre abiotique (climat)Filtre anthropique

EtendueRésolution

Echelle

Echelle locale :



Filtre abiotique : perturbations hydrologiques

Figure 11 : Schéma synthétique des différents résultats obtenus durant cette thèse montrant l’existence de « filtres abiotiques et/ou anthropiques » conditionnant la richesse et la composition des d’espèces non natives de poissons d’eau douce à différentes échelles spatiales

Certains résultats obtenus durant cette thèse (plus particulièrement P1) indiquent que

les espèces invasives peuvent être directement impliquées dans le déclin de populations

natives. Pourtant le rôle direct des invasions biologiques dans l’érosion de la biodiversité est

aujourd’hui remis en cause (p.ex. Gurevitch & Padilla 2004) et fait l’objet d’un débat parmi

les écologues (p.ex. Ricciardi 2004 ; Clavero & Berthou 2005 ; Didham et al. 2005 ; Sagoff

2005 ; Simberloff 2005 ; Light & Marchetti 2007). Un des principaux arguments est que les

invasions biologiques ne seraient qu’une conséquence indirecte des modifications d’habitats,

lesquelles seraient les principales causes de l’érosion de la biodiversité. A mon sens, ce débat

ne peut pas trouver de réponses constructives car la modification des habitats et les invasions

biologiques sont des processus concomitants, qui interagissent dans leurs effets sur la


biodiversité (voir P1 ; Mitchell et al. 2006 ; Didham et al. 2007). En effet, les modifications

d’habitats et les invasions biologiques peuvent avoir des effets interactifs de type additif,

synergique ou antagoniste (P1) sur les espèces natives et sur le fonctionnement des

écosystèmes (voir Didham et al. 2007). Ainsi, l’étude indépendante de l’une de ces deux

causes de l’érosion de la biodiversité ne peut que surestimer ou sous-estimer leur impact

respectif. La compréhension de l’effet combiné des invasions biologiques et des modifications

d’habitats sur la biodiversité (en particulier suite aux changements climatiques et

d’occupation des sols) représente ainsi un enjeu majeur pour les écologues et les gestionnaires

des milieux naturels. Une approche associant des travaux à différentes échelles spatiales, de

l’échelle locale (par le biais d’expérimentations) à l’échelle globale (par le biais de techniques

de modélisation et de systèmes d’information géographique), est selon moi la plus adaptée

pour répondre à un tel enjeu.


Références Allibone RM (2000) Water abstraction impacts on the non-migratory galaxiids of Totara

Creek. Science for Conservation 147: 25–45. Angermeier PL, Winston MR (1998) Local vs. regional influences on local diversity in stream

fish communities of Virginia. Ecology 79: 911-927. Barlow CG, Hogan AE, Rogers LJ (1987) Implication of translocated fishes in the apparent

extinction in the wild of the Lake Eacham rainbowfish, Melanotaenia eachamensis. Aust J Mar Freshwater Res 38: 897-902.

Baxter CV, Fausch KD, Murakami M, Chapman PL (2007) Invading rainbow trout usurp a terrestrial prey subsidy to native charr and alter their behavior, growth, and abundance. Oecologia 153: 461-470.

Blackburn TM, Duncan RP (2001) Establishment patterns of exotic birds are constrained by non-random patterns in introduction. J Biogeogr 28: 927–939.

Bianco PG (1995) Mediterranean endemic freshwater fishes of Italy. Biol Conserv 72: 159–170.

Byers JE (2002) Impact of non-indigenous species enhanced by anthropogenic alteration of selection regimes. Oikos 97: 449–458.

Blanchet S, Loot G, Bernatchez L, Dodson JJ (2007) The disruption of dominance hierarchies by a non-native species: an individual-based analysis. Oecologia 152:569-581.

Boët P (1980) L’alimentation du poisson-chat (Ictalurus melas Raf.) dans le lac de Créteil. Ann Limnol - Int J Limno l16: 255 – 270.

Braig EC, Johnson DL (2003) Impact of black bullhead (Ameiurus melas) on turbidity in a diked wetland. Hydrobiologia 490: 11 – 21.

Brown JH & Maurer BA (1989) Macroecology: the division of food and space among species on continents. Science 243: 1145-1150.

Cassey P, Lockwood JL, Blackburn TM, Olden JD (2007). Spatial scale and evolutionary history determine the degree of taxonomic homogenization across island bird assemblages. Divers Distrib 13: 458-466.

Chown SL, Hull B, Gaston KJ (2005) Human impacts, energy availability and invasion across Southern Ocean Islands. Glob Ecol Biogeogr 14: 521-528.

Clavero M, García-Berthou E (2005) Invasive species are a leading cause of animal extinctions. Trends Ecol Evol 20:110.

Clavero M, García-Berthou E (2006) Homogenization dynamics and introduction routes of invasive freshwater fish in the Iberian Peninsula. Ecol Appl 16: 2313–2324.

Closs GP, Lake PS (1996) Drought, differential mortality and the coexistence of a native and an introduced fish species in a south east Australian intermittent stream. Environ Biol Fishes 47: 17–26.

Cohen AN (2002) Success factors in the establishment of human-dispersed organisms. In: Bullock JM, Kenward RE, Hails RS, editors. Dispersal Ecology. London: Blackwell. pp.374–394.

Colautti RI, MacIsaac HJ (2004) A neutral terminology to define 'invasive' species. Divers Distrib 10: 135-141.

Connor EF, Simberloff D (1979) The assembly of species communities: chance or. competition. Ecology 60:1132-1140

Crivelli AJ (1995) Are fish introductions a threat to endemic freshwater fishes in northern Mediterranean region? Biol Conserv 72: 311–319


Crosby A. (1986) Ecological Imperialism: the Biological Expansion of Europe, 900-1900. Cambridge University Press: Cambridge.

Crowl TA, Townsend CR, McIntosh AR (1992) The impact of introduced brown and rainbow trout on native fish: the case of Australasia. Rev Fish Biol Fish 2: 217–241.

Cucherousset J, Paillisson JM, Carpentier A, Eybert M.-C, Olden JD (2006) Habitat use of an artificial wetland by the invasive catfish Ameiurus melas. Ecol Freshw Fish 15: 589 – 596.

Davey AJH, Kelly DJ, Biggs BJF (2006) Refuge-use strategies of stream fishes in response to extreme low flows. J Fish Biol 69: 1047-1059.

Diamond JR (1983) Laboratory, field and natural experiments. Nature 303: 586-587. Didham RK, Tylianakis JM, Hutchison MA, Ewers RM, Gemmell NJ (2005) Are invasive

species the drivers of ecological change? Trends Ecol Evol 20: 470-474. Didham RK, Tylianakis JM, Gemmell NJ, Rand TA, Robert M (2007) Interactive effects of

habitat modification and species invasion on native species decline. Trends Ecol Evol 20: 470-475.

Duncan JR, Lockwood JL (2001) Spatial homogenization of aquatic fauna of Tennessee: extinction and invasion following land use change and habitat alteration. In: Lockwood JL, McKinney ML, editors. Biotic homogenization. Kluwer Academic/Plenum Publishers: New York. pp. 245–258.

Dunn NR (2003) The effects of extremes in flow on alpine (G. paucispondylus) and Canterbury (G. vulgaris) Galaxias. Unpublished MSc Thesis. University of Canterbury,Christchurch, New Zealand.

Eby LA, Roach WJ, Crowder LB, Stanford JA (2006) Effects of stocking-up freshwater food webs. Trends Ecol Evol 21: 576–584.

Eklöv P, Hamrin SF (1989). Predatory efficiency and prey selection: interactions between pike Esox lucius, perch Perca fluviatilis and rudd Scardinus erythrophthalmus. Oikos 56: 149 – 156.

Eklöv P, VanKooten T (2001) Facilitation among piscivorous predators: effects of prey habitat use. Ecology 82: 2486 – 2494.

Elton CS (1958) The ecology of invasions by animals and plants, Methuen, London. Elvira B (2001) Identification of non-native freshwater fishes established in Europe and

assessment of their potential threats to the biological diversity. Convention on the conservation of European wildlife and natural habitats, 21st meeting Strasbourg, 26– 30 November 2001. 35 p.

Elvira B, Almodóvar A (2001) Freshwater fish introductions in Spain: facts and figures at the beginning of the 21st century. J Fish Biol 59 (Suppl. A): 323–331.

Evans KL, Warren PH, Gaston KJ (2005) Does energy availability influence classical patterns of spatial variation in exotic species richness? Glob Ecol Biogeogr 14: 57-65.

Fattini AO, Petrere M (2007) Which factors determine non-indigenous fish dispersal? A study of the red piranha in tropical Brazilian lakes. In: Gherardi F, editors. Biological invaders in inland waters: profiles, distribution, and threats. Springer: The Netherlands,

Ficetola GF, Thuiller W, Miaud C (2007) Prediction and validation of the potential global distribution of a problematic alien invasive species – the American bullfrog. Divers Distrib 13: 476-485.

Fridley JD, Stachowicz JJ, Naeem S, Sax DF, Seabloom EW et al. (2007) The invasion paradox: reconciling pattern and process in species invasions. Ecology 88: 3-17.

Fuller PL, Nico LG & Williams JD (1999) Nonindigenous fishes introduced into inland waters of the United States. Special Publication 27. American Fisheries Society: Bethesda, Maryland.


García-Berthou E, Alcaraz C, Pou-Rovira Q, Zamora L et al. (2005) Introduction pathways and establishment rates of invasive aquatic species in Europe. Can J Fish Aquat Sci 65: 453–463.

García-Berthou E (2007) The characteristics of invasive fishes: what has been learned so far? J Fish Biol 71 (Supplement D): in press.

Gido KB, Brown JH (1999) Invasion of North American drainages by alien fish species. Freshw Biol 42: 387-399.

Gotelli NJ, Graves GR (1996) Null models in ecology. Smithsonian Institution Press: Washington, DC.

Gowan C, Fausch KD (1996) Mobile brook trout in two high-elevation Colorado streams: re-evaluating the concept of restricted movement. Can J Fish Aquat Sci 53: 1370–1381.

Gozlan RE, St-Hilaire S, Feist SW, Martin P, Kent ML (2005) Disease threats on European fish. Nature 435:1046.

Gregory RS (1993) The effect of turbidity on the predator avoidance behaviour of juvenile Chinook salmon (Oncoryhnchus tshawytscha). Can J Fish Aquat Sci 50: 241 – 246.

Griffen BD (2006). Detecting emergent effects of multiple predator species. Oecologia 148: 702 – 709.

Guéguan JF, Lek S, Oberdorff T (1998) Energy availability and habitat heterogeneity predict global riverine fish diversity. Nature 391: 382-384.

Gurevitch J, Padilla DK (2004) Are invasive species a major cause of extinctions? Trends Ecol Evol 19: 470–474.

Hawkins BA, Field R, Cornell HV, Currie DJ, Guégan JF et al. (2003) Energy, water, and broad-scale geographic patterns of species richness. Ecology 84: 3105–3117.

Hooper DU, Chapin FS, Ewel JJ, Hector A, Inchausti S et al. (2005) Effects of biodiversity on ecosystem functioning: a consensus of current knowledge. Ecol Monogr 75: 3-23.

Hugueny B (1989) West African rivers as biogeographic islands. Oecologia 79: 235-243. Hugueny B & Paugy D (1995) Unsaturated fish communities in African rivers. Am Nat 68:

162-169. Hutchinson GE (1957) Concluding remarks. Cold Springs Harbor Symposium on

Quantitative Biology 22: 423-427. Huston MA (1999) Local processes and regional patterns: appropriate scales for

understanding variation in the diversity of plants and animals. Oikos 86: 393–401. IUCN (2006) 2006 IUCN Red List of Threatened Species. Available:

http://www.iucnredlist.org. Jackson DA, Peres-Neto PR, Olden JD (2001) What controls who is where in freshwater fish

communities: The roles of biotic, abiotic and spatial factors? Can J Fish Aquat Sci 58:157-170.

Jeschke JM, Strayer DL (2005) Invasion success of vertebrates in Europe and North America. Proc Natl Acad Sci U S A 102: 7198-7202.

Keddy PA (1992) Assembly and response rules: two goals for predictive community ecology. J Veg Sci 3: 157-164.

Keith P, Allardi J (2001) Atlas des poissons d'eau douce de France. Patrimoines Naturels 47 : 1 – 387.

Kennard MJ, Arthinghton AR, Pusey BJ, Harch BD (2005) Are alien fish are a reliable indicator of river health? Freshw Biol 50: 174–193.

Kennedy TA, Naeem S, Howe KM, Knops JMH, Tilman D et al. (2002) Biodiversity as a barrier to ecological invasion. Nature 417: 636-638.

Kirchner F, Soubeyran Y (2007) Espèces exotiques envahissantes : vers un état des lieux


global et un réseau d’action coordonné à l’échelle de tout l’outre-mer. 13eme Forum des gestionnaires sur les espèces exotiques envahissantes : une menace majeure pour la Biodiversité. Mnhn - Paris - Vendredi 16 Mars 2007

Kolar CS, Lodge DM (2002) Ecological predictions and risk assessment for alien fishes in North America. Science 298: 1233-1236.

Kreft, H. & W. Jetz (2007): Global Patterns and Determinants of Vascular Plant Diversity. Proc Natl Acad Sci U S A 104: 5925-5930.

La Sorte FA, McKinney ML (2006) Compositional similarity and the distribution of geographical range size for assemblages of native and non-native species in urban floras. Divers Distrib 12: 679-686.

Lawler JJ, Aukema JE, Grant J, Halpern B, Kareiva P et al. (2006) Conservation science: a 20-year report card. Front Ecol Environ 4: 473-480.

Lever C (1996) Naturalized Fishes of the World. Academic Press: London,UK. Levin SA (1992) The problem of pattern and scale in ecology. Ecology 73: 1943–1967 Levine JM (2000) Species diversity and biological invasions: relating local process to

community pattern. Science 288: 852–854. Levine JM, D'Antonio CM (2003) Forecasting biological invasions with increasing

international trade. Conserv Biol 17: 322-326. Light T, Marchetti M (2007) Distinguishing between invasions and habitat changes as drivers

of diversity loss among California’s freshwater fishes. Conserv Biol 21: 434-446. Lodge DM (1993) Biological invasions: lessons for ecology. Trends Ecol Evol 8: 133-137. Lodge DM, Stein RA, Brown KM, Covich AP, Bronmark C et al. (1998) Predicting impact

of freshwater exotic species on native biodiversity: challenges in spatial scaling. Australian Journal of Ecology 23:53-67.

Lockwood JL, Hoopes MF, Marchetti MP (2007) Invasion Ecology. Blackwell Publishing: Oxford. UK.

Loreau M, Naeem S, Inchausti P (2002). Biodiversity and ecosystem functioning: synthesis and perspectives. Oxford University Press: Oxford, UK..

MacArthur RH, Wilson EO (1967) The theory of island biogeography. Princeton University Press: Princeton, New Jersey.

Mack RN, Lonsdale WM (2001) Humans as global plant dispersers: getting more than we bargained for. Bioscience 51:95–103.

Mason N, Lanoiselée C, Mouillot D, Irz P, Argillier C (2007) Functional characters combined will null models reveal inconsistency in mechanisms of species turnover in lacustrine fish communities. Oecologia 153: 441-452.

Matthews KR, Berg NH (1997) Rainbow trout responses to water temperature and dissolved oxygen stress in two southern California stream pools. J Fish Biol 50: 50-67.

McDowall RM (2006) Crying wolf, crying foul, or crying shame: alien salmonids and a biodiversity crisis in the southern cool-temperate galaxioid fishes? Rev Fish Biol Fish 16: 233-422.

McIntosh AR (2000) Habitat and size-related variations in exotic trout impacts on native galaxiid fishes in New Zealand streams. Can J Fish Aquat Sci 57, 2140–2151.

McKinney ML (2005) Species introduced from nearby sources have a more homogenizing effect than species from distant sources: evidence from plants and fishes in the USA. Divers Distrib 11: 367–374.

McKinney ML (2006) Correlated Non-native Species Richness of Birds, Mammals, Herptiles and Plants: Scale Effects of Area, Human Population and Native Plants. Biol Invasions 8: 415-425.


Mercado–Silva N, Olden JD, Maxted JT, Hrabik TR, Vander Zanden MJ (2006) Forecasting the spread of Laurentian Great Lakes region of North America. Conserv Biol 20:1740-1749.

Meyerson LA, Mooney HA (2007) Invasive alien species in an era of globalization. Front Ecol Environ 5: 199–208.

Millennium Ecosystem Assessment (2005) Ecosystems and human well being: current state and trends. In: Rijsberman F, Costanza R, Jacobi P, editors. Freshwater (Vol 1, Chapter 7).World Resources Institute: Washington DC. pp. 165-207.

Miller RR, Williams JD, Williams JE (1989) Extinctions of North American fishes during the past century. Fisheries 14:22-38.

Mitchell CE, Agrawal AA, Bever JD, Gilbert GS, Hufbauer RA et al. (2006) Biotic interactions and plant invasions. Ecol Lett. 9: 726-740.

Mooney HA, Cleland EE (2001) The evolutionary impact of invasive species. Proc Natl Acad Sci U S A 98:5446-5451.

Mouillot D, Dumay O & Tomasini JA (2007) Limiting similarity, niche filtering and functional diversity in coastal lagoon fish communities. Estuar Coast Shelf Sci 71: 443-456.

Moyle PB, Light T (1996) Fish invasions in California: do abiotic factors determine success? Ecology 77: 1666–1670.

Moyle PB (1999) Effects of invading species on freshwater and estuarine ecosystems. In: Sandlund OT, Schei PJ, Viken A, editors). Invasive Species and Biodiversity Management. Kluwer Academic Press: Netherlands. pp. 177–191.

Moyle PJ, Cech (2004) Fishes: An Introduction to Ichthyology - fifth edition. Upper Saddle River, NJ: Prentice-Hall, Inc.

Moyle PB, Marchetti MP (2006) Predicting exotic fishes in freshwater systems: freshwater fishes in California as a model. Bioscience 56: 515–524.

Nekola JC, White PS (1999) The distance decay of similarity in biogeography and ecology. J Biogeogr 26: 867-878.

Oberdorff T, Guégan J.-F, Hugueny B (1995) Global scale patterns of fish species richness in rivers. Ecography 18: 345-352.

Olden JD, Poff NL (2003) Toward a mechanistic understanding and prediction of biotic homogenization. Am Nat 162: 442–460.

Olden JD, Poff NL (2004) Ecological processes driving biotic homogenization: testing a mechanistic model using fish faunas. Ecology 85: 1867–1875.

Olden JD, LeRoy PN, Douglas MR, Douglas ME, Fausch KD (2004) Ecological and evolutionary consequences of biotic homogenization. Trends Ecol Evolution 19: 18-24.

Olden JD, Hogan ZS, Vander Zanden JV (2007) Small fish, big fish, red fish, blue fish: size-biased extinction risk of the world's freshwater and marine fishes. Global Ecol Biogeogr: in press.

Olsson K, Stenroth P, Nyström P, Holmqvist N, McIntosh AR, Winterbourn MJ (2006) Does natural acidity mediate interactions between introduced brown trout, native fish, crayfish and other invertebrates in West Coast New Zealand streams?. Biol Conserv 130:255-267.

O’Neill RV and King AW (1998) Homage to St. Michael; or why are there so many books on scale? In: Peterson DL, Parker VT, editors. Ecological Scale: Theory and Applications. Columbia University Press: New York. pp 3–16.

Palmer ME, Ricciardi A (2004) Physical factors affecting the relative abundance of native and invasive amphipods in the St Lawrence River. Can J Zool 82:1886–1893.


Parker IM, Simberloff D, Lonsdale WM, Goodell K, Wonham M et al. (1999) Impact: toward a framework for understanding the ecological effects of invaders. Biol Invasions 1: 3-19.

Pauchard A, Shea K (2006) Integrating the study of nonnative plant invasions across spatial scales. Biol Invasions 8: 399–413.

Pekcan-Hekim Z, Lappalainen J (2006) Effects of clay turbidity and density of pikeperch (Sander lucioperca) larvae on predation by perch (Perca fluviatilis). Naturwissenschaften 93: 356 – 359.

Perrings C, Dehnen-Schmutz K, Touza J, Williamson, M (2005) How to manage biological invasions under globalization. Trends Ecol Evol 20: 212-215, 2005.

Perry WL, Lodge DM, Feder JL (2002) Importance of hybridization between indigenous and nonindigenous freshwater species: An overlooked threat to North American biodiversity. Systematic Biology 51: 255-275.

Pianka ER (1966) Latitudinal gradients in species diversity: a review of concepts. Am Nat 100:33-46.

Pimentel D, Zuniga R, Morrison D (2005) Update on the environmental and economic costs associated with alien–invasive species in the United States. Ecol Econ 52: 273–88.

Poff NL (1997) Landscape filters and species traits: towards mechanistic understanding and prediction in stream ecology. J N Am Benthol Soc 16: 391–409.

Poff, NL, Olden JD, Vieira NKM, Finn DS, Simmons MP et al. (2006). Functional trait niches of North American lotic insects: traits-based ecological applications in light of phylogenetic relationships. J N Am Benthol Soc 25:730-755

Qian H, Ricklefs RE (2006) The role of exotic species in homogenizing the North American flora. Ecol Lett 9: 1293-1298.

Quist M, Rahel FJ, Hubert WA (2005). Hierarchical faunal filters: an approach to assessing effects of habitat and non-native species on native fishes. Ecol Freshw Fish 14:1-16.

Rahel F (2002) Homogenization of freshwater faunas. Annu Rev Ecol Evol S 33:291-315. Rahel FJ (2007) Biogeographic barriers, connectivity, and biotic homogenization: it’s a small

world after all. Freshw Biol 52: 696–710. Reid SM, Fox MG, Whillans TH (1999). Influence of turbidity on piscivory in largemouth

bass (Micropterus salmoides). Can J Fish Aquat Sci 56: 1362 – 1369. Reyjol Y, Hugueny B, Pont D, Bianco PG, Beier U, et al. (2006). Patterns in species richness

and endemism of European freshwater fish. Global Ecol Biogeogr 16: 65-75. Ricciardi A (2004) Assessing species invasions as a cause of extinction. Trends Ecol Evol 19:

619. Richardson DM, Pysek P, Rejmanek M, Barbour MG, Panetta FD, et al. (2000) Naturalization

and invasion of alien plants : concepts and definitions. Divers Distrib 6: 93-107. Ricklefs RE (1987) Community diversity: relative roles of local and regional processes.

Science 235: 167-171. Ricklefs RE, Schluter D (1993) Species diversity: Regional and historical influences. In:

Ricklefs RE & Schluter D, editors). Species diversity in ecological communities. The University of Chicago Press: Chicago. pp. 350-363.

Rincón PA, Correas AM, Morcillo F, Risueno P, Lobon-Cervia J. (2002). Interaction between the introduced eastern mosquitofish and two autochthonous Spanish toothcarps. J Fish Biol 61: 1560–1585.

Ross ST (1991) Mechanisms structuring stream fish assemblages; are there lessons from introduced species? Env Biol Fishes 30: 359–368.

Sagoff M (2005) Do Non-native Species Threaten the Natural Environment? J Agric Environ Ethics 18: 215-236.


Sakai AK, Allendorf FW, Holt JS, Lodge DM, Molofsky J et al. (2001) The population biology of invasive species. Annu Rev Ecol Evol S 32: 305–332.

Sala OE, Chapin FS, et al. (2000) Global biodiversity scenarios for the year 2100. Science 287:1770-1774.

Sax DF, Gaines SD, Brown JH (2002) Species invasions exceed extinctions on islands worldwide: a comparative study of plants and birds. Am Nat 160: 766–783.

Sax DF, Gaines SD (2003) Species diversity: from global decreases to local increases. Trends Ecol Evol 18:561-566.

Schneider DC (2001) The rise in the concept of scale in ecology. BioScience 51: 545-553. Simberloff D (1986) Introduced insects: A biogeographic and systematic perspective. In:

Mooney HA, Drake JA, editors. Ecology of Biological Invasions of North America and Hawaii. Springer-Verlag: New York. pp. 3-26.

Simberloff D (1989) Which insect introductions succeed and which fail. In: Drake JA, Mooney HA, di Castri F, Groves RH, Kruger FJ, Rejmanek M, Williamson M, editors. Biological Invasions: A global perspective. John Wiley & Sons: New York. pp.61-76.

Simberloff D (1996). Impacts of introduced species in the United States. Consequences: the Nature and Implications of Environmental Changes 2: 13-22.

Simberloff (2005) Non-native species do threaten the natural environment. J Agric Environ Ethics 18- 595-607.

Simon K & Townsend CR (2003) The impacts of freshwater invaders at different levels of ecological organisation, with emphasis on ecosystem consequences. Freshw Biol 48:982-994.

Simpson GG (1953) Evolution and geography – an essay on historical biogeography with special reference to mammals. Condon Lectures. Eugene, OR: Oregon State System of Higher Education.

Smith CL, Powell CR (1971) The summer fish communities of Brier Creek, Marshall County, Oklahoma. Am Mus Novit 2458: 1–30.

Soininen J, McDonald R, Hillebrand H (2007) The distance decay of similarity in ecological communities. Ecography 30: 3-12.

Statzner B, Dolédec S, Hugueny, B (2004) Biological trait composition of European stream invertebrate communities: assessing the effects of various trait filter types. Ecography 27: 470-488.

Stockwell CA, Hendry AP, Kinnison MT (2003) Contemporary evolution meets conservation biology. Trends Ecol Evol 18: 94-101.

Stohlgren TJ, Binkley D, Chong GW (1999) Exotic plant species invade hot spots of native plant diversity. Ecol Monogr 69: 47-68.

Stohlgren TJ, Barnett DT, Kartesz JT (2003) The rich get richer: patterns of plant invasions in the United States. Front Ecol Environ 1: 11–14.

Strauss SY, Lau JA, Carroll SP (2006) Evolutionary responses of natives to introduced species: what do introductions tell us about natural communities? Ecol Lett 9: 357–374.

Stromberg JC, Lite SJ, Marler R, Paradzick R, Shafroth PB (2007) Altered stream-flow regimes and invasive plant species: the Tamarix case. Global Ecol Biogeogr 16: 381-393.

Taylor BW, Irwin RE (2004) Linking economic activities to the distribution of exotic plants. Proc Natl Acad Sci U S A 101: 17725-17730.

Tedesco PA, Oberdorff T, Lasso CA, Zapata M, Hugueny B (2006) Evidence of history in explaining diversity patterns in tropical riverine fish. J Biogeogr 32: 1899-1907.

Tobler WR (1970) A computer movie simulating urban growth in the Detroit region. Econ Geogr 46: 234-240.


Tonn WM (1990) Climate change and fish communities: a conceptual framework. Trans Am Fish Soc 119: 337–352.

Townsend CR, Crowl TA (1991) Fragmented population structure in a native New Zealand fish: an effect of introduced brown trout? Oikos 61: 348–354.

Townsend CR (1996) Invasion biology and ecological impacts of Brown trout (Salmo trutta) in New Zealand. Biol Conserv 78: 13-22.

Townsend CR (2003) Individual, population, community and ecosystem consequences of a fish invader in New Zealand streams. Conserv Biol 17: 38-47.

Thuiller W, Richardson DM, Midgley GF (2007) Will climate change promote alien plant invasions? In: Nentwig W, editors. Biological Invasions, Ecological Studies, Vol. 193, Springer-Verlag: Berlin Heidelberg

Vander Zanden MJ, Casselman JM, Rasmussen JB (1999) Stable isotope evidence for the food web consequences of species invasions in lakes. Nature 401: 464-467.

Vitousek PM, Mooney HA, Lubchenco J, Melillo JM (1997) Human domination of Earth's ecosystems. Science 278: 494-499.

Welcomme RL (1988) International introductions of inland aquatic species. FAO Fish. Tech. Pap. 294, 318 p.

White JL, Harvey BC (2001) Effects of an introduced piscivorous fish on native benthic fishes in a coastal river. Freshw. Biol. 46: 987–995.

Whittaker RJ, Fernández-Palacios JM (2007) Island Biogeography: ecology, evolution, and conservation, 2nd edn. Oxford University Press: Oxford

Wilcove DS, Rothstein D, Bubow J, Phillips A, Losos E (1998) Quantifying threats to imperiled species in the United States. BioScience 48: 607-615.

Williamson M (1996) Biological Invasions. London: Chapman & Hall. Willis KJ, Whittaker RJ (2001) Species diversity. Scale matters. Science 295: 1245–1248 Woodruff DS (2001). Declines of biomes and biotas and the future of evolution. Proc Natl

Acad Sci U S A 98: 5471-5476. Witte F, Goldschmidt T, Goudswaard PC, Lightvoet W, Van Oijen MJP et al. (1992) Species

extinction and concomitant ecological changes in Lake Victoria. Neth J Zool 42: 214-32.

Zambrano L, Martínez-Meyer E, Menezes N, Peterson AT (2006) Invasive potential of common carp (Cyprinus carpio) and Nile tilapia (Oreochromis niloticus) in American freshwater systems. Can J Fish Aquat Sci 69: 1903-1910.


Partie 2 : Publications

Les introductions d’espèces de poissons d’eau douce

P1 Hydrological disturbance benefits a native fish at the expense of

an exotic fish

Leprieur F., Hickey M.A., Arbuckle C.J., Closs G.P., Brosse, S. &

Townsend C.R. (2006)

Journal of Applied Ecology, 43: 930-939.

