Optimisation Des Stratégies De Surveillance Pour La Détection Précoce D’Un Tunicier Envahissant Par L'évaluation Des Mécanismes Et Des Patrons De Recrutement

Optimisation des stratégies de surveillance pour la détection précoce d’un tunicier envahissant par l'évaluation des mécanismes et des patrons de recrutement

Thèse

Samuel Collin

Doctorat interuniversitaire en Océanographie

Philosophiae Doctor (Ph.D.)

Québec, Canada

Résumé

La mondialisation des activités humaines a grandement contribué à la dissémination et l‟introduction anthropique d‟espèces non indigènes (ENI) dans le monde. Le potentiel de dommages est tel qu‟il y a une grande pression sur les gestionnaires environnementaux pour détecter et contrôler les ENI problématiques (espèces envahissantes) avant que des impacts apparaissent. En étudiant les ENI, les écologistes peuvent examiner certains aspects de la survie des espèces, la dispersion et l‟établissement, qui, en plus de répondre à des questions fondamentales en écologie, fournissent des informations essentielles pour optimiser les efforts de gestion. Cependant, les difficultés associées à l‟étude et la détection des populations naissantes ont réduit les études quantitatives sur les processus qui précèdent les envahissements, laissant les gestionnaires de l‟environnement avec peu de directives pour détecter les ENI. Pour soulager ces défauts, cette étude apporte une évaluation quantitative des éléments déterminants du recrutement et de la dispersion du tunicier envahissant notoirement problématique, Ciona intestinalis, à l‟île du Prince-Édouard (IPE), Canada, pendant les phases précoces de l‟envahissement.

Les données de recrutement d‟une population de Ciona ont été collectées sur une période de 2 ans (2008 & 2009), ce qui a permis de modéliser la dissémination (étendue et mode) et d‟examiner les schémas de recrutement pendant l‟établissement. Ces données soulignent l‟importance d‟incorporer la dispersion, aussi bien que la variabilité environnementale, dans les stratégies de monitorage de détection précoce et démontrent comment les facteurs déterminants du recrutement changent quand une population envahissante devient grande et plus répandue. De plus, une série d‟expériences de terrain à petites échelles ont été réalisées pour évaluer les schémas de recrutement pendant la fixation. Les rôles respectifs de la lumière et de la gravité sur le comportement des larves de Ciona ont été identifiés et leur incorporation dans le design du matériel de monitorage a été discutée pour augmenter les taux de fixation et, ainsi, la probabilité de détection. Finalement, la résistance biotique envers les ENI a été examinée en recherchant les interactions les larves de Ciona et deux espèces d‟amphipodes Caprellidae, Caprella linearis (indigène) et C. mutica (non-indigène) que l‟on retrouve à l‟IPE. Cette étude montre comment la présence de caprelles diminue le recrutement larvaire de Ciona et illustre les potentielles interactions négatives entre deux ENI (C. mutica et Ciona), un phénomène rarement documenté. Dans une perspective de gestion, ces interactions négatives peuvent fournir de précieuses connaissances sur de possibles agents de contrôle biologique. De plus, cette étude présente un compte détaillé des mécanismes sous-jacents qui influencent les patrons de recrutement d‟un envahisseur problématique et discute de l‟utilité de ces découvertes pour le monitorage et la gestion future des espèces envahissantes.

iii

Abstract

The globalisation of human activity has contributed greatly to the artificial dispersal and introduction of non-indigenous species (NIS) around the world. The potential for damage is such that there is great pressure on environmental managers to detect and control problematic NIS (i.e., invasive species) before any impacts occur. By studying NIS, ecologists can examine aspects of species survival, dispersal, and establishment, which, in addition to addressing fundamental questions of ecology, provide vital information for optimizing management effort. However, the difficulties associated with studying and detecting nascent populations has restricted quantitative studies on the processes that precede invasion, leaving environmental mangers with little guidance for detecting NIS. To alleviate this shortcoming, this study provides a quantitative assessment of the determinants of recruitment and dispersal of the notoriously problematic invasive tunicate, Ciona intestinalis (henceforth Ciona), in Prince Edward Island (PEI), Canada, during the early stages of invasion.

Recruitment data from a nascent population of Ciona was collected over a two-year period (2008 & 2009), which allowed for dispersal to be modelled (range and peak) and for patterns of recruitment during establishment to be examined. These data highlight the importance of incorporating dispersal, as well as environmental variability, into early-detection monitoring strategies and demonstrate how drivers of recruitment change as the invading population becomes larger and more widespread. Additionally, a series of small-scale manipulative field studies were performed to assess patterns of recruitment during settlement. The respective roles of light and gravity on Ciona larval behaviour were identified and their incorporation into the design of monitoring equipment (to increase settlement rates and, thus, probability of detection) are discussed. Finally, biotic resistance towards NIS was examined by investigating the interactions between Ciona larvae and two species of caprellid amphipod, Caprella linearis (native) and C. mutica (invasive) found in PEI. This study shows how the presence of caprellids reduces Ciona recruitment and illustrates the potential for negative interactions between two NIS (C. mutica and Ciona), a phenomenon rarely documented. From a managerial perspective, these negative interactions can provide valuable insights to potential biocontrol agents. Moreover, this study presents a detailed account of the underlying mechanisms that influence patterns of recruitment of a problematic invader and discusses the utility of these findings for future monitoring and management of invasive species.

Tables des matières

Résumé ...... iii Abstract ...... v Tables des matières ...... vii Liste des tableaux ...... xi Liste des figures ...... xiii Acknowledgements ...... xvii Avant-propos ...... xix Chapitre 1. Introduction générale ...... 1 1.1. Le processus d‟invasion ...... 4

1.1.1. Le transport ...... 4 1.1.2. Introduction ...... 5 1.1.3. L‟établissement ...... 6 1.1.4. L‟expansion de la population ...... 7 1.1.5. Les impacts ...... 8 1.2. Gestion ...... 9

1.3. Le tunicier Ciona intestinalis ...... 11

1.4. Plan et objectifs de la thèse...... 14

Chapitre 2. Optimizing early detection of non-indigenous species: estimating the scale of dispersal of a nascent population of the invasive tunicate Ciona intestinalis (L.) ...... 16 2.1. Résumé ...... 16

2.2. Abstract...... 17

2.3. Introduction ...... 18

2.4. Methods ...... 20

2.4.1. Study site ...... 20 2.4.2. Adult survey ...... 22 2.4.3. Patterns of larval recruitment ...... 22 2.4.4. Dispersal model – kernel fitting ...... 23 2.5. Results ...... 23

2.5.1. Adult distribution ...... 23 2.5.2. Patterns of larval recruitment ...... 24 2.5.3. Dispersal model – kernel fitting ...... 25 2.6. Discussion...... 26

vii

Chapitre 3. From introduction to domination: transitional patterns of recruitment during the rapid establishment of the high-impact invader Ciona intestinalis ...... 30 3.1. Résumé ...... 30

3.2. Abstract ...... 31

3.3. Introduction ...... 32

3.4. Methods ...... 34

3.4.1. Study site ...... 34 3.4.2. Adult survey ...... 35 3.4.3. Larval recruitment ...... 35 3.4.4. Recruitment pattern analysis ...... 35 3.5. Results ...... 37

3.5.1. Adult survey ...... 37 3.5.2. Larval recruitment ...... 39 3.6. Discussion ...... 44

Chapitre 4. The role of phototactic and geotactic behaviours during larval settlement: an experimental field study of the tunicate Ciona intestinalis ...... 49 4.1. Résumé ...... 49

4.2. Abstract ...... 50

4.3. Introduction ...... 51

4.4. Methods ...... 54

4.4.1. Study site ...... 54 4.4.2. Experiment 1 – Timing of settlement ...... 54 4.4.3. Experiment 2 – Light treatments – Daytime ...... 55 4.4.4. Experiment 3 – Light treatments – Nighttime ...... 55 4.4.5. Experiment 4 – Edge effects ...... 56 4.5. Results ...... 56

4.5.1. Timing of settlement ...... 56 4.5.2. Light treatment - Daytime ...... 57 4.5.3. Light treatment – Nighttime ...... 57 4.5.4. Edge effect - Daytime ...... 59 4.5.5. Edge effect – Nighttime ...... 61 4.6. Discussion ...... 61

4.6.1. Light treatments ...... 63 4.6.2. Edge effects ...... 65 viii

Chapitre 5. Established invasive species may increase biotic resistance to new invaders: unexpected negative effect of caprellid amphipods on an invasive tunicate ...... 68 5.1. Résumé ...... 68

5.2. Abstract...... 69

5.3. Introduction ...... 70

5.4. Methods ...... 72

5.4.1. Observational evidence of the Caprella-Ciona interaction...... 72 5.4.2. Experimental demonstration of the Caprella-Ciona interaction ...... 73 5.4.3. Temporal and spatial variation of caprellid colonization….………………………………….75

5.5. Results ...... 75

5.5.1. Observational relationship of the Caprella-Ciona interaction ...... 75 5.5.2. Experimental demonstration of the Caprella-Ciona interaction ...... 76 5.5.3. Temporal and spatial variation of caprellid colonization ...... 77 5.6. Discussion...... 78

Chapitre 6. Conclusions ...... 84 6.1. Résultats principaux ...... 84

6.2. Contributions ...... 87

6.3. Perspectives et considérations futures ...... 89

References ...... 93 Appendice A...... 107 Appendice B ...... 122

Liste des tableaux

Chapitre 3 Table 3.1. Larval recruitment of Ciona intestinalis in Boughton River, Prince Edward Island, Canada, for all 8 grids of settlement plates deployed in August and September 2008 and 2009 ...... 36

Table 3.2. Student‟s t test for Ciona intestinalis larval recruitment densities for all 4 grids of settlement plates deployed in August and September 2008 ...... 41

Table 3.3. Kendall‟s Rank Correlation analysis for Ciona intestinalis larval recruitment for all grids of settlement plates for 2008 and for 2009...... 41

Chapitre 4 Table 4.1. Two-way ANOVA of Ciona intestinalis larval settlement rates on horizontally deployed 20 x 20 cm PVC settlement plates when treated with artificial light sources from above and below ...... 61

Chapitre 5 Table 5.1. Two-way ANOVA results for caprellid recruitment in Boughton River as a function of date and location of the stations throughout July and August, 2009 ...... 79

xii

Liste des figures

Chapitre 1 Figure 1.1. Le processus d‟invasion ...... 5

Figure 1.2. Options de gestion des différentes étapes du processus d‟invasion ...... 9

Figure 1.3. Le tunicier Ciona intestinalis : plans A. externe, B. interne ...... 12

Chapitre 2 Figure 2.1. Distribution of adult Ciona intestinalis in Boughton River (Prince Edward Island, Canada) in June 2008 ...... 21

Figure 2.2. Pattern of Ciona intestinalis larval recruitment on the 88-station grid of settlement plates in Boughton River, Prince Edward Island, Canada ...... 24

Figure 2.3. Predicted vs. observed settlement fitting of Weibull dispersal function for Ciona intestinalis larval recruitment ...... 25

Figure 2.4. Weibull function dispersal kernel fit to Ciona intestinalis larval recruitment ...... 26

Chapitre 3 Figure 3.1. Adult population surveys for Ciona intestinalis in Boughton River, Prince Edward Island, Canada, for June 2008 (A) and 2009 (B) ...... 38

Figure 3.2. Histogram of adult abundance for 50-m transect surveys in Boughton River, Prince Edward Island, Canada for June 2008 and June 2009 ...... 39

Figure 3.3. Larval recruitment of Ciona intestinalis on settlement plates (88 stations) in Boughton River, Prince Edward Island, Canada for Grid 1, August 2008 (A), and Grid 5, August 2009 (B) ...... 42

Figure 3.4. Global Moran‟s I values for all 8 grids of Ciona intestinalis larval recruitment (2008 & 2009) in Boughton River, Prince Edward Island, Canada ...... 43

Figure 3.5. Local Moran‟s I values for 3 grids of Ciona intestinalis larval recruitment (2008) in Boughton River, Prince Edward Island, Canada ...... 45

Figure 3.6. Local Moran‟s I values for 4 grids of Ciona intestinalis larval recruitment (2009) in Boughton River, Prince Edward Island, Canada ...... 46

Chapitre 4 Figure 4.1. Experimental set up for different light and shading treatments for Ciona intestinalis larval settlement plates ...... 56

Figure 4.2. Density of Ciona intestinalis recruits on the undersurface of settlement plates in Boughton River, Prince Edward Island, after sequential 6-hr deployments ...... 58

xiii

Figure 4.3. Average Ciona intestinalis larval settlement on upper and under surfaces of settlement plates in 4 different light treatments ...... 60

Figure 4.4. Distribution of Ciona intestinalis larval settlement on the under-surfaces of settlement plates deployed during the daytime for 2 different treatments ...... 62

Figure 4.5. Distribution of Ciona intestinalis larval settlement on the under-surfaces of settlement plates at night for 4 different light treatments ...... 63

Figure 4.6. Geotactic behaviour of Ciona intestinalis larvae during settlement in response to different lighting conditions...... 66

Chapitre 5 Figure 5.1. Log-scale plot of Ciona intestinalis abundance vs. caprellid abundance on settlement plates for the original bay-wide survey observations in Boughton River, Prince Edward Island ...... 73

Figure 5.2. Caprellid recruitment on settlement plates treated with caprellids and control plates deployed in Boughton River, Prince Edward Island ...... 76

Figure 5.3. Recruitment densities of Ciona intestinalis larvae settlement plates treated with caprellids and control plates ...... 77

Figure 5.4. Time-series of weekly recruitment of caprellids on plastic mesh collectors in Boughton River, Prince Edward Island ...... 78

xiv

For Mum, Dad and the Boyd & Collin clans

xvi

Acknowledgements

Firstly, I would like to thank the Canadian Aquatic Invasive Species Network (CAISN), Quebec Ocean, and the Laval University Biology Department, with whom this project was made possible.

I would like to express my gratitude to my supervisor, Ladd Johnson for all his help throughout this project. Thank you and your family for welcoming me into your home for my first 3 months in Quebec and helping me with finding my place in a new town. Your advice and guidance throughout this project, for both fieldwork and writing, has been invaluable and taught me a lot about the standards required for producing high-level scientific research.

I would also like to offer a special thank you to Brian Leung at McGill University, who essentially acted as my co-supervisor, although not officially, and assisted greatly with project design, data analysis and interpretation, and writing of this thesis. I have greatly enjoyed our collaboration and I hope we will join forces again sometime in the future.

My advisory committee over the last 5 years has been most helpful by providing excellent guidance and assistance with this project. Thank you Maurice Levasseur and John Himmelman of Laval University for your advice from the beginning, Chris McKindsey from the Department of Fisheries and Oceans, Canada for joining my committee late on and organising additional fieldwork to complement my work, Fréderic Maps for agreeing at the last minute to sit on my jury, and Jay Stachowicz of University California at Davis for agreeing to be my external examiner.

A key component of this project has been my collaboration with Paul Edwards at McGill University, with whom I struck not only a great working relationship but also a great friendship. Paul has been an essential part of this project and his contribution towards planning, brainstorming, statistics, editing, and scotch drinking have been outstanding.

Throughout this project there was a very large fieldwork component, which would not have been possible without the assistance of other researchers. Michael Becker, Raphäelle Descôteaux, David Delaney, Mathieu Oreiller, Valery Nearing, Maura Forrest, and Olivia Rhoades have all contributed greatly to this project by spending hours in the field collecting settlement plates, counting endless numbers of Ciona larvae, and spending months with me in the middle of nowhere. No easy task, but carried out superbly.

xvii

Working in the Johnson lab at Laval University over the years has been a fantastic experience. The insights, experience, and skills that each person brings to the table have been invaluable and I wish you all the best with your futures. Outside of the lab, I have made many great friends over the last 5 years, most notably David Paez (Dr. Peas), Joëlle Taillon, Nico Martin, and Juan Diego Urriago Suarez, who all made me feel welcome in Quebec and provided me with lots of fun memories that I will take with me wherever I go. That‟s right! It‟s time for this “fancy English guy” to move on like all of you.

A special thank you to Nicolas Le Corre and Geneviève Parent, who have been my friends since the beginning and helped me throughout my studies. Thank you for taking the time to read through and translate my thesis and if ever you need some English translation (or just feel like drinking beer), my door is always open.

And finally, a very special thank you to Véronique Lemieux for being an enormous part of my life over the last 18 months, and providing me a place of calm during the stresses experienced throughout this endeavor.

xviii

Avant-propos

Cette thèse contient une introduction et une conclusion générales (chapitres 1 et 6) ainsi que 4 chapitres principaux (2 à 5), écrit en anglais, pour soumission à des revues scientifiques. Je suis le premier auteur pour les 4 chapitres. À ce jour, deux chapitres (2 et 5) ont été soumis pour publication et les chapitres restants (3 & 4) sont actuellement en cours de préparation pour soumission. Toutes les données dans cette thèse ont été recueillies et analysées par moi-même (travail sur le terrain et en laboratoire ainsi que des statistiques), à l'exception des travaux de modélisation dans le chapitre 2 (par P. Edwards & B. Leung) et les analyses spatiales dans le chapitre 3 (assisté par P. Edwards & B. Leung). L'écriture et l'édition de tous les chapitres au sein de cette thèse ont été réalisées sous la supervision de mon directeur, le professeur Ladd Johnson.

Le chapitre 2 a été publié sous la référence: Collin SB, Edwards PK, Leung B, Johnson LE (2013). Optimizing early detection of non-indigenous species: estimating the scale of dispersal of a nascent population of the invasive tunicate Ciona intestinalis (L.). Marine Pollution Bulletin. DOI: 10.1016/j.marpolbul.2013.05.040.

Le chapitre 5 a été soumis à la revue Biological Invasions sous la référence: Collin SB, Johnson LE (in review) Established invasive species may increase biotic resistance to new invaders: unexpected negative effect of caprellid amphipods on an invasive tunicate.

Le chapitre 3 est en préparation pour une soumission dans une revue spécialisée en écologie marine. Les co-auteurs sont Paul K. Edwards, Brian Leung, et Ladd E. Johnson.

Le chapitre 4 est en préparation pour une soumission dans une revue spécialisée en écologie marine. Les co-auteurs sont Chris McKindsey et Ladd E. Johnson.

Les travaux présentés dans cette thèse représentent l'aspect empirique des données d'un projet plus large, financé par le Canadian Aquatic Invasive Species Network, dont l'objectif principal était d'intégrer le travail de terrain et de modélisation dans la conception de surveillance afin d‟augmenter la probabilité de détection précoce de l'EAE. La composante de modélisation qui complète cette thèse a été réalisée par Paul Edwards (Université McGill) pour sa thèse de doctorat. Les données du chapitre 2, ainsi que les données physiques environnementales supplémentaires prélevées dans la baie de la rivière Boughton, ont

xix

été utilisé comme modèle de recrutement en ce qui concerne la variabilité de l'environnement et de l'hydrodynamique locale. Ces études visaient à évaluer le rapport coût-efficacité de la collecte des données environnementales supplémentaires nécessaires pour accroître le pouvoir prédictif des modèles de dispersion et de propagation.