Journal of Applied Ecology

2006

43

, 930–939

© 2006 The Authors. Journal compilation © 2006 British Ecological Society

Blackwell Publishing Ltd

Hydrological disturbance benefits a native fish at the expense of an exotic fish

F. LEPRIEUR,* M. A. HICKEY,† C. J. ARBUCKLE,‡ G. P. CLOSS,§ S. BROSSE* and C. R. TOWNSEND§

*

Laboratoire Dynamique de la Biodiversité, UMR 5172, CNRS – Université Paul Sabatier, 118 route de Narbonne, F-31062 Toulouse cedex 4, France;

†

Otago Regional Council, 70 Stafford Street, Dunedin, New Zealand;

‡

Environment Southland, Private Bag 90116, Invercargill, Southland, New Zealand; and

§

Department of Zoology, University of Otago, PO Box 56, Dunedin, New Zealand

Summary

1.

Some native fish in New Zealand do not coexist with introduced salmonids. Previousstudies of disjunct distributions of exotic brown trout

Salmo trutta

and native galaxiidsdemonstrated native extirpation except where major waterfalls prevented upstreammigration of trout. In the Manuherikia River system, we predicted that water abstrac-tion might be a further factor controlling the spatial distribution of both the invader anda native fish.

2.

We applied multiple discriminant function analyses to test for differences in envi-ronmental conditions (catchment and instream scales) at sites with roundhead galaxias

Galaxias anomalus

and brown trout in sympatry and allopatry. We then used a supervisedartificial neural network (ANN) to predict the presence–absence of

G. anomalus

andbrown trout (135 sites). The quantification of contributions of environmental variablesto ANN models allowed us to identify factors controlling their spatial distribution.

3.

Brown trout can reach most locations in the Manuherikia catchment, and oftenoccur upstream of

G. anomalus

. Their largely disjunct distributions in this river aremediated by water abstraction for irrigation, together with pool habitat availability andvalley slope. Trout are more susceptible than the native fish to stresses associated withlow flows, and seem to be prevented from eliminating galaxiid populations from sites inlow gradient streams where there is a high level of water abstraction.

4.

Synthesis and applications

. In contrast to many reports in the literature, our resultsshow that hydrological disturbance associated with human activities benefits a nativefish at the expense of an exotic in the Manuherikia River, New Zealand. Water abstrac-tion is also known to have negative impacts on native galaxiids, therefore we recom-mend restoring natural low flows to maintain sustainable habitats for native galaxiids,implementing artificial barriers in selected tributaries to limit trout predation on nativefish, and removing trout upstream.

Key-words

: biological invasion, disjunct distributions, disturbance, galaxiids, intro-duced trout, water abstraction


(2006)

43

, 930–939doi: 10.1111/j.1365-2664.2006.01201.x

Introduction

Biological invasions along with habitat loss are rec-ognized as major threats to biodiversity world-wide

(Vitousek 1994; Clavero & García-Berthou 2005).Effects of exotic species are well documented and occurfrom individual to ecosystem level (Mack

et al

. 2000;Townsend 2003). At the population level, a compre-hensive understanding of mechanisms leading toinvasion success and impact is necessary to developefficient management tools (Sakai

et al

. 2001). Effectsmay differ across sites (Palmer & Ricciardi 2004) andthe ability to predict impacts requires information

Correspondence: Fabien Leprieur, Laboratoire Dynamiquede la Biodiversité, UMR 5172, CNRS – Université Paul Saba-tier, 118 route de Narbonne, F-31062 Toulouse cedex 4,France (e-mail [email protected]).

931

Hydrological disturbance favours a native fish

© 2006 The Authors. Journal compilation © 2006 British Ecological Society,


,

43

, 930–939

about species’ responses to local abiotic factors as wellas to each other.

Brown trout

Salmo trutta

L. have been introducedfor angling in many countries but have often had neg-ative impacts on native fish populations (Krueger &May 1991; Crowl, Townsend & McIntosh 1992; Morita,Tsuboi & Matsuda 2004). The introduction of browntrout in 1864 to Australia and in 1867 to New Zealandcoincided with declines and local extirpations of nativegalaxiid species through predation by trout (Townsend& Crowl 1991; Closs & Lake 1996) and competitive dis-placement (reviewed by McDowall 1968, 2003; Crowl,Townsend & McIntosh 1992). The strongest evidenceof the effects of introduced brown trout on nativegalaxiids in stream ecosystems is provided by theirnon-overlapping distributions (Crowl, Townsend &McIntosh 1992; Closs & Lake 1996; McIntosh 2000).

The Otago region on the South Island of NewZealand has become recognized as a hotspot of non-migratory galaxiid diversity since the reinstatement ofthe roundhead galaxias

Galaxias anomalus

Stokell, pre-viously confounded with

Galaxias vulgaris

Stokell, andthe recognition of several new species (Department ofConservation 2004). According to the New Zealandthreat classification list (Hitchmough 2002), thesesmall (< 150 mm long) stream-dwelling species are ingradual decline, except for

G. vulgaris

. While habitatdegradation caused by land-use change may beinvolved (Hanchet 1990), Townsend & Crowl (1991)found that land use in the Taieri River catchment(Otago Province) could not account for the observedgalaxiid population fragmentation. Rather, it was largewaterfalls (higher than 3 m) that prevented trout inva-sion and provided upstream refugia for the galaxiids.In the nearby Manuherikia River catchment, where

G.anomalus

is the dominant galaxiid, waterfalls are notsuch a marked feature but the diversion of stream waterfor irrigation is particularly widespread. Water wasoriginally taken for gold mining but the associatedrights were picked up by irrigators and the pattern ofabstraction continued. In the dry Manuherikia region,droughts are a natural feature but have become extendedand aggravated by the increase of water abstraction asa result of agricultural intensification. There is littleprevious work on the effects on fish of anthropogenicallymediated drought (Matthews & Marsh-Matthews 2003).

As local abiotic conditions may mediate the effectsof species’ introductions (Holway, Suarez & Case 2002;Ricciardi 2003), we expected that low flow conditionsassociated with water abstraction might be a furtherfactor controlling the spatial distribution of bothbrown trout and native galaxiids in the ManuherikiaRiver catchment. We analysed a catchment-wide fishpresence–absence data set with the following goals: (i)to identify the environmental factors that may mediatetrout–galaxiid interactions, in order to predict the out-come of brown trout impacts, and thus (ii) to providethe basis for effective management. We made use ofcontrasting modelling techniques, but with special

emphasis on artificial neural networks (ANN), whichhave proved efficient in modelling species’ distributions(Manel, Dias & Ormerod 1999) and predicting en-vironmental impacts (Spitz & Lek 1999).

Method

The gravel-bed tributaries of the Manuherikia River(South Island, New Zealand), with their classic riffle–run–pool structure, rise in steep mountain countrybefore flowing through developed farmland wherewater is taken for stock and irrigation. The main stem ofthe river flows south-west for 85 km to its confluencewith the Clutha River at Alexandra. Its upper reacheshave been dammed for irrigation purposes, while themiddle and lower reaches have large water off-takes.The catchment has a harsh, dry climate, ranging fromover 30

°

C in mid-summer to

−

15

°

C in mid-winter. Itsfish fauna consists of two exotic salmonids (brown troutand the less common brook trout

Salvelinus fontinalis

Mitchill) and several native species: bullies

Gobiomor-phus

spp., longfin eel

Anguilla dieffenbachii

Gray andthe non-migratory alpine galaxias

Galaxias paucispond-ylus

Stokell (recorded at one site) and the yet to benamed

Galaxias

sp. D. However, by far the most wide-spread native species is the roundhead galaxias

G.anomalus

.

The data set contained 135 sites with both biologicaldata (fish species presence–absence) and environmen-tal data (Table 1). We extracted the occurrence of

G.anomalus

and brown trout from the New ZealandFreshwater Fish Database (NZFFD; McDowall &Richardson 1983; Joy & Death 2004). All sites havebeen sampled by electrofishing in summer since 1980,with 70% of samples taken since 2000. The percentagesof the stream bed composed of particular channel units(pool, run, riffle and rapid) were also extracted fromNZFFD. We extended the environmental data set byextracting stream order, valley slope and altitude fromthe River Environmental Classification (REC; Snelder,Biggs & Weatherhead 2004), a river reach-scale geo-graphical information system (GIS). We also deter-mined the relative distance of each site from the mainchannel as described by Schaefer & Kerfoot (2004),assigning a value of 0 to sites on the main channel and1 to sites on headwater streams, and we identified theposition of dams and natural waterfalls (greater than3 m high) that could prevent the upstream migrationof trout. Finally, we calculated an index of waterabstraction in several steps using data from the OtagoRegional Council.

The first step was to incorporate in the GIS all waterabstraction locations in the Manuherikia catchment,each with its maximum permitted rate of water take(MWR; L s

−

1

). Records of actual water abstraction rateswere not available. We also added 50 locations to the

932

F. Leprieur

et al.



,

43

, 930–939

GIS for which hydrological data were available, specif-ically the 7-day mean annual low flow (MALF; L s

−

1

),a measure of the risk of extended low discharge condi-tions (Richter

et al

. 1996). The MALF locations werereasonably evenly distributed in the catchment but didnot correspond to the 135 fish sites. For each of the 48fish sites subject to water abstraction upstream, weattributed a MALF value from the closest hydrologicalmonitoring location. Then we divided the attributedMALF value by the sum of MWR upstream of each ofthe 48 sites. We chose MALF/MWR as the ratio ofinterest because the quantity of water permitted to betaken upstream was always greater than the MALF.Finally, the index of water abstraction was defined asfollows:

This index ranges from 0 to 1 and the closer the indextends to 1, the greater the risk that the stream will dryup because of water abstraction. To include sites whereno abstraction occurred, we transformed this indexinto a categorical series comprising six classes (class 1,no water abstraction; class 2, 0 = IWA < 0·2; class 3,0·2 = IWA < 0·4; class 4, 0·4 = IWA < 0·6; class 5, 0·6 =IWA < 0·8; class 6, 0·8 = IWA = 1).

Two types of analysis (classical multivariate and ANN)were performed to answer two different but comple-mentary questions. First, we applied multiple discrimi-nant function analysis (MDFA) on the environmentalmatrix (135 sites and nine variables; Table 1) to test fordifferences in environmental conditions found instreams with one of three fish groups (group 1, siteswith

G. anomalus

alone; group 2, sites containing both

G. anomalus

and brown trout; group 3, sites with browntrout alone). We also employed a stepwise MDFA toidentify variables most able to discriminate betweenthe fish classification groups. The stepwise procedureconsisted of alternating steps of forward selection and

backward elimination. Wilk’s lambda was used toselect variables and the maximum significance of the

F

to enter and

F

to remove criteria were, respectively, 0·05and 0·1. The selected variables were then used to pre-dict in which pre-defined fish group each site belonged.Finally, we assessed the accuracy of the stepwise MDFAmodel by applying a ‘leave-one-out’ cross-validation test(Efron 1983). This test consists of removing one obser-vation from the original matrix followed by MDFA onthe remaining observations to predict the group mem-bership of the omitted observation. This operation wasrepeated for all the observations of the data matrix.

Secondly, to identify the environmental factors thatmay control the distribution of each species, we used asupervised ANN (Rumelhart, Hinton & Williams1986; Lek & Guégan 1999). The ANN architectureconsists of input, hidden and output layers with a one-way flow of information. The input layer of neuronesrepresents the independent environmental variables(Table 1) and the output layer is a single neurone thatrepresents the dependant variable (i.e. species occur-rence). The number of neurones of the hidden layer (10)and the number of iterations for the back-propagationalgorithm (500) were chosen by comparing differentneural networks with various numbers of hiddenneurones and iterations (Lek & Guégan 1999). Theseparameters were selected to optimize the accuracy ofthe model and minimize trade-off between networkbias and variance. To standardize the scale of measure-ment, independent variables were converted to

z

-scoresprior to training the models (i.e. the variables were cen-tred and reduced to range between 0 and 1). Again weused the ‘leave-one-out’ cross-validation test (Efron1983) to validate the accuracy of each ANN model(Guégan, Lek & Oberdorff 1998). Then we used differ-ent metrics reviewed by Fielding & Bell (1997) andManel, Williams & Ormerod (2001) to evaluate theperformance of the neural network models. The cal-culation of these metrics required the derivation ofmatrices of confusion that identified true positive,false positive, true negative and false negative casespredicted by each model. First, we explored receiver

Table 1. Mean and ranges for environmental variables described at the reach and site scales

Variable Code Minimum Maximum Mean Data transformation

Reach scaleStream order SO 1 7 3·30 NoneValley slope* VS 1 3 2·16 None

Site scaleAltitude (m) ALT 170 1100 586·70 NoneRelative distance from the main channel RDM 0 1 0·53 ArcsinhIndex of water abstraction IWA 1 6 2·56 None% pool %PO 0 100 28·60 Arcsinh% riffle %RF 0 100 21·62 Arcsinh% run %RN 0 100 38·42 Arcsinh% rapid %RP 0 100 11·37 Arcsinh

*Valley slope: 1, high gradient (slope > 0·04); 2, medium gradient (0·02 = slope = 0·04); 3, low gradient (slope < 0·02). The valley slope is based on Euclidean length (m m−1).

IWAMALFMWR

= −

1

933




,

43

, 930–939

operating characteristic (ROC) plots, obtained by plottingthe proportion of true presences (sensitivity) againstthe proportion of false presences (1

−

specificity) forvarying decision thresholds over the entire rangebetween 0 and 1. Two parameters were derived fromthe ROC plots. (i) The area under the curve (AUC),which is a robust indicator of model performance inde-pendent of the threshold probability at which the spe-cies’ presence is accepted. AUC varies from 0·5 for achance performance to 1·0 for a perfect fit. (ii) The opti-mal decision threshold, which maximizes the propor-tion of true presences (sensitivity) and true absences(specificity) that are correctly classified. The con-ventional decision threshold of 0·5 is arbitrary andmay affect the outcome of a model (Manel, Dias &Ormerod 1999). Finally, we used Cohen’s Kappa index,ranging from 0 to 1, to assess whether the performanceof each model differed from expectations based onchance alone. This index is relatively independent ofspecies’ prevalence and values of 0·8–1 are generallyconsidered to indicate excellent model performance(Manel, Williams & Ormerod 2001).

An important issue in model evaluation is determin-ing the relative contribution (i.e. explanatory impor-tance) of each predictive variable. To do this, we usedthe connection weight procedure (Olden, Joy & Death2004), running each model 100 times (Joy & Death2004), and displaying the relationships graphically bymeans of ‘Lowess’ smoothing plots (Trexler & Travis1993) for the most strongly contributing variables(i.e. > 15% of contribution).

The ANN models were generated using Matlab®(Mathworks, Natick, MA, USA) software language.Other analyses were performed with SPSS for windows,version 11·0 (SPSS Inc., Chicago, IL). Data transfor-mations were applied only for multivariate analyses asassumptions about linearity, normality and homoge-neity of variance are not required for ANN methods.

Finally, we checked for spatial autocorrelation inmodel residuals for both MDFA and ANN modellingtechniques. Autocorrelation analyses were based onMantel’s test, which determines linear relationshipsbetween pairwise distance matrices (Mantel 1967). TheEuclidean distance was selected as a geographical dis-tance between sites, and Euclidean distances betweenall pairs of sites were calculated with

x–y

coordinates.This measurement accounted for environmental con-ditions independently of river network structure, as thisstructure has been considered in both MDFA andANN analyses. We measured Euclidean distance forANN (continuous) residuals, and Jaccard distance forMDFA (binary) residuals. For this last analysis, mis-classified sites were coded as one and properly classi-fied sites were coded as zero. Spatial autocorrelationwas then calculated on misclassifications within the overallpattern of sites. The significance of the normalizedMantel statistic (Legendre & Legendre 1998) was eval-uated by comparing the observed value with a refer-ence distribution of 1000 randomly permutated values.

Results

Among the 135 sites, brown trout occurred in 101 sitesand

G. anomalus

in 34 sites, with the two species coex-isting in only 13 sites (Fig. 1). Brown trout were capableof reaching most locations in the Manuherikia catch-ment, and often occurred upstream of

G. anomalus

.There were only three waterfalls and five dams present(Fig. 1); moreover, brown trout were present above andbelow each dam and two of the waterfalls. Therefore wedecided to ignore dams and waterfalls in our analyses.

The MDFA showed that streams containing

G.anomalus

and brown trout in allopatry and sympatrywere characterized by different environmental condi-tions (function 1, Wilk’s lambda = 0·430, chi-square =108·651,

P

< 0·0001; function 2, Wilk’s lambda = 0·769,chi-square = 34·652,

P

< 0·0001). The plot of site scoresfor the first two discriminant functions (Fig. 2a) andthe plot of loadings (Fig. 2b) indicated the environ-mental variables that were most strongly distinguishedamong the fish assemblages. Streams containing only

G.anomalus

were defined by a high percentage of poolhabitat, high risk of drying up as a result of waterabstraction and low to moderate valley slopes. In con-trast, trout streams were characterized by a relativediversity of channel units (riffle, run, rapid) and little orno water abstraction (78% of sites with brown troutalone had no abstraction upstream). Moreover, 20%of sites containing only brown trout were located in

Fig. 1. Map showing the distributions of G. anomalus andbrown trout in allopatry and sympatry in the ManuherikiaRiver catchment.

934

F. Leprieur

et al.



,

43

, 930–939

tributaries far from the main stem of the river and werecharacterized by a high percentage of rapids and highaltitude. The few sites with

G. anomalus

and browntrout in sympatry were characterized by a high risk ofstream drying, low valley slope, good representation ofriffles and runs, and medium to high stream order.

Stepwise discriminant function analysis confirmedthat water abstraction (

F

= 14·183,

P

< 0·0001), per-centage of pool (

F

= 17·785,

P

< 0·0001) and valleyslope (

F

= 4·089,

P

< 0·05) were the most discriminat-ing variables of the streams containing

G. anomalus

andbrown trout in allopatry and sympatry, and most of thesites (76%) were classified correctly to each predefinedfish group using only these variables (Table 2). Thespatial autocorrelation analysis was not significant(Mantel test,

P

= 0·895).The occurrence of both native

G. anomalus

and exoticbrown trout was highly predictable based on neuralnetwork models (> 93% of sites correctly classified;Table 3), with sensitivity (presence correctly predicted)and specificity (absence correctly predicted) bothexhibiting high values. Cohen’s Kappa statistic andAUC were highly significant, with values indicatingexcellent performance of each model. For the threemodels, spatial autocorrelation analysis revealed inde-pendence between site residuals (Mantel test,

P

= 0·182for

G. anomalus

model 1,

P

= 0·439 for

G. anomalus

model2,

P

= 0·111 for brown trout model).The percentage of pool and the index of water abstrac-

tion contributed most to predicting the occurrence ofbrown trout in the ANN models (Fig. 3a). The Lowesssmoothing curve indicated that the probability of occur-rence of brown trout decreased as the percentage of poolor the index of water abstraction increased (Fig. 4a). Inother words, the few sites where brown trout were absentwere characterized by a moderate to high percentage ofpool and a maximum index of water abstraction. In thecase of

G. anomalus

, two variables stood out: the indexof water abstraction and valley slope (Fig. 3b).

Gal-axias anomalus

occurred in sites with low valley slope(< 0·02) and maximal risk of drying up as a result of waterabstraction (Fig. 4b). An increase in percentage poolcorresponded with an increased probability of

G. anom-alus

occurrence (Fig. 4b). Finally, when the occurrenceof brown trout was added as a further independent var-iable to predict

G. anomalus

occurrence, this biotic var-iable contributed most to the ANN model (Fig. 3c).

Discussion

The distributions of

G. anomalus

and brown trout werelargely non-overlapping at the reach scale in theManuherikia River. Both MDFA and ANN modelsindicated that valley slope, the percentage of pool habitat

Fig. 2. (a) Plot of discriminant function scores for each of the135 sites using the first two functions (black circles, 1, G.anomalus only; triangles, 2, G. anomalus and brown trout;white circles, 3, brown trout only; the white circles containinga number correspond to the group centroids). (b) Plot ofcorrelations among discriminating variables and the standardizedcanonical discriminant functions.

Table 2. Classification results obtained by stepwise discriminant function analysis and by ‘leave-one-out’ cross-validation. Thenumber of correctly predicted sites is shown in bold

Group No. of sites

Predicted group membership

Sites correctly predicted (%)1 2 3

1 G. anomalus alone 21 13 5 3 622 G. anomalus and brown trout together 13 2 8 3 623 Brown trout alone 101 9 10 82 81Total 135 24 23 88 76

935




,

43

, 930–939

in the stream reach and, in particular, water abstractioncould best account for their disjunct distributions. Theabsence of spatial autocorrelation in model residualsensures the relevance of the independent variables usedin the ANN models and the randomness of misclassi-fications in the stepwise MDFA model. As a conse-quence, the variables used in both MDFA and ANNmodels were appropriate to classify sites and predictspecies occurrence, respectively.

Brown trout occurred in low, medium and high gra-dient streams, but

G. anomalus

most often occurred inreaches with low valley slopes. This accords with theresults for

G. anomalus

of Allibone & Townsend (1997)for the Taieri River. However, we have no record in theManuherikia River of the extent to which

G. anomalus

may have been restricted to lower gradient sites beforethe arrival of brown trout. Old reports for the nearbyTaieri River suggest that non-migratory galaxiidswere historically much more widespread (Townsend &Crowl 1991).

Table 3. Performance of the ANN models to predict presence–absence of G. anomalus and brown trout according to a ‘leave-one-out’ cross-validation test (see text for the AUC and Kappa index). Brown trout and G. anomalus (model 1) models were built usingthe 10 predictors given in Table 1. The last model (G. anomalus model 2) includes the same input data, plus trout occurrence,considered here as a predictor

SpeciesCorrect classification (%)

Optimal decision threshold

Sensitivity (%)

Specificity (%)

Kappa index P AUC P

Brown trout 94·96 0·60 98·21 88 0·832 < 0·0001 0·909 < 0·0001G. anomalus (model 1) 93·53 0·50 82·22 98·94 0·846 < 0·0001 0·974 < 0·0001G. anomalus (model 2) 100 0·49 100 100 1 < 0·0001 1 < 0·0001

Fig. 3. Relative contribution of the independent variables inthe ANN models. (a) Brown trout model, (b) G. anomalusmodel 1, (c) G. anomalus model 2 (occurrence of brown troutincluded as predictor variable). The direction of the relationshipbetween predictors and predictions, extracted from thecalculation of the connection weight algorithm, is indicatednext to each environmental variable: positive ‘+’, negative ‘–’.Definitions of variables provided in Table 1.

Fig. 4. Bivariate plots of predicted probability of fishoccurrence against the most important contributing variablesin the ANN models. The solid lines represent the Lowesscurves used to fit the data. The proportion of samplesperfectly fitted is indicated by the f-value (i.e. smoothingparameter). The f-values ranged from 0 and to 1 accordingto the sensitivity of the analysis: (a) brown trout ( f = 0·8),(b) G. anomalus model 1, occurrence of trout not included aspredictor ( f = 0·8).

936F. Leprieur et al.

© 2006 The Authors. Journal compilation © 2006 British Ecological Society, Journal of Applied Ecology, 43, 930–939

The development of pastoral irrigation using his-toric gold-mining water races and water rights pre-ceded the arrival of brown trout. The development ofirrigation schemes during the early and mid-1900s waspredominantly in low gradient valleys located in themiddle and downstream reaches of each tributary.More recent expansion of irrigation in the drought-prone Manuherikia catchment has exacerbated thecombined effects of these historical abiotic and bioticdisturbances. At present, the permitted rate of waterabstraction is considerable, with most sites where G.anomalus occurs alone having 7-day minimum annuallow flows ranging from 30 to 300 L s−1, whereas the sumof permitted water abstractions upstream varies from500 to 2300 L s−1. This very high proportion of naturalflow abstracted means that farmers may actually takeall the water from a tributary, and many reaches becomedry during the low flows of summer. These reaches arealso characterized by a high proportion of pool habi-tat, in contrast to the sites containing only brown trout.The predominance of pool habitat and the under-representation of riffles and runs is probably related tothe reduction of stream discharge (Kraft 1972; Elliot 2000).

The resulting periods of low, or even no, flow arecharacterized by the presence of isolated pools andextreme conditions of high temperature (recorded upto 28 °C) and associated low oxygen concentrations,conditions that galaxiid fish tolerate (Richardson,Boubee & West 1994; Dean & Richardson 1999) betterthan salmonids (Closs & Lake 1996). Non-migratorygalaxiids and especially Galaxias eldoni McDowall andG. anomalus have been recorded in Otago streams withtemperatures above 28 °C (C. Arbuckle, personalobservation). During low discharge, moreover, G. vul-garis (Dunn 2003) and Galaxias cobitinis McDowalland Waters (C. Arbuckle, personal observation) arecapable of burrowing and surviving in the stream bed,a behaviour that may be shared by a number of non-migratory galaxiids. The response of brown trout tothese extreme conditions may be death (the upperlethal limit for brown trout is 25 °C; Elliot 1994) ormigration to upstream locations where abstractiondoes not occur. Large-scale migration in response tolow flows is common in salmonid populations (Kraft1972; Gowan & Fausch 1996).

Apart from these irrigation-impacted sites, browntrout occurred throughout the Manuherikia Rivercatchment, even above some large waterfalls and dams.This is because of their introduction for sport fishing,particularly in upstream reaches and more recently inheadwater reservoirs. Headwater introduction seems topromote catchment-wide invasion more than main-stream or low altitude stocking (Adams, Frissel &Rieman 2001). In the Manuherikia catchment, 92% of sitesabove dams contained only brown trout, while only fivesites contained both species. It seems likely that G.anomalus were eliminated from the other sites by acombination of trout predation and competition. Fourlow-order sites, unimpacted by irrigation, contained

only G. anomalus, and these might conform to the pat-tern in the Taieri River where migration barriers leaveupstream refugia for the natives (Townsend & Crowl 1991).One was located above a large waterfall that probablyprevented trout invasion. The other sites were notupstream of large barriers but downstream irrigationwater races comprised dams that may impede migration;stocking of trout may not have occurred in these sites.

The sites where G. anomalus and brown trout occurredin sympatry were associated with low valley slopes inhigh-order streams (in the main stem or close to it),where riffles and runs were well represented in a braidedriver structure. Downstream dispersal from higheraltitude reaches seems not to be important because thegalaxiids are generally not present above sites occupiedby both trout and galaxiids (Fig. 1). More likely, bedinstability in these braided sections promotes fishcoexistence, as noted for similar locations in the TaieriRiver (Townsend 2003) and elsewhere in the SouthIsland (McIntosh 2000). Promotion of the coexistenceof native and exotic species by disturbance has alsobeen reported for other stream fishes (Meffe 1984),amphibians (Doubledee, Muller & Nisbet 2003) andplants (Vujnovic, Wein & Dale 2002).

Overall, the negative association between G. anomalusand brown trout is mainly related to the level of hydro-logical disturbance. Thus local G. anomalus popula-tions have not been excluded by brown trout in lowgradient streams subject to significant water abstrac-tion. Closs & Lake (1996) found that Galaxias olidusGünther were similarly protected from trout predationby severe drought in summer, observing that G. olidussurvived in upstream reaches that tended to stopflowing while brown trout could only survive in down-stream reaches less affected by drought. The pattern isreversed in the Manuherikia, where the risk of streamdrying, because of abstraction for irrigation, occurs inthe middle and downstream reaches.

Our findings run contrary to the idea that anthropo-genic disturbance is more likely to facilitate invasionsof exotic species (Minckley & Meffe 1987; Hobbs &Huenneke 1992; Moyle & Light 1996; Byers 2002).Whether a disturbance will facilitate an invasion dependson whether the disturbance is natural or human-induced(McIntosh 2000) and whether the exotic species (or thenative), by virtue of its evolutionary history, is favouredby the changed conditions (Baltz & Moyle 1993; Moyle& Light 1996; Townsend 2003). Our results are con-sistent with the classical view that exotic species can onlyinvade as far as their physiological tolerances permit(Moyle & Light 1996; Holway, Suarez & Case 2002;Facon et al. 2004). This underlines the need to identifyniche components of exotic species to better forecasttheir distribution and impact on natives.

Exotic species management is particularly sensitivefrom a political point of view when the invader has a

937Hydrological disturbance favours a native fish


high economic value, as is the case with brown trout forsport fishing in New Zealand. Consequently, eradica-tion of brown trout at the catchment scale is not con-ceivable. In addition, management of water abstractionfor irrigation of agricultural land, which has a regionaleconomic importance, is currently focused on improv-ing irrigation efficiency to lessen abstraction require-ments and establishing minimum residual flows tomaintain aquatic habitat. These developments mustproceed with caution to ensure that the re-establishmentof more natural flows, which will favour brown trout,do not threaten the native fish in their physically stressedrefugia.

Key factors in the conservation of G. anomalus, inconjunction with water resource management, are theidentification of locations where galaxiid reserves canbe established, the encouragement of public educationabout the threats faced by the native fish and, wherepossible, the reduction of abstraction. Although waterabstraction has created local abiotic conditions thatseem to prevent G. anomalus extirpation by brown trout,we certainly do not advocate an increase in permittedabstraction for the sake of galaxiid populations. Thistype of hydrological disturbance can be expected tohave negative impacts for G. anomalus by reducing thecarrying capacity of the stream (Allibone 2000a), dis-rupting spawning habitats and juvenile recruitment,and reducing growth rates of larval fish (Allibone 2000b).