Dans le cadre de cette étude, des petits projets parallèles sur la conception de plaques de fixation et la capacité de déploiement (article en cours de révision, le premier auteur David Delaney, Université McGill), ainsi que du recrutement grégaire des larves de Ciona (projet d'initiation de recherche par Mathieu Oreiller, actuellement en cours de préparation pour la soumission dans un journal, premier auteur S. Collin) ont été réalisées. Ces projets parallèles ont été conçus pour s‟assurer que l'équipement le plus efficace était en usage pour ce projet (conception des plaques) et aussi pour aider à interpréter les modes de recrutement des larves.

Chapitre 1

Introduction générale

L‟écologie des invasions, l‟étude des causes et des conséquences de l'introduction d'organismes dans des zones extérieures à leur aire de répartition naturelle (Richardson et al. 2011), est rapidement devenue un domaine d‟étude vaste et diversifié. Le nombre d‟espèces transportées par l‟activité humaine à travers le monde ne cesse d'augmenter chaque année (Ruiz et al. 2000). Bien que les espèces introduites soient omniprésentes et potentiellement très dommageables (Pimentel et al. 2005), il s‟agit d‟un des aspects les plus sous-estimés des changements à l‟échelle planétaire (Occhipinti-Ambrogi 2003 ; Allendorf & Lundquist 2003) et d‟une menace majeure pour la conservation (Walker & Steffen 1997). Les problèmes associés aux espèces envahissantes, c.-à-d. les espèces introduites qui sont répandues et dominantes localement (Colautti & MacIsaac 2004), ont conduit à un mouvement mondial afin d‟identifier les espèces problématiques et les endroits susceptibles à l‟invasion, mais aussi afin d‟approfondir notre compréhension des diffèrentes étapes qui précèdent l‟invasion (Ricciardi & Rasmussen 1998 ; Stachowicz et al. 2002). La structure complexe, c.-à-d. à multiples facettes, du processus d'invasion nécessite une compréhension de plusieurs disciplines afin d‟intégrer les aspects écologiques (Carlton 1993), économiques (Pimentel 2005) et politiques (Gollasch 2007) dans la gestion environnementale et la conservation. La compréhension des rôles et des contributions de chacun de ces domaines est essentielle pour en arriver à une application efficace de la recherche scientifique aux pratiques futures visant à contrôler les mouvements et les impacts des espèces envahissantes.

La répartition naturelle des espèces à travers le monde a été définie par une multitude de processus qui ont conduits au fil du temps à l‟expansion, la contraction et au déplacement de leur répartition géographique (Brown & Sax 2004). Ces changements dynamiques, par l'interaction de processus écologiques, évolutifs, géographiques et climatiques, créent des frontières qui marquent une aire de répartition naturelle temporelle et spatiale. L'introduction d'organismes par l'activité humaine a artificiellement lié et élargi les frontières naturelles des espèces, ce qui a entraîné l'introduction d'espèces dans de nouveaux environnements et de nouvelles communautés, soit hors de leur aire de répartition naturelle (Elton 1958 ; Carlton 1989). Ces espèces non indigènes (ENI) peuvent occasionner des dommages sévères dans la communauté ou l‟industrie locale (Mack 2000 ; Simberloff 2000), ce qui a suscité l'intérêt des milieux universitaires, gouvernementaux et privés (Bax et al. 2001 ; Lodge et al. 2006). Bien que les impacts des ENI puissent être très variables et imprévisibles, allant d'un échec d‟invasion à une invasion dominante du milieu (Colautti & MacIsaac 2004) et à une menace d'extinction des espèces indigènes (Mack 2000), le

risque de préjudices écologiques et économiques est tel qu'il y a un grand intérêt à prévenir l‟établissement et la dissémination des ENI (Pimentel et al. 2005 ; Occhipinti-Ambrogi & Sheppard 2007).

Bien que l‟histoire des invasions ne soit pas récente (p. ex. "Clams before Colombus" [Petersen et al. 1992]), nous en sommes encore à nos débuts concernant l‟étude des invasions ainsi que des processus et des causes d‟invasion. En effet, ce n'est qu'en 1958, quand Charles Elton a publié son ouvrage classique « L‟écologie des invasions par les animaux et les plantes » (traduction libre de « The Ecology of Invasions by Animals and Plants»), que les problèmes potentiels liés aux espèces envahissantes ont été pour la première fois rapportés. À cette époque, de nombreuses études sur les espèces envahissantes ont porté sur les systèmes terrestres et en particulier sur l'introduction de mauvaises herbes et d'autres plantes. Dans les années 1980, la menace causée par les espèces aquatiques envahissantes (EAE) a commencé à attirer plus d'attention (Carlton 1985 ; 1987 ; 1989). Contrairement aux systèmes terrestres, la dispersion de nutriments, de substances, d‟organismes planctoniques et de propagules dans le milieu marin est entraînée par de petits et grands mouvements de masses d‟eau, plutôt que par des mouvements d'air ou d‟animaux. Plusieurs espèces marines ont des cycles de vie complexes qui incluent une croissance et une période de développement dans la colonne d'eau, ce qui augmente le potentiel de dispersion sur de longues distances et crée des niveaux élevés de connectivité entre les populations et les communautés, résultant un système plus « ouvert » (Carr 2003 ; Kinlan & Gaines 2003). Cette différence fondamentale entre les systèmes terrestres et marins limite l'application des pratiques de gestion terrestre dans le contrôle des ENI dans les systèmes marins. Malheureusement, notre compréhension des interactions entre les milieux biologiques, chimiques et marins a été entravée par des difficultés logistiques et des coûts élevés associés au travail dans les systèmes aquatiques. Il en résulte une mauvaise compréhension des déterminants de la dispersion, du recrutement et de la propagation des espèces envahissantes. Sans ces données, les gestionnaires de l'environnement marin ont peu d'indications pour surveiller et contrôler les EAE, ce qui conduit souvent à des efforts de gestion désorganisés et inefficaces (Hulme 2006).

Malgré ces inconvénients, la recherche quantitative sur les EAE s‟intensifie et notre compréhension des processus et des éléments clés de la survie et de l'échec d‟invasion des espèces améliore notre capacité à identifier les espèces potentiellement problématiques, les sites à risque d'invasion, et les principaux vecteurs de propagation (Carlton 1996 ; Ricciardi 1998 ; Kolar & Lodge 2001). Ces données sont essentielles pour la conception de protocoles de gestion efficaces et fiables ainsi que pour des stratégies de surveillance visant à prévenir l'introduction d'EAE, à minimiser la propagation et à réduire les impacts potentiellement négatifs, soit les principaux objectifs de la gestion d'EAE. L‟identification de faiblesses

durant les premiers stades de l'invasion, avant tout impact, qui pourrait être utilisée par des gestionnaires, est un élément clé de la recherche d'EAE. L'étude de ces espèces permet également de se pencher sur des questions de recherche plus fondamentales qui s‟étendent en dehors du domaine de l'écologie des invasions. Nos connaissances sur les origines ainsi que sur les moments et les lieux d'introduction d‟espèces nous permettent d'observer l‟adaptation et la survie de ces dernières dans de nouveaux environnements et les nouvelles interactions entre les espèces, mais aussi de documenter l'établissement de la population en temps réel, à la fois spatialement et temporellement (Lodge 1993 ; Sakai et al. 2001 ; Kinlan & Hastings 2005). Ces observations ont créé un cadre qui a permis aux scientifiques de se poser des questions fondamentales en écologie, évolution et biogéographie (Sax et al. 2005). Avec ces données, nous commençons à comprendre pourquoi certaines espèces sont plus aptes à se reproduire dans des environnements nouveaux et différents, et aussi à mieux saisir le rôle de la résistance biotique de la communauté indigène. Finalement, les résultats de ces études peuvent être réintégrés dans la gestion en améliorant la conception des équipements (Delaney et al. en révision), en optimisant la répartition des efforts de surveillance (Inglis et al. 2006 ; Edwards et al. sous presse), en identifiant les stades de vie vulnérables du cycle pour cibler les efforts de contrôle (Buhle et al. 2005 ; Edwards & Leung 2009), et en identifiant les options possibles pour le contrôle biologique, c.-à-d. l'introduction d'une espèce afin d‟en contrôler une autre (Fagan 2002 ; Secord 2003).

Afin d'élargir notre compréhension des mécanismes fondamentaux qui contribuent à la réussite des invasions, cette thèse examine les déterminants et les modèles de la dispersion et du recrutement d'une population envahissante durant les premiers stades de l'invasion. Les populations naissantes, qui sont encore petites et isolées, fournissent de rares opportunités afin d‟identifier les caractéristiques importantes de dispersion (p. ex. étendue, pic). Ces dernières sont essentielles pour la répartition des efforts de surveillance et la prévision des tendances temporelles et spatiales de propagation (Reed et al. 2000). En outre, des expériences à plus petites échelles sur le comportement des espèces et des nouvelles interactions sont présentées ainsi que leurs applications à la surveillance et au contrôle sont discutées. Tout au long de cette étude, j‟ai ciblé l‟invasion problématique du tunicier envahissant Ciona intestinalis et quand cela était possible, les méthodes et les résultats de ces études ont été discutés dans un contexte plus large à l'égard de leur contribution dans le domaine de l'écologie des invasions et de leur application à la gestion d'EAE, et en particulier, pour l‟optimalisation des méthodes de détection précoce.

1.1. Le processus d’invasion

Le processus d'invasion peut être très variable. En général, les EAE doivent franchir quatre étapes clés avant qu'un risque de préjudice écologique ou économique se produise : 1) le transport, 2) l'introduction, 3) l‟établissement, et 4) l‟invasion, c.-à-d. l‟expansion de la population à un niveau de nuisance (Sakai et al. 2001 ; Colautti & MacIsaac 2004 ; Lockwood et al. 2008 [Fig. 1.1]). Chacune de ces étapes distinctes exige différents traits biologiques pour assurer la survie des individus. Par exemple, la grande majorité des ENI ne parviennent pas à survivre au transport (Lodge 1993). De plus, survivre au transport ne garantit pas le succès de l'invasion (Kolar & Lodge 2001). En effet Selon Williamson & Fitter (1996), seulement une espèce sur dix serait capable de survivre à toutes les étapes. Une meilleure compréhension des caractéristiques requises pour survivre aux différentes étapes pourrait non seulement aider à identifier les EAE potentielles, mais pourrait également identifier où les EAE sont les plus vulnérables lors de l'effort de gestion. Ici, je présente un bref aperçu de chaque étape et identifie les obstacles rencontrés.

1.1.1. Le transport

Le transport est la première étape dans le processus d'invasion et implique la collecte et le déplacement d'une espèce à partir de sa région d'origine vers un nouveau système. Il existe de nombreux vecteurs de transport tels que : la navigation de plaisance (Clarke-Murray et al. 2013) ; l'aquaculture (Minchin 2001), le commerce des aquariums (libération accidentelle ou intentionnelle [Minchin 2001]) et les canaux (p. ex. migration lessepsienne [Zenetos et al. 2005]). Par contre, le vecteur principal est l'industrie maritime (Ruiz et al. 2000 ; Gollasch 2002 ; Minchin et al. 2009). Les organismes marins sont transportés en se fixant à la coque des navires (encrassement) ou sont capturés accidentellement lors de la collecte des eaux de ballast (Carlton 1993). La mondialisation et la modernisation de l‟industrie maritime ont grandement augmenté le nombre de vecteurs possibles (navires et ports) et réduit le temps des trajets entre les sites. Ceci a conduit à des taux de survie plus élevés et à un transport sur de grandes distances des espèces (Carlton 1989 ; Mack et al. 2000 ; Hulme 2009). Il a été estimé que des milliers d'espèces marines sont transportés continuellement (Carlton et Geller 1993).

Durant l‟étape du transport, les organismes marins peuvent être exposés à une variété de facteurs de stress dont l‟écoulement rapide de l'eau (en ce qui concerne l‟encrassement), des mauvaises conditions de l‟eau (dans les eaux de ballast) et le manque de ressources importantes (p. ex. nourriture). La survie dépend grandement de leur capacité à s'adapter et à persister à travers les longs trajets dans des conditions

Figure 1.1 Le processus d‟invasion (adapté d‟après Lockwood et al. [2008]) sous-optimales, ce qui peut nécessiter une tolérance aux conditions environnementales variables (en particulier pour la température et la salinité [Bruijs et al. 2001]) et la capacité à rivaliser avec les autres ENI en transit.

1.1.2. Introduction

À la fin du voyage, les ENI doivent être libérées dans un nouveau lieu où les conditions environnementales sont adéquates pour le recrutement. Les organismes marins sont très sensibles à leur environnement et aux facteurs abiotiques tels que la température, la salinité et le type de substrat. Ces de derniers peuvent déterminer rapidement si une espèce survivra (c.-à-d. résistance abiotique) ou non (Ruiz et al. 2000). Si l'environnement est approprié, une « fenêtre d'invasion » s'ouvrira et le nombre de recrues commencera à augmenter (Carlton, 1996). De plus, les ENI sont soumises à des pressions biotiques telles que la disponibilité alimentaire, la concurrence avec d'autres espèces et la prédation. Ceci peut influer sur 5

les taux de recrutement et, par la suite, sur l'expansion de la population. La résistance au stress biotique des ENI dépend fortement des composantes et de la diversité de la communauté autochtone. Tout comme certaines espèces sont plus efficaces que d'autres à envahir, certaines communautés ont plus de succès à résister aux envahisseurs que d'autres (Kimbro et al. 2013). Il est généralement considéré que les communautés les plus diversifiées sont plus résistantes aux ENI puisque l'utilisation de toutes les ressources est maximisée, laissant peu d'opportunités pour les espèces entrantes (Stachowicz et al. 1999). De même, les communautés moins diversifiées peuvent être plus sensibles à l'invasion puisque les niches potentielles sont plus disponibles. Toutefois, la résistance biotique envers les envahisseurs peut être significativement affectée par les perturbations de l'environnement, à la fois à grande et à petite échelles (Connell & Keough 1985 ; Altman & Whitlatch 2007), telles que les tempêtes, le développement côtier et la prédation. Dans la plupart des cas, les ENI sont introduits dans des environnements artificiels, tels que les ports maritimes et ports de plaisance, où les conditions favorisent les ENI au détriment des espèces indigènes (Glasby et al. 2007). Ces sites agissent comme de petits sanctuaires protégés pour le recrutement des ENI où il y a une faible résistance de la communauté autochtone et une possibilité d‟accumulation des populations envahissantes. À ce stade de l'invasion, les ENI doivent s'adapter à un niveau de conditions environnementales potentiellement sous-optimales (c.-à-d. tolérance aux conditions environnementales variables), utiliser les ressources à leur disposition (c.-à-d. être opportunistes), être plus compétitives que les espèces indigènes dans l‟utilisation des ressources locales et accroitre la taille de la population locale. Par conséquent, les environnements artificiels trouvés dans la plupart des sites d'introduction offrent des conditions idéales pour la survie d'ENI.

1.1.3. L‟établissement

Si toutes les conditions (biotiques et abiotiques) du milieu récepteur permettent la survie, les EAE peuvent établir une population principale plus grande à partir de laquelle elles peuvent se propager. La croissance démographique est un aspect clé de la survie et de l‟expansion de la population (Lockwood et al. 2008) puisqu‟elle apporte une stabilité et réduit les risques associés aux perturbations environnementales (p. ex. les tempêtes), aux pressions démographiques (p. ex. effet d‟Allee soit une densité de population trop faible pour une reproduction efficace [Courchamp et al. 1999]), et aux pressions biotiques (p. ex. la prédation). Pour parvenir à l'établissement, il est nécessaire que les ENI soient introduits pendant le même évènement (pression de propagules [Colautti et al. 2003.]), qu‟il y ait des introductions multiples de la même espèce (Ruiz et al. 1999 ; Forsyth & Duncan 2001), ou encore que les espèces soient capables de capables de se reproduire de façon asexuée, comme la fragmentation par exemple.

La vitesse d'établissement est importante pour la survie de la population, car elle réduit, et élimine en dernier lieu le recours à de futures introductions afin de reconstituer la population. Le temps de latence entre l'introduction et l'établissement est variable et dépend grandement du niveau de résistance au stress biotique de la communauté autochtone et des caractéristiques de reproduction des EAE. Ricciardi (1998) a identifié les principaux traits biologiques que les EAE possèdent fréquemment et qui aident à l'établissement rapide :

 court temps de génération,  taux de croissance rapide,  maturation sexuelle rapide,  capacité de reproduction élevée et  comportement grégaire (attraction conspécifique).

Une fois que l'EAE a été en mesure d‟établir une grande population indépendante, elle peut commencer à disperser un bon nombre de propagules dans le milieu naturel environnant, ce qui augmente la probabilité d'établir des populations satellites à partir du point d'introduction (p. ex. port) et d'interagir davantage avec la communauté autochtone.

1.1.4. L'expansion de la population et la propagation dans l'environnement immédiat

Le point où une EAE devient vraiment envahissante n'est pas facile à définir. Comme le moment de l‟introduction n'est pas toujours évident, le moment de l‟établissement est difficile à distinguer sans surveillance poussée et il est généralement inconnu quand une EAE commence à se répandre dans l'environnement (Kolar & Lodge 2001). Pour beaucoup d‟EAE, il n‟est pas possible d‟établir plus qu‟une population autosuffisante et de se propager loin du point d'introduction, un phénomène connu sous le nom « invasion pinning » (Keitt et al. 2001). Toutefois, pour l'EAE capable d'atteindre un seuil suffisamment élevée de propagules dans l‟environnement, la probabilité d'établir des populations satellites et de se propager plus loin du point d'introduction augmente également. De façon similaire à la phase d'établissement de la population envahissante originale, les populations satellites ont à surmonter la résistance biotique et les problèmes associés à de faibles densités (p. ex. effet d‟Allee). Les populations satellites peuvent aussi dépendre de l‟apport en propagules de la population d'origine. Ces propagules sont soumises à des mécanismes plus stochastiques dans la dispersion naturelle tels que le transport passif par

les courants et les fluctuations des conditions environnementales (Kinlan & Gaines 2003 ; Kinlan & Hastings 2005). Ces derniers peuvent engendrer de la variabilité dans l'apport de propagules, ce qui influe sur le taux de recrutement, sur la croissance démographique et sur la propagation (Ricciardi 1998). Par exemple, la dispersion sur de courtes distances est moins exposée aux courants d'eau et par conséquent peut être plus efficace pour mettre en place rapidement une population locale par l'autorecrutement. Toutefois, pour les espèces à dispersion sur de longues distances, les propagules peuvent entraîner un recrutement de faible densité, mais aussi une plus grande probabilité d'établir une population satellite loin du point d'origine d‟introduction (Pechenik 1999).