In its recovery plan for non-migratory galaxiids, theDepartment of Conservation (2004) emphasizes theneed to maintain and improve fish barriers, to informlandowners of barriers and their importance, and torequest them not to transfer trout above these barriersor allow others to do so. In the case of the ManuherikiaRiver catchment, with only a very few exceptions,brown trout have already been introduced to streamsabove waterfalls and dams. Therefore we suggest that anumber of tributaries should be chosen with habitatfeatures appropriate to the different life-history stagesof G. anomalus, where artificial barriers can be con-structed to impede trout upstream migration. Troutshould be removed by repeated electrofishing above thebarriers, which should permit a G. anomalus popula-tion to recolonize the stream above. This was done suc-cessfully in a montane stream in south-eastern Australia,where a breeding population of G. olidus had becomeestablished 3 years after trout eradication (Lintermans2000). Given the habitat requirements of G. anomalus,such streams are unlikely to support significant stocksof brown trout, so their removal will have little impacton the sports’ angling resource. Only after reserves ofgalaxiid populations have been established, do we recom-mend restoring natural low flows. Indeed, if the nat-ural patterns of low flow are restored to the trout-freestreams that currently support galaxiid populationsprior to the implementation of reserves for galaxiids,we foresee that brown trout will colonize and imperilthe remaining galaxiid populations. Although Morita& Yamamoto 2001, showed that isolation can increase

the extinction probability because of inbreeding instream-dwelling charr Salvelinus leucomaenis Pallas,this risk is probably reduced for small, sedentary fishthat occur at high population densities in small streamreaches, such as G. anomalus. Moreover, when exoticspecies pose an immediate threat to the survival ofnative species, the risk of isolation is justified whenseeking rapid protection of threatened native speciesfrom the negative effects of exotics (Moyle & Sato 1991;Shafer 1995; Novinger & Rahel 2003). Over a longertime scale, inbreeding risk should be properly assessedto ensure the sustainability of threatened populations.

Acknowledgements

We are grateful to Jon Waters (University of Otago)and Mike Tubbs (Department of Conservation) forproviding information during this study. Thanks toSovan Lek for his assistance with artificial neuralnetwork modelling techniques, and to Gael Grenouil-let for his assistance with spatial autocorrelationanalysis. We also thank Kentaro Morita and two anon-ymous referees for their insightful comments on themanuscript.

References

Adams, S.B., Frissel, C.A. & Rieman, B.E. (2001) Geographyof invasion on mountain streams: consequences of head-water lake fish introductions. Ecosystems, 4, 296–307.

Allibone, R.M. (2000a) Assessment techniques for waterabstraction impacts on non-migratory galaxiids of Otagostreams. Science for Conservation, 147, 5–23.

Allibone, R.M. (2000b) Water abstraction impacts on thenon-migratory galaxiids of Totara Creek. Science for Con-servation, 147, 25–45.

Allibone, R.M. & Townsend, C.R. (1997) Distribution of fourrecently discovered galaxiid species in the Taieri River, NewZealand: the role of microhabitat. Journal of Fish Biology,51, 1235–1246.

Baltz, D.M. & Moyle, P.B. (1993) Invasion resistance to intro-duced species by a native assemblage of California streamfishes. Ecological Applications, 3, 246–255.

Byers, J.E. (2002) Impact of non-indigenous species enhancedby anthropogenic alteration of selection regimes. Oikos, 97,449–458.

Clavero, M. & García-Berthou, E. (2005) Invasive species area leading cause of animal extinctions. Trends in Ecology andEvolution, 20, 110.

Closs, G.P. & Lake, P.S. (1996) Drought, differential mortalityand the coexistence of a native and an introduced fish speciesin a south east Australian intermittent stream. EnvironmentalBiology of Fishes, 47, 17–26.

Crowl, T.A., Townsend, C.R. & McIntosh, A.R. (1992) Theimpact of introduced brown and rainbow trout on nativefish: the case of Australasia. Reviews in Fish Biology andFisheries, 2, 217–241.

Dean, T.L. & Richardson, J. (1999) Responses of seven speciesof native freshwater fish and shrimp to low levels of dissolvedoxygen. New Zealand Journal of Marine and FreshwaterResearch, 33, 99–106.

Department of Conservation (2004) New Zealand Non-Migratory Galaxiid Fishes Recovery Plan 2003–13. ThreatenedSpecies Recovery Plan 53. Department of Conservation,Wellington, New Zealand.

938F. Leprieur et al.


Doubledee, R.A., Muller, E.B. & Nisbet, R.M. (2003)Bullfrogs, disturbances regimes and the persistence ofCalifornia red-legged frogs. Journal of Wildlife Manage-ment, 67, 424–438.

Dunn, N.R. (2003) The effects of extremes in flow on alpine(G. paucispondylus) and Canterbury (G. vulgaris) Galaxias.Unpublished MSc Thesis. University of Canterbury,Christchurch, New Zealand.

Efron, B. (1983) Estimating the error rate of a prediction rule:some improvements on cross-validation. Journal of theAmerican Statistical Association, 78, 316–331.

Elliot, J.M. (1994) Quantitative Ecology and the Brown Trout.Oxford University Press, Oxford, UK.

Elliott, J.M. (2000) Pools as refugia for brown trout duringtwo summer droughts: trout responses to thermal andoxygen stress. Journal of Fish Biology, 56, 938–948.

Facon, B., Machline, E., Pointier, J.P. & David, P. (2004)Variation in desiccation tolerance in freshwater snails andits consequences for invasion ability. Biological Invasions,6, 283–293.

Fielding, A.H. & Bell, J.F. (1997) A review of methods for theassessment of prediction errors in conservation presence/absence models. Environmental Conservation, 24, 38–49.

Gowan, C. & Fausch, K.D. (1996) Mobile brook trout in twohigh-elevation Colorado streams: re-evaluating the conceptof restricted movement. Canadian Journal of Fisheries andAquatic Sciences, 53, 1370–1381.

Guégan, J.F., Lek, S. & Oberdorff, T. (1998) Energy availabil-ity and habitat heterogeneity predict global riverine fishdiversity. Nature, 391, 382–384.

Hanchet, S.M. (1990) Effect of land-use on the distributionand abundance of native fish in tributaries of the WaikatoRiver in the Hakaramita Range, North Island New Zealand.New Zealand Journal of Marine and Freshwater Research,24, 159–171.

Hitchmough, R. (2002) New Zealand Threat ClassificationSystem Lists, 2002. Department of Conservation, Welling-ton, New Zealand.

Hobbs, R.J. & Huenneke, L.F. (1992) Disturbance, diversity,invasions: implications for conservation. ConservationBiology, 6, 324–337.

Holway, D.A., Suarez, V.A. & Case, T.J. (2002) Role of abioticfactors in governing susceptibility to invasion: a test withargentine ants. Ecology, 83, 1610–1619.

Joy, M.K. & Death, R.G. (2004) Predictive modelling and spatialmapping of freshwater fish and decapod assemblages usingGIS and neural networks. Freshwater Biology, 49, 1036–1052.

Kraft, M.E. (1972) Effects of controlled flow reduction on atrout stream. Journal of the Fisheries Research Board ofCanada, 29, 1405–1411.

Krueger, C.C. & May, B. (1991) Ecological and genetic effectsof salmonid introductions in North America. CanadianJournal of Fisheries and Aquatic Sciences, 48 (Supplement 1),66–77.

Legendre, P. & Legendre, L. (1998) Numerical Ecology, SecondEnglish Edition. Elsevier, Amsterdam, the Netherlands.

Lek, S. & Guégan, J.F. (1999) Artificial neural networks as atool in ecological modelling, an introduction. EcologicalModelling, 120, 65–73.

Lintermans, M. (2000) Recolonization by the mountaingalaxias Galaxias odilus of a montane stream after theeradication of rainbow trout Oncorhyncus mykiss. Marineand Freshwater Research, 51, 799–804.

McDowall, R.M. (1968) Interactions of the native and alienfaunas of New Zealand and the problem of fish intro-ductions. Transactions of the American Fisheries Society, 97,1–11.

McDowall, R.M. (2003) Impacts of introduced salmonidson native galaxiids in New Zealand upland streams: a newlook at an old problem. Transactions of the American FisheriesSociety, 132, 229–238.

McDowall, R.M. & Richardson, J. (1983) The New ZealandFreshwater Fish Survey: Guide to Input and Output. Fisher-ies Research Division Information Leaflet Number 12.Ministry of Agriculture and Fisheries, Wellington.

McIntosh, A.R. (2000) Habitat and size-related variations inexotic trout impacts on native galaxiid fishes in New Zea-land streams. Canadian Journal of Fisheries and AquaticSciences, 57, 2140–2151.

Mack, R.N., Simberloff, D., Lonsdale, W.M., Evans, H.,Clout, M. & Bazzaz, F.A. (2000) Biotic invasions: causes,epidemiology, global consequences, and control. Ecolo-gical Applications, 10, 689–710.

Manel, S., Dias, J.M. & Ormerod, S.J. (1999) Comparing dis-criminant analysis, neural networks and logistic regressionfor predicting species’ distributions: a case study with aHimalayan river bird. Ecological Modelling, 120, 337–347.

Manel, S., Williams, H.C. & Ormerod, S.J. (2001) Evaluatingpresence–absence models in ecology: the need to accountfor prevalence. Journal of Applied Ecology, 38, 921–931.

Mantel, N. (1967) The detection of disease clustering and ageneralized regression approach. Cancer Research, 27, 209–220.

Matthews, W.F. & Marsh-Matthews, E. (2003) Effects of droughton fish across axes of space, time and ecological complexity.Freshwater Biology, 48, 1232–1253.

Meffe, G.K. (1984) Effect of abiotic disturbance on coexist-ence of predator–prey fish species. Ecology, 65, 1525–1534.

Minckley, W.L. & Meffe, G.K. (1987) Differential selection byflooding in stream-fish communities of the arid AmericanSouthwest. Community and Evolutionary Ecology of NorthAmerican Stream Fishes (eds W.J. Matthews & D.C. Heins),pp. 93–104. University of Oklahoma Press, Norman, OK.

Morita, K., Tsuboi, J. & Matsuda, H. (2004) The impact ofexotic trout on native charr in a Japanese stream. Journal ofApplied Ecology, 41, 962–972.

Morita, K. & Yamamoto, S. (2001) Effects of habitat frag-mentation by damming on the persistence of stream-dwellingcharr populations. Conservation Biology, 16, 1318–1323.

Moyle, P.B. & Light, T. (1996) Fish invasions in California: doabiotic factors determine success? Ecology, 77, 1666–1670.

Moyle, P.B. & Sato, G.M. (1991) On the design of preserves toprotect native fishes. Battle Against Extinction: Native FishManagement in the American West (eds W.L. Minckley &J.E. Deacon), pp. 155–173. University of Arizona Press,Tucson, AZ.

Novinger, D.C. & Rahel, F.J. (2003) Isolation managementwith artificial barriers as a conservation strategy for cut-throat trout in headwater streams. Conservation Biology,17, 772–781.

Olden, J.D., Joy, M.K. & Death, R.G. (2004) An accuratecomparison of methods for quantifying variable impor-tance in artificial neural networks using simulated data.Ecological Modelling, 178, 389–397.

Palmer, M.E. & Ricciardi, A. (2004) Physical factors affectingthe relative abundance of native and invasive amphipods inthe St Lawrence River. Canadian Journal of Zoology, 82,1886–1893.

Ricciardi, A. (2003) Predicting the impacts of an introducedspecies from the invasion history: an empirical approachapplied to zebra mussels invasions. Freshwater Biology, 48,972–981.

Richardson, J., Boubee, J.A.T. & West, D.W. (1994) Thermaltolerance and preference of some native New Zealandfreshwater fish. New Zealand Journal of Marine and Fresh-water Research, 28, 399–407.

Richter, B.D., Baumgartner, J.V., Powell, J. & Braun, D.P.(1996) A method for assessing hydrologic alteration withinecosystems. Conservation Biology, 10, 1163–1174.

Rumelhart, D.E., Hinton, G.E. & Williams, R.J. (1986) Learningrepresentations by back-propagating error. Nature, 323,533–536.

939Hydrological disturbance favours a native fish


Sakai, A.K., Allendorf, F.W., Holt, J.S., Lodge, D.M., Molofsky,J., With, K.A., Baughman, S., Cabin, R.J., Cohen, J.E., Ellstrand,N.C., McCauley, D.E., O’Neil, P., Parker, I.M., Thompson, J.N.& Weller, S.G. (2001) The population biology of invasive species.Annual Review of Ecology and Systematics, 32, 305–332.

Schaefer, J.F. & Kerfoot, J.R. (2004) Fish assemblage dynamicsin an adventitious stream: a landscape perspective.American Midland Naturalist, 151, 134–145.

Shafer, C.J. (1995) Values and shortcomings of small reserves.Bio-Science, 45, 80–88.

Snelder, T., Biggs, B.J.F. & Weatherhead, M. (2004) NewZealand River Environment Classification User Guide.Ministry for the Environment, Wellington, New Zealand.www.niwa.co.nz/ncwr/rec/.

Spitz, F. & Lek, S. (1999) Environmental impact predictionusing neural network modelling. An example in wildlifedamage. Journal of Applied Ecology, 36, 317–326.

Townsend, C.R. (2003) Individual, population, communityand ecosystem consequences of a fish invader in NewZealand streams. Conservation Biology, 17, 38–17.

Townsend, C.R. & Crowl, T.A. (1991) Fragmented popula-tion structure in a native New Zealand fish: an effect ofintroduced brown trout? Oikos, 61, 348–354.

Trexler, J.C. & Travis, J. (1993) Non-traditional regressionanalyses. Ecology, 74, 1629–1637.

Vitousek, P.M. (1994) Beyond global warming: ecology andglobal change. Ecology, 75, 1861–1876.

Vujnovic, K., Wein, R.W. & Dale, M.R.T. (2002) Predictingplant species response to disturbance magnitude in grasslandremnants of central Alberta. Canadian Journal of Botany,80, 504–511.

Received 19 July 2005; final copy received 13 April 2006 Editor: Paul Giller


P2

Impact of the invasive black bullhead (Ameirius melas Raff.) on the predatory efficiency of northern pike (Esox lucius L.)

Kreutzenberger K., Leprieur F., & Brosse, S.

Journal of Fish Biology (en révision mineure)

1

Impact of the invasive black bullhead (Ameiurus melas Raf.) on the

predatory efficiency of northern pike (Esox lucius L.)

K. Kreutzenberger, F. Leprieur, S. Brosse

Laboratoire Evolution & Diversité Biologique, U.M.R 5174, C.N.R.S -Université

Paul Sabatier, 118 route de Narbonne, F-31062 Toulouse cedex 4, France.

Running title: Impact of black bullhead on northern pike

2

Abstract

The influence of the invasive black bullhead (Ameiurus melas Raf.) on the predatory

efficiency of the northern pike (Esox lucius L.) was investigated using an additive

experimental design. Pike predatory success on 0+ roach (Rutilus rutilus L.) was significantly

reduced in the presence of bullhead. Among the different hypotheses that may explain such a

pattern, the hypothesis of direct competition between pike and bullhead was not verified, as

bullhead hardly fed on roach. Similarly, pike predatory efficiency did not decrease with

turbidity, rejecting therefore the hypothesis of an indirect effect through bullhead-generated

turbidity. Therefore, the reduced predatory efficiency of pike was probably related to

behavioural interference between pike and bullhead. These laboratory results confirm the

potential negative impact of bullhead on native European fauna, with a particular emphasis on

pike, which is a top predator considered as vulnerable in some European regions. These

results based on laboratory experiments need to be tested in natural environments to allow

generalization.

Key words: invasive species, predatory efficiency, multipredator, turbidity

3

Introduction

Freshwater ecosystems have received many fish invaders (Welcomme, 1988), and

these invasive species have been recognised as a major threat to biodiversity and ecosystem

integrity (Vitousek et al., 1997; Mack et al., 2000). Non-native fishes can modify the strength

of biotic interactions (competition, predation) within native communities (Townsend, 2003;

Blanchet et al., 2007). They can also play a role in the introduction of parasites and diseases,

contribute to genetic deterioration, and modify the environment (Taylor et al., 1984).

According to Holčík (1991), 134 non-native freshwater fish were introduced in Europe and

almost all large European river basins are now invaded by non-native species (Clavero &

Garcia Berthou, 2006; Leprieur et al., 2007). However, the impact of most fish introductions

on the native European fish fauna is still unknown (Elvira, 2001).

The black bullhead (Ameiurus melas Rafinesque 1818, hereafter called bullhead), an

ictalurid fish native to North America, is one of the most abundant non-native fish species in

European freshwater ecosystems (Declerck et al., 2002; Cucherousset et al., 2006). Bullhead

can account for more than 30% of fish abundance (Boët, 1980; Cucherousset et al., 2006),

with biomasses ranging from 5 to 50 kg ha-1 (Louette & Declerck, 2006). Most European

policies therefore consider this species as liable to cause biological disequilibrium (e.g. Elvira,

2001; Keith & Allardi, 2001). However, the impact of bullhead on native fish populations has

never been quantified.

The bullhead is a benthivorous fish inhabiting standing waters with soft bottom

substrata (Keith & Allardi, 2001), and its activity is known to generate turbidity (Braig &

Johnson, 2003). Although usually considered as detritivorous, its diet may include live fish

(Boët, 1980). Bullhead may therefore affect the native fauna in three distinct ways. First, it

may prey directly on some species, therefore reducing the amount of available prey for native

predators. Second, bullhead may have an indirect impact by generating turbidity (Braig &

4

Johnson, 2003), that can modify the feeding efficiency of visual predators (Reid et al., 1999;

Utne-Palm, 2002). Third, due to their high local abundance, bullhead behaviour may interfere

with accompanying species and hence negatively affect the behavioural feeding phases of

native predators and/or the anti-predator behaviour of native prey.

In this context, the direct (i.e. predation), indirect (i.e. turbidity) and interference

effects of bullhead on the predatory efficiency of pike (Esox lucius L. 1758) were examined in

the laboratory; and more specifically whether bullhead in the presence of pike led to a

predation risk reduction or enhancement for prey in clear and turbid waters. The pike was

selected as it frequently co-occurs with bullhead in Europe (e.g. Cucherousset et al., 2007).

Moreover, the two species commonly prey on roach (Rutilus rutilus L. 1758) (Boët, 1980;

Hart & Connellan, 1984) and may therefore compete for food. In addition, pike is a visual

predator (Casselman & Lewis, 1996) that may be affected by the turbidity generated by

bullhead activity. In the present study, an additive experimental design (sensu Griffen, 2006)

was conducted at two turbidity levels (i.e. Clear and Turbid Water), that consisted in

comparing predation by each species separately to predation when the species were combined.

This design is commonly employed to detect predation risk reduction or enhancement for

prey subject to consumption by multiple predators (reviewed by Sih et al., 1998).

Material and Methods

Experimental design

Experiments were carried out in autumn 2006. Wild fish were used exclusively to

avoid potential bias due to behavioural changes between farmed and wild strains (Johnsson et

al., 2001). Bullhead and pike were 1+ fish, 143.1 ± 1.1 S.E. mm LT and 30.2 ± 0.7 S.E. g; and

267.4 ± 4.0 S.E. mm LT and 77.9 ± 5.6 S.E. g, respectively. 0+ roach were selected as they are

5

a prey for pike and bullhead in Europe (Boët, 1980; Bruslé & Quignard, 2001). The

size/weight of roach (83.0 ± 0.8 S.E. mm LT and 4.2 ± 0.1 S.E. g) were consistent with those

found in the stomach content of both pike and bullhead (Hart & Connellan, 1984; Declerck et

al., 2002). Prior to the experiments, each species was kept for 2 to 6 weeks in separate 600 l

tanks. Roach were fed with fish pellets, bullhead were fed with 0+ roach and fish pellets and

pike were fed with 0+ roach. Pike and bullhead were starved for a week before the beginning

of each experiment.

Experiments took place in 200 l tanks (100 x 40 x 50 cm) at a temperature of 18 ±

0.5 °C. The bottom of each tank was filled with 5 cm fine sandy substratum (grain size < 1

mm). Diffuse light conditions (1600 ± 10 Lux) were provided by four fluorescent tubes

mounted 15 cm above the tank, which reproduced sunlight with a natural photoperiod (light

was automatically turned on at dawn and off at dusk). The additive experimental design

consisted in one control treatment (no predators) and three predator treatments: pike alone,

bullhead alone and the two predators together. The prey density was identical in each

treatment (10 0+ roach): (i) 10 roach were introduced in the control treatment; (ii) one pike

and 10 roach in the pike treatment; (iii) three bullheads and 10 roach in the bullhead treatment

and (iv) one pike, three bullheads and ten roach in the multipredator treatment. Introducing

three bullheads per tank gave a similar predator biomass in the pike and bullhead treatments

and respected the gregarious habits of this species (Bruslé & Quignard, 2001; Keith & Allardi,

2001). Likewise, introducing no more than one pike per tank is consistent with the territorial

habits of this species (Eklöv & Hamrin, 1989).

Each treatment was run at two turbidity levels: low turbidity, hereafter called clear

water (CW, 1.5 ± 0.04 S.E. NTU), and high turbidity, hereafter called turbid water (TW, 72.5

± 0.29 S.E. NTU). Turbidity was stabilised using an aquarium water pump. It was measured 5

times a day with a Hach 2100 P portable turbidimeter that quantifies the amount of light from

6

an incandescent bulb, scattered at a 90° angle, in nephelometric turbidity units (NTU). In TW

experiments turbidity was controlled by adding 40 g of bentonite clay to the experimental

tanks. TW turbidity was about 70 NTU, a level frequently observed in stagnant lowland water

bodies invaded by bullhead (e.g. dyked wetlands, Braig & Johnson, 2003).

Before each experiment (in both CW and TW experiments), fish were acclimated for

12 h. During this period, tanks were separated into two equal parts by a Plexiglas sheet to

keep predators away from prey. The separation was carefully removed at the end of the

acclimatizing period. Each experiment lasted 3 days (72 h). At the end of each experiment,

the number of remaining roach was counted to deduce the predatory efficiency of each single

predator. In the multipredator treatment, only bullheads were killed and their stomach

contents analysed since pike is classified as vulnerable in France (Keith & Allardi, 2001). The

number of remaining roach was also counted at the end of the experiment. This allowed us to

determine the number of roach consumed by bullheads and by pike. At the end of each

experiment, all the fish were removed from the tank and a new set of fish used for the

following replicate to avoid pseudoreplication. Between each replicate, the tank was emptied

and the water was changed to avoid potential bias due to chemical cues. All experiments were

replicated 8 times. All the pikes were released after the experiments in the same area they

were caught.

Data analysis

The multiple predator effect was determined by comparing the number of prey

remaining for the four treatments (i.e. pike; bullhead; pike + bullhead; no-predator control) at

two turbidity levels. To do this, a three-way ANOVA was applied on log-transformed prey

abundance at the end of each experiment, with the presence/absence of each predator species

treated as a separate factor (Sih et al., 1998; Griffen, 2006). A significant two-way interaction

7

(pike x bullhead) indicates the presence of a non-additive effect of combining the two

predator species and significant three-way interaction (pike x bullhead x turbidity) indicates

that the effect of the two predators changes with turbidity. Then, the predator efficiencies of

pike and bullhead at two turbidity levels were compared using two-way ANOVA on log-

transformed number of prey consumed. Multiple post-hoc comparisons were conducted with

Tukey’s HSD tests. Data were log-transformed prior to each analysis to meet the assumptions

for parametric statistical analysis (i.e. normality and homoscedasticity).

Results

The number of roach remaining for each treatment revealed that turbidity did not

affect prey vulnerability (Table 1). Then, a lower number of remaining prey was found in the

pike-alone treatment than both in the no-predator control (Tukey’s test, P < 0.001, Fig. 1) and

in the bullhead-alone treatment (Tukey’s test, P < 0.001, Fig. 1). In contrast, the number of

remaining prey when pike and bullhead were combined was greater than expected by the

additive experimental design (significant pike x bullhead interaction, Table 1). Indeed, the

number of remaining prey in the multipredator treatment was significantly greater than in the

pike-alone treatment (Tukey’s test, P < 0.05, Fig. 1), but did not significantly differ from that

observed in both the no-predator treatment (Tukey’s test, P = 0.246, Fig. 1) and the bullhead-

alone treatment (Tukey’s test, P = 0.395, Fig. 1). Last, the number of prey remaining in the

bullhead-alone treatment did not differ from that observed in the no-predator control (Tukey’s

test, P = 0.991, Fig. 1).

Considering the number of roach consumed by each predator revealed consistent

results (Table 2). Moreover, a significant effect of bullhead on the predation efficiency of pike

was found (Table 2), resulting in a significant decrease in the number of roach consumed

[Tukey’s test, P < 0.01, Fig. 2(a)]. In contrast, pike did not affect the roach consumption by

8

bullhead [Table 2 and Fig. 2(b)]. Last, turbidity did not affect the predatory efficiency of

either pike or bullhead (Table 2).

Table I: Three-way analysis of variance applied to compare the number of remaining prey in the multiple and single predator treatments at two levels of turbidity (clear water or turbid water). Degrees of freedom, Df.

Source of variation Df SS F P

Pike 1 3.685 21.206 0.000

Bullhead 1 0.653 3.758 0.058

Turbidity 1 0.069 0.394 0.533

Pike x Bullhead 1 0.970 5.583 0.022

Pike x Turbidity 1 0.127 0.728 0.397

Bullhead x Turbidity 1 0.003 0.018 0.894

Pike x Bullhead x Turbidity 1 0.019 0.109 0.743

Error 56 9.730 Table II: Two-way analysis of variance of roach prey consumption by single predators in the presence of another predator at two turbidity levels (clear water or turbid water). Degrees of freedom, Df.

Source of variation Df SS F P Pike predation efficiency 1 Bullhead 1 2.791 12.986 0.001 Turbidity 1 0.006 0.029 0.865 Bullhead x Turbidity 1 0.000 0.001 0.971 Error 28 6.017 Bullhead predation efficiency 1 Pike 1 0.212 0.902 0.350 Turbidity 1 0.013 0.055 0.817 Pike x Turbidity 1 0.001 0.004 0.951 Error 28 6.571

9

Pike Bullhead Pike +Bullhead

Control

0

2

4

6

8

10

Figure 1: Mean number of remaining prey (± standard error, n = 8) in each experiment for clear water (□) and turbid water (■) environments.

Bullhead Bullhead / Pike0

1

2

3

4

5

NS

**

Pike Pike / Bullhead

a) b)

0

2

4

6

8

10

Figure 2: Mean number of prey (± standard error, n= 8) consumed by each predator in clear water (□) and turbid water (■) environments. a) Prey consumed by pike alone and in the presence of bullhead; b) Prey consumed by bullhead alone and in the presence of pike. ** P < 0.01; ns: P > 0.05.

10

Discussion

In this study, the effect of a turbidity level (c.a. 70 NTU) frequently observed in

standing waters invaded by bullhead (Braig & Johnson, 2003) was tested on the predator

efficiency of pike. The predatory success of pike was not affected by turbidity. This result

contrasts with previous studies that showed that turbidity reduced the feeding efficiency of

visual predators such as Micropterus salmoides Lacepède 1802 (Reid et al., 1999) and Perca

fluviatilis L. 1758 (Pekcan-Hekim & Lappalainen, 2006). However, these results parallel

those of Mauck & Coble (1971) on the independence between pike feeding efficiency and

water turbidity. Although the ability to detect prey by visual predators, such as pike, is

probably affected by turbidity, this may be compensated by an equivalent decrease of prey’s

ability to detect predators (Gregory, 1993).

Whatever the turbidity level, no significant effect of multiple predator treatment on the

number of remaining prey compared to no-predator control was observed. In other words pike

predatory efficiency was significantly reduced by the presence of bullhead. Three main

processes can account for this decrease in pike predation efficiency: (i) direct competition

between pike and bullhead for roach prey; (ii) an interaction other than competition between

roach and bullhead interfering with pike foraging success; and (iii) an interaction between

pike and bullhead reducing pike foraging success.

A direct competition between pike and bullhead for roach prey is unlikely as the

number of prey consumed by bullhead did not differ from the mortality of roach in the

absence of any predator. This means that bullhead fed little on roach in the experiments.

Although bullhead is considered as an opportunistic predator (Bruslé & Quignard, 2001), able

to prey on roach (Boët, 1980), bullhead predation was mainly directed towards dead or

injured fish lying on the bottom. It can therefore be considered that direct predation of

bullhead hardly affected roach abundance, and consequently that bullhead do not directly

11

compete with pike. That result is probably influenced by roach size and although using

smaller roach would probably increase the predatory success of bullhead, such a fish

combination would not have been realistic in regard to the size structure of wild roach

populations during the period selected to run the experiments (autumn). It also seems unlikely

that the reduction in pike predation efficiency was related to interactions between roach and

bullhead. Indeed, prey movement generally increases in the presence of multiple predator

species (e.g. Eklöv & VanKooten, 2001) leading to an increase in predator-prey encounter

rates, which therefore pushes prey to adopt riskier behaviour (Soluk & Collins, 1988;

Wissinger & Mc Grady, 1993). If this were the case, bullhead would have increased the

number of roach encounters with pike, and hence led to an increase in pike predation

efficiency.

Finally, the hypothesis of behavioural interference between pike and bullhead is the

most likely explanation of the reduction of pike predatory efficiency. According to Sih et al.

(1985), predatory species may interfere with each other, thus decreasing their combined

effects on prey populations. In this study a non-additive predation effect of pike and bullhead

on roach was detected, corresponding to a reduction of the predation risk for roach. Indeed,

pike predation tactics consists in a complex succession of behavioural components after prey

selection, which consists in a slow approach of the prey preceding attack, capture and

ingestion (Harper & Blake, 1990). Interference during this succession of behavioural phases

in pike feeding strongly reduces its foraging success (Nilsson et al., 2006). Because 1+

bullhead (i) have an activity peak during the day (Darnell & Meierotto, 1965) that

corresponds to the feeding period of pike (Bruslé & Quignard, 2001) and (ii) are known to

exhibit aggressive behaviour against all the species they encounter (e.g. Karp & Tyus, 1990),

the repeated nips of bullhead against pike (bullhead nips against pike were observed several

times each day) probably disturbed the foraging behaviour of the pike and led to a decrease in

12

their combined success through pike predation. Bullhead nips against pike were frequently

observed in this study, but were not quantified as only observable in CW experiments (in TW

experiments, water was not sufficiently clear to enable continuous behavioural observations).