1.1.5. Les impacts

L‟établissement de populations larges et dominantes peut avoir de multiples répercussions sur la communauté indigène. En plus de la compétition avec les espèces indigènes pour les ressources telles que la nourriture et l'espace, les ENI sont capables d'hybridation, d‟introduction de parasites, d‟altération de la structure de l'environnement (p. ex. les ingénieurs écologiques tels que les moules [Crooks 2002 ; Cuddington & Hastings 2004]) et de modifications de la composition démographique de la communauté autochtone en s'attaquant aux espèces indigènes et en fournissant d'autres proies pour les prédateurs indigènes (Ruiz et al. 1997 ; Sakai et al. 2001). Les processus à long terme de la succession de la communauté visent à atteindre un état d'équilibre entre les interactions proie-prédateur et l‟allocation des ressources (Connell & Slatyer 1977). Cela peut être perturbé par la présence des ENI (p. ex. prédateurs des poissons-lions dans les Caraïbes [Albins & Hixon 2008]), dont les effets ne sont pas immédiatement visibles, mais qui peuvent avoir des effets irréversibles qui sont souvent reconnus trop tard par les gestionnaires (Bax et al. 2003).

En ce qui concerne les EAE en particulier, leurs impacts ne se limitent pas à la communauté autochtone. De nombreuses industries maritimes ont également subi un impact économique (Pimentel 2005). Dans la plupart des cas, l'impact le plus important vient de l'EAE sessile qui se fixe sur les surfaces dures et qui est capable d'établir des populations denses très rapidement. Ces espèces peuvent être particulièrement problématiques pour les ports, l'aquaculture et les plaisanciers qui doivent assumer des coûts supplémentaires pour le nettoyage de l'équipement et l'élimination de ces espèces. Le niveau élevé de variabilité dans le processus d‟invasion rend difficile la prédiction des impacts qu‟elles auront sur le milieu récepteur. Bien que des tentatives aient été faites (p. ex. Branch & Steffani 2004), il est difficile pour les gestionnaires d‟identifier les espèces qui seront problématiques et où elles seront problématiques.

1.2. Gestion

À l'heure actuelle, la plupart des études sur les EAE mettent l‟accent sur la documentation des impacts (p. ex. le coût des dommages environnementaux) plutôt que d'identifier des solutions qui aideraient la prise de décisions des gestionnaires (Hulme 2006). En ciblant l'identification des facteurs de stress et des faiblesses associées à chaque étape de l'invasion, il est possible d‟identifier les points clés pour la gestion des EAE, là où elles sont particulièrement vulnérables et sensibles. La nature « ouverte » des systèmes marins, avec peu de barrières à la dispersion, rend le contrôle de la dispersion et du recrutement d'EAE difficiles (Carr 2003), et ce, même si quelques succès ont été rapportés (voir Hewitt & Campbell 2007). Cela est particulièrement vrai lorsque la population envahissante devient considérablement répandue avec de multiples populations satellites puisqu‟elle devient plus difficile à localiser et à circonscrire. La pratique la plus efficace pour la gestion consiste donc à adopter une approche proactive en empêchant l'introduction initiale de l'EAE en traitant les vecteurs de transport (Finnoff et al. 2007).

Figure 1.2. Options de gestion des différentes étapes du processus d‟invasion (Olenin et al. 2011)

L‟identification de régions contenant beaucoup d‟EAE potentielles et de vecteurs principaux d'EAE est essentielle pour la maîtrise et la prévention des introductions (Ricciardi 1998), mais en dépit des mesures de prévention actuelles, p. ex. l'échange de l'eau médio-océanique de ballast et le nettoyage des coques, les

introductions se produisent encore (Lodge et al. 2006). Par conséquent, la deuxième meilleure option s‟avère la détection précoce et la réponse rapide. Cela nécessite la détection de la présence d'une EAE au point d'introduction, soit avant la dispersion généralisée, et avant que le recrutement ne se soit produit. La détection précoce de l‟EAE, c.-à-d. lorsque la population est encore petite et relativement isolée, peut réduire les coûts et les efforts requis pour une gestion réussie et ainsi créer des possibilités de confinement, de contrôle et potentiellement d'éradication (Fig. 1.2 [Myers et al. 2000 ; Simberloff 2003 ; 2011 Olenin et al.]). Pour augmenter la probabilité de détection précoce, les gestionnaires doivent s‟adresser les questions suivantes avant qu‟une stratégie de surveillance soit mise en œuvre :

 Quelles sont les régions avec un grand risque d‟invasion (p. ex. grands ports)?  Il y a-t-il un historique problématique d‟EAE susceptible d'être transportées dans cette région?  Si c'est le cas, le milieu récepteur favorise-t-il le recrutement des espèces identifiées?  Que connaissons-nous de l‟histoire de vie de l‟espèce cible et est-ce que certains traits de cette dernière peuvent être utilisés dans une stratégie de surveillance?

Toutefois, la conception de méthodes optimales pour la détection précoce est un processus complexe (Hulme 2006) en raison des difficultés inhérentes à la détection de quelques individus et de la variabilité naturelle chez les espèces marines. Évidemment, en augmentant les efforts, nous augmentons également la probabilité de détection, mais aussi de beaucoup les coûts. Par conséquent, nous devons trouver un équilibre où les efforts de conservation et les coûts sont minimums tout en conservant une probabilité de détection très forte (Baxter & Possingham 2010). Pour y parvenir, nous devons veiller à ce que la répartition spatiale de l'effort de surveillance soit correctement établie et que l‟équipement utilisé soit conçu de façon optimale (p.ex. matériaux, structure) et déployé (p. ex. profondeur, durée) pour favoriser le recrutement de l'espèce cible. Cela nécessite une bonne compréhension du comportement, des déterminants de recrutement et du potentiel de dispersion de l‟espèce. Comme scientifiques des invasions, il est de notre responsabilité de rendre ces données disponibles aux gestionnaires de l'environnement dans un format pratique. Il est donc important d‟augmenter le nombre d'études quantitatives sur les EAE actuelles, qu‟elles soient problématiques ou non, qui se penchent sur des aspects plus fondamentaux de la dispersion et du recrutement des populations envahissantes plutôt que sur des études d‟observation.

La nature idiosyncrasique du processus d'invasion crée des événements spécifiques au site et à l‟espèce (Lonsdale 1999) : une invasion réussie sur un site ne garantit pas une invasion réussie à un autre, de même qu‟un site sensible à l'invasion d'une espèce ne signifie pas qu'il sera sensible à toutes les espèces. Par

contre, il est un bon indicateur de la capacité d‟invasion des espèces. Par conséquent, une approche « one- size-fits-all » (taille unique) ne peut pas être appliquée à la gestion de l'EAE. Une meilleure compréhension des mécanismes plus fondamentaux impliqués dans l'invasion doit être acquise. Ces données sont aussi essentielles pour améliorer les aspects plus théoriques de gestion environnementale (p. ex. la modélisation) qui sont de plus en plus importants pour prédire plus particulièrement la propagation des EAE (Lockwood et al. 2008).

1.3. Le tunicier Ciona intestinalis

Le tunicier solitaire Ciona intestinalis, ci-après nommé Ciona, est un invertébré marin sessile, connu pour être une espèce envahissante à succès mondial (Lambert et Lambert 1998 ; McDonald 2004; Robinson et al. 2005 ; Uribe et Etchepare 2002 ; Howes et al. 2007). En dépit d'être bien étudiée, le Ciona est une espèce cryptogénique (une espèce d'origine inconnue), généralement considérée comme étant originaire du nord de l‟Europe. Au cours des 20 dernières années, la présence de Ciona dans l'est du Canada, particulièrement en Nouvelle-Écosse et à l'Île-du-Prince-Édouard (Î.-P.-É.), s'est avérée particulièrement problématique pour l'industrie de la mytiliculture. En effet, le Ciona est capable de former des colonies denses sur les coquilles des moules d'élevage et des équipements associés, ce qui entraîne une perte de moules (« sloughing »), une réduction de la croissance des moules en raison de la concurrence directe pour la nourriture, des dommages et pertes matérielles ainsi que des coûts de nettoyage élevés (Thompson & McNair 2004 ; Locke et al. 2007 ; Sephton et al. 2011). Le risque continu de propagation dans l'est du Canada a porté Ciona en première ligne. De nombreux efforts ont été déployés afin de contrôler les impacts (Paetzold et al. 2011 ; Davidson et al. 2012) et d'identifier les régions à risque d'invasion (Therriault & Herborg 2007).

Les Ciona ont un corps gélatineux souple (15 cm de longueur et 3 cm de diamètre [Carver et al. 2003]) qui se compose de deux siphons inhalant et exhalant dont la principale fonction est de filtrer l'eau pour s‟alimenter (Fig. 1.3). Ils ont un taux de fécondité élevé, estimé à 10 000 ovules par adulte (Petersen & Svane 1995) et sont classés comme géniteurs par diffusion. Les gamètes (spermatozoïdes et ovules) sont relâchés séparément dans la colonne d'eau, généralement à l'aube (Lambert & Brandt 1967), et la fécondation est externe. Les ovules sont libérés dans des chaînes collantes qui se fixent facilement sur les structures avoisinantes (Petersen & Svane 1995). Les œufs fécondés se développent en simples larves de têtard (Bone 1992) dont environ 50% s'échappent de la chaîne d‟ovules et se libèrent dans la colonne d'eau

Figure 1.3. Le tunicier Ciona intestinalis (Cirino et al. 2002) : plans A. externe B. interne.

(Petersen & Svane 1995) devenant ainsi des larves planctoniques. Ces larves planctoniques sont ensuite soumises à des courants qui leur permettent de se disperser plus loin de la population adulte. Toutefois, les larves de Ciona sont lécithotrophes et ne s‟alimentent pas, ce qui entraine une source d'énergie finie et par conséquent, une durée maximum de 6 jours (Svane & Havenhand 1993) dans la colonne d'eau afin de trouver un substrat approprié pour la fixation.

Pendant la période planctonique, les larves de Ciona ont développé une suite de comportements en réponse à différents stimuli environnementaux, qui favorisent la dispersion et augmentent la probabilité de localiser des substrats appropriés. Bien qu‟elles soient de petites tailles (environ 1 mm de longueur), les

larves de Ciona sont capables de nager activement (Bone 1992), cette propriété leur permettrait de modifier leur position verticale dans la colonne d'eau (Nakagawa et al. 1999). Durant les premières étapes de la dispersion des larves, elles préfèrent nager à la proximité de la surface de l'eau probablement pour favoriser leur dispersion. Cependant, lorsque les larves deviennent « compétentes » (prêtes à se fixer), leurs comportements changent et elles commencent à descendre vers le fond à la recherche d'obscurité sous des surfaces en surplomb (Rius et al. 2010). Ces comportements sont contrôlés par deux organes : l‟otolithiques (récepteur de la pesanteur) et l‟ocelle (récepteur de la lumière) qui permettent aux larves de faire la distinction entre le haut et le bas et la détection des variations de l'intensité lumineuse (Nakagawa et al. 1999). Il est généralement accepté que les larves de Ciona sont positivement phototactique (c.-à-d. attirées par la lumière) et négativement géotactique (c.-à-d. mouvement vers le haut) près de la surface de l‟eau, mais deviennent pour négativement phototactique et positivement géotactique lorsqu‟elles sont à la recherche d‟une surface sombre pour s'installer. Après leur fixation, les larves commencent leur processus de métamorphose, soit la résorption de la queue et le développement des siphons péribranchiaux (Berrill 1947). Ciona a un taux de croissance rapide et devient mature sexuellement dès deux mois (Dybern 1965). Cette phénologie permet aux populations de Ciona d'avoir deux pics de recrutement au sein de la même saison de reproduction. Les premières larves fixées deviennent matures sexuellement dans la même saison, ce qui peut entraîner une croissance rapide de la population.

Ciona possède de nombreuses caractéristiques essentielles pour un établissement réussi, soit une grande fécondité, un taux de croissance rapide, une maturation rapide et une courte période de dispersion. Ceci a d‟ailleurs contribué à son succès mondial comme espèce envahissante. Son statut actuel dans l'est du Canada comme espèce envahissante a fait de lui une espèce importante et pertinente à étudier. En plus, les caractéristiques de son cycle de vie en font une espèce facile à étudier puisque les larves peuvent se fixer sur des plaques suite à une courte phase planctonique et les adultes sont sessiles, ce qui permet un suivi temporel facile. La simplicité des méthodes requises pour la surveillance des adultes et des larves permet donc la collecte de grandes quantités de données relativement facilement.

1.4. Plan et objectifs de la thèse

L'objectif principal de cette étude était d'évaluer les mécanismes sous-jacents qui influencent les modes de recrutement larvaires et de dispersion de Ciona, à la fois à grande et à petite échelles, au cours des premiers stades de l'invasion et d‟appliquer ensuite ces données à des stratégies actuelles de surveillance. Pour ce faire, une série d‟observations et de manipulations ont été réalisées à Boughton River (Î.-P.-É.), Canada. Il s‟agit d‟un site d'une population envahissante de Ciona récemment détectée (2007, comm. pers. Garth Arsenault, UPEI). Ce travail a été réalisé en étroite collaboration avec un groupe de recherche de l'Université McGill (Paul Edwards et le professeur Brian Leung), qui a apporté modélisation pour l'analyse du recrutement à grande échelle. Bien que cette étude ne porte que sur une seule population de Ciona, dans un site unique, les approches de la collecte de données et leurs applications pour la gestion des Cionas dans l'est du Canada sont discutées, ainsi que les contributions élargies pour la gestion et l'écologie des espèces envahissantes.

Chapitre II :

Optimisation de la détection précoce des espèces non indigènes : estimation de l'ampleur de la dispersion d'une population naissante du tunicier envahissant Ciona intestinalis (L.)

L'objectif principal de cette étude était de modéliser le potentiel de dispersion (c.-à-d. étendue et pic) de Ciona en déterminant la répartition des adultes et en recueillant des données de recrutement de larves d'une population naissante. Ces données ont été utilisées afin d‟identifier les stratégies optimales de surveillance pour la détection précoce, plus spécifiquement la répartition spatiale de l'effort, et d'évaluer la probabilité de contenir la propagation du Ciona autour de l‟ Î.-P.-É., Canada.

Chapitre III :

De l'introduction à la dominance : les modèles transitoires de recrutement au cours de l’établissement rapide et problématique de l’envahisseur Ciona intestinalis.

L'objectif principal de cette étude était de documenter les tendances spatiales et temporelles de la répartition des adultes et du recrutement des larves d'une population envahissante de Ciona à partir du point d'introduction (petit et isolé) jusqu‟à l'établissement généralisé. Ces données ont été utilisées pour identifier les facteurs importants de recrutement durant les premiers stades de l'invasion, avant l'établissement, et informer sur la répartition spatiale de l'effort de surveillance. 14

Chapitre IV :

Le rôle des comportements phototactiques et géotactiques pendant la colonisation larvaire : une étude expérimentale sur le terrain sur le tunicier Ciona intestinalis.

L'objectif principal de cette étude était d'identifier les rôles respectifs des comportements phototactiques et géotactiques des larves de Ciona lors de la fixation et d'évaluer l‟intégration de ces comportements dans la conception d‟équipement de surveillance afin d‟améliorer la probabilité de détection précoce.

Chapitre V :

L’établissement des espèces envahissantes peut augmenter la résistance biotique de nouveaux envahisseurs : effet négatif inattendu d'amphipodes caprelles sur un tunicier envahissant.

L'objectif principal de cette étude était d'évaluer le potentiel des ENI à contribuer et à renforcer la résistance au stress biotique de la communauté autochtone vers d'autres ENI. Les interactions entre deux espèces d‟amphipodes caprelles, Caprella linearis (indigène) et C. mutica (envahissante), et des larves de Ciona ont été examinées au cours d‟une observation initiale à grande échelle et ultérieurement, lors de manipulations à petite échelle.

Chapitre 2

Optimizing early detection of non-indigenous species: estimating the scale of dispersal of a nascent population of the invasive tunicate Ciona intestinalis (L.)

2.1. Résumé

La connaissance de la dissémination et de l‟établissement au cours des stades précoces d‟envahissement est essentielle pour distribuer l‟effort de monitorage, détecter les populations naissantes et prédire la dispersion. Cependant, la rareté de ce type de données fournit peu d‟indications pour les programmes de monitorage. Dans cette étude, nous présentons des données sur la distribution des adultes et les schémas subséquents du recrutement larvaire d‟une population naissante du tunicier envahissant Ciona intestinalis à l‟île du Prince-Édouard, Canada. Les modèles de niche existants indiquent que l‟ensemble du site d‟étude est propice au recrutement, suggérant une probabilité égale de détection sur le site. Au contraire, nous avons trouvé un schéma de recrutement larvaire hétérogène, incluant des zones sans recrutement. En modélisant la forme de la dispersion, nous avons montré que Ciona n‟est pas capable de se disséminer naturellement entre les baies, empêchant une dispersion supplémentaire et fournissant des directives pour de futurs monitorages. Nos résultats soulignent également comment les modèles à larges échelles, bien qu‟importants, manquent les schémas à petites échelles qui sont essentiels pour le monitorage et la détection précoce des espèces envahissantes.

2.2. Abstract

Knowledge of dispersal and establishment during the early stages of invasion is essential for allocating monitoring effort, detecting nascent populations and predicting spread. The scarcity of these data, however, provides little guidance for monitoring programs. Here we present data on the adult distribution and the subsequent pattern of larval recruitment from a nascent population of the invasive tunicate Ciona intestinalis in Prince Edward Island, Canada. Existing niche models indicate the entire study site is suitable for recruitment, suggesting an equal probability of detection throughout the site. In contrast, we found a heterogeneous pattern of larval recruitment, including areas of zero recruitment. By fitting a dispersal kernel, we show Ciona is not capable of naturally dispersing between bays, restricting further spread, and provide guidance for future monitoring. Our results also highlight how large-scale models, although important, lack the small-scale patterns essential for monitoring and early detection of invasive species.

2.3. Introduction

Non-indigenous species (NIS) have created serious challenges for ecosystem conservation and human industries across the planet (Mack et al. 2000; Hooper et al. 2005). The impacts of NIS are highly variable and unpredictable, but such is the potential for ecological and economic harm that there is great interest among ecologists to prevent the establishment and spread of NIS (Pimental et al. 2005; Occhipinti- Ambrogi and Sheppard 2007). The invasion process is often considered to consist of four stages: (1) transport and introduction; (2) local establishment; (3) regional spread; and (4) population growth to nuisance levels (MacIsaac et al. 2002; Lockwood et al. 2005). Intervention at the first stage is obviously the best strategy because once established, eradication or control of NIS is costly at best or impossible at worst (Myers 2000). Still, in spite of efforts aimed at preventing transportation of potential invasive species, introductions continue to occur (Lodge et al. 2006). Unfortunately, the subsequent stages of such invasions (i.e., local establishment and spread) usually pass unnoticed due to a lack of effective monitoring programs (Kraft and Johnson 2000; Vander Zanden and Olden 2008). Such programs are critical as they can provide information at a variety of levels, including patterns of invasions (e.g., identification of vectors and invasion “hotspots”) and rates of range expansion. Most importantly, early detection can provide a first warning of an initial invasion and offers the best opportunity for any possible eradication efforts as populations are both small and localized (Simberloff et al. 2005). Moreover, identification of nascent populations is important for controlling NIS by detecting and suppressing fringe populations (Moody and Mack 1988), thereby slowing the rate of spread (Sharov et al. 1998) and improving the efficacy of preventative measures (e.g., quarantines). Studying the transitional phase between introduction and establishment can thus provide key information for management, especially if new NIS can be detected early.

Although conceptually a simple idea, the early detection of nascent populations is inherently difficult, and the temporal window between introduction and widespread dispersal can be short, adding additional pressure on monitoring strategies. These same factors have also made research in this area difficult, and thus there is a paucity of studies conducted during these early stages (Marsico 2010). In the simplest terms, the problem is one of sampling. There are two basic aspects to any monitoring effort: the overall effort (e.g., number of samples) and the design of the sampling effort (i.e., the spatial and temporal distribution of effort). With regard to the former, more effort creates a greater probability of detection, but with obvious associated costs. The latter is less obvious and requires knowledge of both the likely time and location of introductions. However, it is unlikely that the initial colonists will arrive in numbers

sufficiently high to be detected, and thus most monitoring efforts functionally focus on detecting the subsequent first generations (e.g., Kraft and Johnson 2000; Hayes et al. 2005). Although much literature focuses on searching where initial introductions are most likely to occur (see review by Campbell et al. 2007), knowledge of the intergenerational dispersal of propagules becomes essential information for optimizing the sampling design, especially when points of potential introduction are known or suspected.

In aquatic systems, the unintentional introduction of most NIS occurs through a limited number of vectors, principally though shipping, boating and aquaculture activities. Although introduction can occur while vectors are in transit (e.g., sloughing of hull-fouling organisms, release of ballast water) or through less predictable vectors of secondary spread (e.g., recreational boating; Kelly et al. 2013), the principal vectors typically operate at or from specific locations (e.g., ports, boat launches) and thus, the possible points of introduction are generally well known. Natural dispersal is often considered an unmanageable vector (Kanary et al. 2011), but if the dispersal potential is known, sampling stations can then be positioned at appropriate distances away from suspected points of introduction. However, patterns of dispersal are, in reality, very poorly known, even for sessile benthic organisms due to the small size of the dispersive larval stage (but see Olson 1983; Stoner 1992). From a theoretical standpoint, dispersal should assume a kernel function (e.g., integro-difference models, Kot et al. 1996) that describes patterns of recruitment in relation to the location of the adult population. This approach has been successfully used in landscape ecology to model dispersal in plants (Fitt et al. 1987) and insects (Roques et al. 2008), albeit usually over large spatial scales (Siegel et al. 2003). However, it is not clear how applicable these techniques are when working in the marine environment or at the finer spatial scales over which the dispersal of some newly-introduced AIS typically occurs. Moreover, the extent to which water motion (e.g., tides, currents) influences the distribution of propagules is unknown (but see Kanary et al. 2011). For effective monitoring, it is thus essential to know the scales of effective dispersal (i.e., the range of dispersal) as well as the degree and predictability of any heterogeneity within that range in order to appropriately assign sampling effort. Otherwise, sampling will be haphazardly assigned (i.e., no guidance from predictable heterogeneity in recruitment within the dispersal range) or take place at too large a scale (i.e., outside the dispersal range).

To investigate the importance of dispersal when designing monitoring strategies for NIS, we studied the very early stages of invasion (1 year after known introduction) of a known problematic invasive tunicate, Ciona intestinalis (Lambert and Lambert 1998; McDonald 2004; Robinson et al. 2005; Uribe and Etchepare 2002; Howes et al. 2007), hereafter after referred to as Ciona, in Prince Edward Island (PEI), Canada. Ciona is of particular importance in PEI as it heavily fouls local aquaculture facilities, creating

additional costs to mussel farming practices (Thompson and MacNair 2004). A large-scale species distribution model (SDM) for Ciona has already been applied to Canadian waters (Therriault and Herborg 2008), which identified the eastern and northern coastlines of PEI as at-risk from Ciona invasion based upon temperature, salinity, dissolved oxygen and chlorophyll a data. Although this work correctly predicted the eastern coastline of PEI as at-risk (Ciona was first detected in Brudenell River, eastern PEI in 2004 [Locke et al. 2007] and has since spread to adjacent bays), it provides little in terms of guidance for managers monitoring for Ciona within this region or elsewhere. Without further knowledge on the dispersal range and patterns of recruitment, monitoring efforts will be poorly allocated and lack the accuracy required for early detection.

In this study we used information on the distribution of a small adult population and subsequent larval recruitment within a small bay in eastern PEI to investigate the dispersal potential and patterns of recruitment of Ciona. Using these data we assess the problems associated with monitoring allocation at small scales and offer guidance on key dispersal characteristics of Ciona to future monitoring programmes in areas at risk of invasion by Ciona. At present there are other similar bays within the same at-risk region that are not invaded and, therefore, information to guide the design of monitoring programs is required. More generally, we feel our approach can be applied not only to other species of tunicates but to other invasive species as well.

2.4. Methods

2.4.1. Study Site

Fieldwork was conducted on the eastern coast of PEI in Boughton River, which, despite its name, is a small semi-enclosed bay, approximately 6 km in length, widening from west to east (maximum width ~ 2 km) but with only a small entrance (approximately 100 m wide) to the open sea (Fig. 2.1). Ciona was first detected in Boughton River in the fall of 2007 (2 recruits, G. Arsenault, pers. comm.), which provided a rare opportunity to collect high-resolution recruitment data during the very early stages of an invasion. The water regime within the bay is governed by a semi-diurnal tidal system (tidal period of approximately 12 hr), which creates a bi-directional, east-west water flow within the bay. Towards the mouth of the bay, current flow follows a man-made channel that navigates a shallow region towards the middle of the bay. However, west of this shallow region, the cross-section of the bay is more typical of a river cross-section, with deeper water found in the middle of the bay. Due to the frequency of tidal changes (every 6 hr) and

Figure 2.1. Distribution of adult Ciona intestinalis in Boughton River (Prince Edward Island, Canada) in June 2008 based on 50-m underwater surveys (n = 33). Larger circles represent larger populations and clear circles represent an absence of adults. Light grey rectangles within the bay represent mussel leases, the scored area represents a shallow region (1-2 m depth), and the black arrow signifies the man-made channel.

the larval planktonic duration of Ciona (minimum of 24 hrs), we expected larvae to be exposed to several tidal changes and thus be fairly diffused throughout the bay.

Temperature and salinity data (collected on 22 Aug 2012 within the bay using an SBE 19plus V2 Seacat CTD Profiler at 2-m depth) ranged from 20.5 – 24.4 ºC (mean = 22.2 ºC) and 25.4 – 27.7 ppt (mean = 27.1 ppt). Both of these ranges fall well within the published temperature (-1 to 30ºC) and salinity (12 – 40ppt) tolerances of adult Ciona (Dybern 1965, 1967; Carver et al. 2006; Therriault and Herborg 2007), indicating the entirety of Boughton River was suitable for Ciona recruitment. The general pattern for temperature was an increase towards the western end of the bay, away from the mouth of the bay, and salinity values showed the opposite trend with higher values to the east. However, the profiler casts also indicate that freshwater input from surface runoff was minimal, affecting only the top meter of the water

column in the western end of the bay. There were no obvious effects of freshwater input in the main body of the bay or at the 2-m sampling depths.

The natural substrata in Boughton River is predominantly sand and silt, which are unsuitable for Ciona - therefore Ciona was only observed on artificial substrata, as is seen generally in PEI (Locke et al. 2007). The majority of suitable artificial substrata in Boughton River is provided by the mussel aquaculture facilities (i.e., “leases”, including anchors, buoys, lines, and mussel “socks”) located in the centre of the bay, a small loading dock towards the west, and a small harbour near the entrance of the bay (Fig. 2.1). The limited substrata available for settlement constrained the possible locations for colonization, which improved our ability to locate and map the adult population.

2.4.2. Adult Survey

Underwater surveys were performed in 2008 to map the distribution of the adult population on mussel leases. Thirty-three surveys were conducted in June before Ciona normally becomes reproductively active (Ramsay et al. 2009) and thus documented the distribution of the first reproductively-active adult population of the year. Each survey consisted of a 50-m transect within a mussel lease surveyed by two divers starting from a haphazardly-chosen location. Each transect ran horizontally through the water following rows of vertically-hung mussel socks, which were completely inspected, including the natural substratum below the mussel lines. The total number of Ciona present was recorded by both divers, and the two counts averaged. If mussel leases had recently been harvested (i.e., mussel socks removed), 50 m of the remaining in situ equipment (e.g., ropes, buoys and anchor lines) was surveyed.

2.4.3. Patterns of Larval Recruitment

Unlike previous studies using SDMs where different habitats were surveyed (e.g., Inglis et al. 2006), we deployed a large grid of quasi-systematically-placed settlement plates (88 stations) at a constant depth (2 m) to collect recruitment data. Plates were deployed for 2 weeks beginning on 8 Aug 2008 both within and outside of the mussel leases (Fig. 2.2). Each station contained a 20x20 cm2 PVC settlement plate (roughened with #50 sandpaper to enhance settlement) suspended horizontally from a surface buoy and anchored to the bottom by a 9-mm polypropylene cord attached to a cement block; an intermediate weight kept the top half of the line vertical and thus the plate horizontal. After collection, the plates were

transported to the laboratory in seawater and the total number of larvae settled on the underside of each plate counted using a stereomicroscope at 40 × magnification within 48 h of collection.

2.4.4. Dispersal Model - kernel fitting

Dispersal kernels describe the probability of dispersal of propagules at different distances from adult sources (Nathan 2006). To determine if the pattern of larval recruitment could be modelled as a function of distance and adult population size, we fit:

||A Pj c A i f( d i j ; v) i1

Where Pj is the recruitment on plate j, c is a fecundity scalar, |A| is the total number of adult populations in the whole bay, Ai is the size of the adult population i, di-j is the distance between plate j and adult population i and v is the vector of free parameters to the function f(d). For the function f(d) we used the Weibull distribution, as it has a very flexible shape that may approximate a variety of dispersal kernels (Morales and Carlo 2006), including both hump-shaped and exponential decay shapes. The c and v parameters were fit automatically using R (R Development Core Team 2012). We used bootstrapping to generate the 95% confidence intervals for the fit parameters. From the original set of sampling stations, we drew 10,000 random subsamples with replacement. We refit the kernel parameters for each random subset to generate a distribution of fit parameters.

2.5. Results

2.5.1. Adult Distribution

Ciona was observed in 10 of the 33 surveys with a mean number of 2.5 ± 1.4 (S.E.) per transect, with a maximum of 40. Although Ciona was present in several locations, the highest densities were found to be concentrated in the northern part of the bay (Fig. 2.1). Ciona were never observed on the natural substrata within and surrounding the mussel leases.

Figure 2.2. Pattern of Ciona intestinalis larval recruitment on the 88-station grid of settlement plates in Boughton River, Prince Edward Island, Canada. Settlement plates (400cm2) were deployed on 8 August 2008 for 2 weeks. The size of the circle increases with larger counts of larvae/plate, and clear circles represent zero recruitment. Grey rectangles within the bay represent mussel leases.

2.5.2. Patterns of Larval Recruitment

Of the 88 settlement plates deployed, a total of 82 were recovered, and Ciona recruitment occurred on 66. The average number of recruits was 9.1 (S.E. = 1.0) per plate, with a maximum count of 49. Recruits were found on all but 4 of the 63 stations located within the mussel leases and on all 6 stations located west of the leases. In contrast, Ciona was present on only one of the 13 stations east of the mussel leases, i.e., towards the mouth of the bay. Overall the pattern of recruitment was found to be heterogeneous (Global Moran‟s I = 0.266, P = <0.01), with higher levels of recruitment generally occurring towards the western end of the bay (Fig. 2.2).

Figure 2.3. Predicted settlement vs. observed settlement fitting of Weibull dispersal function (Pearson‟s r = 0.64) for Ciona intestinalis larval recruitment, using adult surveys and larval recruitment data in Boughton River (Prince Edward Island, Canada).

2.5.3. Dispersal Model - kernel fitting

The Weibull distribution fit the recruitment data (parameters c = 0.33 +/- 0.0006, λ (scale) = 1.75 +/- 0.0038 and k (shape) = 1.39 +/- 0.0019). The overall fit was good with a Pearson r of 0.64 (Fig. 2.3). The fitted Weibull function showed that larval recruitment was not highest close to adult populations, but peaked at a distance of 0.86 km (Fig. 2.4) with 95% of the recruitment occurring within 3.8 km from the source.

Figure 2.4. Weibull function dispersal kernel fit to Ciona intestinalis larval recruitment based on values of larval recruitment with distance from adult populations in Boughton River (Prince Edward Island, Canada). “P” represents the proportion of settling larvae.

2.6. Discussion

The dispersal of tunicates is widely believed to be limited due their short planktonic period, and the dispersal kernel derived from our data demonstrates clearly that Ciona’s dispersal range is limited to just a few kilometres. More surprising is the successful fitting of the Weibull function which, in contrast to the commonly-assumed negative exponential decay curve (Kot et al. 1996), highlights how patterns of recruitment can take unexpected forms, i.e., the lack of recruitment near adult source populations. From a management perspective, sampling at mid-range distances (~1 km), rather than close to the adult source, would thus be more effective for detecting invasions. The shape of the dispersal kernel can be attributed to certain characteristics of Ciona’s life cycle. Ciona is an external fertilizer, and therefore there can be time lags between egg release (up to 30 h [Svane and Havenhand 1993]), fertilization (24 h to become competent [Berrill 1947]) and settlement, which would result in low levels of settlement close to adults.

Overall, estimates of the dispersal period after larvae become competent range from a few hours to 6 d (Svane and Havenhand 1993), but generally settlement is thought to occur at the shorter end of this scale – minutes to hours (Berrill 1947; Svane and Havenhand 1993), which would explain the peak in settlement around 1 km. The high temperatures recorded in Boughton River, between 20ºC and 24ºC could shorten the development process (Svane and Young 1989), which would also contribute the 1-km peak and the limited dispersal range.

Ciona’s dispersal potential is too short for inter-bay dispersal in PEI, since bays are typically separated by distances of 10s of km, suggesting the recent spread between bays was most likely human-mediated, e.g., aquaculture activity and/or recreational boating (Darbyson et al. 2009). These findings may seem intuitive after the fact, but these data have not been available to managers in a practical format. Had they been made available earlier, local managers could have acted more promptly in containing Ciona once it had first been detected by controlling human activities to prevent further spread to adjacent bays outside of the natural dispersal range (Gertzen & Leung 2011; Clarke Murray et al. 2011). This applies not only to Ciona but to other species of invasive tunicates in PEI and likely elsewhere.

The over-dispersion of data in the Weibull observed-versus-predicted plot (Fig. 2.3) is indicative of un- modelled predictors (Hilbe 2007) and suggests that the simple, symmetrical distance from adults is not the only factor driving the pattern of recruitment. Our standardized sampling method reduced the influence of other likely environmental factors (e.g., substratum, depth), but other unidentified factors may have influenced recruitment patterns. The shape of the bay and tidal regime suggest that east-west water movement may be greater than north-south, though our dispersal kernel implicitly assumes symmetry. Other dynamics, such as turbulence and random diffusion will contribute to some north-south movement but, nonetheless, elongation of dispersal along the east-west axis may account for some of the over-dispersion of data. Despite this, the fit of the Weibull function is still remarkably good and captures well the dispersion of larval recruitment.

The small size of Boughton River, the planktonic larval duration of Ciona (<1-6 d) and the frequently changing tides might be expected to lead to a widespread dispersal of larvae throughout the bay. In spite of these homogenising factors, a signal of the adult population was seen clearly in the distribution of recruitment. Not only did we find a heterogeneous pattern of recruitment, we also found regions of zero settlement within the bay, which was surprising given the long deployment time of the settlement plates (2 weeks). We suspect that circulation might have contributed significantly to the unexplained variation we observed. For example, the general lack of settlement in the eastern end of the bay could have been caused

by changes in water direction and current magnitude created by the channel, the shallow area, and the outlet to the sea. It is also possible that the current speeds in the eastern end of the bay are too high for larval settlement (Abelson and Denny 1997). Regardless, our intention here was to minimize the effect of environmental variation (e.g., standardized plates at a fixed depth) to characterize dispersal patterns statistically, and not to determine the precise mechanisms underlying them. At a larger scale, there is likely some larval dispersal outside of the bay similar to that predicted by Kanary et al. (2011) elsewhere in PEI and observed by Petersen and Svane (1995) in Denmark. However, unfavourable conditions and low densities (i.e., Allee effects [Courchamp et al. 1999]) will undoubtedly limit the establishment of new populations and thus secondary spread.

The SDM approach applied by Therriault and Herborg (2008), one that has been effective in both aquatic and terrestrial systems (Peterson and Vieglais 2001; Thuiller et al. 2005; Inglis et al. 2006; Trebitz et al. 2009), successfully identified regions at risk in Canada, but the resolution was far too large to capture the variations in recruitment observed within this study. A key problem with the SDM approach for early detection of invasive species is that it is static in nature, with dispersal generally not considered limiting, and assumes species presence in all potential habitats suitable for recruitment. This assumption may be appropriate for species that are in ecological equilibrium and widely dispersed, i.e., native or well- established invasive species, but during the early stages of invasion, nascent populations are usually small and isolated, occupying only a small portion of suitable habitat available (Rouget and Richardson 2003). For example, prior to this study, it would have seemed reasonable to place a monitoring station anywhere in Boughton River but, even in this small and simple bay, recruitment is highly variable, and poorly- placed monitoring stations could return a false negative result. Therefore, the uses of large-scale predictions alone may not be accurate enough to effectively guide the detection of nascent populations.