No other disturbing behaviour by bullhead toward pike that may affect the results was

observed.

This study is the first to demonstrate a negative effect of the invasive black bullhead

on the predatory efficiency of pike through direct inter-species interaction that probably

occurs in the form of behavioural interference. Reducing predatory efficiency may affect pike

growth rate and/or survival as well as modify prey selection (Eklöv & Hamrin, 1989). The

results therefore confirm the potential negative impact of black bullhead on European native

fauna, and particularly on pike which is a top predator considered as vulnerable in some

European regions (e.g. Povž, 1996; Keith & Allardi, 2001). However, the strength of biotic

interactions is known to be influenced by environmental characteristics such as fish density,

structure of the environment or resource availability (Eklöv & VanKooten, 2001; Blanchet et

al., 2006). Although laboratory experiments cannot reproduce the complexity of the natural

environment, the experiments were designed to fit the environmental conditions found in

most European reservoirs and lakes. The autumn period was selected as it corresponds to a

low water period in most South European reservoirs and lakes due to water withdrawal for

agriculture and/or power generation (Brosse, 2000; Brosse et al., 2007). Hence fish density

increases a lot due to the drastic reduction of the water volume, increasing encounter rates

between fish. This is particularly true for bullhead that occurs in high biomass and densities in

most European lowland lakes (Boët, 1980; Cucherousset et al., 2006; Louette & Declerck,

2006). Moreover, the water level decrease leads to the disappearance of aquatic vegetation,

and hence strongly reduces habitat complexity. This means that the spatial fish assemblage

patterns known during summer no longer exist (Brosse, 2000; Brosse et al., 2007) and all fish

13

share the same habitat. Such a homogeneous environment as well as the high fish density is

consistent with the laboratory design. Finally, the sizes for the fish in this study are those

found during autumn in the natural environment. Nevertheless, the results based on laboratory

experiments need to be tested in natural environments to allow generalization. In addition,

behavioural observations would provide interesting insights into the interactions between pike

and bullhead. Combining field and laboratory results would enable management priorities to

be established based on the best scientific assessment of the impact of bullhead on pike

predatory efficiency, prey selection, growth and survival and hence on the structure of native

fish assemblages.

Acknowledgment

We are grateful to Simon Blanchet and to two anonymous referees for helpful comments on

the manuscript. This study was supported by the ANR "Freshwater fish diversity" (ANR -06-

BDIV-010, French Ministry of Research).

References

Blanchet, S., Dodson, J.J. & Brosse, S. (2006). Influence of habitat structure and fish density on Atlantic salmon Salmo salar L. territorial behaviour. Journal of Fish Biology 68, 951 – 957.

Blanchet, S., Loot, G., Grenouillet, G. & Brosse, S., (2007). Competitive interactions between native and exotic salmonids: a combined field and laboratory demonstration. Ecology of Freshwater Fish 16, 133-143.

Boët, P. (1980). L’alimentation du poisson-chat (Ictalurus melas Raf.) dans le lac de Créteil. Annales de Limnologie 16, 255 – 270.

Braig, E.C. & Johnson, D.L. (2003). Impact of black bullhead (Ameiurus melas) on turbidity in a diked wetland. Hydrobiologia 490, 11 – 21.

Brosse, S. (2000). Habitat, spatial dynamics and fish community structure in lake Pareloup (Aveyron, France). Cybium 24, 311-312.

Brosse, S., Grossman, G.D. & Lek, S. (2007). Fish assemblage patterns in the littoral zone of a European reservoir. Freshwater Biology 52, 448 - 458.

Bruslé, J. & Quignard, J.P. (2001). Biologie des poisons d’eau douce européens. Ed. Tec & Doc, Paris, France.

14

Casselman, J.M. & Lewis, C.A. (1996). Habitat requirements of northern pike (Esox lucius). Canadian Journal of Fisheries and Aquatic Sciences 53, 161 – 174.

Clavero, M. & Garcia Berthou, E. (2006). Homogenization dynamics and introduction routes of invasive freshwater fish in the Iberian Peninsula. Ecological Applications 16, 2313 – 2324.

Cucherousset, J., Paillisson, J.M., Carpentier A. & Chapman L.J. (2007). Fish emigration from temporary wetlands during drought: the role of physiological tolerance. Fundamental and Applied Limnology - Archiv für Hydrobiologie 168, 169 – 178.

Cucherousset, J., Paillisson, J.M., Carpentier A., Eybert, M.- C & Olden, J.D. (2006). Habitat use of an artificial wetland by the invasive catfish Ameiurus melas. Ecology of Freshwater Fish 15, 589 – 596.

Darnell, R.M. & Meierotto R.R. (1965). Diurnal periodicity in the black bullhead, Ictalurus melas (Rafinesque). Transactions of the American Fisheries Society 94, 1 – 8.

Declerck, S., Louette, G., De Bie, T. & De Meester, L. (2002). Patterns of diet overlap between populations of non-indigenous and native fishes in shallow ponds. Journal of Fish Biology 61, 1182 – 1197.

Eklöv, P. & Hamrin, S.F. (1989). Predatory efficiency and prey selection: interactions between pike Esox lucius, perch Perca fluviatilis and rudd Scardinus erythrophthalmus. Oikos 56, 149 – 156.

Eklöv, P. & VanKooten, T. (2001). Facilitation among piscivorous predators: effects of prey habitat use. Ecology 82, 2486 – 2494.

Elvira, B. (2001). Identification of non-native freshwater fishes established in Europe and assessment of their potential threats to the biological diversity. Convention on the conservation of European wildlife and natural habitats, 21st meeting Strasbourg, 26 – 30 November 2001. 35 p.

Gregory, R.S. (1993). The effect of turbidity on the predator avoidance behaviour of juvenile Chinook salmon (Oncoryhnchus tshawytscha). Canadian Journal of Fisheries and Aquatic Sciences 50, 241 – 246.

Griffen, B.D. (2006). Detecting emergent effects of multiple predator species. Oecologia 148, 702 – 709.

Harper, D.G. & Blake, R.W. (1990). Prey capture and the fast –performance of northern pike Esox lucius. Journal of Experimental Biology 155, 175 – 192.

Hart, P.J.B. & Connellan, B. (1984). Cost of prey capture, growth rate and ration size in pike, Esox lucius L., as functions of prey weight. Journal of Fish Biology 25, 279 – 292.

Holčík, J. (1991). Fish introductions in Europe with particular reference to its central and eastern part. Canadian Journal of Fisheries and Aquatic Sciences 48, 13 – 23.

Johnsson, J.I., Höjesjö, J. & Fleming, I.A (2001). Behavioural and heart rate responses to predation risk in wild and domesticated Atlantic salmon. Canadian Journal of Fisheries and Aquatic Sciences 58, 788 – 794.

Karp, C.A. & Tyus H.M. (1990). Behavioral interactions between young Colorado squawfish and six fish species. Copeia 1, 25 – 34.

Keith, P. & Allardi, J. (2001) Atlas des poissons d'eau douce de France. Patrimoines Naturels 47, 1 – 387.

Leprieur, F., Beauchard, O., Hugueny, B., Grenouillet, G. & Brosse, S. (2008). Null model of biotic homogenization: a test with the European freshwater fish fauna. Diversity & Distributions In press. doi: 10.1111/j.1472-4642.2007.00409.x

Louette, G. & Declerck, S. (2006) Assessment and control of non-indigenous brown bullhead Ameiurus nebulosus populations using fyke nets in shallow ponds Journal of Fish Biology 68, 522-531.

15

Mack, R.N., Simberloff, D., Lonsdale, W.M., Evans, H., Clout, M. & Bazzaz, F.A. (2000). Biotic invasions: causes, epidemiology, global consequences, and control. Ecological Applications 10, 689 – 710.

Mauck, W.L. & Coble, D.W. (1971). Vulnerability of some fishes to northern pike (Esox lucius) predation. Journal of the Fisheries Research Board of Canada 28, 957 – 969.

Nilsson, P.A., Turesson, H. & Brönmark, C. (2006). Friends and foes in foraging: intraspecific interactions act on foraging-cycle stages. Behaviour 143, 733 – 745.

Pekcan-Hekim, Z. & Lappalainen, J. (2006). Effects of clay turbidity and density of pikeperch (Sander lucioperca) larvae on predation by perch (Perca fluviatilis). Naturwissenschaften 93, 356 – 359.

Povž, M. 1996. The red data list of the freshwater lampreys (Cyclostoma) and fish (Pisces) of Slovenia. In Conservation of Endangered Freshwater Fish in Europe (Kirchhofer, A. & Hefti, D. eds), pp 63 – 72. Basel: Birkhäuser verlag.

Reid, S.M., Fox, M.G. & Whillans T.H. (1999). Influence of turbidity on piscivory in largemouth bass (Micropterus salmoides). Canadian Journal of Fisheries and Aquatic Sciences 56, 1362 – 1369.

Sih, A., Crowley, P., McPeek, M., Petranka, J., & Strohmeier, K. (1985). Predation, competition, and prey communities. Annual Review of Ecology and Systematics 16, 269 – 311.

Sih, A., Englund, G. & Wooster, D. (1998). Emergent impacts of multiple predators on prey. Trends in Ecology and Evolution 13, 350 – 355.

Soluk, D.A. & Collins, N.C. (1988). Synergistic interactions between fish and invertebrate predators: facilitation and interference among stream predators. Oikos 52, 94 – 100.

Taylor, J.N., Courtenay, W.R. & McCann, J.A. (1984). Known impacts of exotic fishes in the continental United States. In Distribution, Biology, and Management of Exotic Fishes (Courteney, W.R. & Stauffer, J.R., eds), pp 322 – 374. Baltimore MD: John Hopkins University Press.

Townsend, C.R. (2003). Individual, population, community and ecosystem consequences of a fish invader in New Zealand streams. Conservation Biology 17, 38 – 17.

Utne-Palm, A.C. (2002). Visual feeding of fish in a turbid environment: physical and behavioural aspects. Marine and Freshwater Behaviour and Physiology 35, 111 – 128.

Vitousek, P.M., Mooney, H.A., Lubchenco, J. & Mellilo, J.M. (1997). Human domination of Earth’s ecosystems. Science 277, 494 – 499.

Welcomme, R.L. (1988). International introductions of inland aquatic species. FAO Fisheries Technical Paper 294.

Wissinger, S. & Mc Grady, J. (1993). Intraguild predation and competition between larval dragonflies: direct and indirect effects on shared prey. Ecology 74, 207 – 218.


P3

Null model of biotic homogenization: a test with the European freshwater fish fauna

Leprieur F., Beauchard O., Hugueny B., Grenouillet G. & Brosse, S.

Diversity and Distribution (sous presse)

© 2007 The Authors DOI: 10.1111/j.1472-4642.2007.00409.xJournal compilation © 2007 Blackwell Publishing Ltd www.blackwellpublishing.com/ddi

1

Diversity and Distributions, (Diversity Distrib.)

(2007)

BIODIVERSITYRESEARCH

ABSTRACT

In recent years, there has been growing concern about how species invasions andextinctions could change the distinctiveness of formerly disparate fauna andflora, a process called biotic homogenization. In the present study, a null model ofbiotic of homogenization was developed and applied to the European freshwater fishfauna. We found that non-native fish species led to the greatest homogenization insouth-western Europe and greatest differentiation in north-eastern Europe. Comparingthese observed patterns to those expected by our null model empirically demonstratedthat biotic homogenization is a non-random ecological pattern, providing evidencefor previous assumptions. The place of origin of non-native species was also consideredby distinguishing between exotic (originating from outside Europe) and translocatedspecies (originating from within Europe). We showed that exotic and translocatedspecies generated distinct geographical patterns of biotic homogenization acrossEurope because of their contrasting effects on the changes in community similarityamong river basins. Translocated species promoted homogenization among basins,whereas exotic species tended to decrease their compositional similarity. Quantifyingthe individual effect of exotic and translocated species is therefore an absoluteprerequisite to accurately assess the spatial dynamics of biotic homogenization.

Keywords

biotic homogenization, exotic species, freshwater fish, null model, translocated

species.

INTRODUCTION

The introduction of non-native species and the extinction of

native species have together caused loss of taxonomic regional

distinctiveness among formerly disparate faunas and floras

(reviewed by Olden & Rooney, 2006). This decrease in beta-

diversity, also called biotic homogenization (BH) by McKinney

& Lockwood (1999), is expected to have important evolutionary

and ecological consequences (Olden

et al

., 2004). BH is now an

important research agenda for ecologists as it represents a process

including both species invasions and extirpations, two key

components of the modern biodiversity crisis (Olden, 2006).

Three distinct forms of BH (genetic, taxonomic and

functional) were defined by Olden

et al

. (2004). Among them,

taxonomic homogenization (TH) has been empirically studied

for various taxonomic groups (reviewed by Olden, 2006), and

has been explicitly formalized by Olden & Poff (2003, 2004) into

a mechanistic model incorporating scenarios of invasion and

extinction. These scenarios show how species invasions and/or

extinctions can lead to TH (i.e. increase in community similarity)

or to taxonomic differentiation (i.e. decrease in community

similarity). Although TH patterns have been commonly related

to environmental and human factors in homogenization studies

(e.g. Marchetti

et al

., 2001; Rooney

et al

., 2004; Olden

et al

.,

2006), few have tested whether TH is geographically structured

(Smith, 2006). In addition, none of these studies addressed

whether the place of origin of non-native species influenced TH

patterns. Recently, McKinney (2005) found that translocated

species (i.e. species introduced within their native biogeographical

zone in localities where they did not historically occur) have a

greater homogenization effect than exotic species (i.e. species

originating from another biogeographical area). However, tests

of this assumption are scarce (La Sorte & McKinney, 2005), and

to our knowledge the joint and individual effects of translocated

and exotic species on TH have never been compared.

In this context, this study aims (i) to identify the relative roles

of exotic and translocated species in driving TH patterns and

(ii) to test whether these patterns are geographically structured.

We explored fish homogenization and differentiation in the 25

major European river basins as extended information is available

on native and non-native fish species in these basins. Moreover,

almost all homogenization studies have been conducted in North

1

Laboratoire « Evolution and Diversité

Biologique », UMR 5174, CNRS – Université

Paul Sabatier, 118 route de Narbonne,

F-31062 Toulouse cedex 4, France,

2

Ecosystem

Management Research Group, Department of

Biology, Faculty of Sciences, University of

Antwerp, Universiteitsplein 1, BE-2610

Antwerpen (Wilrijk), Belgium,

3

Laboratoire

d’Ecologie des Hydrosystèmes Fluviaux,

Université Claude Bernard, 43 Bd. du 11

novembre 1918, 69622 Villeurbanne cedex 05,

France

*Correspondence: Fabien Leprieur, Laboratoire « Evolution and Diversité Biologique », UMR 5174, CNRS – Université Paul Sabatier, 118 route de Narbonne, F-31062 Toulouse cedex 4, France. Tel. 00-33-5-61-55-67-47; E-mail: [email protected]

Blackwell Publishing Ltd

Null model of biotic homogenization: a test with the European freshwater fish fauna

F. Leprieur

1

*, O. Beauchard

2

, B. Hugueny

3

, G. Grenouillet

1

and S. Brosse

1

F. Leprieur

et al.

© 2007 The Authors

2


, Journal compilation © 2007 Blackwell Publishing Ltd

America, and homogenization processes need evidence from

other continents to allow general validation. To quantify TH, we

applied the quantitative framework of Olden & Poff (2003, 2004)

and among the 14 scenarios proposed by these authors, we tested

those without extinction events. Indeed, these scenarios

correspond to the European situation as no basin-scale extinctions

among non-migratory freshwater fish were reported in large

European river basins (e.g. Keith & Allardi, 2001; Clavero &

García-Berthou, 2006).

Previous studies exploring BH assumed that non-native species

were not randomly distributed across localities as species

introductions are primarily related to human purposes

(Blackburn & Duncan, 2001; Jeschke & Strayer, 2006). This led to

the assumption that BH is a non-random ecological pattern (e.g.

Duncan & Lockwood, 2001; Olden

et al

., 2004). However, these

studies did not compare observed patterns to those expected

under a null hypothesis. Then, the quantification of TH is

commonly based on similarity indices, such as the Jaccard

index (Olden & Rooney, 2006), that are notoriously difficult to

interpret without knowing their expected values under null

hypotheses of random distribution of species among localities

(Henderson & Heron, 1977; Connor & Simberlorff, 1978; Raup

& Crick, 1979). Last, the change in community similarity among

pairwise localities is negatively related to the initial similarity

(Olden & Poff, 2004), preventing direct comparison between

pairwise values. In this context, we generated random assemblages

of non-native species using Monte Carlo simulations to compare

observed TH patterns with regard to null models. These models

are widely used to test specific hypotheses about patterns in

nature by creating artificial data sets that could be expected if

a given null hypothesis is true (e.g. Gotelli, 2000). Compared to

other modelling approaches, a null model deliberately excludes

the mechanism on interest being tested. In this study, we expect

that if TH is generated by a non-random distribution of non-

native species, observed TH patterns should differ from

those expected by chance alone. Particularly, if the slope of the

relationship between the change in community similarity and

the initial similarity among basins is lower than predicted by the

null model, this would imply that basins that are initially similar

in their species composition are more likely than expected by

chance to be invaded by the same species.

METHODS

Data sources

Freshwater fish occurrences were compiled from published data

on the major European river basins. We selected the 25 basins

(Fig. 1a) for which sufficient information on the fish fauna is

available at the basin scale (see Appendix S1 in Supplementary

Material for a full list of references). This avoided potential bias

in our analysis due to incomplete surveys of both native and

non-native fish species. For each basin, we distinguished three

categories of species: natives, exotics (i.e. species originating

from outside Europe) and translocated (i.e. species native to

Europe introduced into drainages where they did not historically

occur). We considered as non-native a species with self-

reproducing populations or populations artificially maintained

by regular and long-term restocking. Only strictly freshwater fish

were considered because (i) migratory and brackish species

would introduce potential bias in the analysis as we considered

each basin as a biogeographical island, and (ii) information

availability on the distribution of migratory and brackish species

is much more limited than for resident species.

Quantifying homogenization/differentiation

For each pair of basins (

n

= 300), we calculated the Jaccard

similarity index corresponding to two time situations, i.e.

initial (

J

initial

) and final (

J

final

) situations. This similarity index is

commonly applied in homogenization studies dealing with

presence/absence matrices (Olden & Rooney, 2006). The initial

situation only included native species that represented the

historical pool of species (Olden & Rooney, 2006). On the contrary,

the final situation included native species plus non-native species

(i.e. the contemporary species pool). TH was quantified from

differences in Jaccard index (expressed as percentage) for each

pair of basins between the final and initial situations (i.e. the

change in similarity among pairwise basins: Pairwise

∆

CS; Rahel

2000; Olden & Poff, 2003). Positive values of Pairwise

∆

CS indicate

a TH among pairwise basins, whereas negative values indicate

a taxonomic differentiation among pairwise basins.

We explored TH patterns using different metrics and

quantitative representations. First, we investigated the relation-

ship between Pairwise

∆

CS and the initial similarity among

basins (i.e. initial situation,

J

initial

). As geographical distance and

species similarity are generally inversely related at large spatial

scales (Nekola & White, 1999), we also considered

J

initial

as a

surrogate of geographical distance among basins (e.g. Reyjol

et al

., 2006). The relationship between Pairwise

∆

CS and

J

initial

permitted us to relate observed patterns to the prediction of

the two invasion-only scenarios of Olden & Poff (2003, 2004):

(scenario I1) the same species invade driving TH; (scenario I2)

different species invade driving taxonomic differentiation.

Second, to understand how each basin changed relative to all

others in Europe, we computed the average of Pairwise

∆

CS

between each basin and the 24 other basins (i.e. Basin

∆

CS or

rate of homogenization/differentiation per basin). We then

mapped the Basin

∆

CS to explore geographical patterns of

TH across Europe and applied a Mantel’s test to assess whether

the Basin

∆

CS was geographically structured (i.e. to determine if

the basins that are close together had more similar rates of

homogenization/differentiation than distant ones). The Mantel’s

test consists of testing the correlation between two distance

matrices using a randomization procedure (10,000 permutations;

see Legendre & Legendre, 1998). We used the Euclidean distance

to compute the matrices of (i) geographical distance (based on

mean latitude and longitude of each basin) and (ii) Basin

∆

CS

distance (based on the rates homogenization/differentiation per

basin). Last, we quantified the continental level of homogenization

or differentiation in Europe by averaging the 300 Pairwise

∆

CS

(Continental

∆

CS).

Null model of biotic homogenization

© 2007 The Authors



3

Null model of homogenization/differentiation

Monte Carlo simulations were developed to generate 10,000

matrices of Pairwise

∆

CS expected by chance alone (i.e. Pairwise

∆

CS generated by a random distribution of non-native species).

We first tested whether the observed values of Basin

∆

CS and

Continental

∆

CS were generated by a non-random distribution of

non-native species. Observed values of Basin

∆

CS and Continental

∆

CS were compared to their null distributions that derived from

the simulated Pairwise

∆

CS (two-tailed test,

α

= 5%). Then, we

determined whether the relationship observed between Pairwise

∆

CS and

J

initial

differed from those expected by chance alone by

calculating the regression parameters (i.e. slope and intercept)

from the 10,000 simulations and comparing the observed values

to their null distributions (two-tailed test,

α

= 5%).

Our Monte Carlo simulations consisted in randomly assigning

each non-native fish species into the 25 basins (translocated

species were randomly assigned only in basins where they do not

naturally occur). We applied a fixed-equiprobable algorithm

(Gotelli, 2000) to generate the random matrices of non-native

species occurrences. This algorithm implied that occurrences of

non-native species were conserved as in the original matrix (i.e.

the number of basins in which each non-native species occurs is

fixed), whereas the total number of non-native species per basin

was allowed to vary randomly (i.e. columns equiprobable).

Non-native species occurrences were maintained constant

during simulations to account for interspecific differences in

colonization ability and/or human induced propagule pressure.

An equiprobable total of columns means that (i) all the basins are

equiprobably sustainable for all the non-native species, and

(ii) the non-native species are distributed randomly among the

basins as all of them can colonize all the basins. According to

Gotelli (2000), the fixed-equiprobable algorithm is efficient

to avoid type I and II errors concerning statistically significant

patterns for a random matrix. When selecting a null model, every

feature of the randomized data would be preserved as in the

observed data, except the feature that the study aims to test

(Tokeshi, 1986). This ensures that the model does not become

biologically too unrealistic. In our null model, we did not

maintain the number of non-native species per basin constant

as in the original matrix because it is well accepted that most

communities in nature are not saturated (e.g. Hugueny & Paugy,

1995; Smith & Shurin, 2006). This means that all communities

may be susceptible to invasion by non-native species regardless

of native species richness (e.g. Moyle & Light, 1996; Gido &

Brown, 1999; Smith & Shurin, 2006). Then, Olden & Rooney

(2006) argued that BH should not be systematically confused

with patterns of species invasions (i.e. number of invaders) as is

commonly done in the literature. Therefore, allowing the

number of non-native species to vary in each basin permitted

Figure 1 Geographical distribution of the rates of homogenization/differentiation per basin (Basin ∆CS;%). (a) Map of the 25 major European river basins (1: Guadalquivir; 2: Tagus; 3: Douro; 4: Ebro; 5: Garonne; 6: Loire; 7: Seine; 8: Rhône; 9: Pô; 10: Rhine; 11: Weser; 12: Elbe; 13: Oder; 14: Wisla; 15: Danube; 16: Dniestr; 17: Dniepr; 18: Don; 19: Volga; 20: Ural; 21: Petchora; 22: Dniva; 23: Neva; 24: Kemijoki; 25: Glomma). (b) Basin ∆CS (%) based on non-native species (both exotic and translocated species). (c) Basin ∆CS (%) based on exotic species alone. (d) Basin ∆CS (%) based on translocated species alone. The graduation of green indicates a taxonomic differentiation and the graduation of red indicates a taxonomic homogenization. See Methods for more details.

F. Leprieur

et al.

© 2007 The Authors

4



us to test the null hypothesis that the rates of homogenization/

differentiation per basin (Basin

∆

CS) were dependent of the

number of non-natives in each basin. We quantified TH and

applied our null model by first considering overall non-native

species (exotic and translocated species were pooled) and then by

distinguishing between exotic and translocated species. The null

model program was computed by the authors with the open

source

software (Ihaka & Gentleman, 1996).

RESULTS

General trends in native and non-native species richness

The native non-migratory freshwater fish fauna of the 25 major

European river basins was composed of 136 species. We

identified 38 exotic species and 40 translocated species with

a large variation in species richness between basins (Table 1).

Native species richness was independent of the number of

non-native species (Pearson’s correlation:

r

= 0.151,

P

= 0.236).

However, when distinguishing between exotic and translocated

species, native and exotic species richness was positively correlated

(Pearson’s correlation:

r

= 0.575,

P

= 0.001), whereas native

and translocated species richness was negatively correlated

(Pearson’s correlation: r = –0.449,

P

= 0.012). This relationship

was strongly influenced by the basin area, as when controlling for

this variable with partial regressions, these correlations became

marginal (partial Pearson’s correlation:

r

= 0.406,

P

= 0.049) or

non-significant (partial Pearson’s correlation:

r

= –0.321,

P

= 0.126) for exotic and translocated species, respectively.

Pairwise change in community similarity (Pairwise

∆∆∆∆

CS)

Pairwise

∆

CS and the initial similarity among basins displayed

a negative linear relationship as predicted by the invasion-only

scenarios of Olden & Poff (2004) (Fig. 2). When first analysing

the joint effect of exotic and translocated species, both homoge-

nization (i.e. 60% of Pairwise

∆

CS > 0%) and differentiation

(i.e. 40% of Pairwise

∆

CS < 0%) among basins were observed

(Fig. 2a). Then, when analysing the effect of exotic and trans-

located species separately, we noticed (i) a general trend of

differentiation among basins for exotic species (i.e. 75% of

Pairwise

∆

CS < 0%), except for basins sharing few native species

(i.e. low

J

initial

) that became more similar (Fig. 2b), and (ii) a

general trend of homogenization among basins (i.e. 93% of

Pairwise

∆

CS > 0%) within the entire range of

J

initial

for trans-

located species (Fig. 2c). The slopes of these relationships were

less steep than those expected by the null model (two-tailed test,

P

= 0.000, Fig. 2). The intercepts were significantly lower than

those expected by the null model (two-tailed test,

P

= 0.000).

Fish homogenization/differentiation in Europe (Continental

∆∆∆∆

CS, Basin

∆∆∆∆

CS)

A continental level of homogenization was observed when

analysing the joint effect of exotic and translocated species

(Continental

∆

CS = 2.2%,

n

= 300), which was greater than

expected by the null model (two-tailed test,

P

< 0.0001). A

general trend of homogenization was also observed at the

basin scale (i.e. 17 basins out of 25; Fig. 1b). The Basin

∆

CS were

spatially autocorrelated (Mantel test,

r

= 0.357,

P

= 0.006) and

differed from those expected by the null model in nine basins

(Table 2).

Then, analysing exotic species alone revealed a continental

level of differentiation (Continental

∆

CS = –1.6%,

n

= 300),

which was lower than expected by the null model (two-tailed

test,

P

= 0.0003). A general trend of differentiation was also

observed at the basin scale (i.e. 19 out of 25 basins, Fig. 1c). The

Basin

∆

CS

were spatially autocorrelated (Mantel test,

r

= 0.3218,

P

< 0.001) and differed from those expected by the null model

in most basins (i.e. 20 out of 25, Table 2). No significant linear

relationship was established between the observed Basin

∆

CS

and the number of exotic species per basin (

R

2

= 0.015,

P

= 0.565).

Contrary to exotic species, translocated species have led to

a continental level of homogenization (Continental

∆

CS = 5% in

average,

n

= 300), which was greater than expected by the null

model (two-tailed test,

P

< 0.0001). Fish homogenization was

also recorded at the basin scale (Fig. 1d). The Basin

∆

CS were not

Table 1 Number of native, exotic and translocated freshwater fish in the 25 major European river basins. The basin numbers (Code) are those used in Fig. 1a.

Code Basin Native Exotic Translocated

1 Guadalquivir 12 5 5

2 Tagus 18 6 7

3 Douro 13 8 7

4 Ebro 19 8 11

5 Garonne 18 11 13

6 Loire 21 9 11

7 Seine 22 8 10

8 Rhône 31 11 9

9 Pô 28 11 10

10 Rhine 31 16 11

11 Weser 29 7 7

12 Elbe 34 8 4

13 Oder 39 12 5

14 Wisla 31 11 3

15 Danube 67 18 2

16 Dniestr 48 12 0

17 Dniepr 49 10 0

18 Don 45 13 1

19 Volga 51 16 2

20 Ural 35 4 0

21 Petchora 20 0 0

22 Dvina 25 1 0

23 Neva 35 0 1

24 Kemijoki 17 2 0

25 Glomma 17 2 1

[Correction added after online publication 28 August 2007: values ofNative, Exotic and Translocated freshwater fish are corrected as above].


© 2007 The Authors



5

spatially autocorrelated (Mantel test,

r

= 0.107,

P

= 0.15) and

differed from those expected by the null model in only four

basins (Table 2). A significant linear relationship was established

between the observed Basin

∆

CS and the number of translocated

species per basin (

R

2

= 0.6441,

P

< 0.0001).

DISCUSSION

As mentioned by Schoener (1987), although an ecological

pattern might be statistically significant, its features may not

differ significantly from the output of a null model. To our

knowledge, this is the first study that aimed to test whether the

observed TH patterns differed from those expected by a null

model. To generate null assemblages of non-native species, we

did not maintain constant the number of non-native species as in

the original matrix. We considered that each basin was equivalent

in its susceptibility to invasion independently of the number of

native species present. As expected, we did not observe a strong

relationship between the number of native species and the

number of exotic and translocated species, respectively, when

controlling for the basin area. This confirms that the positive

correlation between the number of native and non-native species

that is commonly observed on large spatial scales, may be related

to covarying factors (e.g. Davies

et al

., 2005).

The successive introductions of non-native fish species (i.e.

the joint effect of exotic and translocated species) increased on

Figure 2 Change in community similarity (Pairwise ∆CS,%) of the freshwater fish fauna among 300 pairwise comparisons of the 25 major European river basins in relation with their initial similarity (Jinitial,%). Solid black lines represent observed relationships and dashed grey lines represent the average simulated relationship (n = 10,000). (a) Non-native species (both exotic and translocated species), observed relationship: Pairwise ∆CS = –0.21Jintial + 0.06, average simulated relationship: Pairwise ∆CS = –0.37Jintial + 0.08. (b) Exotic species alone, observed relationship Pairwise: ∆CS = –0.19; Jinitial + 0.023; average simulated relationship: Pairwise ∆CS = –0.26Jintial + 0.03. (c) Translocated species alone, observed relationship: Pairwise ∆CS = –0.04 Jinitial + 0.055; average simulated relationship: Pairwise ∆CS = –0.18Jintial + 0.08.

F. Leprieur et al.

© 2007 The Authors6 Diversity and Distributions, Journal compilation © 2007 Blackwell Publishing Ltd

average the taxonomic similarity among the 25 major European

basins (Continental ∆CS = 2.2%), which is consistent with other

empirical case studies analysing TH at the regional and continental

scales as reviewed by Olden (2006). This continental level of

homogenization was significantly greater than those expected

under the null hypothesis, indicating that random assemblages

of non-native species have a higher differentiation effect than

actually observed. This was predictable as our null model

allowed all non-native species to colonize all the basins. Overall,

these results indicate that fish homogenization in Europe was not

random in regards to the geographical distribution of both exotic

and translocated species. This finding is supported by previous

studies suggesting that the geographical distribution of non-

native species was not random due to (i) differences or similarities

in human-selected species and propagule pressure (e.g. Blackburn

& Duncan, 2001); (ii) dispersal abilities and environmental

tolerances of the introduced species (e.g. Kennard et al., 2005);

and (iii) the environmental and biological attributes of the

recipient region (i.e. climate, human-modified habitats and

biotic resistance, e.g. Moyle & Marchetti, 2006). Several studies

clearly point out intentional human activities (e.g. angling,

aquaculture, biological control) as being the main determinants

of fish introductions in European states (e.g. Vooren, 1972;

Holcík, 1991). For example, Gambusia affinis Baird & Girard and

Gambusia holbrooki Girard were mainly introduced in southern

Europe for mosquito control (Keith & Allardi, 2001; Doadrio,

2002). Similarly, exotic fish assemblages were spatially structured

along a latitudinal gradient in the Iberian Peninsula, with species

related to sport fishing being characteristic of northern basins

(Clavero & García-Berthou, 2006).

Distinguishing between exotic and translocated species

revealed that translocated species generated a higher continental

level of homogenization (Continental ∆CS = 5%) than overall

non-native species. Indeed, exotic species decreased, on

average, the taxonomic similarity among basins (Continental

∆CS = –1.6%), counteracting therefore the homogenization

effect of translocated species. These opposite effects of exotic and

translocated species on TH were clearly distinguished when

Table 2 Observed rates of homogenization/differentiation per basin (Basin ∆CS percentage) compared to those expected by the null model (two-tailed test). The results are indicated for non-native species (i.e. exotic and translocated species were pooled) and for exotic and translocated species alone, respectively. The alphabetical codes in parentheses correspond to the conclusion of a two-tailed test: (HS) observed rates of homogenization are significantly smaller than expected by the null model; (HG) observed rates of homogenization are significantly greater than expected by the null model; (DS) observed rates of differentiation are significantly smaller than expected by the null model; (DG) observed rates of differentiation are significantly greater than expected by the null model. No indications in parentheses mean that observed rates of homogenization/differentiation did not differ from those expected by the null model. *P < 0.001 (Bonferroni correction), ns = non-significant. The basin numbers (Code) are those used in Fig. 1a.

Code Basin

Basin ∆CS (%)

Non-native species Exotic species Translocated species

1 Guadalquivir 3.55 ns 1.52 (HG) * 2.77 ns

2 Tagus 4.95 ns 2.24 (HG) * 3.68 ns

3 Douro 6.19 ns 2.44 (HG) * 5.14 ns

4 Ebro 8.98 ns 2.27 (HG) * 8.66 ns

5 Garonne 9.08 (HG) * 0.34 (HG) * 12.28 (HG) *

6 Loire 4.64 (HG) * –2.10 (DS) * 9.08 (HG) *

7 Seine 3.53 (HG) * –2.43 (DS) * 7.76 ns

8 Rhône 3.56 (HG) * –1.23 (DS) * 6.26 ns

9 Pô 6.61 (HG) * 1.69 (HG) * 6.4 ns

10 Rhine 0.21 ns –3.80 (DS) * 5.24 ns

11 Weser –0.91 ns –3.19 ns 2.59 ns

12 Elbe 1.09 (HG) * –3.04 (DS) * 5.42 ns

13 Oder 1.91 (HG) * –2.19 (DS) * 5.35 (HG) *

14 Wistula –0.84 ns –4.51 ns 4.36 ns

15 Danube 2.07 (HG) * –0.62 (DS) * 3.41 (HG) *

16 Dnestr 1.02 ns –2.09 ns 4.14 ns

17 Dnepr 1.62 ns –1.61 (DS) * 4.09 ns

18 Don 0.37 ns –2.97 ns 4.62 ns

19 Volga –1.80 (DG) * –4.41 (DG) * 3.7 ns

20 Ural –0.01 ns –3.21 (DG) * 3.84 ns

21 Petchora –0.49 ns –2.28 (DG) * 1.91 ns

22 Dvina –1.23 ns –3.66 (DG) * 2.88 ns

23 Neva –0.73 ns –4.51 (DG) * 4.57 ns

24 Kemijoki –1.20 ns –2.00 ns 0.68 ns

25 Glomma 3.77 ns –1.84 (DG) * 6.47 ns


© 2007 The AuthorsDiversity and Distributions, Journal compilation © 2007 Blackwell Publishing Ltd 7

plotting the pairwise change in community similarity against the

initial similarity among basins (Fig. 2). Translocated species

produced homogenization among both neighbouring basins

(high Jinitial) and distant ones (low Jinitial), whereas exotic species had

an overall differentiation effect (i.e. decreased the compositional

similarity among basins with an initial similarity ranging from

0.1 to 0.6). These results support the predictions of the two

invasion-only scenarios of Olden & Poff (2004). Then, they are

consistent with recent works of La Sorte & McKinney (2005),

suggesting that differences in homogenization effect between

exotic and translocated species may be related to their differences

in geographical distribution patterns. Indeed, different sets of

exotic species were introduced in different sets of neighbouring

basins, i.e. 4.8 basins per species on average (e.g. Lepomis gibbosus

Linnaeus, Ameirus melas Rafinesque in western Europe and

Perccottus glenii Dybowski, Mylopharyngodon piceus Richardson

in eastern Europe), which led to an overall decrease in species

similarity among basins (scenario I2, Olden & Poff, 2004). In

contrast, most translocated fish species in Europe are native to

eastern basins (e.g. Sander lucioperca Linnaeus, Silurus glanis

Linnaeus) and were widely introduced in less speciose drainages

of western and southern Europe (Keith & Allardi, 2001; Doadrio,

2002). This led to an increase in the size of their geographical

ranges (i.e. from an average historical range of 8.3 basins per

species to a current range of 12.3) and hence in species similarity

among basins (scenario I1, Olden & Poff, 2004). Such a pattern

of homogenization has also been reported in the USA, where

increased similarity among states is partly due to the expansion

of cosmopolitan US fish from eastern to western basins (Fuller

et al., 1999; Rahel, 2000).

The fact that the observed slope of the relationship between

Pairwise ∆CS and Jinitial was lower than predicted by the null

model means that basins that are initially similar in their species

composition are more likely than expected by chance to be

invaded by the same species. Many factors may lead to this

pattern but two are probably important: (i) environmental filters

(e.g. Mediterranean basins are likely to be naturally inhabited

and invaded by drought resistant species) and (ii) geographical

structure in introduction pathways (i.e. neighbouring basins

having high initial faunal similarity are likely to receive similar

non-native species, see Clavero & García-Berthou, 2006). While

in both cases (translocated and exotics), the observed slope was

significantly lower than those expected under the null model, the

deviation from the null expectation was much more pronounced

for translocated species. Indeed, contrary to exotic species, the

observed distribution of translocated species is mainly asymmetric

(i.e. introduction pathway from eastern to western Europe). In

contrast, our simulations generated symmetrical distribution

patterns by allowing all the translocated species to colonize

all the basins where they did not naturally occur (i.e. in both

western and eastern Europe). This produced therefore different

‘null subsets’ of translocated species across Europe and hence an

overall decrease in community similarity among initially similar

basins (see Rahel, 2002; Olden & Poff, 2004).

Although TH is commonly presented as the average change in

community similarity among regions of a given biogeographical

area (e.g. Rahel, 2000; Taylor, 2004; Olden, 2006), complementary

information can be obtained by quantifying TH at the regional

scale (as expressed by Basin ∆CS). Indeed, when considering

both exotic and translocated species, our spatial autocorrelation

analysis revealed that neighbouring basins tend to display more

similar rates of homogenization/differentiation than distant

basins. This results in greatest rates of homogenization for

south-western basins and greatest rates of differentiation for

north-eastern basins (see Fig. 1b). Such a latitudinal pattern of

TH is consistent with the introduction pathways of fish species in

Europe, recently analysed by García-Berthou et al. (2005). These

authors reported a higher ratio of received to given introductions

in southern countries and a lower one in northern countries.

Particularly, numerous non-European and European fish species

were introduced from France to the Iberian Peninsula (García-

Berthou et al., 2005; Clavero & García-Berthou, 2006), which

differs from the rest of Europe by its low number of native species

and high level of fish endemism (Doadrio, 2002). This explains

why French and Iberian basins experienced similar changes in

their species composition (i.e. homogenization, see Fig. 1b).

Similarly, a significant latitudinal gradient of TH was observed

when analysing the individual effect of exotic species (see

Fig. 1c), except that fish fauna homogenization occurred only in

southern Europe (i.e. the Garonne and Pô river basins and the

basins of the Iberian Peninsula). Overall, the European river

basins were homogenized or differentiated independently of the

number of exotic species as almost all of the observed rates of

homogenization/differentiation per basin differed from those

expected by the null model. This was confirmed by the lack of a

significant linear relationship between the observed rates of

homogenization/differentiation per basin and the number of

exotic species. In contrast, our spatial autocorrelation analysis

revealed that the homogenization pattern resulting from

translocated species did not match a latitudinal gradient as it

did for exotics (Fig. 1d). However, we noticed that the nearby

Ebro, Pô, and French river basins displayed the highest rates of

homogenization. These basins are also characterized by numerous

translocated species (9.8 species on average, n = 6) compared to

other basins (2.4 species on average, n = 19). This suggests that

the number of translocated species strongly influenced the rate

of homogenization in each basin. Indeed, we found that (i)

almost none of the rates of homogenization per basin differed

from those expected by the null model and (ii) the rate of

homogenization per basin was significantly related to the

number of translocated species. Although we showed that

the rate of homogenization in a given basin could be accurately

predicted by the number of translocated species, we do not

encourage future studies to explore geographical patterns of TH

only on the basis of the number of invaders. Indeed, we clearly

demonstrated that for exotics, this conclusion was not accurate.

This implies that tracing the identity of species (and not the

number of species) is a fundamental prerequisite for quantifying

the changes in community similarity among localities (e.g. Olden

& Rooney, 2006; Qian & Ricklefs, 2006; Smart et al., 2006).

Overall, our results are consistent with previous studies

exploring patterns of BH over large spatial scales (i.e. region,

F. Leprieur et al.


continent), especially for plants (e.g. Rejmánek, 2000; Schwartz

et al., 2006). Indeed, we showed that non-native fish species

produced differentiation among neighbouring basins (i.e. with

high initial similarity) and homogenization among distant ones

(i.e. with low initial similarity). According to Marchetti et al.

(2001), these opposite patterns (homogenization vs. differen-

tiation) can be also explained by the spatial scale of the study.

Here we highlight that when considering a single spatial scale,

distinguishing between exotic and translocated species produced

opposite patterns whatever the distance is between basins.

When discussing BH, the place of origin of non-native species

constitutes therefore a crucial consideration, together with

distance between sites and spatial scale.

The overall differentiation effect of exotic freshwater fish in

Europe contrasts with that recently reported on plants by Qian &

Ricklefs (2006). Indeed, exotic plant species (i.e. originating

from outside North America) introduced in North American

provinces had an overall homogenization effect, due to a lower

spatial turnover rate than the natives. Such a difference between

our results and those of Qian & Ricklefs (2006) can be related to

intrinsic differences between fish and plants. Indeed, freshwater

fish can hardly disperse from one basin to another without

human transports as river basins are separated by barriers

insurmountable for fish. In contrast, plants can naturally

disperse after introduction due to well-known passive dispersal

mechanisms (e.g. winds, animals). The role of dispersion should

therefore be considered in future homogenization studies by

conducting cross-taxonomic comparisons within and between

regions.

Recently, Rooney et al. (2007) addressed whether measures of

BH were relevant to conservation efforts. They highlighted that

conservation significance of BH depends on the scale of the

study. The introduction of non-native fish species in European

basins promoted the greatest homogenization of species

composition in south-western Europe. This should be inter-

preted with caution as it can result from two distinct scenarios:

(i) invasion of non-native species and extinction of native species

leading to a biotic impoverishment or (ii) invasion without

extinction of native species (see Olden & Poff, 2003, 2004). The

current situation in the major European river basins follows the

second scenario as no basin-scale extinctions were recorded.

However, in southern Europe, several studies reported declines

and extirpations of native and endemic fish species at local scales

(i.e. in localities within a river basin) due to the spread of

non-native species (Bianco, 1995; Elvira & Almodóvar, 2001).

Moreover, the process of extinction itself may occur on a much

longer timescale than invasions, which would make the perceived

impact of invasions dependent on the timescale of observation

(Sax et al., 2002). Although our basin scale approach does not

permit to quantify the risks of biotic impoverishment, it

indicates that southern European basins are the most prone to

homogenization. Indeed, ongoing fish invasions (Clavero &

García-Berthou, 2006), combined with the spread of the highly

seasonal Mediterranean climate in southern Europe, may

increase the risk of extinction for endemic fish that are already

threatened (Reynolds et al., 2005; Griffiths, 2006). A particular

attention should therefore be given to the outcome of fish

invasions in southern European basins that are recognized as

hotspots of fish diversity in Europe (Reyjol et al., 2006).

CONCLUSION

In this study, we clearly showed that exotic and translocated fish

species generated distinct geographical patterns of BH across

Europe because of their contrasting effects on the changes in

community similarity among river basins. Therefore, pooling

translocated and exotic species as is commonly done in

homogenization studies (e.g. Marchetti et al., 2001; Rooney

et al., 2004; Castro et al., 2006; Smith, 2006) can introduce a

major drawback in the quantification of the geographical pattern

of TH. We therefore recommend that future efforts in homogeni-

zation studies focus on making a clear distinction between exotic

and translocated species to accurately assess the spatial dynamics

of BH.

Comparing the observed TH patterns to those expected by

a null model empirically demonstrated that BH is a non-random

ecological pattern, therefore providing evidences in favour of

previous assumptions (McKinney & Lockwood, 1999; Duncan &

Lockwood, 2001; Olden et al., 2004). Because species invasions

and extinctions are likely to continue increasing over time with

increasing human activities (Sala et al., 2000), we expect that

homogenization of the world biota will also continue to intensify.

We feel that the null model approach presented here has useful

implications in the field of conservation biology and biogeography.

Indeed, null models were lacking in the exploration of BH,

whereas these models have long been applied to testing

large-scale ecological patterns (e.g. Connor & Simberlorff, 1978).

We invite biogeographers and ecologists to extend our null

model approach to other empirical data involving both species

invasions and extinctions. This will enable a relationship to be

established between each scenario of the mechanistic model of

Olden & Poff (2003) and a rigorous null model. We also encourage

future researches to apply other algorithms generating null

distributions of non-native and extinct species such as those with

fixed rows and columns sums that account for both interspecies

differences and environmental variability among localities (i.e.

the swap algorithm, Gotelli, 2000). However, such an algorithm

cannot be easily achieved as it requires reshuffling translocated

species only in the localities where they did not naturally occur.

Specific algorithms should be developed in this aim.

ACKNOWLEDGEMENTS

We are grateful to Simon Blanchet and Peter Winterton for their

helpful comments on this manuscript. This study was supported

by the ANR ‘Freshwater fish diversity’ (ANR –06-BDIV-010,

French Ministry of Research).

REFERENCES

Bianco, P.G. (1995) Mediterranean endemic freshwater fishes of

Italy. Biological Conservation, 72, 159–170.


© 2007 The AuthorsDiversity and Distributions, Journal compilation © 2007 Blackwell Publishing Ltd 9

Blackburn, T.M. & Duncan, R.P. (2001) Establishment patterns

of exotic birds are constrained by non-random patterns in

introduction. Journal of Biogeography, 28, 927–939.

Castro, S.A., Munoz, M. & Jaksic, F.M. (2006) Transit towards

floristic homogenization on oceanic islands in the south-

eastern Pacific: comparing pre-European and current floras.

Journal of Biogeography, 34, 213–222.

Clavero, M. & García-Berthou, E. (2006) Homogenization

dynamics and introduction routes of invasive freshwater fish in

the Iberian Peninsula. Ecological Applications, 16, 2313–2324.

Connor, E.F. & Simberlorff, D. (1978) Species number and

compositional similarity of the Galapagos flora and avifauna.

Ecological Monographs, 48, 219–248.

Davies, K.F., Chesson, P., Harrison, S., Inouye, B.D., Melbourne

B.A. & Rice, K.J. (2005) Spatial heterogeneity explains the scale

dependence of the native-exotic diversity relationship. Ecology,

86, 1602–1610.

Doadrio, I. (2002) Atlas y libro rojo de los peces continentales de

España. Ministerio de Medio Ambiente, Madrid, Spain.

Duncan, J.R. & Lockwood, J.L. (2001) Spatial homogenization of

aquatic fauna of Tennessee: extinction and invasion following

land use change and habitat alteration. Biotic homogenization

(ed. by J.L. Lockwood and M.L. McKinney), pp. 245–258.

Kluwer Academic/Plenum Publishers, New York.

Elvira, B. & Almodóvar, A. (2001) Freshwater fish introductions

in Spain: facts and figures at the beginning of the 21st century.

Journal of Fish Biology, 59 (Suppl. A), 323–331.

Fuller, P.L., Nico, L.G. & Williams, J.D. (1999) Nonindigenous

fishes introduced into inland waters of the United States.

Special Publication 27. American Fisheries Society, Bethesda,

Maryland.

García-Berthou, E., Alcaraz, C., Pou-Rovira, Q., Zamora, L.,

Coenders, G. & Feo, C. (2005) Introduction pathways and

establishment rates of invasive aquatic species in Europe.

Canadian Journal of Fisheries and Aquatic Sciences, 65, 453–

463.

Gido, K.B. & Brown, J.H. (1999) Invasion of North American

drainages by alien fish species. Freshwater Biology, 42, 387–399.

Gotelli, N.J. (2000) Null model analysis of species co-occurrence

patterns. Ecology, 81, 2606–2621.

Griffiths, D. (2006) Pattern and process in the ecological

biogeography of European freshwater fish. Journal of Animal

Ecology, 75, 734–751.

Henderson, R.A. & Heron, M.L. (1977) A probabilistic method

of paleobiogeographic analysis. Lethaia, 10, 1–15.

Holcík, J. (1991) Fish introductions in Europe with particular

reference to its central and eastern part. Canadian Journal of

Fisheries and Aquatic Sciences, 48 (Suppl. 1), 13–23.

Hugueny, B. & Paugy, D. (1995) Unsaturated fish communities

in African rivers. American Naturalist, 68 (162), 169.

Ihaka, R. & Gentleman, R. (1996) R: a language for data analysis

and graphics. Journal of Computational and Graphical Statistics,

5, 299–314.

Jeschke, J.M. & Strayer, D.L. (2006) Determinants of vertebrate

invasion success in Europe and North America. Global Change

Biology, 12, 1608–1619.

Keith, P. & Allardi, J. (2001) Atlas des poissons d’eau douce de

France. Patrimoines Naturels, 47, 1–387.

Kennard, M.J., Arthinghton, A.R., Pusey, B.J. & Harch, B.D.

(2005) Are alien fish are a reliable indicator of river health?

Freshwater Biology, 50, 174–193.

La Sorte, F.A. & McKinney, M.L. (2006) Compositional similarity

and the distribution of geographical range size for assemblages

of native and non-native species in urban floras. Diversity and

Distributions, 12, 679–686.

Legendre, P. & Legendre, L. (1998) Numerical ecology, 2nd edn

(English). Elsevier, Amsterdam, the Netherlands.

Marchetti, M.P., Light, T., Feliciano, J., Armstrong, T., Hogan, Z.,

Viers, J. & Moyle, P.B. (2001) Homogenization of California’s

fish fauna through abiotic change. Biotic homogenization (ed.

by J.L. Lockwood & M.L. McKinney), pp. 259–278. Kluwer

Academic/Plenum Publishers, New York,.

McKinney, M.L. (2005) Species introduced from nearby sources

have a more homogenizing effect than species from distant

sources: evidence from plants and fishes in the USA. Diversity

and Distributions, 11, 367–374.

McKinney, M.L. & Lockwood, J.L. (1999) Biotic homogenization:

a few winners replacing many losers in the next mass extinction.

Trends in Ecology & Evolution, 14, 450–453.

Moyle, P.B. & Light, T. (1996) Biological invasions of freshwaters:

empirical rules and assembly theory. Biological Conservation,

78, 149–161.

Moyle, P.B. & Marchetti, M.P. (2006) Predicting exotic fishes in

freshwater systems: freshwater fishes in California as a model.

Bioscience, 56, 515–524.

Nekola, J.C. & White, P.S. (1999) The distance decay of similarity

in biogeography and ecology. Journal of Biogeography, 26, 867–

878.

Olden, J.D. (2006) Biotic homogenization: a new research agenda

for conservation biogeography. Journal of Biogeography, 33,

2027–2039.

Olden, J.D. & Poff, N.L. (2003) Toward a mechanistic under-

standing and prediction of biotic homogenization. American

Naturalist, 162, 442–460.

Olden, J.D. & Poff, N.L. (2004) Ecological processes driving

biotic homogenization: testing a mechanistic model using fish

faunas. Ecology, 85, 1867–1875.

Olden, J.D. & Rooney, T.P. (2006) On defining and quantifying

biotic homogenization. Global Ecology and Biogeography, 15,

113–120.

Olden, J.D., Poff, N.L., Douglas, M.R., Douglas, M.E. & Fausch,

K.D. (2004) Ecological and evolutionary consequences of

biotic homogenization. Trends in Ecology & Evolution, 19, 18–24.

Olden, J.D., Poff, N.L. & McKinney, M. (2006) Forecasting

faunal and floral homogenization associated with human

population geography in North America. Biological Conserva-

tion, 127, 261–271.

Qian, H. & Ricklefs, R.E. (2006) The role of exotic species in

homogenizing the North American flora. Ecology Letters, 9,

1293–1298.

Rahel, F.J. (2000) Homogenization of fish faunas across the

United States. Science, 288, 854–856.

F. Leprieur et al.


Rahel, F.J. (2002) Homogenization of freshwater faunas. Annual

Review of Ecology, Evolution and Systematics, 33, 291–315.

Raup, D.M. & Crick, R.E. (1979) Measurement of faunal similarity

in paleontology. Journal of Paleontology, 53, 1213–1227.

Rejmánek, M. (2000) A must for North American biogeographers.

Diversity and Distributions, 6, 208–211.

Reyjol, Y., Hugueny, B., Pont, D., Bianco, P.G., Beier, U., Caiola, N.,

Casals, F., Cowx, I., Economou, A., Ferreira, T., Haidvogl, G.,

Noble, R., De Sostoa, A., Vigneron, T. & Virbickas, T. (2006)

Patterns in species richness and endemism of European

freshwater fish. Global Ecology and Biogeography, 16, 65–75.

Reynolds, J.D., Webb, T.J. & Hawkins, L.A. (2005) Life history

and ecological correlates of extinction risk in European

freshwater fishes. Canadian Journal of Fisheries and Aquatic

Sciences, 62, 854–862.

Rooney, T.P., Wiegmann, S.M., Rogers, D.A. & Waller, D.M.

(2004) Biotic impoverishment and homogenization in

unfragmented forest understory communities. Conservation

Biology, 18, 787–798.

Rooney, T.R., Olden, J.D., Leach, M.K. & Rogers, D.A. (2007)

Biotic homogenization and conservation prioritization.

Biological Conservation, 134, 447–450.

Sala, O.S., Chapin, F.S., Armesto, J.J., Berlow, E., Bloomfield, J.,

Dirzo, R., Huber-Sanwald, E. et al. (2000) Global biodiversity

scenarios for the year 2100. Science, 287, 1770–1774.

Sax, D.F., Gaines, S.D. & Brown, J.H. (2002) Species invasions

exceed extinctions on islands worldwide: a comparative study

of plants and birds. American Naturalist, 160, 766–783.

Schoener, T.W. (1987) Axes of controversy in community.

Community and evolutionary ecology of North American stream

fishes (ed. by W.J. Matthews and D.C. Heins), pp. 8–16.

University of Oklahoma Press, Norman, Oklahoma.

Schwartz, M.W., Thorne, J.H. & Viers, J.H. (2006) Biotic

homogenization of the California flora in urban and urbaniz-

ing regions. Biological Conservation, 127, 282–291.

Smart, S.M., Thompson, K., Marrs, R.H., Le Duc, M., Lindsay, G.,

Maskell, C. & Firbank, L.G. (2006) Biotic homogenization and

changes in species diversity across human-modified ecosystems.

Proceedings of the Royal Society of London Series B, Biological

Sciences, 273, 2659–2665.

Smith, K.G. (2006) Patterns of nonindigenous herpetofaunal

richness and biotic homogenization among Florida counties.

Biological Conservation, 127, 327–335.

Smith, S. & Shurin, J.B. (2006) What do invasions tell us about

saturation? Conceptual ecology and invasions biology (ed. by M.

Cadotte, S. McMahon and T. Fukami), pp. 423–444. Springer,

Dordrecht, the Netherlands.

Taylor, E.B. (2004) An analysis of homogenization and differen-

tiation of Canadian freshwater fish faunas with an emphasis on

British Columbia. Canadian Journal of Fisheries and Aquatic

Sciences, 61, 68–79.

Tokeshi, M. (1986) Resource utilization, overlap and temporal

community dynamics: a null model analysis of an epiphytic

chironomid community. Journal of Animal Ecology, 55, 491–

506.

Vooren, C.M. (1972) Ecological aspects of the introduction of

fish species into natural habitats, in Europe, with special

reference to the Netherlands. A literature survey. Journal of

Fish Biology, 4, 565–583.

SUPPLEMENTARY MATERIAL

The following supplementary material is available for this article:

Appendix S1 Bibliographic sources used to set up the freshwater

fish database of the 25 major European river basins.

This material is available as part of the online article from

http://www.blackwell-synergy.com/doi/abs/10.1111/

j.1472-4642.2007.00409.x

(This link will take you to the article abstract).

Please note: Blackwell Publishing is not responsible for the

content or functionality of any supplementary materials supplied

by the authors. Any queries (other than missing material) should

be directed to the corresponding author for the article.

http://www.blackwell-synergy.com/doi/abs/10.1111/j.1472-4642.2007.00409.x


P4

Patterns and mechanisms of the distance decay of similarity in the European freshwater fish fauna: contrasting native and exotic

species

Leprieur F., Olden J.D., Lek S. & Brosse S.

(en préparation)

1

Patterns and mechanisms of the distance decay of similarity in the

European freshwater fish fauna: contrasting native and exotic species

F. Leprieur 1, J.D. Olden 2, S. Lek 1, S. Brosse 1

1 Laboratoire Evolution et Diversité Biologique, U.M.R 5174, C.N.R.S -

Université Paul Sabatier, 118 route de Narbonne, F-31062 Toulouse cedex 4,

France.

2 University of Washington, School of Aquatic & Fishery Sciences, Box 355020, Seattle,

Washington 98195-5020, USA.

Running title: the distance decay of similarity in the European freshwater fish fauna.

2

Abstract

Explaining the processes that contribute to spatial variability in assemblage structure

remains one of the fundamental themes of contemporary community ecology and was

particularly achieved by exploring the distance decay of similarity in ecological communities.

As for native species, recent studies reported a decline in exotic species similarity (i.e. species

introduced from outside the biogeographic zone considered) with increasing distance between

localities. However, it remains unclear whether the distance decays of similarity in native and

exotic assemblages are governed by the same mechanisms (i.e. environmental filtering vs.

dispersal limitation). In the present study, we analysed the distributional patterns of native and

exotic fish species across the major European river basins and related these patterns to spatial,

environmental and anthropogenic factors. We showed that exotic species distributional

patterns relative to geographical distance overall match those found with natives (i.e. low

difference in species turnover along geographical distance). Our results suggest that both

environmental filtering (relative to climate) and dispersal limitation (relative to historical

factors for native species and to selective human-mediated introductions for exotics) are

important in explaining large-scale distributional patterns of native and exotic freshwater

species. Complementary analyses incorporating species traits of both native and exotic would

be particularly valuable to test the hypothesis that environmental filtering of exotic species

generates regional exotic assemblages that functionally converges on native assemblages.