The most effective method for monitoring would be to use the larger-scale SDM and the smaller-scale dispersal kernel approaches in tandem to identify locations that are both highly suitable to a given NIS (environmental aspect) and likely to receive a high number of propagules (dispersal aspect). However, a limitation exists in that SDMs are based upon the ecology of adult stages (e.g., environmental tolerances), whereas for many AIS, such as Ciona, monitoring targets the planktonic larvae stage, which can have different environmental constraints. Larval behaviour is influenced by a variety of environmental stimuli (Crisp 1961; Olson 1983; Pawlik 1992; Rodriguez et al. 1993; Abelson and Denny 1997; Kingsford et al. 2002), which occur on very small scales and cannot be accurately captured at large scales. Therefore, using environmental tolerance estimates from adult distributions to design detection strategies that target

the dispersive larval stage may be misleading, and where possible, monitoring strategies need to integrate multi-scale analyses that incorporate both life stages.

We were fortunate in discovering this invasion in its earliest stages as the patterns of recruitment within the bay became less clear over time (Collin 2013). However, it only represents a single observation of early larval recruitment, driven by an isolated adult population. Thus, extrapolating our result to other systems could be misleading, especially for other taxon and other coastal conditions. Moreover, even for the particular species examined here, the patterns could vary spatially and temporally. We feel nevertheless that without further information, it still provides a useful starting point for similar situations (e.g., another embayment in PEI). Clearly, additional such studies are needed and strongly encouraged as replication is needed for validating the specifics of this situation as well as exploring applications to other systems.

Our study demonstrates how a small-scale approach can complement the larger-scale guidance currently available. Large-scale surveys that identify and prioritize areas at risk (i.e., SDMs) are essential but should be considered as a preliminary filter. The next logical step is to pinpoint the most likely sites of introduction and, working on a more local scale, assess the suitability of the surrounding environment. Finally, this information needs to be coupled with an understanding of the life history and dispersal potential of the focal species to identify the appropriate scale for monitoring in order to optimize the allocation of sampling efforts, especially when targeting propagules produced by a nascent population (e.g., Kraft and Johnson 2000). It is essential then for managers to recognize that by incorporating multi- scale approaches to surveillance and monitoring, coupled with behavioural and dispersal characteristics, the early detection of nascent populations of nuisance species becomes a more realistic objective in the management of biological invasions.

Chapitre 3

From introduction to domination: Transitional patterns of recruitment during the rapid establishment of the high-impact invader Ciona intestinalis

3.1. Résumé

L‟information concernant les populations naissantes durant les premiers stades d‟invasion est possiblement la plus pertinente afin d‟allouer adéquatement l‟effort de surveillance et de détection précoce (éradication, ralentissement de la propagation). Malheureusement, nous estimons généralement l‟établissement et la dissémination chez une population déjà bien établie (p. ex. Therriault & Herborg 2008) et il est difficile de définir dans quelle mesure ces estimations se traduisent par la détection de nouvelles populations, petites et isolées. En une rare occasion, il a été possible de documenter les modes de recrutement précoce d'une population naissante du tunicier envahissant Ciona intestinalis dans la baie de la rivière Boughton (Île-du-Prince-Édouard, Canada) un an après sa détection et de tester si les prédictions d'établissement et de propagation (noyaux de dispersion) changent lorsque l'invasion progresse. L‟étude de la population adulte et du recrutement larvaire a été réalisée durant deux étés (2008 et 2009) par l'intermédiaire de la plongée et de plaques de fixation. Les premières observations en 2008 ont révélé un noyau de dispersion distinct qui était toutefois indécelable en 2009. Le patron de recrutement est devenu plus homogène au cours de l‟année 2008, peut-être en raison d'une population plus large et plus répandue, et est étonnamment revenu nettement hétérogène en 2009, malgré l'augmentation du recrutement. Bien qu‟il soit demeuré hétérogène, le patron de recrutement en 2009 est néanmoins spatialement stable avec des aires significatives de recrutement faible et élevé. Nous attribuons l‟hétérogénéité au début de 2008 à la nature petite et restreinte de la population fondatrice et la stabilité ainsi que l‟hétérogénéité de 2009 à une plus grande influence des facteurs abiotiques sur le recrutement.

3.2. Abstract

Information on nascent populations during the early stages of invasion is potentially the most relevant for allocating monitoring effort and for early detection when management can be the most effective (e.g., eradication or slowing the spread). Unfortunately, we typically estimate establishment and spread from already well-established populations (e.g., Therriualt & Herborg 2008), and it is unclear how well such estimates translate to detecting new (small and isolated) populations. From a rare opportunity, we documented the early recruitment patterns of a nascent population of the invasive ascidian Ciona intestinalis in Boughton River (Prince Edward Island, Canada) one year after detection to test whether predictions of establishment and spread (e.g., dispersal kernels) change as the invasion progresses. Surveys of the adult population and larval recruitment were performed over the following two summers (2008 & 2009) via scuba transects and settlement plates, respectively. Initial observations in 2008 found a distinct dispersal kernel, which, however, became undetectable in 2009. The pattern of recruitment in 2008 became more homogeneous with time, possibly due to a larger and more widespread population, but surprisingly, returned to a distinctly heterogeneous pattern in 2009, despite an order of magnitude increase in recruitment. Although consistently heterogeneous, the pattern of recruitment in 2009 was nevertheless spatially stable with significant clusters of high and low recruitment. We attribute early heterogeneity in 2008 to the small and restricted nature of the founding population, but the latter heterogeneity and stability in 2009 to a greater influence of abiotic factors on recruitment.

3.3. Introduction

The globalization of human activity and industry has led to an increase in the frequency of introduction events of non-indigenous species (NIS) to novel habitats, in both terrestrial and marine systems (Carlton and Geller 1993; Lodge 1993; Mack et al. 2000; Pimental et al. 2005). The impacts associated with these introductions, on the native community and environment, have created serious challenges for conservation and human industries, due, in part, to the unpredictability of the invasion process (Mack et al. 2000; Hooper et al. 2005). Moreover, the difficulty in identifying potentially problematic species, gauging the speed of establishment, and predicting the early patterns of spread have restricted our ability to detect, control and potentially eradicate NIS prior to any detrimental impacts (Hulme 2006). Addressing these concerns requires an improvement in our understanding of the drivers of dispersal and recruitment during the early stages of the invasion process, particularly the events that take place immediately after introduction when the invading population is still small and isolated. Further understanding of these processes can guide monitoring strategies and subsequently improve our ability to detect and respond to NIS rapidly (Simberloff et al. 2005).

The invasion process consists of four key stages: transportation; introduction; establishment; and further spread to nuisance levels (Colautti & MacIsaac 2004; Lockwood et al. 2005). All four stages present a wide and variable array of challenges for population persistence, particularly after transportation as the recipient environmental and/or native community are unique at every site of introduction (Lonsdale 1999, Lodge 1993). Once introduced, NIS are subject to variations in resource availability (e.g. food, space) and species interactions (e.g. biotic resistance), both of which have the potential to either encourage fast establishment and spread, possibly resulting in a problematic invasion, or hinder survival, often resulting in a failure to establish and disappearance. In the majority of cases, NIS fail to reach a critical density required for survival (Williamson & Fitter 1999), leaving themselves susceptible to establishment failure by either natural processes, such as Allee effects (Courchamp et al. 1999; Taylor & Hastings 2005), or human intervention, e.g., physical removal (see Hewitt & Campbell 2007). Therefore, to establish a new population, it is imperative for an NIS to quickly form a persistent population.

The difficulties associated with early detection have created a paucity of studies that focus on nascent populations (Marsico 2010; Kraft and Johnson 2000; Vander Zanden and Olden 2008). At present, targeted monitoring strategies rely heavily on studies from the species native ranges or on already established invasive populations. Although these studies can provide valuable information for identifying

suitable habitats for NIS (e.g., Inglis et al. 2006), they identify the potential distribution of an invading species, which may take many generations to occupy. In reality, nascent populations occupy only a small percentage of the suitable habitat available to them (Rouget and Richardson 2003). By designing monitoring strategies on a species potential distribution, we risk assigning monitoring effort in areas outside the target species dispersal potential, leading to a failure to detect a new population or detecting an NIS once it has already become established. Delayed detection of NIS greatly reduces the management options available, as the cost and required effort for successful management increases (Finnoff et al. 2007; Olenin et al. 2011).

In the marine environment, the window of opportunity for detecting and effectively managing (e.g., eradicating) nascent populations of aquatic invasive species (AIS) can be very short, as dispersal mechanisms in aquatic environments can promote fast rates of spread (Carr et al. 2003; Kinlan & Hastings 2005). For sessile AIS that have a sedentary adult stage, local retention of a sufficient number of propagules is essential for population survival, particularly when there is a single, isolated population (Swearer et al. 2002), as it replenishes the original founding population, increases the local density, and increases the probability of next-generation reproduction (Sponaugle et al. 2002; Kinlan & Gaines 2003). Short effective dispersal ranges also contribute to the difficulty of detecting new populations of AIS, as detection can only occur within this range. Therefore, if a probable site of introduction is known, managers can optimize monitoring strategies by incorporating aspects of small-scale dispersal and assign monitoring effort to adequately survey within the dispersal ranges of suspected AIS.

To increase the probability of detecting nascent populations and optimize monitoring strategies, it is imperative that we investigate in more detail the very early stages of an invasion and document the evolution of the population. In this study, we capitalised on an early detection of a nascent population of the highly-problematic invasive tunicate Ciona intestinalis (Lambert and Lambert 1998; McDonald 2004; Robinson et al. 2005; Uribe and Etchepare 2002; Howes et al. 2007) in Prince Edward Island (PEI), Canada, and assessed whether the dynamics and predictors of well-established populations are applicable to nascent populations in the early stages of invasion. We document in detail the rate of population expansion and concurrent patterns of recruitment during the very early stages of invasion, from just a few individuals to a widespread and dominant population.

Ciona intestinalis, hereafter Ciona, has a history of successful invasions worldwide, most recently in PEI where it has proven to be highly problematic by fouling mussel aquaculture (Thompson and MacNair

2004; Locke et al. 2007). Despite the well-documented distribution and impacts of Ciona in eastern Canada (Locke 2007; Sephton et al. 2011; Sargent et al. 2013), and the identification of areas at risk from further spread (Therriault and Herborg 2008), there has been little attention to the processes leading to the dense populations that cause impacts, leaving little guidance for early detection monitoring. Here we provide key information on how patterns of recruitment evolve during the early stages of invasion and how these data can improve monitoring strategies aimed at detecting nascent populations.

3.4. Methods

3.4.1. Study Site

Boughton River is a small semi-enclosed inlet on the eastern coastline of PEI, approximately 6 km in length and 2 km wide, which is governed by a semi-diurnal tidal system (tidal period of approximately 12 hours). The substratum in the bay is predominantly sand and silt, but the large presence of mussel aquaculture, a small harbour at the mouth of the bay and a small loading dock towards the west provide suitable solid substrata for Ciona.

The seasonal temperature range for Boughton River varies dramatically from ~5ºC in December to 23ºC in August and September (2009 data, DFO Mussel Monitoring Program, PEI). However, due to the small entrance to the bay (approximately 100 m wide) and the subsequent restrictions on water flow, spatial variation in temperature and salinity within the bay is low; during August 2008 and October 2010, temperature and salinity data collected at multiple locations (P. Edwards, unpubl. data) showed a temperature range of 20.5 – 24.4 ºC (Aug 2008) and 10.4 – 12.7 ºC (Oct 2010) and a salinity range of 25.4 – 27.7 ppt (Aug 2008) and 27.2 – 29.1 ppt (Oct 2010) at 2-m depth, the depth at which we placed settlement plates (see larval recruitment below). The general pattern for temperature was an increase towards the west, with salinity values showing the opposite trend; the flushing of water in the east of the bay during tidal changes and surface run-off towards the west most likely created this pattern. All of these values fall within the published tolerance ranges for Ciona (Dybern 1965, 1967; Carver et al. 2006; Therriault and Herborg 2008), which suggests both temperature and salinity should not physiologically restrict Ciona recruitment within the bay.

3.4.2. Adult Survey

SCUBA surveys were performed along mussel farm installations and the total number of observed Ciona was counted separately by two divers and subsequently averaged (Fig. 3.1). Each 50-m transect ran horizontally through the water and every vertically-hung mussel „sock‟ was completely inspected (i.e., top to bottom, ~3 m in length), including the natural substratum below the mussel lines. If mussel leases had recently been harvested (i.e., mussel socks removed), 50 m of the remaining in situ equipment (e.g., ropes, buoys and anchor lines) was surveyed. A total of 33 dive surveys were performed in June 2008 and 35 in June 2009, at the start of Ciona’s reproductive season (Ramsay et al. 2009), to document the distribution of the first reproductively-active adult population of the year.

3.4.3. Larval Recruitment

A quasi-systematic „grid‟ of 88 settlement plates was deployed throughout Boughton River, within and outside the mussel leases (Fig. 3.2). At each station, a 20 x 20-cm grey PVC settlement plate was deployed at a 2-m depth for approximately two weeks, where possible. On collection, the plates were stored separately in seawater and the total number of settled larvae on each plate was counted (using a stereomicroscope at 40x magnification) within 48 h of collection. The grid of settlement plates was repeated 8 times: four in 2008 and four in 2009, during August and September.

3.4.4. Recruitment Pattern Analysis

Spatial variations in larval recruitment were analysed using both global and local Moran‟s i analyses. The global Moran‟s i analysis provides an indication of whether the overall pattern of recruitment is negatively (homogeneous) or positively (heterogeneous) spatially autocorrelated, whereas the local Moran‟s i identifies specific areas of similarity (clustering) or dissimilarity (outliers) (Boots 2002). The local Moran‟s i values were determined by comparing each station with the nearest three neighbours (stations). Three neighbours was the maximum allowed while keeping the distance between compared stations low (< 1 km; as Boughton River is approximately 6 km in length, we considered distances greater than 1 km between compared stations too large for identifying local clusters).

Table 3.1. Larval recruitment values for Ciona intestinalis in Boughton River, PEI for all 8 grids of settlement

plates deployed in August and September 2008 and 2009.

14 81 79

165

2320 2320 2870

± 676

(98%) 40120

21/09/09

14 81 75

210

2950 2950 2630

± 680

(93%) 36800

07/09/09

14 82 80

208

2920 2920 1540

± 481

(98%) 21600

24/08/09

14 82 80

149

2090 2090 1620

± 419

(98%) 22740

10/08/09

82 72

5.0

45.2

1041

115.7

(88%)

± 13.4

27/09/08

27 33 23

6.9

98.3

2655

187.5

(70%)

± 95.3

18/09/08

14 54 24 72

0.7

2655

(44%)

22/08/08 9.62.7 ±

14 82 66 35

0.6 2.5

(80%)

08/08/08 8.50.9 ±

ate ate

ona

ymentduration (days)

Grid Collection date Depl Total stations collected Stations with (%total) Average larval abundance(± S.E.) perplate Average settlementr (larvae/day) perplate Max. settlementper plate Max. settlementrate (larvae/day) perplate

In addition to the Moran‟s analysis, a Kendall tau rank correlation coefficient analysis was performed to assess whether the rank order of stations (recruit abundance) for each grid was significantly correlated.

We compared all grids in 2008 and separately for 2009. If all the pair-wise comparisons within each year were correlated, we considered this to be an indication of stationarity in the pattern of larval recruitment and, therefore, a well-established population. Likewise, a non-significant or significantly independent result was considered an indication of instability and, therefore, occupation of a smaller, initial range rather than the potential range.

3.5. Results

3.5.1. Adult Survey

Ciona was detected in 10 of the 33 surveys performed in 2008 (Fig. 3.1.); averaging 2.5 ± 1.4 (S.E.) individuals per 50-m transect (max = 41). Ciona was found to be more abundant in the north-west part of the bay, but, although it was detected in several locations, only 2 surveys recorded counts higher than 10 adults. In general, the mussel lines were clean with few fouling organisms, although other species of invasive tunicate (Botryllus schlosseri, Botrylloides violaceus, and Styela clava) were observed in small numbers. There was no observed Ciona recruitment on the bay floor surrounding the mussel leases.

In 2009, Ciona was detected in 30 of the 35 surveys; averaging 393.8 ± 142.5 (S.E.) individuals per 50-m transect (max = 4030), in considerably higher numbers than the 2008 adult surveys (Fig. 3.2.). The highest concentrations were found in the most westerly mussel leases, although large densities in the centre of the bay were also detected (Fig. 3.1). The adult distribution was more widespread, compared to 2008, but there were still areas where Ciona was absent. Ciona was observed occasionally surviving on the bay floor directly beneath the mussel leases, but this was most likely due to the sloughing of mussels already colonised by Ciona rather than larval recruitment on the bay floor.

Figure 3.1. Adult population surveys (50-m transects) for Ciona intestinalis in Boughton River, Prince Edward Island for June 2008 (A) and 2009 (B).

Figure 3.2. Histogram of adult Ciona intestinalis abundance along 50-m transects in Boughton River, Prince Edward Island for June 2008 (n = 33) and June 2009 (n = 35)

3.5.2. Larval Recruitment

In 2008, 4 grids of settlement plates were deployed in August and September; unfortunately many stations were lost or excluded from analysis, particularly in grids 2 & 3. For Grid 2, 34 stations were lost due to an unforeseen wearing of the rope line, resulting in the loss of buoys and settlement plates. The lost stations were immediately replaced with new equipment for the third grid. However, during the collection of Grid 3, we found high recruitment levels of caprellid amphipods (Caprella linearis & C. mutica) on our settlement plates, whose presence was found to negatively affect Ciona larval recruitment (see Chapter 5). The high caprellid abundance was most likely the result of a longer deployment (27 days) and an already existing fouling community on the „old‟ stations from Grid 2 (replaced stations would have been free of fouling organisms on deployment). Further investigation revealed that Ciona larvae were either absent or in very low numbers when caprellid abundance was greater than 150 ind.plate-1 (see Chapter 5). Therefore, we only used Ciona recruitment data for plates with less than 150 caprellids.

In 2008, all four grids recorded the presence of Ciona (Table 3.1). The average settlement rate recorded in each grid increased towards the end of the summer, reaching a peak in Grid 3 (6.9 larvae.plate-1.day-1), although this value was largely influenced by high settlement on a single plate (2665 larvae, almost half the 6188 larvae total for Grid 3). Additionally, the maximum larval settlement rates steadily increased throughout the study period, from just 2.5 larvae.plate-1.day-1 in Grid 1 to 115.7 larvae.plate-1.day-1 in Grid 4, suggesting the fecundity (and size) of the adult population had increased throughout the late summer and early fall. The percentage of settlement plates that recorded Ciona settlement varied greatly in 2008, from 44% in Grid 2 to 88% in Grid 4. Although Ciona detection was fairly widespread throughout the bay (see Grid 1 example [Fig. 3.3]), the global pattern of recruitment varied between grids, from a more heterogeneous pattern at the beginning of the summer (Grid 1), to a more random pattern (Moran‟s i value closer to 0) for grid‟s 2, 3, and 4 (Fig. 3.4). The low number of stations for Grid 3 (n = 33) may have influenced the low Moran‟s i value, but when the n for Grid 1 was also reduced to 33, we continue to find a significantly heterogeneous pattern (i = 0.19, p = <0.01). This suggests that, despite the low n in Grid 3, the pattern of larval settlement became more homogeneous later in the summer. Throughout the summer, larval recruitment density increased, with significant differences detected between grids 1 & 4, and 2 & 4 (Table 3.2).