Key words: distance decay of similarity, freshwater fish, native and exotic species,

environmental filtering, dispersal limitation.

3

Introduction

Enhancing our knowledge of the processes that shape spatial variability in assemblage

structure remains one of the fundamental challenges in contemporary community ecology and

conservation biogeography (Tuomisto et al. 2003; Gilbert & Lechowicz 2004; Whittaker et

al. 2005; Olden 2006; Gaston et al. 2007). This is best exemplified by the increasing focus on

exploring the patterns of distance decay of similarity in ecological communities, i.e. the

decrease in compositional similarity (i.e., increasing species turnover) between localities with

increasing geographical distance separating them (Nekola & White 1999; Soininen et al.

2007).

The recent synthesis by Soininen et al. (2007) suggests that the distance decay of

similarity is caused by at least two, not necessary mutually-exclusive, mechanisms. First, the

environmental filtering hypothesis predicts that community composition change is a result of

species-specific niche differences in evolved adaptive responses along spatially-structured

environmental gradients. Therefore, the environment may act as a selective filter removing

species which lacks ecological or life-history traits that confer persistence under a given set of

abiotic conditions (Keddy 1992). Second, the dispersal limitation hypothesis predicts that (i)

differences in species dispersal capabilities produce patterns of decay in community similarity

with distance even in homogeneous environments (neutral processes); and (ii) spatial

configuration of the landscape (e.g. the size and isolation of habitats) influence species

turnover by controlling the movement of organisms and hence limit the dispersion of

organisms among localities. A landscape with major geographical barriers to movement

would produce greater decays of community similarity compared to an open and more

homogeneous landscape.

One additional mechanism may explain the distance decay of similarity in ecological

communities. Habitat disturbances associated to human activities (i.e. anthropogenic

4

disturbances), such as land conversion to urban or agriculture use, may act as a selective filter

by favouring well-adapted floras and faunas to artificial environments (e.g. King & Buckney

2000; Kennard et al. 2005; Clergeau et. al 2006). Under this particular case of the

environmental filtering hypothesis, patterns of similarity decays with distance would be the

result of greater similarity in anthropogenic disturbances between neighbouring localities

Recent studies have reported a decline in exotic species similarity (i.e. species

established from outside the biogeographic zone considered) with increasing distance between

localities (McKinney 2004; LaSorte & McKinney 2006; Qian & Ricklefs 2006; LaSorte et al.

2007); a pattern also observed for a number of native communities (Nekola & White 1999;

Soininen et al. 2007). What remains unclear, however, is whether the distance decay of

similarity in native and exotic assemblages are governed by the same mechanisms?

Biogeographical patterns of exotic species offer a unique opportunity to test among the

competing hypotheses of environmental filtering and dispersal limitation for shaping patterns

of distance decay in community similarity. Indeed, the distributions of exotic species are not

constrained by historical factors (e.g. glacial and geological events) that partly explain the

current distribution of native species (Ricklefs & Schluter 1993; Mandrak 1995; Tedesco et

al. 2006; Montoya et al. 2007). Then, exotic species are likely to be less dispersally-limited

than natives due to human-assisted introductions that breach natural geographic barriers to

movement (e.g. Qian & Ricklefs 2006; Rahel 2007). In that context, we may expect that

dispersal limitation for native species will play a more important role in driving the distance

decay of similarity than for exotics. Hence exotic species would exhibit a lower rate of

similarity decay with distance (i.e. a lower species turnover along geographical distance) than

natives.

Whereas decline in compositional similarity with distance has been widely reported

for diverse set of taxa, few studies have explored this phenomenon for freshwater fishes

5

(Soininen et al. 2007). Yet, freshwater ecosystems present numerous advantages as river

basins are separated from one another by barriers insurmountable for fish (land or ocean) and

they form therefore biogeographical islands whose space is perfectly delimited (Hugueny

1989; Reyjol et al. 2006). In the present study, we explored distributional patterns of native

and exotic fish species across the major European river basins (see Leprieur et al. 2007). We

tested (i) whether native and exotic species exhibited similar spatial patterns of composional

similarity across Europe (i.e. the distance decay of similarity) and (ii) whether these patterns

were driven by the same mechanisms (environmental filtering vs. dispersal limitation).

Material and Methods

Data sources

Our study explores patterns of freshwater fish biogeography from the Iberian

Peninsula and France in the West to the Ural Mountains in the East (see Fig 1). Freshwater

fish occurrences for native species and exotic species (i.e. originating from outside Europe)

were compiled for the major European river basins from published data (see Leprieur et al.

2007). In this study, we also compiled fish occurrences for the Dälalven river basin (Sweden)

to have an equal sample size between northern and southern river basins. This led to a total of

26 major river basins with 138 native and 38 exotic species. We considered only strictly

freshwater fish because migratory and brackish species would introduce potential bias in the

analysis as we considered each river basin as a biogeographic island (Hugueny 1989).

We collated environmental data related to habitat heterogeneity, climate and

anthropogenic disturbance for each river basin, which were summarized in 11 explanatory

variables. Several climatic variables were derived for each river basin from a global scale

meteorological database (0.5° square grid coverage) (Leemans & Cramer 1991), which has

6

often been used in large scale ecological studies (e.g. Guégan et al. 1998; Beauchard et al.

2003). Climatic variables included annual average precipitation (mm), coefficient of variation

of monthly precipitation (i.e. standard deviation of monthly precipitation/annual average

precipitation; calculated from the twelve monthly values), annual number of rainy days (i.e.

the annual number of days with precipitation), annual average temperature (°C), coefficient of

variation of monthly temperature. Following Guégan et al. (1998), we characterized basin-

level habitat heterogeneity according to basin surface area (km2) and mean annual discharge

(m3/year) at the river mouth. We also used the altitudinal range (m) (calculated from a

geographical atlas) as it is well known that the wider the altitudinal range in a basin, the

greater the habitat heterogeneity. Lastly, we compiled human population density, the

percentage of urban area and the number of cities with more than 100,000 habitants to

characterize the degree of human-related habitat disturbance (i.e. anthropogenic disturbance,

see Olden et al. 2007). For each river basin, the surface area, the annual discharge and the

human variables were compiled from the World Resources Institute (2003).

Data analysis

Geographical patterns of compositional similarity

Patterns of compositional similarity between localities have, in large part, been studied

using species similarity metrics such as the Jaccard and Sorenson indices (e.g. Connor &

Simberloff 1978; Nekola & White 1999; Olden & Rooney 2006; Qian & Ricklef 2006).

Although these indices have been criticized for their dependence with species richness

gradients (Koleff et al. 2003), we used Jaccard’s index because of its wide application in

ecology and more particularly in studies exploring the distance decay of similarity (see review

of Soininen et al. 2007). This allowed us to make cross-comparisons between recent studies in

7

the literature. The Jaccard index of similarity ranges from 0 to 1 and is defined quantitatively

as J = a/(a + b + c), where a is the number of taxa shared between two localities, and b and c

are the numbers of taxa unique to either locality (Legendre & Legendre 1998).

To identify geographical patterns in compositional similarity, we applied a

hierarchical cluster analysis (Ward method, Legendre & Legendre 1998) on the similarity

matrix between river basins for the native (136 species in 26 basins) and exotic species pools

(38 exotic species in 24 basins). We tested whether the similarity matrices of exotic and

native species were significantly correlated by using a Mantel test, i.e. whether basins that are

similar in their native species composition are also similar in their exotic species composition.

A Mantel test quantifies the magnitude of the linear relationship between two distance

matrices (Legendre & Legendre 1998). Because all pairs of observations in a given distance

matrix are not independent, Monte Carlo permutations (10,000) were used to test if the

observed value of the Mantel-test statistic (rM) differed from those expected under the null

hypothesis (i.e., no correlation between the two distance matrices).

Distance decay of similarity

We investigated the distance decay of similarity by plotting the pairwise similarity in

species composition between basins against the geographical distance separating them. We

quantified the geographical distance between basins (based on mean latitude and longitude of

each basin) according to Euclidean distance. Both pairwise similarity and geographical

distance values were log-transformed to improve the linearity of the distance decay plot for

both exotic and native species.

We applied a randomization procedure to test for differences in the rate of decay in

species similarity with geographical distance (i.e. a measure of species turnover) between

native and exotic species (Nekola & White 1999, Steinitz et al. 2006). The procedure

8

involved the followed steps: (1) species similarity values of the two data sets (native vs.

exotic species) were rescaled to a common mean; (2) for each pair of basins, the two values of

species similarity (exotic and native species similarity) along with the corresponding distance

were randomly reassigned to the two data sets; (3) then, linear regression was used to

determine the slope of the distance decay of similarity for the two randomized data sets and

the absolute difference between the two slopes was calculated; (4) steps (2) and (3) were

repeated 9,999 times; (5) finally, the difference between the slopes of the two original data

sets was compared with the distribution of the differences between the slopes of the 9,999

randomized data sets in order to determine its significance level (two-tailed test, α=5%).

Relating distance-decay patterns of similarity to environmental and human factors

To test the competing environmental filtering and dispersal limitation hypotheses, we

applied simple and partial Mantel tests (Legendre & Legendre 1998) and two complementary

variance decomposition techniques: (i) variation partitioning applied to regression on distance

matrix (Borcard 1992; Legendre & Legendre 1998) and (ii) hierarchical partitioning applied

to multiple regression models (Chevan & Sutherland 1991; Heikkinen et al. 2005).

We first used simple Mantel tests to analyse the relationship between the exotic and

native similarity matrices respectively and each distance matrix of environmental variables

related to habitat heterogeneity, climate and anthropogenic disturbance. To account for spatial

autocorrelation, partial Mantel tests (9,999 permutations) were then applied to assess the

importance of the explanatory variables in influencing patterns of compositional similarity

after having removed the effect of geographic distance between basins (Legendre & Legendre

1998). Simple and partial Mantel tests were performed by using the “vegan package”

implemented in the open source R software (Ihaka & Gentleman 1996).

9

We then applied multiple linear regression on distance matrix (using PERMUTE!

3.4.9, Casgrain 2001) to quantify the variation in native and exotic compositional similarity

explained by four models based on different groups of explanatory variables: (i) geographical

distance, habitat heterogeneity, climate and anthropogenic disturbance-related variables

combined, i.e. the full model; (ii) geographical distance and habitat heterogeneity-related

variables combined; (iii) geographical distance and climatic variables combined; (iiii)

geographical distance and anthropogenic disturbance-related variables combined. Multiple

regression on distance matrices is conceptually similar to traditional linear regression except

that the dependent and independent variables are square distance matrices instead of single

vectors (Legendre & Legendre 1998). For each of the four models considered, a variance

partitioning method described in Borcard et al. (1992) was applied to distinguish between the

effect of geographical distance and the effect of the habitat heterogeneity, climate and

anthropogenic disturbance-related variables. This method consists in decomposing the total

variation in species similarity into four fractions: (i) uniquely explained by geographical

distance; (ii) coexplained by the explanatory variables and geographical distance; (iii)

uniquely explained by the explanatory variables; (iiii) unexplained fraction.

Finally, we applied hierarchical partitioning (Chevan & Sutherland 1991) because

multi-collinearity among environmental variables (expressed as distance values, see Table 1)

may lead to misleading inferences about the mechanisms driving patterns of compositional

similarity. Hierarchical partitioning is based on the theorem of hierarchies in which all

possible models (2k models for k explanatory variables, i.e. 4096 submodels for 12

explanatory variables including geographical distance) in a multiple regression setting are

considered jointly to attempt to identify the most likely causal factors (Chevan & Sutherland

1991; McNally 2002; Heikkinen et al. 2005). In contrast to variation partitioning (Borcard et

al. 1992), hierarchical partitioning provides, for each explanatory variable separately, an

10

estimate of the independent and joint contribution with all other variables. The independent

and joint contributions of each variable are distinguished by comparing the increase of the fit

of all submodels with a particular variable compared to equivalent submodels without that

factor (Chevan & Sutherland 1991). Hierarchical partitioning was conducted using multiple

linear regression and R2 as the goodness-of-fit measure. (‘hier.part package’ version 0.5–1

implemented in the open source R software (Ihaka & Gentleman 1996).

Results

Geographical patterns of compositional similarity

For both native and exotic fish species, patterns of compositional similarity followed a

geographical trend across Europe. Geographically close river basins showed the greater level

of community similarity (i.e., classified close together by the hierarchical cluster analysis)

whereas more distant basins showed little agreement in species membership (Fig. 2). Four

groups of river basins were distinguished when analysing the distribution of native species

(Ward linkage = 1.5; Fig. 2A): (i) river basins of the Iberian Peninsula in south-western

Europe (1-4); (ii) river basins of the Ponto-Caspian sea in eastern Europe (15-20); (iii) river

basins of western-central Europe (21,22, 24-26) and (iiii) river basins of northern Europe (5-

14, 23). Overall, we found similar groupings of basins based on the distribution of exotic

fishes (Fig 2B). This pattern was confirmed by a Mantel test showing that the native and

exotic similarity matrices were significantly correlated (RM=0.3607, P=0.003), i.e. basins that

were similar in their native species composition tended to be similar in their exotic species

composition. However, this correlation was no longer significant after eliminating the effect

of geographical distance (partial Mantel test: RMP=0,130, P=0.149).

11

Distance decay of similarity

Pairwise species similarity decreased significantly with increasing distance between

river basins for both native (Mantel test: RM= -0.417, P<0.001) and exotic species (RM= -

0.747, P<0.001) (Table 2 and Fig. 3). The randomization test revealed that native species

presented a lower rate of decay in similarity with increasing distance between basins than

exotics (regression slopes= -0.090 and -0.156 for native and exotic species respectively, two-

tailed test, P=0.0001).

12

1

23

45

67

8 9

10

1112 13 14

15

16

17 18

19

20

21

2223

24

25

1000Km

N

26

Figure 1: A) Map of the 26 major European River basins (1: Guadalquivir; 2: Tagus; 3: Douro; 4: Ebro; 5: Garonne ; 6: Loire; 7: Seine; 8: Rhône; 9: Pô; 10: Rhine; 11: Weser; 12 : Elbe; 13: Oder; 14: Wisla; 15: Danube; 16: Dniestr; 17: Dniepr; 18: Don; 19: Volga; 20: Ural; 21: Petchora; 22: Dniva; 23: Neva; 24: Kemijoki; 25: Glomma; 26 : Dälalven).

13

2- Tagus20- Ural18- Don

23- Neva

10- Rhine

12- Elbe13- Oder

9- Po

6- Loire

1- Guadalquivir

15- Danube

8- Rhone

5- Garonne

16- Dniestr

25- Glomma21- Petchora

24- Kemijoski14- Wisla

11- Weser

22- Dvina

19- Volga

3- Douro

17- Dniepr

4- Ebro

7- Seine

26- Dalalven

2.5 2.0 1.5 1.0 0.5 0.0

Native species (A)

Distance (Ward linkage) Distance (Ward linkage)

4- Ebro9- Po

12- Elbe

13- Oder

6- Loire

19- Volga18- Don

22- Dvina

24- Kemijoski

16- Dniestr

5- Garonne

15- Danube14- Wisla

11- Weser

8- Rhone

17- Dniepr

10- Rhine

3- Douro2- Tagus

7- Seine

1- Guadalquivir

26- Dalalven25- Glomma20- Ural

2.5 2.0 1.5 1.0 0.5 0.0

Exotic species (B)

Figure 2: Clustering of the European river basins based on their native (A) and exotic (B) freshwater fish compositional similarity according to Jaccard’s index and Ward’s linkage method.

14

0

0.2

0.4

0.6

0.8

0 1 2 3 4 5

0 1 2 3 4 50

0.2

0.4

0.6

0.8

Log geographical distance

Native species (A)

Exotic species (B)

Figure 3: Relationship between pair-wise native (A) and exotic (B) fish compositional similarity (Jaccard’s index) and geographical distance among river basins. Solid black lines represent observed relationships.

15

Relating distance-decay patterns of similarity to environmental and human factors

Simple Mantel tests (Table 2) showed that pairwise similarity in native species

decreased significantly with increasing difference between river basins in number of rainy

days (RM=-0.558, P<0.001) and annual average temperature (RM=-0.458, P<0.001). After

controlling for the effect of geographical distance, these correlations remained significant for

the number of rainy days (RPM=-0.430, P=<0.001, Table 2). Pairwise similarity in exotic

species (Table 2) decreased significantly with increasing difference between river basins in

number of rainy days (RM=-0.556, P<0.001), annual temperature (RM=-0.607, P<0.001) and

variation of monthly temperature (RM=-0.649, P<0.001). These correlations remained weakly

significant after accounting for geographical distance (Table 2).

Among the three regression models associated to different groups of explanatory

variables (i.e. habitat heterogeneity, climate and anthropogenic disturbance), the model

incorporating both geographical distance and climate accounted for the highest explained

variance for both native and exotic species (R2= 0.430 and 0.489 for native and exotic species,

respectively, Table 3). Variation partitioning applied to this model for both native and exotic

species showed that the total variation in species similarity explained by the pure effect of

geographical distance was much less than that explained by climatic conditions in isolation

and in combination with geographical distance.

For exotic species, the regression models incorporating both geographical distance and

the habitat heterogeneity or anthropogenic disturbance-related variables explained a non

negligible amount of variation in species similarity (around 40%, see Table 3). However,

more than half of this variation was explained by the pure effect of geographical distance

(23.4 % and 28.1% for the habitat heterogeneity and anthropogenic disturbance-related

models, respectively, Table 3). Moreover, the amount of variance uniquely explained by the

habitat heterogeneity or anthropogenic disturbance-related variables was very weak (Table 3).

16

Hierarchical partitioning provided consistent results with those found in previous

analyses. The 4096 submodels computed in hierarchical partitioning explained in average

31.8% and 50.8% of the variation in native and exotic species similarity between river basins,

respectively (±12.4% and 11.9% SD for native and exotic species, respectively). The

geographical distance and the climatic variables expressed as distance values (i.e. annual

average temperature and number of rainy days) displayed the highest independent and joint

contribution to the total explained variance in native fish similarity (i.e. total contribution ≥

10%, Fig.4A). These variables had a high joint contribution with all other variables in

explaining the variation in native species similarity between river basins (Fig 4A), because

they were spatially autocorrelated and highly collinear (see Table 1). The number of rainy

days accounted for the highest independent contribution to the total explained variance (i.e.

10%, Fig 4A).

Results for exotic species similarity showed that the geographical distance and the

climatic variables (i.e. annual average temperature, variation in monthly temperature and

number of rainy days) displayed the highest independent and joint contribution to the total

explained variance (total contribution ≥ 10%, Fig.4B). As for native fishes, these variables

had a high joint contribution with all other variables in explaining the variation in exotic

species similarity between river basins (Fig 4B), because they were spatially autocorrelated

and highly collinear (see Table 1). The geographical distance, annual temperature, variation in

monthly temperature and number of rainy days displayed similar independent contributions to

the total explained variance (ranging from 3% to 5%, Fig 4B).

17

Table 1: Mantel correlation (RM) between each distance-based variable1. P-values are indicated under parenthesis.

1; SA: surface area; ALT: altitudinal range; PL: annual average precipitation (mm), CVPL: coefficient of variation of monthly precipitation; NRD: annual number of rainy days; TP: annual average temperature (°C), CVTP: coefficient of variation of monthly temperature); POP: population densities; CITIES: number of large cities; URB: urban area (%); DIST: geographical distance.

DIS SA ALT PL CVPL NRD TP CVTP POP CIT URB DIST

DIS 1

SA 0.862 (p<0.001) 1

ALT -0.030 (0.630)

-0.000 (0.503) 1

PL -0.056 (0.577)

-0.065 (0.554)

0.471 (<0.001) 1

CVPL -0.122 (0.04)

-0.117 (0.07)

-0.041 (0.259)

0.042 (0.272) 1

NRD 0.084 (0.245)

0.016 (0.373)

0.221 (0.009)

0.081 (0.224)

0.062 (0.188) 1

TP 0.133 (0.164)

0.037 (0.323)

0.132 (0.051)

0.168 (0.084)

0.093 (0.103)

0.771 (<0.001) 1

CVTP 0.176 (0.117)

0.137 (0.133)

0.110 (0.082)

0.214 (0.052)

0.073 (0.177)

0.450 (<0.001)

0.746 (<0.001) 1

POP -0.079 (0.651)

-0.119 (0.122)

0.067 (0.166)

0.223 (0.054)

0.088 (0.143)

-0.035 (0.566)

0.096 (0.159)

0.171 (0.103) 1

CITIES 0.613 (0.008)

0.543 (0.039)

0.073 (0.226)

-0.069 (0.592)

-0.079 (0.821)

-0.131 (0.130)

-0.127 (0.160)

-0.085 (0.452)

0.046 (0.243) 1

URB 0.076 (0.223)

0.026 (0.303)

0.191 (0.013)

0.471 (<0.001)

0.044 (0.256)

0.217 (0.011)

0.364 (<0.001)

0.389 (<0.001)

0.575 (<0.001)

-0.023 (0.508) 1

DIST 0.196 (0.076)

0.205 (0.057)

0.278 (0.001)

0.347 (0.001)

-0.028 (0.604)

0.591 (<0.001)

0.728 (<0.001)

0.564 (<0.001)

0.028 (0.305)

-0.118 (0.123)

0.515 (<0.001) 1

18

Table 2: Results from simple (RM) and partial Mantel test (RPM, geographical distance partialled out) between native and exotic fish compositional similarity respectively and environmental and human variable. * p<0.05; ** p<0.004 (bonferroni corrected family-wide α=0.004)

Native species Exotic species Variables RM P RPM P RM P RPM P Geographical distance -0.417** <0.001 - - -0.747** < 0.001 - - Habitat heterogeneity Surface area 0.109 0.199 0.302** 0.001 -0.370** 0.001 -0.178 0.080 Discharge 0.036 0.465 0.131 0.200 -0.213 0.120 -0.070 0.318 Altitudinal range -0.157* 0.036 -0.077 0.353 -0.178* 0.010 -0.012 0.610 Climate Annual precipitation -0.079 0.236 0.073 0.279 -0.237 0.013 -0.002 0.575 CV montly precipitations -0.027 0.289 -0.088 0.148 -0.091 0.469 -0.030 0.718 Number of rainy days -0.558** <0.001 -0.430** <0.001 -0.456** < 0.001 -0.185* 0.045 Annual average temperature -0.458** <0.001 -0.254* 0.007 -0.607** < 0.001 -0.201* 0.010 CV monthly temperature -0.027 0.325 0.281** 0.001 -0.649** < 0.001 -0.278* 0.009 Antropogenic disturbance Population density 0.107 0.200 0.130 0.137 -0.015 0.417 -0.019 0.370 Number of large cities 0.140 0.181 0.144 0.157 -0.035 0.339 -0.083 0.283 Urban area (%) -0.029 0.387 0.207* 0.007 -0.281** 0.003 -0.065 0.242

19

Table 3: Results from multiple matrix regression models applied to different groups of explanatory variables (see Material and Methods for more details) and variation partitioning to decompose the total variance of each model in four fractions: (G) uniquely explained by geographical distance; (VG) coexplained by the explanatory variables and geographical distance; (V) uniquely explained by the explanatory variables; (U) unexplained fraction. These fractions are expressed in percentage. The total explained variance (R2) of each model is significant (* p<0.001). The full model contains both geographical distance and the climatic, habitat heterogeneity and human activities-related variables.

Model R2 G VG V U Native species Full model 0.478* 2.2 30.7 14.9 52.2 Habitat heterogeneity 0.186* 16.4 1.5 0.7 81.4 Climate 0.430* 1.3 25.9 15.8 57.0 Antropogenic disturbance 0.218* 19.0 2.0 1.0 78.0 Geographical distance 0.171* Exotic species Full model 0.503* 2.1 36.8 11.4 49.7 Habitat heterogeneity 0.395* 23.4 15.6 0.6 60.4 Climate 0.489* 6.6 32.4 9.9 51.1 Antropogenic disturbance 0.396* 28.1 10.8 0.7 60.3

Geographical distance 0.389*

20

-10

0

10

20

30

40

50 Joint contributionIndependent contribution

-10

0

10

20

30

40

50

Climate Anthropogenicdisturbance

HabitatHeterogeneity

Distance

Exotic species (B)

Native species (B)

Figure 4: Results from hierarchical partitioning analysis illustrating the independent and joint contributions of the explanatory variables in explaining the variation in native (A) and exotic (B) fish species similarity across Europe. Values are presented as the percentage of the total explained variance. Variable abbreviations are DIST: geographical distance; DIS: Discharge; SA: surface area; ALT: altitudinal range; TP: annual average temperature (°C); CVTP: coefficient of variation of monthly temperature; PL: annual average precipitation (mm), CVPL: coefficient of variation of monthly precipitation; NRD: annual number of rainy days; POP: population densities; URB: urban area (%); CITIES: number of large cities.

21

Discussion

Whether the regional distribution of species is limited by dispersal or by

environmental conditions has been debated for a long time by biogeographers and ecologists

(Ricklefs & Schuter 1993; Jackson & Harvey 1989; Nekola & White 1999; Chust et al. 2006).

Our results revealed that the compositional similarity in native fish species decreased with

both geographical distance and difference in climatic conditions (i.e. the annual average

temperature and the number of rainy days) between river basins. However, these factors

displayed a high joint contribution in explaining the variation in native species similarity

between river basins. This is not surprising as climatic conditions in Europe show a distinct

geographic signature (see Table 1) with (i) northern latitudes characterized by low

temperature and high number of rainy days, and (ii) southern latitudes characterized by high

temperature and low number of rainy days. This suggests that native species similarity

between river basins decrease along spatially-structured climatic gradients and therefore

supports the hypothesis of environmental filtering (Tonn et al. 1990; Keddy 1992; Mouillot et

al. 2007). However, it is important to recognize that the strong spatial autocorrelation in

climatic conditions found in Europe (i.e. the fact that neighbouring river basins have similar

climatic conditions) makes difficult to clearly distinguish between the relative role of

geographical distance per se (i.e. the role of dispersal limitation) and climate (Jackson &

Harvey 1989; Gilbert & Lechowicz 2004). Hence the role of dispersal limitation in shaping

the distance decay of similarity in the European native freshwater fish fauna can not be

understated.

Recently, Reyjol et al. (2006) provided strong evidence that the European river basins

can be considered as non-equilibrated islands in which species extinctions were not fully

balanced by colonization from neighbouring river basins. Indeed, small river basins located in

western and northern Europe experienced higher rates of extinction than the large Ponto-

22

Caspian river basins during the Pleistocene glaciations (i.e. the Danube, Dniestr, Dniepr, Don

and Volga river basins, Reyjol et al. 2006; Griffiths 2006). Postglacial colonization rates were

probably highest for the river basins close to the Black Sea, which is recognized as a refuge

zone during the Pleistocene glaciations. The dual processes of differential species extinction

and colonization may explain why Ponto-Caspian river basins support the highest native fish

species richness in Europe (Reyjol et al. 2006). In that context, based on our results and those

of Reyjol et al. (2006), we can conclude that the present-day distribution of the native

European freshwater fish fauna is likely the result of both dispersal limitation (associated with

past historical events) and species-specific responses to spatially-structured climatic gradients

(environmental filtering). The combined effect of post-glacial dispersal limitation and climatic

gradients in structuring regional fish communities have been also reported in North America

(e.g. Jackson & Harvey 1989; Mandrak 1995; Oswood et al. 2000; Hoagstrom & Berry 2006).

As for native species, our results revealed for exotic species provide support for the

environmental filtering hypothesis. Indeed, climatic conditions in isolation and in

combination with geographical distance explained together much of the variation in exotic

species similarity between river basins (Table 3). However, climate and geographical distance

in combination explained a greater fraction of the variation in exotic species similarity

between river basins (32.4 %) than climate in isolation (9.9%). This suggests that dispersal

limitation is also important in explaining the composition of regional exotic fish assemblages.

For example, Garcia-Berthou et al. (2005) have recently identified pathways of freshwater

species introductions in Europe (involving essentially fishes) and showed that these ones were

spatially-structured. Midlatitude western European countries (e.g. France and Germany)

received many exotic species from North America and provided many of them to both

southern (Spain) and northern (Sweden) countries. In contrast, midlatitude eastern European

countries (e.g. Romania and Poland) received some exotic species from former Soviet Union

23

but provided no species to other countries. Such a geographical structure in introduction

pathways results from the intentional human-selected nature of fish introductions (Copp et al.

2005; García-Berthou et al. 2005). More generally, this agrees with a number of studies

showing that differences or similarities in human-selected species (e.g. Blackburn & Duncan

2001; Clavero & García-Berthou 2006) may explain the geographical distributions of non-

native species. In that context, the distribution of exotic fish species across Europe is likely

the result of both (i) dispersal limitation relative to selective human-mediated introductions

and (ii) environmental filtering that remove species which lacks traits for persisting under a

given set of climatic conditions (e.g., Fausch et al. 2001; Marchetti & Moyle 2004; Copp et

al. 2005).

Among the climatic variables considered, the annual temperature and the number of

rainy days were found to be important in explaining both the variation in native and exotic

species similarity between river basins. These two climatic variables are related to broad-scale

physiological and ecological requirements of freshwater fish species. First, water temperature

directly influences the metabolic rates, physiology, and life-histories of aquatic species.

Freshwater fishes and the organisms on which they feed respond to thermal heterogeneity and

require specific temperature ranges to survive and reproduce (Magnuson et al. 1979). A

number of studies have shown that the thermal tolerances of freshwater fish determine the

distributional limits to their range (e.g. Jackson & Harvey 1989; Shuter & Post 1990; Minns

& Moore 1995; Lappalainen & Soininen 2006). For example, the northern Arctic Glomma,

Dälalven, Kemijoski river basins were only invaded by two introduced cold water salmonid

species (Oncorhynchus mykiss and Salvelinus fontinalis) because the harsh environmental

conditions of these rivers (i.e. low temperature, long period of ice coverage) probably impede

the successful establishment of many exotic fish species. Then, in the present broad-scale

context, the number of rainy days may account for the temporal persistence of suitable

24

habitats for freshwater fish. Research suggest that patterns of rainy days influences stream

flow constancy as the number of rainy days is negatively correlated with its coefficient of

variation (Beauchard et al. 2003). Therefore we expect an increasing stability of water supply

with an increasing number of annual rainy days. For example, Mediterranean-type streams in

southern Europe experience strong seasonal patterns of flow (i.e. low flows in summer that

restrict aquatic habitats to small isolated pools, and high flows in winter and spring) and long

period of high temperature. These fluctuating environmental conditions are likely to prevent

the invasion of large fish species more adapted to more stable aquatic habitats (Vila-Gispert et

al. 2005) and to favour exotic warm water species adapted to drought prone river systems

(e.g. Gambusia holbrooki, Ameirius melas, Lepomis gibbosus). Such a species sorting process

according to hydrological stability has been reported for both native and exotic freshwater

fishes in North America (Moyle & Light 1996; Marchetti & Moyle 2001; Hoeinghaus et al.

2006).

Finally, our results indicated that exotic species displayed a higher rate of similarity

decay (i.e. species turnover rate) with distance than native fishes. However, the difference

between native and exotic decay rates was weak (Slopes= -0.090 and -0.156 for native and

exotic species respectively). Therefore, it is not surprising that geographically close river

basins that were similar in their native species composition also tended to be similar in their

exotic species composition. In addition, this result agrees with two recent studies for

European (Leprieur et al. 2007) and Australian river basins (Olden et al. 2007) showing that

exotic fish introductions have led to a low decrease in fish fauna similarity between adjacent

river basins (i.e. a low taxonomic differentiation).

Overall, our results contrast with our initial prediction and with previous large-scale

studies on exotic plant species that exhibited a lower rate of similarity decay than native

species (LaSorte & McKinney 2006; Qian & Ricklef 2006). These studies suggested that

25

exotic plants were less dispersally-limited than natives due to intentional or/and accidental

introductions over long distance. In addition, Qian & Ricklefs (2006) found that exotic plant

distributions in states and provinces of the USA and Canada showed less association with

climatic conditions compared to native species; a result that contrasts our findings for

European freshwater fishes. Such differences between our results and those of Qian &

Ricklefs (2006) are likely related to several factors. First, Qian & Ricklef (2006) analysed

plant distributional patterns using state- and province-level geographical units which may not

be a relevant biogeographical scale as river basin for freshwater fish. Indeed, the regional

species pool of plants is maybe not accurately defined using provinces and states because

these political units do not take into account biogeographical barriers (e.g. mountain ranges,

large rivers) that define the historical floral distinctiveness of a region. Second, exotic plants

are probably less dispersally-limited than exotic freshwater fish due to intrinsic differences in

dispersal capabilities between freshwater fish and plants. Last, in contrast to plants, much of

freshwater fish introductions are intentional (and therefore geographically-localized) since

they are strongly associated to human uses (e.g. sport fishing; see García-Berthou et al. 2005).

Overall, these not mutually-exclusive factors may explain why exotic plants in North America

were found to be (i) less related to climatic gradients; and (ii) more widely distributed over

large spatial scale than exotic fish species in Europe.

Conclusion

We showed that exotic species distributional patterns relative to geographical distance

overall match those found with natives (i.e. low difference in species turnover along

geographical distance). Our results suggest that both environmental filtering (relative to

climate) and dispersal limitation (relative to historical factors for native species and to

selective human-mediated species introduction for exotic species) are important in explaining

26

large-scale distributional patterns of native and exotic species. First, this confirms that exotic

and native species tend to exhibit similar biogeographical patterns (e.g. Sax 2001; Labra et al.

2005). Then, it agrees with much of recent studies highlighting that environmental filtering

and dispersal limitation were not mutually exclusive in explaining spatial variability in

assemblage structure (e.g. Tuomisto et al. 2003; Gilbert & Lechowicz 2004; Cottenie 2005;

Beisner et al. 2006; Beck & Vun Khen 2007). However, complementary analyses

incorporating species traits of both native and exotic would be particularly valuable. For

example, species-level analyses (see Hoeinghaus et al. 2006; Mason et al. 2007; Mouillot et

al. 2007) would permit to test the hypothesis that environmental filtering of exotic species

generates regional exotic assemblages that functionally converges on native assemblages.

Finally, because species invasions is a temporal dynamic process, future studies should

concentrate on (i) comparing the slope of the distance decay of similarity over time and (ii)

testing if the main determinants of the distance decay of similarity in exotic species

assemblages are temporally consistent.

Acknowledgments

This study was supported by the ANR "Freshwater fish diversity" (ANR -06-BDIV-010,

French Ministry of Research).

References

Beauchard, O., Gagneur, J. & Brosse, S., (2003). Macroinvertebrate richness patterns in North African streams. Journal of Biogeography, 30, 1821-1833.

Beck, J. & Vun Khen, C. (2007) Beta-diversity of geometrid moths from northern Borneo: effects of habitat, time and space. Journal of Animal Ecology, 76, 230-237.

Beisner, B. E., P. R. Peres-Neto, E. S. Lindstrom, A. Barnett & M. L. Longhi. (2006) The role of environmental and spatial processes in structuring lake communities from bacteria to fish. Ecology, 87, 2985-2991.

Borcard, D. Legendre, P. & Drapeau, P (1992) Partialling out the spatial component of ecological variation. Ecology, 73, 1045-1055.

27

Casgrain, P. (2001) Permute 3.4.9. User’s Manual. Montreal, Canada http://www.fas.umontreal.ca/biol/casgrain

Chevan, A. & Sutherland, M. (1991) Hierarchical Partitioning. American Statistician, 45, 90–96.

Chust, G., Chave, J. Condit, R., Aguilar, S., R. Perez, S. Lao, S. (2006) Determinants and spatial modelling of tree beta-diversity in a tropical forest landscape in Panama. Journal of Vegetation Science, 17, 83-92.

Clavero, M. & García-Berthou, E. (2006) Homogenization dynamics and introduction routes of invasive freshwater fish in the Iberian Peninsula. Ecological applications, 16, 2313–2324.

Connor, E.F. & Simberlorff, D. (1978) Species number and compositional similarity of the Galapagos flora and avifauna. Ecological Monographs, 48, 219-248.

Copp, G.H., Bianco, P.G., Bogutskaya, N., Eros, T., Falka, I., Ferreira, M.T., Fox, M.G., Freyhof, J., Gozlan, R.E., Grabowska, J., Kováç, V., Moreno-Amich, R., Naseka, A.M., Peñáz, M., Povz, M., Przybylski, M., Robillard, M., Russell, I.C., Stakéenas, S., Íumer, S., Vila-Gispert, A. & Wiesner, C. (2005) To be, or not to be, a non-native freshwater fish? Journal of Applied Ichthyology, 21, 242–262.

Cottenie, K (2005) Integrating environmental and spatial processes in ecological community dynamics. Ecology Letters, 8, 1175-1182.

Fausch, K. D., Taniguchi, Y., Nakano, S., Grossman, G. D. & Townsend, C. R. (2001). Flood disturbance regimes influence rainbow trout invasion success among five Holarctic regions. Ecological Applications, 11, 1438–1455.

Fuller, P.L., Nico, L.G. & Williams, J.D. (1999) Nonindigenous fishes introduced into inland waters of the United States. Special Publication 27. American Fisheries Society, Bethesda,Maryland.

García-Berthou, E., Alcaraz, C., Pou-Rovira, Q., Zamora, L.,Coenders, G. & Feo, C. (2005) Introduction pathways and establishment rates of invasive aquatic species in Europe. Canadian Journal of Fisheries and Aquatic Sciences, 65, 453–463.

Gaston, K.J. Evans, KL & Lennon, J.L. (2007) The scaling of spatial turnover: pruning the thicket. In: Scaling biodiversity (Storch, D., Marquet, P. Brown, J.M. eds). Cambridge university press.

Gilbert, B. & Lechowicz, M. J. (2004) Neutrality, niches, and dispersal in a temperate forest understory. Proceedings of the National Academy of Sciences USA, 101, 7651-7656.

Griffiths, D. (2006) Pattern and process in the ecological biogeography of European freshwater fish. Journal of Animal Ecology, 75, 734-751.

Guégan, J.F., Lek, S., Oberdorff, T., (1998) Energy availability and habitat heterogeneity predict global riverine fish diversity. Nature, 391, 382-384.

Heikkinen RK, Luoto M, Kuussaari M, & Poyry J. (2005) New insights into butterfly-environment relationships using partitioning methods. Proc. R. Soc. Lond. Ser. B-Biol. Sci., 272, 2203-2210.

Hoagstrom, C.W. & Berry, C.R. (2006) Island biogeography of native fish faunas among Great Plains drainage basins: basin scale features influence composition. American Fisheries Society Symposium, 48, 221-264.

Hoeinghaus, D.J., Winemiller, K.O. & Birnbaum, J.S. (2007) Local and regional determinants of stream fish assemblage structure: inferences based on taxonomic vs. functional groups. Journal of Biogeography, 34, 324-338.

Hugueny, B. (1989) West African rivers as biogeographic islands. Oecologia, 79, 235-243. Ihaka, R. & Gentleman, R. (1996) R: a language for data analysis and graphics. Journal of

computational and graphical statistics, 5, 99-314.

28

Jackson, D.A. & Harvey, H.H. (1989) Biogeographic associations in fish assemblages: local vs. regional processes. Ecology, 70, 1472-1484.

Keddy, P.A. (1992) Assembly and response rules: two goals for predictive community ecology. Journal of Vegetation Sciences, 3, 157-164.

Kennard, M.J., Arthinghton, A.R., Pusey, B.J. & Harch, B.D. (2005) Are alien fish are a reliable indicator of river health? Freshwater Biology, 50, 174-193.

King, S.A. & Buckney, R.T. (2000) Urbanization and exotic plants in northern Sydney streams. Austral Ecology, 25, 455-461.

Koleff, P, Gaston, K.J & Lennon J.J. (2003) Measuring beta diversity for presence-absence data. Journal of Animal Ecology, 72, 367-382.

Labra, F.A., S.R. Abades & P.A. Marquet (2005) Distribution and abundance. Scaling patterns in exotic and native bird species. In: Species Invasions. Insights into Ecology, Evolution and Biogeography (D.F. Sax, J.J. Stachowicz & S.D. Gaines, eds), pp. 421-446. Sinauer Associates Inc, Massachusetts.

Leprieur, F., Beauchard, O., Hugueny, B., Grenouillet, G. & Brosse, S. (2007) Null model of biotic of homogenization: a test with the European freshwater fish fauna. Diversity and Distributions (doi:10.1111/j.1472-4642.2007.00409.x).

Lappalainen, J. & Soininen, J. (2006) Latitudinal gradients in niche breadth and position - regional patterns in freshwater fish. Naturwissenschaften, 93, 246-250

La Sorte, F.A. & McKinney, M.L. (2006) Compositional similarity and the distribution of geographical range size for assemblages of native and non-native species in urban floras. Diversity and Distributions, 12, 679-686.

La Sorte, F.A., McKinney, M.L., Pyšek, P., Klotz, S., Rapson, G.L. Celesti-Grapow, L., and Thompson, K. (2007) Distance decay in similarity among European urban floras: the impact of anthropogenic activities on beta diversity. Global Ecology and Biogeography, in press.

Leemans, R. & Cramer, W. (1991) The IIASA database for mean monthly values of temperature, precipitation and cloudiness of a global terrestrial grid. International Institute for Applied System Analysis (IIASA), RR-91–18.

Legendre, P. & Legendre, L. (1998) Numerical Ecology, Second English Edition. Elsevier, Amsterdam, the Netherlands.

Mac Nally, R. (2002) Multiple regression and inference in conservation biology and ecology: further comments on identifying important predictor variables. Biodiversity and Conservation, 11, 1397–1401.

Magnuson, J.J., L.B. Crowder and P.A. Medvick. 1979. Temperature as an ecological resource. American Zoologist, 19, 331-43.

Mandrak, N. E. (1995) Biogeographic patterns of fish species richness in Ontario lakes in relation to historical and environmental factors. Canadian Journal of Fisheries and Aquatic Sciences, 52, 1462-1474.

Marchetti, M.P. & Moyle, P.B. (2001) Effects of flow regime and habitat structure on fish assemblages in a regulated California stream. Ecological Applications, 11, 530-539.

Marchetti, M.P., Moyle, P.B., & Levine, R. (2004) Invasive species profiling? Exploring the characteristics of non-native fishes across invasion stages in California. Freshwater Biology, 49, 646–661.

Mason, N., Lanoiselée, C., Mouillot, D., Irz, P., Argillier, C. (2007) Functional characters combined will null models reveal inconsistency in mechanisms of species turnover in lacustrine fish communities. Oecologia, 153, 441-452.

McKinney, M.L. (2004) Do exotics homogenize or differentiate communities? Roles of sampling and species richness. Biological Invasions, 6, 495-504.

29

Minns, C.K. & Moore, J.E. (1995) Factors limiting the distributions of Ontario’s freshwater fishes: the role of climate and other variables, and the potential impacts of climate change. In: Beamish RJ (ed) Climate change and northern fish populations. Canadian Special Publication in Fisheries and Aquatic Sciences, 121,137–160.

Montoya, D., Rodríguez, M.Á., Zavala, M.Á. & Hawkins, B.A. (2007) Glacial dynamics and contemporary richness patterns of Holarctic trees. Ecography, 30, 173-182.

Mouillot, D., Dumay, O. & Tomasini, J.A. (2007) Limiting similarity, niche filtering and functional diversity in coastal lagoon fish communities. Estuarine Coastal and Shelf Science, 71, 443-456.

Moyle, P.D. & T. Light. (1996) Fish invasions in California: do abiotic factors determine success? Ecology, 77, 1666-1670.

Nekola, J. C. & White, P. S. (1999) The distance decay of similarity in biogeography and ecology. Journal of Biogeography, 26, 867-878.

Olden, J. D. (2006) Biotic homogenization: a new research agenda for conservation biogeography. Journal of Biogeography, 33, 2027-2039.

Olden, J. D., & T. P. Rooney. (2006) On defining and quantifying biotic homogenization. Global Ecology and Biogeography, 15, 113-120.

Olden, J.D., Kennard, M.J., & Pusey, B.J. (2007) Species invasions and the changing biogeography of Australian freshwater fishes. Global Ecology and Biogeography (doi: 10.1111/j.1466-8238.2007.00340.x).

Oswood, M.W., Reynolds, J.B., Irons III, J.B. & Milner, A.M. (2000) Distributions of freshwater fishes in ecoregions and hydroecoregions of Alaska. Journal of the North American Benthological Society, 19, 405-418.

Qian, H & Ricklefs, R.E. (2006) The role of exotic species in homogenizing the North American flora. Ecology Letters, 9, 1293-1298.

Rahel, F.J. (2007) Biogeographic barriers, connectivity, and biotic homogenization: it’s a small world after all. Freshwater Biology, 82, 696–710

Reyjol Y., Hugueny, B., Pont, D., Bianco, P. G., Beier, U., Caiola, N., Casals, F., Cowx, I., Economou, A., Ferreira, T., Haidvogl, G., Noble, R., De Sostoa, A., Vigneron, T. & Virbickas, T. (2006). Patterns in species richness and endemism of European freshwater fish. Global Ecology and Biogeography, 16, 65-75.

Ricklefs, R. & Schluter D. (1993) Species diversity in ecological communities: historical and geographical perspectives. Chicago: University of Chicago Press.

Sax, D.F. (2001) Latitudinal gradients and geographic ranges of exotic species: implications for biogeography. Journal of Biogeography, 28, 139-150.

Shuter, B.J. & Post, J.R. (1990). Climate, population variability, and the zoogeography of temperate fishes. Transactions of the American Fisheries Society, 119, 314-336.

Soininen, J., McDonald, R. & Hillebrand, H. (2007) The distance decay of similarity in ecological communities. Ecography, 30, 3-12.

Steinitz, O., Heller, J., Tsoar, A., Rotem, D. & Kadmon, R. (2006) Environment, dispersal and patterns of species similarity. Journal of Biogeography, 33, 1044-1054.

Tedesco, P.A., Oberdorff, T, Lasso, C.A., Zapata, M. & Hugueny, B. (2006) Evidence of history in explaining diversity patterns in tropical riverine fish. Journal of Biogeography, 32,1899-1907.

Tonn, W. M., Magnuson, J. J., Rask, M. & Toivonen, J. (1990) Intercontinental comparison of small-lake fish assemblages: the balance between local and regional processes. The American Naturalist, 136, 345-375.

Tuomisto H., Ruokolainen, K. & Yli-Halla M. (2003) Dispersal, environment, and floristic variation of western Amazonian forests. Science, 299, 241-244.

30

Vila-Gispert, A., Alcaraz, C. & García-Berthou, E. (2005) Life-history traits of invasive fish in small Mediterranean streams. Biological Invasions, 7, 107-116.

Whittaker, R. J., M. B. Araujo, P. Jepson, R. J. Ladle, J. E. M. Watson, & K. J. Willis. (2005) Conservation Biogeography: assessment and prospect. Diversity and Distributions, 11, 3-23.

World Resources Institute (2003) Watersheds of the World: An Earth Trends Special Collection.http://earthtrends.wri.org/maps_spatial/watersheds/i


P5

Fish invasions in the world’s river systems: when natural processes are blurred by human activities

Leprieur F., Beauchard O., Blanchet S., Oberdorff T. & Brosse S.

PLoS Biology (accepté)

1

Fish Invasions in the World’s River Systems: When Natural Processes are

Blurred by Human Activities

Fabien Leprieur 1, Olivier Beauchard 2, Simon Blanchet 3, Thierry Oberdorff 4 and Sébastien

Brosse 1

1 Laboratoire Evolution & Diversité Biologique, U.M.R 5174, C.N.R.S -

Université Paul Sabatier, 118 route de Narbonne, F-31062 Toulouse cedex 4,

France.

2 University of Antwerp, Faculty of Sciences, Department of Biology, Ecosystem

Management Research Group, Universiteitsplein 1, BE-2610 Antwerpen (Wilrijk), Belgium.

3 Département de Biologie, Centre Interuniversitaire de Recherche sur le Saumon Atlantique

(CIRSA) and Québec-Océan, Université Laval, Sainte-Foy, Quebec City, Quebec, Canada.

4 Institut de Recherche pour le Développement (UR131), Antenne au Muséum National

d'Histoire Naturelle, 43 rue Cuvier, 75231 Paris cedex, France.

Running title : Fish Invasions in the World’s River Systems

2

Abstract

Because species invasions are a principal driver of the human-induced biodiversity

crisis, the identification of the major determinants of global invasions is a prerequisite for

adopting sound conservation policies. Three major hypotheses, which are not necessarily

mutually exclusive, have been proposed to explain the establishment of non-native species:

the “human activity” hypothesis, which argues that human activities facilitate the

establishment of non-native species by disturbing natural landscapes and by increasing

propagule pressure; the “biotic resistance” hypothesis, predicting that species-rich

communities will readily impede the establishment of non-native species; and the “biotic

acceptance” hypothesis, predicting that environmentally suitable habitats for native species

are also suitable for non-native species. We tested these hypotheses and report here a global

map of fish invasions (i.e., the number of non-native fish species established per river basin)

using an original worldwide dataset of freshwater fish occurrences, environmental variables,

and human activity indicators for 1,055 river basins covering more than 80% of Earth’s

surface. First, we identified six major invasion hotspots where non-native species represent

more than a quarter of the total number of species. According to the World Conservation

Union, these areas are also characterised by the highest proportion of threatened fish species.

Second, we show that the human activity indicators account for most of the global variation in

non-native species richness, which is highly consistent with the “human activity” hypothesis.

In contrast, our results do not provide support for either the “biotic acceptance” or the “biotic

resistance” hypothesis. We show that the biogeography of fish invasions matches the

geography of human impact at the global scale, which means that natural processes are

blurred by human activities in driving fish invasions in the world’s river systems. In view of

our findings, we fear massive invasions in developing countries with a growing economy as

3

already experienced in developed countries. Anticipating such potential biodiversity threats

should therefore be a priority.

Introduction

The deliberate or accidental introduction of species outside their native range is a key

component of the human-induced biodiversity crisis, harming native species and disturbing

ecosystems processes [1–3]. The greater the introduction of non-natives in a region, the

higher the probability that some of them become invasive and will hence cause ecological or

economic damage [4,5]. Patterns of non-native species richness are therefore relevant in

forecasting the overall impact of invasions on a global scale [5] and should help management

authorities to adopt sound, effective conservation policies [5–7].

The process of species invasion consists of three successive stages: initial dispersal,

establishment of self-sustaining populations, and spread into the recipient habitat. The last

two stages are contingent upon the first one, i.e., if initial dispersal is interrupted,

establishment and spread do not occur [8]. Three major hypotheses, which are not necessarily

mutually exclusive, have been proposed to explain invasion patterns: the “human activity”

[9], “biotic acceptance” [10], and “biotic resistance” [11] hypotheses. The “human activity”

hypothesis refers to the three stages of the invasion process (initial dispersal, establishment,

and spread), whereas the “biotic resistance” and “biotic acceptance” hypotheses address only

the establishment and spread stages [12]. With regards to the establishment stage, the “human

activity” hypothesis predicts that, by disturbing natural landscapes and increasing propagule

pressure (i.e., the number of individuals released and the frequency of introductions in a given

habitat), human activities facilitate the establishment of non-native species [9,13,14].

Everything else being equal, a positive relationship is therefore expected between non-native

species richness and quantitative surrogates of propagule pressure and habitat disturbance

4

(e.g., gross domestic product [GDP], percentage of urban area, and human population density

[5]). Then, the “biotic acceptance” hypothesis predicts that the establishment of non-native

species would be greatest in areas that are rich in native species and with optimal

environmental conditions for growth (i.e., “what is good for natives is good for non-natives

too” [10]). Everything else being equal, native and non-native species richness should co-vary

positively with environmental factors such as energy availability and habitat heterogeneity,

which are already recognised as the primary global determinants of native species richness

[15,16]. In contrast, the “biotic resistance” hypothesis predicts that species-poor communities

will host more non-native species than species-rich communities, the latter being highly

competitive and hence readily impede the establishment of non-native species [11,17].

Therefore, a negative relationship is expected between native and non-native species richness.

To date, the relative importance of these hypotheses in explaining the variation in non-native

species richness had never been tested at the global scale.

We tested these hypotheses and report a global map of fish invasions (i.e., the number

of non-native fish established per river basin) by using an extensive worldwide dataset of

freshwater fish occurrences (i.e., more than 40,000 occurrences of 9,968 fish species) on the

river basin scale (1,055 basins covering more than 80% of Earth’s surface). Freshwater fish

offer a unique opportunity to identify factors that are responsible for large-scale gradients in

non-native species richness for at least two main reasons. First, among vertebrate groups,

freshwater fish have been widely introduced over the world [18], which often had subsequent

negative consequences on native species and ecosystems integrity [19–23]. Second, as rivers

are separated from one another by barriers insurmountable for freshwater fish (land or ocean),

they form kind of “biogeographical islands”, whose space is delimited [15]. This implies that

the natural and human factors shaping global patterns of non-native species richness can be

easily separated.

5

Results

Our results revealed six global invasion hotspots where non-native species represent

more than a quarter of the total number of species per basin: the Pacific coast of North and

Central America, southern South America, western and southern Europe, Central Eurasia,

South Africa and Madagascar, and southern Australia and New Zealand (Figure 1A).

According to The World Conservation Union (IUCN) Red List [25], these areas were also

characterised by the highest proportion of fish species having a high risk of extinction in the

wild (Figure 2).

Analysing the absolute number of species, we found that river basins of the Northern

Hemisphere host the highest number of non-native fish species (Figure 1B). The human

factors considered here to test the “human activity” hypothesis (GDP, population density,

percentage of urban area) were found to be positively related to non-native species richness

(Table 1), after controlling for the effects of environmental conditions and native species

richness. In contrast, the positive correlation between native and non-native richness that was

expected by the “biotic acceptance” hypothesis was not significant after controlling for the

effects of propagule pressure and habitat disturbance (Table 1). Indeed, the environmental

factors displayed either no (net primary productivity) or a weak positive correlation

(altitudinal range and basin area) with non-native species richness, after controlling for the

effects of propagule pressure and habitat disturbance (Table 1). The negative correlation

between native and non-native richness, expected by the “biotic resistance” hypothesis, was

not significant after controlling for the effects of environmental conditions, propagule

pressure, and habitat disturbance (Table 1).

Then, we applied hierarchical partitioning [26–28] that aims to quantify the

independent explanatory power of each variable by considering all possible submodels. The

deviance explained by the 128 submodels computed in hierarchical partitioning accounted in

6

average for 52% of the total deviance (±7% standard deviation [SD], min = 37%, max =

67%). The human factors had together the greatest independent effect on non-native species

richness (70%, Table 2). Among the human factors, the GDP (an economical index of human

activities [9]) had the greatest independent explanatory power (43%; Table 2). To a lesser

extent, the habitat heterogeneity (i.e., basin area and altitudinal range) and the number of

native species also contribute to the variation in non-native species between river basins

(Table 2).

To test for potential bias in our results due to differences in sampling effort between

continents, bootstrap analysis was performed by applying hierarchical partitioning to 1,000

random subsets of 100 basins. For each variable, the independent effect observed did not

differ from the 95% bootstrap percentile confidence interval (Table 2), testifying that potential

differences in sampling effort between continents hardly affected the results.

7

B

[ 0 % - 5 % ]] 5% - 25% ]

] 25 % - 95 % ]

A

[ 0 - 5 ]] 5 - 20 ]

] 20 - 70 ]

Figure 1. Worldwide Distribution of Non-Native Freshwater Fish. (A) The percentage of non-native species per basin (i.e., the ratio of non-native species richness/total species richness) and (B) the non-native species richness per basin. Each basin was delimited by a GIS using 0.5° × 0.5° unit grid. The maps were drawn using species occurrence data for 9,968 species in 1,055 river basins covering more than 80% of continental areas worldwide. Invasion hotspots are defined as areas where more than a quarter of the species are non-native (red areas on map (A)), leading to define six invasion hotspots: the Pacific coast of North and Central America, southern South America, western and southern Europe, central Eurasia, South Africa and Madagascar, Southern Australia, and New Zealand.

8

Table 1. Spearman Rank Correlation (rs) between the Number of Non-Native Fish Species (Residuals) and Each Explanatory Variable Related to the “Human Activity,” “Biotic Acceptance,” and “Biotic Resistance” Hypotheses (n = 597).

For each hypothesis, the relationship between the number of non-native fish species and the explanatory variables considered was quantified by controlling for the effects of the explanatory variables relevant to the other hypotheses (see Materials and Methods for more details). * p < 0.006 (Bonferroni correction, α = 0.006).

rs p

Human activity hypothesis Gross domestic product 0.550* <0.0001 Percentage of urban area 0.556* <0.0001 Population density 0.306* <0.0001

Biotic acceptance hypothesis Number of native species 0.093 0.062

Altitudinal range 0.264* <0.0001

Basin area 0.175* <0.0001

Net primary productivity -0.008 0.842 Biotic resistance hypothesis Number of native species -0.034 0.400

9

Table 2. Independent Effect of Each Environmental and Human Activity–Related Variable on the Number of Non-Native Species per Basin

Independent effect (%) (n=597)

95% boostrap confidence interval

(n=100)

Gross domestic product 43.06 [36.68 ; 45.08]

Percentage of urban area 13.94 [10.93 ; 16.63]

Population density 13.36 [11.63 ; 14.91]

Number of native species 5.16 [3.57 ; 7.46]

Altitudinal range 7.11 [4.24 ; 10.23]

Basin area 15.08 [9.96 ; 19.06]

Net primary productivity 2.26 [1.35 ; 4.63]

Hierarchical partitioning was applied to the 597 basins for which the seven variables selected to test the “human activity,” “biotic acceptance,” and “biotic resistance” hypotheses were available. The independent effect of a variable was expressed as a percentage of the total independent contribution associated with the seven variables. To test potential bias due to sample size, hierarchical partitioning was run on 1,000 random subsets of 100 basins among the total of 597 basins. For each variable, the independent effect based on 597 basins did not differ from the 95% bootstrap percentile confidence interval, testifying that sample size hardly affected the results. Both analyses underline the predominant role of the three human variables that together represent more than 70% of the independent effect.

10

0%

1%

2%

3%

4%

5%

6%

Criticallyendangered

Endangered Vulnerable

[ 0% - 5% ]

] 5% - 25% ]

] 25% - 95% ]

Figure 2. Percentage of Threatened Species for the Three Invasion Levels. Threatened species were identified from the IUCN Red List (vulnerable, endangered, critically endangered). We calculated the percentage of threatened species, listed in the IUCN Red List, for the three invasion levels considered in Figure 1A. Each invasion level expessed as the percentage of non-native species. ([ 0%–5% ], ]5%–25%], ]25%–95%]) account for 8,363, 2,257, 1,241 native species and 544, 240, 271 river basins, respectively.

11

Discussion

By using an explanatory modelling approach, we showed that the human activity

indicators of the world’s river basins were positively related to the number of established non-

native fish species. In addition, they account for most of the global variation in non-native

species richness, giving support for the “human activity” hypothesis. More particularly, we

highlight that the level of economic activity of a given river basin (expressed by the GDP)

strongly determines its invasibility. Three non-exclusive mechanisms may account for this

pattern. First, economically rich areas are more prone to habitat disturbances (e.g., dams and

reservoirs modifying river flows) that are known to facilitate the establishment of non-native

species [7,23,29]. Second, high rates of economic exchanges increase the propagule fluxes of

non-native species [6,9] via ornamental trade, sport fishing, and aquaculture [18]. Third, the

increased demand for imported products associated with economic development increase the

likelihood of unintentional introductions through the import process [6].