In 2009, the total number of larvae found settling in all grids significantly increased from 2008 (Wilcoxon rank sum test: w = 7747.5, p < 0.01). The average number of recruiting larvae ranged from 2088.3 ± 419.2 per plate (S.E.) in Grid 5 to 2951.3 ± 680.17 per plate (S.E.) in Grid 7, but there were no significant differences between larval recruitment between the 4 grids (Kruskal-Wallis multi-comparison test [p- value: 0.05]). Similar to 2008, average larval recruitment in 2009 was highest at the beginning of September (Grid 7: 210.8 larvae.plate-1.day-1) and the maximum settlement rate recorded at the end of September (Grid 8: 2865.7 larvae.plate-1.day-1). The lowest settlement rates were recorded in Grid 5 but were still approximately 13 × higher than the highest settlement rates recorded in 2008, indicating that the initial adult population of 2009 (June population that survived winter) was larger than the reproductive population in September 2008. In all four grids of 2009, Ciona was recorded on at least 93% of the plates deployed, indicating larval recruitment was widespread. However, despite the dramatic increase in larval recruitment in 2009 (2009 maximum was 15× the 2008 maximum) and the higher percentage of stations recording Ciona recruitment, there were still plates on which no recruitment occurred (eastern end of the bay) and the overall pattern of recruitment remained heterogeneous (Fig 3.4).

Table 3.2. Student‟s t test results for Ciona intestinalis larval recruitment densities for all 4 grids of settlement plates deployed in August and September 2008.

2008 2 3 4 1 t = -0.38 t = 1.88 t = -2.74 p = 0.7 p = 0.07 p < 0.01 2 t = 1.87 t = -2.61 p = 0.07 p = 0.01 3 t = 1.48 p = 0.15

Table 3.3. Kendall‟s Rank Correlation analysis for Ciona intestinalis larval recruitment for all grids of settlement plates for 2008 and 2009 (τ value: 1 = perfect correlation, 0 = independent, -1 = perfect dissimilarity). *Comparison 2 - 3 (2008) is absent due to lack of comparable stations between grids.

2008 2 3 4

1 τ = 0.075 τ = 0.231 τ = 0.173 p = 0.487 p = 0.178 p = 0.04 2 τ = 0.308 n/a* p = 0.006 3 τ = 0.331 p = 0.015 2009 6 7 8

5 τ = 0.591 τ = 0.524 τ = 0.432 p <0.001 p <0.001 p <0.001 6 τ = 0.585 τ = 0.474 p <0.001 p <0.001 7 τ = 0.455 p <0.001

Figure 3.3. Larval recruitment of Ciona intestinalis on settlement plates (88 stations) in Boughton River, Prince Edward Island, for Grid 1, August 2008 (A) and for Grid 5, August 2009 (B).

*Maps of all 8 grids can be found in Appendice A

Figure 3.4. Global Moran‟s i values for all 8 grids of Ciona intestinalis larval recruitment (2008 & 2009) in Boughton River, Prince Edward Island (-1 = perfect dispersion, 0 = random, 1 = perfect correlation). (Standard Deviation bars).

The results from the local Moran‟s i analysis indicate that in 2008 there were areas of significant clustering in grids 1, 2, & 4, but not in Grid 3, possibly due to the reduced number of stations for the latter (Fig. 3.5). The locations of these areas of clustering were not consistent and varied between the three grids: the mouth and centre of the bay (Grid 1); the western end of the bay with two outliers in the mussel leases (Grid 2); and a single point of clustering and a single outlier in the west of the mussel leases (Grid 4). This variable pattern of recruitment was also detected in the Kendall‟s rank analysis (Table 3.3), with significant correlations found only in grids 1 & 4 and 2 & 4 with low τ values, indicating a low level of correlation.

For 2009, all four grids showed similar patterns of recruitment, with significantly high areas of clustering at the western end of the bay, west of the mussel leases (Fig. 3.6). Only Grid 5 found significant clustering within mussel leases (still towards the western end of the bay), and only Grid 8 had a significant outlier (this station recorded an abnormally low amount of settlement in an area normally associated with very

high settlement). This stationarity in the pattern of recruitment was also evident in the Kendall‟s rank analysis (Table 3.3), with all four grids (5, 6, 7 & 8) showing significant correlations with high τ values.

3.6. Discussion

The introduction and subsequent invasion of Ciona in PEI emphasises the speed at which AIS can become problematic, when presented with suitable conditions, and the importance of early detection for effective management. The intensity of the mussel farming industry in PEI has provided an abundance of available substrata, which likely contributed greatly to the rapid establishment of Ciona in Boughton River as there is little competition for space. In addition to the vast availability of substrate, Ciona’s high fecundity (approximately 10,000 eggs per adult [Petersen & Svane 1995]), relatively short planktonic larval period (1 – 6 days [Berrill 1947; Svane and Havenhand 1993]), and rapid maturation period (1 - 2 months [Dybern 1965]) would also have contributed to the rapid establishment of a dense and dominant population.

The rapid population growth observed in 2008 and 2009 in Boughton River shows that Ciona was capable of widespread establishment in just 2-3 years and suggests the initial introduction occurred in 2006 or 2007. Not only does this demonstrate the damaging potential of Ciona as an invasive species, it also highlights just how short the window is for early detection and subsequent managerial action. The period of high variability observed in 2008, when the population size was low, indicates an unstable population that may have been vulnerable to an eradication attempt (i.e., physical removal of the population). By 2009, however, when the population was much larger and widespread, eradication would have been too difficult and costly to be a viable option (but see Edwards & Leung 2009), leaving control as the most effective method for management; at present, mussel farmers in PEI use high-pressure water sprays to remove Ciona from mussel shells and equipment to reduce their impact on mussel growth and post- harvesting cleaning costs.

Although Ciona’s population increased dramatically in 2009, the resulting pattern of larval recruitment was unexpected. Considering the environmental homogeneity of Boughton River, the frequent tidal changes (every 6 h), and the planktonic duration of Ciona larvae (minimum of 24 h), we expected the larvae to be well dispersed throughout the bay and, therefore, the pattern of recruitment to become more homogeneous as the population grew. However, not only did we continue to find a heterogeneous pattern

Figure 3.5. Local Moran‟s i values for 3 grids of Ciona intestinalis larval recruitment (2008) in Boughton River, Prince Edward Island. Open circles indicate a negative i value (outlier). Filled circles indicate areas of similarity (clustering). Grid 3 is absent as there were no significant clusters or isolation detected.

Figure 3.6. Local Moran‟s i values for 4 grids of Ciona intestinalis larval recruitment (2009) in Boughton River, Prince Edward Island. Open circles indicate a negative i value (outlier). Filled circles indicate areas of similarity (clustering). Maps show western end of bay only, as no other points of significance were detected in the rest of the bay.

of recruitment with significant areas of clustering, we also found areas of no settlement towards the mouth of the bay, which remained constant throughout 2009. The driving factors for this pattern are unknown, but possible explanations include: variation in the concentration of larvae caused by local circulation; faster water currents towards the entrance of the bay reducing settlement rates (Abelson & Denny 1997); the maintenance of a concentration of larvae west of the mussel leases resulting from a lack of suitable substrata „upstream‟; a possible settlement cue created by higher phytoplankton production caused by surface run-off west of the mussel leases (phytoplankton would be filtered out as water passes through the mussel leases, reducing concentrations to the east); higher temperatures to the west could increase the rate of larval development (Sanford et al. 2006), which in turn would increase the number of larvae ready to settlement; an area of optimal temperature and/or salinity for settlement west of the mussel leases increases settlement rates.

In Chapter 2, the 2008 adult survey data and the Grid 1 larval recruitment data were used to fit a dispersal kernel, which suggested the early pattern of recruitment was driven by adult distribution. If dispersal was the key driver for recruitment, we expected that in 2009, when the adult population is larger and more widespread, the bay to become saturated with larvae and the pattern to become more homogeneous. However, the highly skewed pattern of larval recruitment found in 2009 (high towards the west, low towards the east) indicates that the adult distribution had little effect on the overall pattern of recruitment; if adult distribution was a key driver, we would have expected to find a more equal east-west distribution

of larval recruitment. The disjoint between adult distribution and larval recruitment in 2009 indicates a change in the drivers of recruitment, shifting from limited dispersal from a small source population in 2008 to control by abiotic variables in 2009, i.e. environmental and/or hydrodynamic factors. This shift in drivers should create a more stable pattern of recruitment (if competition and predation are nominal) as abiotic variables will generally remain more stable than the more dynamic, shifting distribution of nascent populations; during the establishment phase of the invasion process, the location of the adult (source) population is changeable as it expands into the surrounding available habitat. This would be particularly evident if the AIS had a short maturation period and first settlers became reproductive later in the same season (as seen in this study in 2008).

Identifying the key factors that determine the patterns of recruitment, for both nascent and established populations, can be of value to early detection and monitoring programs. If a potential point of introduction for NIS is suspected, the spatial allocation of monitoring effort should be designed around the dispersal potential of a targeted species, in addition to the surrounding habitat (Edwards et al. in review). However, if a targeted species is detected outside of its dispersal potential, it could be an indication that the NIS has already begun to establish a more widespread population, with additional reproductive populations outside the suspected point of introduction. The shifting patterns in recruitment seen in Boughton River clearly show that both dispersal and environmental factors drive patterns of recruitment, but the influence of each shifts as the invasion progresses. This highlights the value of integrating both approaches when assigning monitoring effort aimed at early detection.

Although a very dense population of Ciona was quickly established in Boughton River, it remained localised and spread was restricted to the confines of the bay. The lack of suitable natural habitat/substrata between bays and the inability of larvae to disperse the distance required for inter-bay spread most likely restricted the spread of Ciona. The short dispersal ability of Ciona is encouraging for management in PEI, as populations can be largely contained by controlling human activity (e.g., boating, aquaculture equipment exchange) between bays; there is potential for avian transportation, although the probability of Ciona being introduced in high enough numbers to establish a new population would be low.

The results from this study present an extreme example of an invasion event, but it is important to recognize that this example is one species within a single site. Invasion events are very much species and site specific and, although we have presented a detailed assessment of a Ciona invasion, without further studies from different sites, the broader applications of these data are limited. Ciona is a worldwide

invasive species, yet the invasion of PEI remains exceptional, both in terms of speed of establishment and the size/density of the established populations. The success of Ciona in PEI begs the questions: why was the invasion so dramatic; what is controlling Ciona populations in other sites where the invasion is less problematic; and, are the less-problematic invasive populations of Ciona still in a transitional phase preceding establishment? Without further studies from multiple sites of invasion, we limit our ability to identify more general patterns of Ciona invasion that can be applied to monitoring strategies in sites currently at-risk from invasion.

The ubiquity of NIS allows us to perform real-time observations of single species in multiple environments and make direct comparisons of factors that can promote high-impact invasions and those that hinder establishment. Information on failed or casual NIS (species unable to form self-replacing populations, but persist through repeated introductions [Richardson et al. 2011]) can be equally as valuable as information on successful invaders, particularly if the species in question has been a successful invader elsewhere. This study identified a distinct temporal shift in the pattern of recruitment as the population increased in size and distribution, highlighting the importance of incorporating both dispersal and environmental preference into monitoring strategies. Additionally, areas of consistently high and no recruitment („hotspots‟ and „notspots‟) were identified, which can provide essential information for ensuring the allocation of monitoring effort in currently Ciona-free bays is not poorly assigned (i.e., monitoring in an area where no recruitment will occur). Moreover, this study raises concerns over our current tendency to focus effort on documenting the impacts of NIS, rather than examining the more fundamental aspects of dispersal and recruitment of invading populations. The lack of data on establishing populations ultimately limits our ability to predict patterns of spread (temporally and spatially) and, more importantly, design effective and reliable methods of early detection.

Chapitre 4

The role of phototactic and geotactic behaviours during larval settlement: An experimental field study on the tunicate Ciona intestinalis

4.1. Résumé

Le comportement des larves peut avoir d‟importantes implications pour le succès du recrutement, particulièrement pour les espèces marines sessiles. Pour augmenter la probabilité de recruter dans des habitats convenables, les larves ont évolué pour devenir hautement sensibles à leur environnement, modifiant leur comportement en fonction des différents signaux. Dans cette étude, nous avons manipulé les conditions de lumière pour évaluer les rôles relatifs des comportements phototactiques (lumière) et géotactiques (gravité) des larves du tunicier Ciona intestinalis pendant la colonisation. Notre étude confirme que les larves de Ciona ont un comportement de phototactisme négatif, avec une densité plus importante sur la surface inférieure sombre des plaques de recrutement par rapport aux surfaces inférieures qui sont illuminées. Cependant, des expériences nocturnes supplémentaires avec manipulation de la direction de la lumière ont montré que les densités de larves restaient faibles sur les surfaces supérieures, qu‟elles soient ombragées ou non. Ces résultats suggèrent que l‟orientation de la surface a plus d‟influence que les conditions de lumière. En effet, quand la lumière est absente, les densités de larves restent élevées sur les surfaces inférieures, indiquant que le comportement géotactique joue un rôle plus important que la lumière pour déterminer le schéma de colonisation. Cependant, quand la surface inférieure est ombragée et que les comportements photo- et géotactiques agissent conjointement, le taux de colonisation augmente significativement. Après contact avec la surface, les larves continuent de montrer un comportement phototactique négatif et montrent des densités de colonisation plus importantes vers le centre plus sombre des plaques de fixation. En général, les résultats de cette étude suggèrent que les larves de Ciona continuent de montrer un comportement phototactique négatif pendant et après la colonisation, mais que le comportement géotactique négatif est une réponse conditionnelle aux conditions sombres. Par conséquent, les comportements photo- et géotactiques ne sont pas redondants, mais complémentaires pendant la colonisation.

4.2. Abstract

Larval behaviour can have important implications for recruitment success, particularly for sessile marine species. To increase the probability of recruiting in suitable habitats, larvae have evolved to become highly sensitive to the surrounding environment, altering their behaviour according to various cues. In this study, we manipulated light conditions to assess the relative roles of phototactic (light) and geotactic (gravity) behaviours of larvae of the tunicate Ciona intestinalis during settlement. Our study confirms Ciona larvae exhibit negatively phototactic behaviour, with higher densities recruiting to the darker undersurfaces of settlement plates than on illuminated undersurfaces. However, additional night-time experiments that manipulated the direction of light show larval densities remain low on upward facing surfaces, whether shaded or not. These results suggest surface orientation is more influential than light conditions. Indeed, when light is absent, larval densities remain high on under-surfaces, indicating that geotactic behaviour plays a bigger role than light in determining the pattern of settlement. However, when the undersurface is shaded and both phototactic and geotactic behaviours work conjointly, the rate of settlement is significantly elevated. After surface contact, larvae continue to show a negatively-phototactic behaviour and settle in higher densities towards the darker centre of the settlement plate. In general, the results from this study suggest Ciona larvae remain negatively-phototactic during and after settlement, but that negatively-geotactic behaviour is a conditional response to dark conditions. Therefore, phototactic and geotactic behaviours are not redundant, but complementary during settlement.

4.3. Introduction

For many marine invertebrates, events that take place during the dispersive planktonic larval stage have important implications on population distribution and abundance (Roughgarden et al. 1988; Cowen and Sponaugle 2009). Patterns of dispersal and the processes that drive these are fundamental in determining patterns of recruitment. How larvae respond to changes in the surrounding environment during the free- swimming and settlement stages influence these patterns but also have important implications on recruitment success, post-settlement mortality and ultimately adult survival and fitness (Keough and Downes 1982). Initially, dispersal is fairly passive; some larvae have the capacity to control their depth in the water column, exposing them to different currents that can increase or decrease dispersal distance (Morgan et al. 2009; Kunze et al. 2013), but ultimately dispersal is largely driven by water movement (Mileikovsky 1973). This period of the larval stage can influence the range and shape of dispersal trajectories. However, once larvae become competent and ready to settle, the role of larval behaviour over small scales becomes more apparent as they search for a suitable substratum (Rodriguez et al. 1993). How larvae respond to small-scale variations in the surrounding environment can influence patterns of recruitment within their dispersal range.

The search for suitable habitats is guided by the surrounding environment, and larvae depend on both physical and chemical cues to detect appropriate substrata, avoiding sub-optimal or dangerous environments (Rittschof et al. 1998, Pawlik 1992). Larval behaviour can vary among species, depending on the environmental conditions required for the juvenile and/or adult stages. Behaviour can be influenced by hydrodynamic variables (Butman 1987; Abelson et al. 1993; Havenhand and Svane 1991; Pawlik et al. 1991), light (Bingham 1993; Feng 2010), gravity (Vermeij 2006; Park 2004; Mogami 1988), and a variety of chemical traces; e.g., pheromones from conspecifics (Kingsford 2002), competitors (Grosberg 1981; Bakus et al 1986), predators (Sih 1982; Holomuzki 1986), and food sources (Sebens 1981). Once a suitable substratum has been detected, there can be an additional period of small-scale exploration to ensure appropriate surface characteristics, e.g., stability and complexity (Mullineaux and Butman 1991; Glasby and Connell 2001; Lemire and Bourget 1996; Pech 2002).

The planktonic larval stage has particular importance for sessile organisms, such as ascidians, which have a sedentary adult stage. As adult forms have no direct means of movement, ascidians are reliant on the ability of larvae to disperse and locate suitable habitats. To achieve this, ascidian larvae have evolved sensitivity to a variety of environmental stimuli, both chemical (e.g., Bakus et al. 1986; Davis 1991;

Durante 1991; Stoner 1994) and physical (e.g., Davis 1987; Flores and Faulkes 2008). Of these environmental cues, light is generally considered one of the most influential factors (Thorson 1964; Young and Chia 1985; Svane and Dolmer 1995), and the majority of ascidian larvae possess a photoreceptive organ (ocellus), which can distinguish between light and dark conditions (Kusabe and Tsuda 2007). Ascidian larvae have shown ontogenetic changes in behaviour towards light (phototaxis) and gravity (geotaxis) throughout the planktonic stage, shifting from an initial positively phototactic and negatively geotactic behaviour to a negatively phototactic and positively geotactic behaviour during settlement (Young and Chia 1982; Miller and Hadfield 1986, Hurlbut 1993; McHenry 2003). This alternation in behaviour is thought to promote early dispersal and, more importantly, increase the probability of a larva locating a shaded surface in shallow water during settlement. Adult ascidians are commonly found on overhanging surfaces (e.g., under rocks, boat hulls [e.g., Collin et al. 2010; Murray et al. 2012]), which is generally considered to be the result of larval behaviour rather than post-settlement mortality (Crisp and Ghobashy 1971; Young and Chia 1985; Svane and Dolmer 1995). Shaded overhanging surfaces are optimal for newly settled larvae and juveniles because there is less competition for space from algae, no threat of smothering from sedimentation, and the dark conditions reduce predation (Young and Chia 1984).