The “biotic resistance” hypothesis cannot explain the pattern of fish invasions

observed, because no negative relationship between native and non-native species richness

was found after controlling for the effects of environmental conditions, propagule pressure,

and habitat disturbance. This means that regional species-rich communities are not necessarily

a barrier against the establishment of non-native species [17]. Our results are consistent with

several studies showing that species-rich fish communities can support higher species

richness if the pool of potential colonisers is increased by species introductions [24,30,31].

More generally, our results agree with studies on various taxa that do not report biotic

resistance at broad spatial scales [10,11]. Then, we provide no real support for the alternative

“biotic acceptance” hypothesis [10] even if native and non-native species richness do respond

similarly to some of the environmental gradients tested (i.e., altitudinal range and basin area).

Actually, the absence of a significant positive relationship between native and non-native

12

species richness implies that species-rich river basins do not support more non-native species

than basins with a low native species richness (i.e., “the rich do not get richer”). This contrasts

with numerous continental and regional-scale studies on plants and animals that report a

strong matching between native and non-native species richness [10,32–35]. More generally,

our results do not agree with the expectation that native and non-native species richness

covary positively at macroecological scales [36].

The interpretation of the exact role of human activities (i.e., propagule pressure and

habitat disturbance) in driving broad-scale patterns of non-native species richness faced major

difficulties in previous continental and regional-scale studies due to covariations between

human and natural factors [9,13,34,35]. Indeed, because humans may have preferred to settle

in areas providing diverse natural resources, human population was found to be largest in

regions with high levels of habitat heterogeneity and energy availability that favour species-

rich native fauna and flora [34,37]. This therefore makes it difficult to determine whether the

often-reported positive relationship between native and non-native species richness is driven

by (i) common responses to habitat heterogeneity and energy availability or (ii) increased

propagule pressure and habitat disturbance. Such difficulties were probably related to the

spatial extent considered (i.e., a continental or regional extent). Indeed, we found a weak

covariation between environmental and human descriptors of the world’s river basins at the

global scale (Pearson’s correlation coefficients: r < 0.35, Table S1). This allowed us to clearly

disentangle the relative roles of human activities and environmental conditions in shaping the

global pattern of fish invasions. We show that the biogeography of fish invasions at the global

scale matches the geography of human impact but not the biogeography of native species.

Because increasing the number of non-native species increases the risk of biodiversity

loss [4,5], our results have two major implications for future conservation strategies. First, the

six global invasion hotspots identified here account for the highest proportion of threatened

13

fish species listed on the IUCN Red List [25]. These areas are also recognised as being

biodiversity hotspots (particularly southern Europe, South Africa and Madagascar, southern

Australia, and New Zealand [38,39]). Although species classified on the IUCN Red List are

threatened by various sources of disturbance (e.g., habitat loss, pollution, species invasion,

and overexploitation [25]), non-native species are recognised as a major threat to biodiversity

after habitat loss [25,40]. For example, 20% of the 680 species extinctions listed by the IUCN

were directly caused by species invasions [2]. Freshwater fish follow the same tendency, as

20% of the species listed by the IUCN are threatened by non-native species [41]. In that

context, we recommend that non-native species importations in the six invasion hotspots be

prohibited without detailed risk and long term cost-benefits assessments [42]. Special

attention should also be given to these areas to design efficient control programs of already-

established non-native species.

Second, as we provide strong evidence for the “human activity” hypothesis (with a

special emphasis on economic activity), we expect that river basins of developing countries

will host an increasing number of non-native fish species as a direct result of economic

development. This constitutes a serious threat to global biodiversity, because rivers of most

developing areas (e.g., southern Asia, western and central Africa) are characterised by high

levels of endemism [38]. Anticipating potential biodiversity threats should therefore be a

priority, because once they are established, the eradication of a non-native species is

extremely difficult and result in high economic costs [43].

Despite the increasing literature on non-native species, this study is, to our knowledge,

the first to provide a global map of species invasions for a given taxonomic group and should

stimulate others to test the generality of these findings for other taxa at this spatial scale. Such

broad-scale analyses would help local researches to focus on non-native species control in the

most sensitive areas (e.g., the six invasion hotspots we identified here for freshwater fish).

14

This study should also stimulate researches on freshwater ecosystems by combining the

existing global scale databases of physical disturbances [44,45] and the global pattern of fish

invasions given here. This would permit to quantify river basins threats by considering

simultaneously different sources of disturbance. Such an approach is urgently needed as rivers

are among the most threatened ecosystems of the world [46] and as freshwater fish constitute

a major source of protein for a large part of the world population [46].

Materials and Methods

Databases

We conducted an extensive literature survey of native and non-native freshwater fish

species check lists. Only complete species lists at the river basin scale were considered, and

we discarded incomplete check lists such as local inventories of a stream reach or based only

on a given family. The resulting database was gathered from more than 400 bibliographic

sources including published papers, books, and grey literature databases (references available

upon request). Our species database contains species occurrence data for the world’s

freshwater fish fauna at the river basin scale (i.e., 80% of all freshwater species described [47]

and 1,055 river basins covering more than 80% of Earth’s surface). It constitutes the most

comprehensive global database for freshwater fish occurrences at the river basin scale and, to

our knowledge, the largest database for a group of invaders. We considered as non-native a

species (i) that did not historically occur in a given basin and (ii) that was successfully

established, i.e., self-reproducing populations. Estuarine species with no freshwater life stage

were not considered in our analyses.

The environmental and human databases contain seven variables selected to test (i) the

“human activity” hypothesis: human population density (number of people km–2), percentage

15

of urban area and purchase power parity GDP (in US$); (ii) the “biotic acceptance”

hypothesis: number of native fish species, basin area (km2), altitudinal range (m), net primary

productivity (NPP in kg-carbon m–2 year–1), and (iii) the “biotic resistance hypothesis”:

number of native fish species. The area of each river basin was taken from published and

unpublished data. The altitudinal range for each river basin was determined from a

geographical atlas. We calculated the mean value of NPP, human population density, GDP,

and percentage of urban area over the surface area of each basin from 0.5° × 0.5° grid data

available in the Center for International Earth Science Information Network (CIESIN) and the

Atlas of Biosphere [48,49]. The surface area and altitudinal range at the river basin scale are

used as quantitative surrogates for habitat heterogeneity [16], which is known to influence

native freshwater fish species richness [15,16]. Net primary productivity is used as a

quantitative surrogate to river basin energy availability [16] and strongly correlates to native

freshwater fish species richness [15,16]. This is verified in our data, as we found that both

basin area and NPP are positively correlated to native species richness (partial Pearson’s

correlation coefficient: r = 0.592 and p < 0.0001 for basin area while controlling for the effect

of NPP; r = 0.514 and p < 0.0001 for NPP while controlling for the effect of the basin area).

Then, the human population density, percentage of urban area, and GDP were used as

quantitative surrogates for propagule pressure and habitat disturbance [5,9,33]. The GDP

measures the size of the economy and is defined as the market value of all final goods and

services produced within a region in a given period of time.

Fish invasions mapping

We first mapped the worldwide distribution of (i) the non-native species richness per

basin and (ii) the percentage of non-native species per basin (i.e., the ratio of non-native

species richness/total species richness). To do that, each basin was delimited by a geographic

16

information system (GIS) using a grid reference of 0.5° latitude and 0.5° longitude and then

reported on a world map. We used three classes of percentage (Figure 1A) and richness

(Figure 1B) of non-native species to draw colour maps. Other maps with more classes were

tried and provided similar results. We selected the one that minimised differences in sample

size (i.e., number of river basins) between classes. The percentage of non-native species per

basin was used to define invasion hotspots where more than a quarter of the species are non-

native (i.e., the third class of percentage of non-native species; red areas in Figure 1A). It was

preferred to the richness in non-natives due to its independence from native richness and basin

area. For each of the three levels of fish invasion ([ 0%–5% ], ]5%–25%], ]25%–95% ]), we

determined the percentage of species facing a high to extremely high risk of extinction in the

wild, i.e., the vulnerable, endangered, and critically endangered fish species according to the

IUCN Red List [25]. The percentage of threatened species should be regarded with caution,

because the IUCN Red List for freshwater fish is still incomplete. The percentages of

threatened species for the three levels of fish invasion are therefore probably underestimated.

Although we recognise the potential biases and limitations of the IUCN listing procedure, the

IUCN Red List of threatened species remains the most objective and authoritative system for

classifying species in terms of the risk of extinction at the global scale [41,50]. The list of

basins for the three levels of invasion is provided in Dataset S1.

Modelling method

In this study, to test the three hypotheses (i.e., “human activity”, “biotic acceptance”,

and “biotic resistance”), we did not build the best single and parsimonious model by using

stepwise selection of a subset of independent variables having a significant effect on the

number of non-native species per basin (i.e., predictive approach). Indeed, a single best model

is not necessarily the best explanatory model, because minimizing the overall difference

17

between the observed and predicted values does not necessarily equate to determining

probable influence in a multivariate setting [26–28,51,52]. In addition, a simple regression

model cannot identify situations in which potentially important independent variables are

suppressed by other variables due to their high colinearity. When there is colinearity between

independent variables, the direct response of the dependent variable to a independent variable

may in fact only be an indirect effect owing to high dependence of the considered variable

with one or many others [27].

In our dataset, the seven environmental and human variables are not independent

(Pearson’s correlation coefficient values ranging from –0.25 to 0.79, Table S1). We therefore

evaluated the independent explanatory power of each environmental and human variable by

using hierarchical partitioning [26–28,51,52], a method based on the theorem of hierarchies in

which all possible models in a multiple regression setting are considered jointly to attempt to

identify the most likely causal factors (explanatory approach).

If we consider k, the number of explanatory variables (X1, Xi,…, Xk), there are 2k

possible models (i.e., 128 submodels by considering the seven explanatory variables),

including the null model (M0). The Ri is a measure of fit between one independent variable Xi

and the dependent variable Y. The fit between each of the seven explanatory variables and the

dependent variable Y (number of non-native fish species per basin) was measured by the

reduction of deviance generated by introducing a given variable into all of the possible

models built with the six other variables within the considered hierarchies. We used a

generalised linear model (GLM) with a Poisson error to treat our count data (i.e., the number

of non-native fish species per basin). Each explanatory variable was log-transformed to meet

the assumptions of normality and homoscedasticity.

We consider k! hierarchical orderings of models that always begin with M0 and end

with Mx123…k. For any given initial variable Xi, there are (k – 1)! possible hierarchies

18

containing k(k – 1)! models in which Xi appears. For each hierarchy, we evaluate the influence

of Xi on each of the k models including Xi (increase in model fit generated by including the

variable Xi within each model). The independent influence (Ii) of Xi on Y was obtained by

averaging all of the k(k – 1)! increases of fit. This averaging alleviates multicolinearity

problems that are ignored by using a simple regression model [26–28]. The joint component Ji

(effect caused jointly with the k – 1 other variables) is obtained by subtracting Ii from Ri, with

Ri = Ii + Ji. If all explanatory variables were completely independent of one another, there

would be no joint contributions [26]. For each variable, the independent and joint

contributions are expressed as the percentage of the total explained deviance (R)

1 1 1

k k k

i i ii i i

R I J R I J= = =

= + = = +∑ ∑ ∑ .

In our models, the total independent contribution accounts for 75% of the total

explained deviance, which means that the joint contribution of each explanatory variable was

weak in explaining the global variation in non-native species richness (Figure S1). We

therefore quantified the independent effect (IEi) of each variable on the dependent variable Y

as the percentage of the total independent contribution, i.e.

1

ii k

ii

IIEI

=

=

∑. The significance of

the independent effect (IEi) of each variable was determined by a randomization approach (n

= 100) which yielded Z-scores [52]. Statistical significance was based on an upper confidence

limit of 0.95. Each variable display a significant independent effect.

We applied hierarchical partitioning to a subsample of 597 basins (Afrotropical: 72;

Australian: 94; Nearctic: 127; Neotropical: 68; Oriental: 29; Palearctic: 207) for which all

seven environmental and human variables used were available. To test potential bias due to

differences in sampling effort between continents, hierarchical partitioning was run on 1,000

19

random subsets of 100 basins among the total of 597 basins. For each variable, we calculated

the 95% bootstrap percentile confidence interval of the independent effect (IEi). Hierarchical

partitioning was conducted using the ‘hier.part’ package [52] version 1.0-1 implemented on

the open source R software [53]. Hierarchical partitioning implemented for linear

relationships was relevant to our data, because preliminary analyses did not detected any

significant effect of polynomial terms. The hierarchical partitioning results were compared

with those obtained with another method (i.e., variation partitioning, [54]). Overall, the results

of the two methods were similar, and the variables highlighted as significant by the two

approaches were the same.

Hierarchical partitioning does not provide information on the form of the relationship

(positive or negative) between the number of non-native species and each explanatory

variable. To test the “human activity” hypothesis, we analysed the form and the significance

of the relationship between each variable related to the “human activity” hypothesis (GDP,

percentage of urban area, and population density) and the residuals from a GLM with a

Poisson error. This model explains the number of non-native species by using independent

variables related to the “biotic resistance” and “biotic acceptance” hypotheses (number of

native species, altitudinal range, basin area, and net primary productivity). This allowed us to

control for the effects of environmental conditions and native species richness. Then, to test

the “biotic acceptance” hypothesis, we analysed the form and the significance of the

relationship between each variable related to the “biotic acceptance” hypothesis (i.e., number

of native species, altitudinal range, basin area, and net primary productivity) and the residuals

from a GLM explaining the number of non-native species by using the human activity–related

variables (i.e., GDP, percentage of urban area, and population density). This allowed us to

control for the effects of propagule pressure and habitat disturbance. Lastly, to test the “biotic

resistance” hypothesis, we analysed the form and the significance of the relationship between

20

the number of native species and the residuals from a GLM explaining the number of non-

native species by using independent variables related to the “biotic acceptance” and “human

activity” hypotheses (i.e., altitudinal range, basin area, net primary productivity, GDP,

percentage of urban area, and population density). This allowed us to control for the effects of

environmental conditions, propagule pressure and habitat disturbance. To test the relationship

between the model residuals and each explanatory variable, we performed a Spearman rank

correlation test, because the model residuals were not normally distributed.

Abbreviations: GDP, gross domestic product; GLM, generalised linear model; IUCN, The

World Conservation Union; NPP, net primary productivity

Supporting Information

Dataset S1. Names and Invasion Levels of the 1,055 River Basins. The three invasion levels

are those used in Figure 1A (i.e., the percentage of non-native species per basin). (i) [ 0%–5%

]; (ii) ]5%–25%]; (iii) ]25%–95% ]. Longitude and latitude at the river mouth was also

provided for the 1,055 river basins.

Figure S1. Results from Hierarchical Partitioning Analysis Illustrating the Independent and

Joint Contributions of the Explanatory Variables in Accounting for the Variation in Non-

Native Species Richness between River Basins (n = 597).Values are presented as the

percentage of the total explained deviance extracted from a GLM with a Poisson error. The

total independent contribution of the explanatory variables accounts for 75% of the total

explained deviance.

21

Table S1. Pearson’s Correlation Coefficient (r) between Each Explanatory Variable. NSR:

native species richness; AR: altitudinal range; BA: basin area; NPP: net primary productivity;

GDP: gross domestic product; PUA: percentage of urban area, PD: population density. Bold

values indicate a significant correlation p < 0.002 (Bonferroni correction, α = 0.002).

Acknowlegments

We thank J. Chave, E. Danchin, C.R. Townsend and P. Winterton for their insightful

comments, which have improved the manuscript. This work was supported by the National

Research Agency (ANR) Freshwater Fish Diversity (ANR-06-BDIV-010).

References

1. Vitousek PM, Mooney HA, Lubchenco J, Melillo JM (1997) Human domination of Earth's ecosystems. Science 278: 494–499.

2. Clavero M, García-Berthou E (2005) Invasive species are a leading cause of animal extinctions. Trends Ecol Evol 20: 110.

3. Byrnes JE, Reynolds PL, Stachowicz JJ (2007) Invasions and extinctions reshape coastal marine food webs. PLoS ONE 2(3): e295. doi:10.1371/journal.pone.0000295

4. Jeschke JM, Strayer DL (2005) Invasion success of vertebrates in Europe and North America. Proc Natl Acad Sci U S A 102: 7198–7202.

5. Pyšek P, Richardson DM (2006) The biogeography of naturalization in alien plants. J Biogeogr 33: 2040–2050.

6. Levine JM, D'Antonio CM (2003) Forecasting biological invasions with increasing international trade. Conserv Biol 17: 322–326.

7. Williamson M (1996) Biological invasions. London: Chapman & Hall. 244 p. 8. Puth M, Post DM (2005) Studying invasion: have we missed the boat? Ecol Lett 8: 715–

721. 9. Taylor BW, Irwin RE (2004) Linking economic activities to the distribution of exotic

plants. Proc Natl Acad Sci U S A 101: 17725–17730. 10. Fridley JD, Stachowicz JJ, Naeem S, Sax DF, Seabloom EW, et al. (2007) The invasion

paradox: reconciling pattern and process in species invasions. Ecology 88: 3–17. 11. Levine JM (2000) Species diversity and biological invasions: relating local process to

community pattern. Science 288: 852–854. 12. Lockwood JL, Hoopes MF, Marchetti MP (2007) Invasion Ecology. Oxford: Blackwell

Publishing. 304 pp. 13. Chown SL, Gremmen NJM, Gaston KJ (1998) Ecological biogeography of southern

ocean islands: species–area relationships, human impacts, and conservation. Am Nat 152: 562–575.

14. Meyerson LA, Mooney HA (2007) Invasive alien species in an era of globalization. Front Ecol Environ 5: 199–208.

22

15. Oberdorff T, Guégan JF, Hugueny B (1995) Global scale patterns of fish species richness in rivers. Ecography 18: 345–352.

16. Guéguan JF, Lek S, Oberdorff T (1998) Energy availability and habitat heterogeneity predict global riverine fish diversity. Nature 391: 382–384.

17. Kennedy TA, Naeem S, Howe KM, Knops JMH, Tilman D et al. (2002) Biodiversity as a barrier to ecological invasion. Nature 417: 636–638.

18. Lever C (1996) Naturalized Fishes of the World. London: Academic Press. 408 p. 19. Vander Zanden MJ, Casselman JM, Rasmussen JB (1999) Stable isotope evidence for the

food web consequences of species invasions in lakes. Nature 401: 464–467. 20. Kolar CS, Lodge DM (2002) Ecological predictions and risk assessment for alien fishes

in North America. Science 298: 1233. 21. McDowall RM (2006) Crying wolf, crying foul, or crying shame: alien salmonids and a

biodiversity crisis in the southern cool-temperate galaxioid fishes? Rev Fish Biol Fish 16: 233–422.

22. Leprieur F, Hickey M, Arbuckle CJ, Closs G, Brosse S et al. (2006) Hydrological disturbance benefits a native fish at the expense of an exotic fish. J Appl Ecol 43: 930–939.

23. Light T, Marchetti M (2007) Distinguishing between invasions and habitat changes as drivers of diversity loss among California’s freshwater fishes. Conserv Biol 21: 434–446.

24. Gido KB, Brown JH (1999) Invasion of North American drainages by alien fish species. Freshw Biol 42: 387–399.

25. IUCN (2006) 2006 IUCN Red List of Threatened Species. Available: http://www.iucnredlist.org.

26. Chevan A, Sutherland M (1991) Hierarchical partitioning. Am Stat 45: 90–96. 27. Mac Nally R (2002) Multiple regression and inference in conservation biology and

ecology: further comments on identifying important predictor variables. Biodivers Conserv 11: 1397–1401.

28. Heikkinen RK, Luoto M, Kuussaari M, Pöyry J (2005) New insights to butterfly–environment relationships with partitioning methods. Proc R Soc Lond B 272: 2203–2210.

29. Havel JE, Lee CE, Vander Zanden MJ (2005) Do reservoirs facilitate invasions into landscapes? BioScience 55: 518–252.

30. Smith SA, Bell G, Bermimgham E (2004) Cross-Cordillera exchange mediated by the Panama Canal increased the species richness of local freshwater fish assemblages. Proc R Soc Lond B 271: 1889–1896.

31. Leprieur F, Beauchard O, Hugueny B, Grenouillet G, Brosse S (2007) Null model of biotic homogenization: a test with the European freshwater fish fauna. Divers Distrib: In press. doi: 10.1111/j.1472-4642.2007.00409.x.

32. Stohlgren TJ, Barnett DT, Kartesz JT (2003) The rich get richer: patterns of plant invasions in the United States. Front Ecol Environ 1: 11–14.

33. Marchetti MP, Light TS, Moyle PB, Viers J (2004) Invasion and extinction in California fish assemblages: testing hypotheses using landscape patterns. Ecol Appl 14:1507–1525.

34. Evans KL, Warren PH, Gaston KJ (2005) Does energy availability influence classical patterns of spatial variation in exotic species richness? Glob Ecol Biogeogr 14: 57–65.

35. Chown SL, Hull B, Gaston KJ (2005) Human impacts, energy availability and invasion across Southern Ocean Islands. Glob Ecol Biogeogr 14: 521–528.

23

36. Melbourne BA, Cornell HV, Davies KF, Dugaw CJ, Elmendorf S et al. (2007) Invasion in a heterogeneous world: resistance, coexistence or hostile takeover? Ecol Lett 10: 77–94.

37. Rejmánek M (2003) The rich get richer – responses. Front Ecol Environ 1: 122–123 38. Moyle PB, Cech JJ (2004) Fishes: An introduction to ichthyology. New Jersey: Prentice-

Hall. 726 p. 39. Wilcove DS, Rothstein D, Bubow J, Phillips A, Losos E. 1998. Quantifying threats to

imperiled species in the United States. BioScience 48: 607–615. 40. Myers N, Mittermeier RA, Mittermeier CG, Da Fonsega GAB, Kent J (2000).

Biodiversity hotspots for conservation priorities. Nature 403: 853–858. 41. Olden JD, Hogan ZS, Vander Zanden JV (2007) Small fish, big fish, red fish, blue fish:

size-biased extinction risk of the world's freshwater and marine fishes. Global Ecol Biogeo 16: 694–701.

42. Lodge DM, Williams S, MacIaac HJ, Hayes KR, Leung B et al. (2006) Biological invasions: recommendations for U.S. policy and management. Ecol Appl 16: 2035–2054.

43. Pimentel D, Zuniga R, Morrison D (2005) Update on the environmental and economic costs associated with alien–invasive species in the United States. Ecol Econ 52: 273–88.

44. Vörösmarty CJ, Green P, Salisbury J, Lammers RB (2000) Global water resources: vulnerability from climate change and population growth. Science 289: 284–288.

45. Nilsson C, Reidy CA, Dynesius M, Revenga C (2005) Fragmentation and flow regulation of the World’s large river systems. Science 308: 405–408.

46. Millennium Ecosystem Assessment (2005) Ecosystems and human well being: current state and trends. Freshwater (Vol 1, Chapter 7). In : Rijsberman F, Costanza R, Jacobi P editors. Washington (D.C.): World Resources Institute. pp. 165–207.

47. Lêvèque C, Oberdorff T, Paugy D, Stiassny M, Tedsco PA (2007) Global diversity of fish (Pisces) in freshwater. Hydrobiologia: In press.

48. Center for International Earth Science Information Network (CIESIN) (2005) Gridded Gross Domestic Product (GDP). Available: http://islscp2.sesda.com/ISLSCP2_1/html_pages/groups/soc/gdp_xdeg.html.

49. Centre for Sustainability and the Global Environment (SAGE) (2002) Atlas of the Biosphere. Available: http://www.sage.wisc.edu/.

50. Rodrigues ASL, Pilgrim JD, Lamoreux JF, Hoffmann M, Brooks TM (2006) The value of the IUCN Red List for conservation. Trends Ecol Evol 21: 71–76.

51. Pont D, Hugueny B, Oberdorff T (2005) Modelling habitat requirement of European fishes: do species have similar responses to local and regional environmental constraints? Can J Fish Aquat Sci 62: 163–173.

52. MacNally R, Walsh CJ, (2004) Hierarchical partitioning public-domain software. Biodivers Conserv 13: 659–660.

53. Ihaka R, Gentleman RJ (1996) R: a language for data analysis and graphics. J Comput Graph Stat 5: 299–314.

54. Borcard D, Legendre P, Drapeau P (1992) Partialling out the spatial component of ecological variation. Ecology 73: 1045–1055.

Auteur : Fabien LEPRIEUR Titre : Les introductions d’espèces de poissons d’eau douce : distribution spatiale, déterminants et impacts sur les espèces natives Directeur de thèse : Sébastien BROSSE Lieu et date de soutenance : Toulouse, le 7 décembre 2007 Résumé : Bien que les espèces non natives de poissons d’eau douce soient bien identifiées, les facteurs déterminant leur distribution spatiale ainsi que leurs impacts sur la biodiversité sont encore peu connus, en particulier à de larges échelles spatiales. Dans ce contexte, cette thèse vise : (i) à une meilleure compréhension de l’impact des espèces non natives de poissons d’eau douce sur les espèces natives ; et (ii) à identifier les facteurs qui contrôlent la distribution spatiale des espèces non natives. Pour cela, différents niveaux de perception du processus d’introduction d’espèces et différentes échelles spatiales ont été considérés. Les résultats obtenus à l’échelle locale (135 stations au sein d’un bassin hydrographique de Nouvelle Zélande) ont permis de mettre en évidence que l’impact d’une espèce invasive (la truite, Salmo trutta L.) sur une espèce native (Galaxias anomalus Stockell) peut varier spatialement en fonction des caractéristiques abiotiques locales. En particulier, les perturbations anthropiques, telles que la modification des habitats résultant des variations de débit, ne favorisent pas forcément les espèces invasives. Ainsi, la conservation d’une espèce native menacée nécessite des mesures de gestion adaptées au contexte environnemental local. Enfin, une étude expérimentale souligne le fait qu’une espèce considérée à priori comme invasive et nuisible (le poisson chat, Ameiurus melas Raff.) doit faire l’objet d’études quant à son impact réel sur les espèces natives ; ceci afin de mettre en place des mesures de gestion adaptées aux caractéristiques comportementales et écologiques de l’espèce impactée. Les résultats obtenus à l’échelle régionale (bassin hydrographique) montrent que les introductions d’espèces de poissons d’eau douce en Europe ont conduit à une augmentation de la diversité alpha des bassins hydrographiques (c'est-à-dire une augmentation du pool régional d’espèces), mais ont provoqué une diminution de la diversité beta (homogénéisation taxonomique). L’augmentation du pool régional de poissons d’eau douce en Europe ne doit pas forcément être interprétée comme bénéfique pour la biodiversité, car les extinctions d’espèces se déroulent généralement à des échelles de temps plus grandes que le phénomène d’introduction d’espèces. Ensuite, il semblerait que la distribution actuelle des poissons d’eau douce exotiques en Europe (c.-à-d. les espèces non européennes) soit le résultat combiné d’une limitation de leur dispersion liée aux activités humaines et d’un contrôle environnemental associé aux contraintes climatiques. Enfin, il est montré que le niveau d’anthropisation d’un bassin hydrographique, et plus particulièrement sa richesse économique, est le principal déterminant de la richesse régional en espèces non natives de poissons d’eau douce. Mots clés : espèces non natives, poissons d’eau douce, assemblages d’espèces, macroécologie, homogénéisation biotique, modèles nuls, filtres environnementaux, hotspots d’invasion.

Freshwater fish invasions: spatial distribution, determinants and impacts on native species

Abstract

Although non-native fish species are well identified, the determinants of their spatial distribution and their impacts on biodiversity are poorly documented, especially at large spatial scales. In that context, this thesis aims (i) at improving our knowledge on the potential impacts of non-native fish species and (ii) at identifying the factors controlling their spatial distribution. This was achieved by considering different spatial scales.

The local-scale approach (stream reach within a river basin) first shows that local

abiotic conditions can influence the spatial distribution of an invasive species (brown trout, Salmo trutta L.) in a New Zealand river basin and hence can mediate its impact on a native species (Galaxias anomalus Stockell). Especially, anthropogenic disturbances (such as water abstraction for agricultural purposes) do not necessarily promote species invasions as reported by most previous studies. Therefore, the effective conservation of threatened native species implies the implementation of management strategies adapted to the local environmental context. Last, an experimental study reveals that a species considered as invasive (such as brown bullhead, Ameiurus melas Raff.) should be systematically studied in regards to its impact on native species. This is necessary to set up management strategies that account for the behavioural and ecological characteristics of the impacted native species.

The regional-scale approach (river basin) first shows that the introductions of non-native fish species in Europe led to (i) an increase of the size of the regional pool of species (i.e. an increase in alpha diversity) and (ii) a decrease of the taxonomic similarity between river basins (i.e. a decrease in beta diversity corresponding to a taxonomic homogenization). Such an increase of the regional pool of species should not be interpreted as beneficial for the European biodiversity. Indeed, the process of extinction itself may occur on a much longer timescale than invasions, which makes the perceived impact of invasions at the regional scale dependent on the timescale of observation. Then, the results suggest that the spatial distribution of exotic fish species across Europe (i.e. species originating from outside Europe) is related to both (i) dispersal limitation relative to selective human-mediated introductions; and (ii) environmental filtering. Last, human activities and more particularly economic activity are found to be the main determinants of fish invasions in the world’s river systems.

Key words Non-native species, freshwater fish, assemblages, macroecology, biotic homogenization, null models, environmental filtering, invasion hotspots.

Documents

UNIVERSITE PAUL SABATIER-TOULOUSE IIIthesesups.ups-tlse.fr/151/1/Leprieur_Fabien.pdfJe remercie vivement Colin Townsend de m’avoir accueilli dans son laboratoire à l’université