This preference for overhanging surfaces has driven studies on how phototactic and geotactic cues influence larval behaviour, usually by exposing them to different lighting conditions and substratum orientations. Ascidian larvae have been found to show photokinetic behaviour, by increasing their swimming activity when exposed to dark conditions. This behaviour is known as the shadow response and is considered to aid larvae to locate the darker undersurfaces. This behaviour was originally observed in Botryllus schlosseri larvae on eelgrass blades (Woodbridge 1924) and has since been investigated in 8 different species of ascidians (Young and Chia 1985), although their results did not confirm the presence of shadow response in all species.

Light has been identified as an important cue for Ciona intestinalis (hereafter referred to as Ciona), a species not studied by Young and Chia (1985), by inducing gamete release in adults (Lambert and Brandt 1967) and influencing swimming behaviour (Zega et al. 2006) and settlement distribution of larvae (Rius et al 2010). Ciona have small (0.7 – 1.1 mm), simple tadpole larvae but are capable of directed swimming (Bone 1992). During the very early planktonic stages, newly hatched larvae swim upwards towards the water surface, which is thought to be caused by geo-negative rather than photo-positive behaviour as the ocellus is not functional for the first 4h after hatching (Nakagawa et al. 1999). Laboratory studies by Rius

et al. (2010) showed that Ciona larvae have a preference for dark (shaded) surfaces, when presented with dark and light conditions, but not for a specific surface orientation (upward, vertical, and downward). However, when larvae were kept in constant light or dark conditions only, they showed a preference for downward-facing surfaces. Although these results suggest that Ciona larvae respond to both light and surface orientation, there has been little or no effort to tease apart the specific roles of geotactic and phototactic behaviours and identify exactly how Ciona larvae locate overhanging surfaces. The generally accepted theory is that, throughout the larval period, light is the primary driver of behaviour and that Ciona‟s preference for overhanging surfaces is a consequence of shade rather than surface orientation. Results of past laboratory experiments support this theory and, in doing so, suggest a redundancy in geotactic behaviour when light is present. A major fault in previous experiments of larval behaviour is a failure to manipulate the direction of light, e.g., facing upwards and/or downwards, simultaneously with surface orientation. Without this additional light treatment, the separate roles of phototactic and geotactic behaviour cannot be fully investigated.

To date, studies on ascidian recruitment have relied heavily on adult distribution surveys and laboratory experimentation on larval settlement behaviour, usually on very small scales (i.e., settlement distribution within small acrylic boxes or petri-dishes [Young and Chia 1985, Svane 1987, Durante 1991, Svane and Dolmer 1994, Feng et al. 2010, Rius et al. 2010]). The extrapolation of small-scale laboratory experiments to describe and/or explain natural phenomena should be applied with caution as, under the controlled conditions of a laboratory, the behaviour of larvae can only give a rough approximation of what occurs in nature (Young and Chia 1985). To alleviate these concerns, we performed a series of field studies that kept the number of controlled variables to a minimum (substratum, depth, light), resulting in a more realistic assessment of larval behaviour.

To assess the relative influences of light and surface orientation on Ciona during settlement, we manipulated the direction of light sources around settlement plates, using artificial light, and compared variations in larval densities on the upper and under surfaces of the settlement plate. Additionally, we investigated the influence of light on larval behaviour after initial surface contact. Low density edge effects have previously been observed on the under-surfaces of settlement plates (pers. obs. S. Collin), a phenomenon that has of yet not been explained, but could be attributed to variations in light concentration from the plate edge to the plate centre.

4.4. Methods

4.4.1. Study site

All experiments were carried out in Boughton River, a small, semi-enclosed bay on the eastern coastline of Prince Edward Island (PEI), eastern Canada, in 2010 and 2012. In this location, adult populations of Ciona can be found growing in dense aggregations on mussel aquaculture installations. The size of the adult population, coupled with its high fecundity (10,000 – 12,000 eggs per adult [Carver 2006]) has created areas of extremely high settlement rates (~100/cm2) in 2 weeks, with Ciona often covering 100% of settlement plates [S. Collin, pers. obs.]). This area of high settlement provided an excellent and rare opportunity to perform short manipulative experiments (between 6 h and 24 h) on phototactic and geotactic larval behaviours in the field with little competition from other species.

4.4.2. Experiment 1 - Timing of settlement

To assess whether light is required for larval settlement, we first set out to identify what time of day larval settlement rates peak. Ciona larvae have a planktonic period of between 24 h and 6 d (Svane & Havenhand 1993), which means they are exposed to both day and night conditions. Therefore, timing of settlement is influenced by larval behaviour, not necessarily by the timing of gamete release by the adult (as seen for Botryllus schlosseri by Olson 1983). However, if larval competency is reached within 24 h of gamete release, larvae will be ready to settle by sunrise the next day, suggesting adult behaviour could influence on the timing of settlement, if fertilization took place immediately after gamete release. If no difference between day and night settlement rates is observed, we could assume that light is not a key factor during settlement and the larvae are using a different cue or coming into contact with the settlement plates by chance.

To test this, we deployed five 20 × 20 cm grey PVC settlement plates at a depth of 2 m (same approach used in Chapters 1 & 2), which were replaced every 6 hours for a 24 hr period, giving a total of four sets of five plates. Plates were sanded on the underside to encourage settlement (generally true for marine invertebrates, but see Flores and Faulkes 2008). The first set of plates was deployed at 12:00, resulting in deployment periods between 12:00-18:00, 18:00-00:00, 00:00-06:00, and 06:00-12:00. This allowed us to compare settlement rates during the morning, afternoon, evening and night. In conjunction with the 6 hr plates, we deployed an additional set of five plates for the total 24 hr period to assess whether the 54

combined settlement recorded during the short 6-hour deployments mirrors the total settlement over the sampling period. On collection, the total number of settled larvae on the underside of the plate was counted in the laboratory using a 40 x magnification microscope. The study was repeated three dates in the summer of 2010 (Aug-25 – Aug-26, Aug-28 – Aug 29, and Sep-18 – Sep-19). During this period the timing of sunrise ranged from 06:24 and 06:54 and the timing of sunset ranged from 20:04 and 19:18 (August and September, respectfully). Additionally, tidal cycles varied between the three replicates with low tide occurring at 03:55 and 05:42 for Aug-25 and Aug-28, respectively, and low tide occurring at 12:25 on Sep-18.

4.4.3. Experiment 2 - Light Treatments - Daytime

To assess whether Ciona larvae have shadow response behaviour, we deployed a set of twelve 20 × 20 cm PVC settlement plates, six with three flashlights fixed 20 cm below the settlement plate and six with 3 wooden blocks fixed (replicating the shape of the flashlights) set out at a depth of 2 m for 12 h (10:00 – 22:00). The “faux” flashlights were installed to ensure the physical presence of the flashlights, which could potentially alter the flow around the plate, did not influence larval behaviour. Plates were deployed in a spatially random order, separated by an approximate distance of 25 m. On collection, the total number of settled larvae on the underside of the plates was counted.

4.4.4. Experiment 3 - Light Treatments - Nighttime

Using the same set-up as Experiment 2, we deployed the plates during the night with flashlights located above and below the settlement plates. By deploying the plates at night, we could control the direction of light (e.g., from below only) and, therefore, create shaded conditions on the upper-surfaces, as well as the undersurfaces, of the settlement plate, which allowed for the separation of shade and downward-facing surfaces. We compared four different light treatments (Fig. 4.1): above light and below light (ALBL); above dark and below light (ADBL); above light and below dark (ALBD); and above dark and below dark (ADBD). Six plates for each treatment were deployed from 20:00 to 08:00, approximately 30 minutes before and after sunset and sunrise, respectively, at a depth of 1.3 m. The short deployment period (12 hours) meant sediment build-up on the upper-surfaces was minimal and unlikely to influence larval settlement, a problem previously highlighted by McDougal (1943). Both upper and under surfaces of the settlement plates were sanded to encourage settlement and on collection, the total number of larvae settled on the upper and under-surfaces was counted.

Figure 4.1. Experimental set up for different light and shading treatments for larval settlement plates in Experiment 3: ADBL - lighting on under-side only; ALBD - Lighting on upper-side only; ALBL - lighting on both upper and under-sides; ADBD - no lighting

4.4.5. Experiment 4 - Edge Effects

To assess the small-scale behaviour of Ciona larvae once surface contact had been made, we also mapped their final location on settlement plates. Using the plates from experiments 2 and 3 we divided each plate into 4 zones (2 cm wide) radiating from the center to the edge of the plate. This created 5 zones, although the central zone of the plate (4×4 cm) was not included in the analysis due to proximity to the supporting rope. All recruits were counted in each zone and converted to density (larvae.cm-2) as each zone consisted of a different area (e.g., the inner zone has an area of 48 cm2 and the outer zone has an area of 144 cm2).

4.5. Results

4.5.1. Timing of Settlement

The settlement rates of Ciona over a 24-hour period in Boughton River were very high (mean ± SD: 3390 ± 1142.31, for Replicate 2) but also varied considerably between the three replicates (mean ± SD: 18.6 ± 12.82, for Date 1) (Fig. 4.2). The peaks in settlement during the 24-hour period also varied, but for all three replicates the highest settlement rates were observed during the daytime (morning or afternoon). For

all three dates, settlement rates varied significantly with time of day (One-way ANOVA test of log- transformed settlement data: Date 1 - F = 8.18, p < 0.01; Date 2 - F = 20.38, p < 0.01; Date 3 - F = 125.8, p < 0.01). On Date 1, settlement rates during the afternoon were significantly higher than in the morning, evening, or at night (Tukey HSD multi-comparison test: p = 0.006; 0.009; and 0.003, respectively). On Date 2, settlement rates during the morning were significantly higher than in the afternoon, evening, or night (p < 0.001 for all three comparisons). On Date 3, settlement rates did not differ between the morning and afternoon (p = 0.97) but both were significantly higher than that observed in the evening and at night (p < 0.01 for all comparisons). For all dates, settlement rates did not differ significantly between the evening and night (p = 0.93, 0.92, and 0.06 for Dates 1, 2, and 3, respectively).

It is important to note that weather records for Boughton River indicate that cloud conditions were similar during all three replicates, and therefore the low settlement rates during Date 1 (Aug 25-26) cannot be explained by variations in sunlight (http://www.weatheroffice.gc.ca). Additionally, the tidal cycle for Date 1 was similar to Date 2, suggesting that the timing of low and high tide do not influence the rates of larval settlement; this is also evident for Date 3 as the rates of larval settlement remained similar to Date 2 despite the opposite tidal cycle. However, despite the lower settlement rates on Date 1, we still observed the same pattern of settlement, with higher rates observed during daylight hours.

4.5.2. Light Treatment – Daytime

Ciona larvae were found in significantly higher densities on the undersurfaces of control plates than on the light treated plates (One-way ANOVA, F = 24.68, p < 0.001), averaging 1742.5 ± 698.4 (SD) larvae for the control plates and 240 ± 246.9 (SD) larvae for the light treated plates.

4.5.3. Light Treatment – Nighttime

For all four treatments, settlement was considerably higher on the under-surfaces of the settlement plates than on the upper-surfaces; a total of just 7 larvae observed on the upper-surfaces of all settlement plates (three larvae in ALBD and four larvae in ADBD). These larvae were found on both light-treated and dark surfaces. Therefore, due to low settlement rates and frequent zero counts, these data were not statistically analysed further.

Figure 4.2. Density of Ciona intestinalis recruits on the undersurface of settlement plates in Boughton River, Prince Edward Island, Canada, after sequential 6-hr deployments starting at 12:00 and the cumulative 24-hr deployment. Values represent means of 5 settlement plates, and bars are standard errors.

Consistent with our first experiment, the number of Ciona larvae settling during the night time experiment was considerably lower than that recorded in the daytime (Fig. 4.3), with the exception of ALBD (mean ± SD = 1387 ± 523.55), which re-created similar conditions to those found in the daytime (i.e., light above, dark below) and had a similar larval densities to the control plates in Experiment 2 (1742.5 ± 698.4).

On the under surfaces of the settlement plates, larvae were found recruiting in all 4 treatments, with the highest settlement recorded in ALBD (approx. 7x higher than ADBD, the next highest). The direction of the light treatment had a significant effect on larval recruitment rates (Table 4.1), e.g., the lowest settlement rates were observed on the plates treated with light from below (ADBL & ALBL). However, there was also a significant interaction detected between the two treatments (Table 4.1), suggesting a conditional affect (e.g., light from above increases settlement rates only when the undersurface is dark).

4.5.4. Edge Effect – Daytime

Ciona recruits were found in all 4 zones of the plates, on both the control and light-treated plates (Fig. 4.4). Larval densities were generally higher towards the middle of the plate (Zone 1), for both light-treated and control plates, and decreased towards the plate edge. However, this effect was not statistically significant for the light-treated plates (One-way ANOVA, F = 0.49, p = 0.693). On the control plates, density did not differ significantly between Zones 3 and 4, the two outer zones (Tukey HSD test; p = 0.169), but did between Zones 1 and 4 and Zones 2 and 4 (p = 0.02 and p = 0.03, respectively). Settlement density did not differ significantly among zones 1, 2, and 3.

Figure 4.3. Average Ciona intestinalis larval settlement on upper and under surfaces settlement plates in 4 different light treatments: ADBL - lighting on under-side only; ALBD - Lighting on upper-side only; ALBL - lighting on both upper and under-sides; ADBD - no lighting, deployed in Boughton River, PEI over a 12 hour period during the night.

Table 4.1. 2-way ANOVA of Ciona intestinalis larval settlement rates on horizontally deployed 20x20 cm settlement plates when treated with artificial light sources from above and below.

Df Mean Sq F value P value

Above 1 2100417 29.696 <0.001

Below 1 3307838 46.767 <0.001

Above:Below 1 2006817 28.373 <0.001

Residuals 20 70730

4.5.5. Edge Effect – Nighttime

Larvae settled in all 4 zones for all 4 treatments, albeit in low densities for ADBL and ALBL (Fig. 4.5). Settlement density only differed significantly among zones for ADBL (Kruskal Wallis test, P < 0.01) and ALBL (One-way ANOVA, p < 0.01), but not for ALBD (p = 0.097) and ADBD (p = 0.07). For ADBL, settlement was lower towards the middle of the plate (Zone 1) than for the edge of the plate (Zone 4). In contrast, settlement in ALBL showed the opposite effect such that larval densities in Zone 4 were significantly lower than zones 1, 2 and 3 (p < 0.01 for all three comparisons), although recruit densities were very low. Settlement in Zones 1, 2 and 3 did not differ significantly. Although settlement densities did not differ significantly among zones for ALBD, the distribution of settled larvae did show a clear decrease towards the plate edge (Fig. 4.5).

4.6. Discussion

Our results confirm the importance of light during the larval settlement phase of Ciona intestinalis by influencing the timing of settlement, the rate of settlement, and larval behaviour post-surface contact. However, our results also demonstrate that surface orientation has an important influence on larval distribution, more so than light, as settlement remains denser on under-surfaces in complete darkness and low on upper-surfaces when shaded.

Figure 4.4. Distribution of Ciona intestinalis larval settlement on the under-surfaces of settlement plates deployed during the daytime for 2 different treatments; light on under-side, and no light. Settlement plates were divided into 4 radiating zones of 2 cm width; 1 = inner zone, 4 = outer zone (plate edge)

In previous studies, light has been found to induce gamete release (Lambert and Brandt 1967), influence swimming behaviour (Zega et al. 2006) and substratum choice (Rius et al. 2010) in Ciona, but whether light influences the timing of settlement had not been investigated. In Experiment 1, we found significantly higher rates of settlement during the daytime (morning and afternoon) than in the evening or night, suggesting Ciona larvae require lighter conditions during daylight hours to locate shaded surfaces for settlement. There was, however, settlement at night, indicating settlement is not exclusive to daylight hours. Therefore, the higher settlement rates during the daytime may not necessarily be directly driven by the presence of sunlight but by the creation of shadows, making overhanging surfaces easier to detect (as suggested by Olson 1983). As all surfaces are dark during the night, they are harder to locate and their presence can go undetected. Therefore, we consider nighttime settlement is the result of „chance encounters‟ rather than surface detection. However, upon contact, the larva still needs to „decide‟ whether to attach or seek an alternative, suggesting settlement during the night is not a completely passive process.

Figure 4.5. Distribution of Ciona intestinalis larval settlement (± S.E.) on the under-surface of settlement plates at night for 4 different light treatments: ADBL - Light on from below only; ALBD - Light from above only; ALBL - Light from above and below; ADBD - No light. Settlement plates were divided into 4 radiating zones of 2 cm width: 1 = inner zone; 4 = outer zone (plate edge).

4.6.1. Light Treatments

The results from Experiment 2 confirm Ciona‟s preference for shaded surfaces, with higher densities found on the control plates. Despite creating “brighter” conditions on the under-surface of the settlement plates, reducing the shadow effect, we still observed larval settlement on all plates, indicating that the presence of light does not completely deter settlement. There are three possible explanations for this: 1) any downward facing surface is adequate but shaded surfaces are more easily detected; 2) the intensity of light produced by the flashlights was not sufficient enough to deter settlement; or 3) the “desperate larvae” theory (Elkin 2007), where larvae with limited and short planktonic periods may settle in less than optimal conditions to ensure settlement occurs. In this case, any hard substrata shaded or not, may do.

Analysis of the settlement plates in Experiment 3 confirmed our findings for Experiment 2: settlement on undersurfaces treated with light (ADBL and ALBL) was significantly lower than the control plates (ADBD) and the plates light-treated from above only (ALBD). The higher settlement rates on the under- surfaces of the control plates in dark conditions (ADBD), again indicates that Ciona larvae do not require the presence of light to detect downward facing surfaces and that geotactic behaviour alone is sufficient. However, when light is introduced from above the plate, creating a shaded under-surface (similar to daylight conditions), we found a dramatic increase in larval settlement, indicating that although larvae do not necessarily require light to locate downward facing surfaces (ADBD), its presence certainly acts as a guide and amplifies settlement rates. Light is clearly influential during settlement, but, if light was the principal driving factor for locating a shaded surface, irrespective of orientation, we would expect to find increased levels of settlement on the upper-surfaces of the settlement plates treated with light from below only (ADBL) – the reverse conditions of ALBD. However, we continued to find zero settlement of recruits on the upper-surfaces. There are two possible explanations for this: 1) Ciona’s preference for overhanging surfaces is far stronger than its preference for shaded surfaces, or 2) the rate of larvae supply to under-surfaces is higher than the upper-surface. If the probability of larvae coming into contact with both upper and under-surfaces was equal, the “desperate” larvae theory (Elkin 2007) suggests we would find more larval settlement on the shaded upper-surfaces. However, the extremely low settlement observed on upper-surfaces suggests swimming behaviour prior to settlement is conducive to downward facing surfaces and reduces the probability of contact with upward facing surfaces.

Woodridge (1924) suggested that the shadow response for B. schlosseri assisted with the detection of shaded surfaces, resulting in higher settlement rates on downward facing surfaces. This is also true for Ciona larvae, but it does not explain the high settlement on under-surfaces when light is absent (ADBD). Svane and Dolmer (1995) also found surface orientation remained significant when light was absent and suggested that geo-negative behaviour was more influential for locating downward facing surfaces than light. Our results suggest that geotactic behaviour is more influential than phototactic behaviour for Ciona larvae also (but see Rius et al. 2010), but, additionally, that geo-negative behaviour is a conditional response to darker conditions. This conditional response would increase the probability of locating the shaded undersurface of the settlement plates in treatment ALBD, but reduce contact probability with the upper-surfaces in treatment ADBL (Fig. 4.6). However, when light is absent (ADBD), we propose that Ciona larvae return to a default geo-negative behavioural response, the same behaviour observed during the very early larval stages prior to ocellus development (Nakagawa et al. 1999). This behaviour would result in geo-negative swimming behaviour throughout the night, which increases the probability of coming into contact with an overhanging surface and avoid upward facing surfaces. This behaviour would

also ensure that if settlement occurs during the night, the chosen substratum will most-likely be shaded during the daytime. It is clear from these results that both phototactic and geotactic behaviours play important roles during settlement, but ultimately we believe it is a conditional geo-negative swimming response to dark conditions that drives the resulting pattern of recruitment, rather than photo-negative behaviour alone.

4.6.2. Edge Effects

The small-scale distribution analysis of larval densities in Experiment 2 confirmed the presence of an edge effect, where a decrease in recruitment was found in Zone 3 and particularly Zone 4 of the control plates. Assuming the probability of contact is equal for all four zones, the low recruit densities recorded in Zone 4 suggest there is a period of migration away from the plate edge after contact is made. This pattern of larval distribution was not present on the light-treated plates, where a more even distribution of larval density was observed. This change in larval distribution indicates that the observed edge effect was a response to variations in light concentration, with darker conditions found towards the centre of the plate where larvae were concentrated. How the larvae are able to locate the darker surfaces is not exactly clear. There are two possible mechanisms: 1) the larvae can detect a gradient in light concentration and move towards darker conditions; or 2) there is a period of exploration after first contact, where the larva searches until the light conditions are adequately low.

For the plates deployed at night (Experiment 3), we did not find any significant edge effects, despite a distinct decrease in larval density towards the plate edge in treatment ALBD. It is important to note that, as a result of the experimental design, the distribution of light from the flashlights was concentrated towards the centre of the plate and therefore would have been weaker towards the plate edge. Some light will have been diffused towards the plate edge and possibly underneath, providing enough light to create a slight edge effect, but it would not have been as strong as that created by natural light during the daytime. For the other light treatments, the differences in larval densities between zones were either not significant or the densities of larvae we too low (<1.larva.cm-2) to make any clear conclusions.

Figure 4.6. Geotactic behaviour of Ciona intestinalis larvae during settlement in response to different lighting conditions: 1) during the daytime and treatment ALBD; 2) treatment ADBL (light from below); and 3) during the night time (ADBD).

The presence of an edge effect can have important ramifications for choice of settlement plate size. As the edge effect is caused by light, the width of the affected area can be assumed to remain constant for different plate sizes and, therefore, the percentage area affected would be larger for smaller plates. For example, if the affected outer-band of the settlement plate remained at 2-4 cm in thickness, the percentage area affected for a 20 x 20 cm plate would be between 36% and 64%, whereas for a 10 x 10 cm plate the percentage area affected would be between 64% and 94%, which supports the findings of Delaney et al. (in review [Appendice B]), who found that both plate size and deployment depth can influence Ciona larval densities on the under-surfaces of settlement plates. As settlement plates are a common method used for early detection of introduced populations of Ciona, we recommend that the size of plate being used be as large as possible to increase the area of the plate unaffected by light and subsequently increase the probability of detection.

We have shown that both phototactic and geotactic behaviours play important roles during settlement of Ciona larvae. However, our results strongly suggest that a geo-negative swimming behaviour, in response to the detection of dark conditions, has the greater influence on the distribution of larval settlement by ensuring the detection of overhanging surfaces, even when light is absent. Once surface contact is made, larvae continue to behave photo negatively and avoid bright conditions. During daylight hours, the interchanging phototactic and geotactic behaviours throughout the planktonic stage (dispersal and settlement), combine to guide larvae to optimal conditions for recruitment. However, during the night when light is absent, larvae rely on a geo-negative swimming behaviour to increase the probability contacting overhanging surfaces. Once surface contact is made, without light to guide behaviour, there is no exploratory period and, therefore, no edge effect is present.

Understanding the behavioural mechanisms that influence larval settlement is important for further our understanding of environmental preferences, post-settlement survival, and ultimately patterns of recruitment. With this knowledge, we can improve our ability to accurately survey and monitor the size and distribution of marine populations, which have important consequences for environmental management, such as species detection, reserve design, and conservation.

Chapitre 5

Established invasive species may increase biotic resistance to new invaders: unexpected negative effect of caprellid amphipods on an invasive tunicate

5.1. Résumé

Comme le nombre d‟introductions d‟espèces non indigènes (ENI) continue d‟augmenter, les écologistes font face à des opportunités nouvelles et uniques d‟observer des interactions entre des espèces qui ne coexistent pas naturellement. Ces interactions peuvent avoir d‟importantes implications pour les processus d‟envahissement, car elles déterminent potentiellement quand les ENI deviennent répandues et abondantes, survivent en faibles nombres, ou ne réussissent pas à s‟établir et disparaissent. Bien que de nombreuses études aient naturellement mis l‟accent sur les interactions entre les ENI et les espèces indigènes pour examiner leurs impacts et la résistance des communautés qui subissent des invasions, peu d‟études ont examiné les effets des ENI les unes sur les autres. Dans certains cas, les interactions peuvent faciliter les processus d‟envahissement pour l‟une ou les deux espèces (c.-à-d. « effondrement de l‟envahissement »), mais la compétition ou la prédation peuvent également mener à des interactions négatives. La récente introduction de l‟ascidie jaune, Ciona intestinalis, à l‟île du Prince-Édouard (Canada) menace l‟aquaculture des moules via d‟importantes biosalissures sur l‟équipement et les moules. En effectuant une surveillance à grande échelle et des expériences de terrain à petite échelle, nous avons montré que le recrutement de Ciona diminue drastiquement en présence d‟amphipodes Caprellidae, incluant l‟ENI Caprella mutica. Ces amphipodes Caprellidae ne sont pas connus pour être des prédateurs des larves de tuniciers et cette étude apporte donc un exemple imprévu et frappant d‟une espèce envahissante qui contribue à la résistance biotique d‟une communauté native envers un envahisseur secondaire et d‟effets non-additif lors d‟invasions multiples.

5.2. Abstract

As the number of introductions of non-indigenous species (NIS) continues to rise, ecologists are faced with new and unique opportunities to observe interactions between species that do not naturally co-exist. These interactions can have important implications on the invasion process, potentially determining whether NIS become widespread and abundant, survive in small numbers, or fail to establish and disappear. Although many studies have naturally focused on the interactions between NIS and native species to examine their interactions and the biological resistance of the recipient community to invasion, few have examined the effects that NIS have on each other. In some cases, interactions can facilitate the invasion process of one or both species (i.e., “invasional meltdowns”), but competition or predation can lead to negative interactions as well. The recent introduction of the vase tunicate, Ciona intestinalis, in Prince Edward Island (Canada) has harmed mussel aquaculture via heavy biofouling of equipment and mussels. Through both a broad-scale survey and small-scale field experiments, we show that Ciona recruitment is drastically reduced by caprellid amphipods, including the NIS Caprella mutica. This study provides an exciting example of how established invasive species can negatively impact the recruitment of a secondary invader, highlighting the potential for non-additive effects of multiple invasions.

5.3. Introduction

The threats posed by non-indigenous species (NIS) have been well documented (Elton 1958; Carlton 1993; Lodge 1993; Mack 2000; Occhipinti-Ambrogi and Savini 2003), and their potential impact on native species and community structure has raised concern among scientists and conservationists (Lodge 2006; Occhipinti-Ambrogi 2007). Although the majority of introduced species fail to establish or affect the native community (Williamson 1996), the potential for damage caused by the few that do become truly invasive (i.e., widespread and dominant [Colautti and MacIsaac 2004]) is such that there is keen interest in understanding the determinants of a successful invasion (Ricciardi and Rasmussen 1998; Stachowicz et al. 2002; Pimental 2005). As the number of introduction events continues to rise (Ruiz et al. 2000), scientists are presented with increasing opportunities to investigate the invasion process and identify why some species are globally successful while others stumble at the first hurdle (Sakai et al. 2001).

The invasive process involves multiple steps (Colautti and MacIsaac 2004), including survival during transport and after introduction, but ultimately a successful invasion (i.e., either a locally-dominant or a widespread NIS) relies on the ability of the NIS to survive interactions with other species already present, overcoming the “biotic resistance” of the recipient community (Elton 1958; Stachowicz et al. 1999; deRivera et al. 2005). Typically, this resistance is thought to come from native species, but as the communities accumulate NIS over time, the likelihood of interactions between NIS increases. The first thinking in this regard focussed on the potentially positive effects that established NIS could have on subsequent arrivals by outcompeting or preying on native species or by modifying the environment (e.g., ecological engineers [Cuddington and Hastings 2004]). This process of positive feedback, where the impact of one NIS facilitates the establishment of another (O‟Dowd et. al. 2003; Helms et. al. 2011) has become known as an “invasional meltdown” (Simberloff and Von Holle 1999) and is now a popular notion in invasion ecology. The ubiquity of this phenomenon, however, remains unknown (Simberloff 2006) and to date there are few examples that support this idea (Ricciardi 2001; Heimpel et al. 2010; Relva et. al. 2010; Green et al. 2011; Montgomery et al. 2012).

In contrast, the possibility of negative interactions between NIS has drawn little attention, perhaps due to an implicit assumption that NIS interactions are unlikely due to the relatively low numbers of NIS within many communities. There have been, however, reports from both terrestrial and marine systems where NIS interact negatively, either through predation (e.g., Lohrer and Whitlatch 2002; Platvoet et al. 2009; Buric et al. 2009; Dick and Platvoet 2000; La Pierre et al. 2010) or resource competition (Uygun et al.

1994; Ricciardi and Whoriskey 2004; Griffen et al. 2008). Although there are no records of such interactions resulting in the local extinction of an NIS, they certainly have the potential to influence the speed and pattern of invasion as well as the overall impact of the respective invaders.

This lack of attention to the prevalence and potential of negative NIS interactions could be due to a general tendency to focus on invasion “horror stories” and to identify worst-case scenarios for predicting impacts on native species and industry. Negative interactions between NIS then become a lower priority for conservation and management, especially when a newly-arrived NIS negatively affects an established invader (e.g., the replacement of the zebra mussel, Dreissena polymorpha, by the quagga mussel, D. bugensis; Riccardi and Whoriskey 2004). In the contrary sense, they may have an unexpectedly important role in the invasion process by ultimately reducing the probability of establishment or eventual impacts of newly-arriving NIS. In other words, if one invader can negatively impact the establishment and survival of another, would this not benefit the native community? By understanding the extent and potential of such negative interactions, scientists could provide new insights, not only in the theoretical framework of invasion ecology, but also for novel management options. More precisely, such knowledge would expand our ability to carry out multispecies risk assessments, i.e., the net effect of multiple invaders on a recipient community, where information on both positive and negative NIS interactions and the circumstances within which they arise is essential (Leung et al. 2012).

In this study, we examine the interactions between two species of caprellid amphipods (including the NIS Caprella mutica) and the invasive tunicate Ciona intestinalis (hereafter referred to as Ciona) in Prince Edward Island (PEI), Canada. Ciona is a notorious invasive species that has highly impacted aquaculture in PEI (and elsewhere in the world [Millar, 1958; Kang et al., 1978; Uribe and Etchepare, 1999; Tan et al., 2002]) and has been a focus of management effort since its arrival there in 2004 (Locke et al. 2007). In PEI, the reproductive season of Ciona starts in summer and peaks in fall (Ramsay et al. 2009), and recruitment can occur extensively (S. Collin, unpubl. data). C. mutica is less problematic, and in spite of its global success as an invasive species (Willis et al. 2004; Bushbaum and Gutow 2005; Ashton et al. 2007; Locke et al. 2007), it has received little attention, both in PEI (where it was first noted in 1997 [Locke et al. 2007]) and elsewhere. Before our study, little was known about the predatory potential of C. mutica, and although it was suspected to be important (Epelbaum et al. 2009), only very recent work has confirmed that they are able to feed on tunicate larvae in the laboratory (Rius et al. in press). No field evidence has been available, either to confirm the reality of this interaction or to estimate the demographic conditions under which it can occur. Here we specifically investigate, by means of a broad-scale survey,

small-scale field experiments, and a time-series study, the potential for caprellids to reduce the recruitment of Ciona and generally discuss the possibility of established NIS contributing to the biotic resistance of the native community towards a secondary invader.

5.4. Methods

5.4.1. Observational evidence of the Caprella-Ciona interaction

In August 2008, as part of a larger study on the dispersal of a nascent Ciona population, we setup a grid of 88 stations that covered the entirety of Boughton River, a 12-km2 inlet in eastern PEI used extensively for mussel aquaculture (for further details see Collin et al. 2013). Each station consisted of a grey 20x20-cm PVC plate suspended horizontally at a fixed 2-m depth on a line anchored to the bottom. Plates were replaced every two weeks to examine the progression of the invasion. A design flaw, however, led to the loss of approximately half the stations during the second sampling period. These stations were quickly replaced with new equipment for the third sampling period. Unfortunately, the original stations were not cleaned at that time, leading to the development of a fouling community consisting primarily of blue mussels (Mytilus spp.), another invasive tunicate (Botryllus schlosseri), and caprellid amphipods (Caprella spp.) on the supporting structure (i.e., buoy, line and anchor). Thus, upon recovery, we found widely-varying levels of caprellid recruitment on the plates themselves (0 - ~9300 ind.plate-1), presumably due to a combination of migration from the fouling community on the adjacent anchor line and de novo colonisation.

As initial field observations suggested that Ciona recruitment was very low on plates heavily colonized by caprellids, we quantified both variables. Plates were brought back to the lab in separate plastic containers where the caprellids were then removed and preserved in 95% alcohol. Ciona recruitment was measured by counting the total number of recruits on the underside of the plate using a stereomicroscope at 40x magnification. To estimate caprellid abundance, samples were filtered to remove alcohol and left to “drip- dry” for 10 minutes. The total wet biomass was measured, and then a sub-sample was removed by hand, weighed and counted. The weight and number of individuals from the sub-sample was then used to estimate the total number in the sample. The caprellids samples were a mix of adult and juvenile C. linearis and C. mutica, but, due to the similarities between the smaller juveniles (< 10 mm in length) of both species, accurate morphological identification of all specimens was impossible, and we thus did not

Figure 5.1. Log-scale plot of Ciona intestinalis abundance (+1) vs. caprellid abundance (+1) for the original bay- wide survey observations on 20x20-cm PVC settlement plates in Boughton River, Prince Edward Island, collected Sept 18 2008.

attempt to quantify the proportions of the two species. The relationship between Ciona recruitment and caprellid abundance was examined using a log-log regression of all stations (original and replaced) and original and replaced stations separately.

5.4.2. Experimental demonstration of the Caprella-Ciona interaction

To assess whether the observed variation in Ciona recruitment seen in the synoptic survey (see Results) was indeed caused by the presence of caprellids we set up a small-scale field experiment where caprellid densities on the settlement plates were manipulated. As caprellids are small and mobile, they are exceedingly difficult to manipulate in the field. Thus, we devised an indirect method that involved the collection of caprellids and their subsequent transplantation in close proximity to settlement plates, which

in turn enhanced caprellid colonization of the plates – a technique that mimicked the problem that led to our synoptic survey (see above). Although this method increased the probability of caprellid colonization, we could not completely control caprellid densities or which species of caprellid recruited onto individual plates and, therefore, a mixture of both C. linearis and C. mutica in various abundances was examined.

The precise technique involved collecting caprellids using circles of black plastic mesh (25-cm diameter; 0.3-cm mesh size), attached to a line at a depth of 1.5 m (the approximate depth of the top of the submerged mussel farms). The mesh circles were kept flat using cable ties to attach them to plastic circular frames of the same diameter. These collectors were deployed vertically in the water column in close proximity (approximately 10 m) to mussel aquaculture sites (“farms”), where caprellids are often found in large numbers (S. Collin, pers. obs.). After two weeks the mesh circles were transferred to a new set of five replicate lines (one mesh circle per line) and attached at a depth of 2.2 m, with a 20 x 20 cm plate attached 20 cm above (i.e., at 2 m, the depth of the original synoptic survey). As a control for this “caprellid-enhanced” treatment, mesh circles that had not been previously colonised by caprellids were attached to five other lines with plates and interspersed among the caprellid-enhanced plates, all within an area of approximately 0.1 km2. This “control” treatment was not intended to prevent natural caprellid colonization, but rather to contrast with levels found on the enhanced-caprellid plates. The experiment ran for four days in an area of Boughton River where Ciona settlement rates are very high (~ 2800 larvae.plate-1.day-1 [S. Collin, unpubl. data]). This period was the minimum amount of time considered for both sufficient movement of caprellids to the treated plates and natural Ciona larval recruitment to occur while also short enough to keep natural colonization of caprellids on the “control” plates low. The experiment was repeated twice, beginning on Sep 14th and Sep 29th 2010. On collection, caprellids were immediately removed from the plates, counted and identified to species when possible; Ciona recruits on the underside of each plate were counted using a stereomicroscope at 40x magnification within 24 h of collection. A non-parametric Mann-Whitney U test was used (due to non-normal distribution of data) to compare both caprellid and Ciona abundances between the two treatments (i.e., caprellid-enhanced vs. control plates).

5.4.3. Temporal and spatial variation of caprellid colonization

We measured variation in caprellid recruitment in July and August 2009 at four locations within and outside a mussel farm (approximately 100 m away). Because caprellids brood their young and have no larval stage, they have a short dispersal range, limited to crawling and swimming short distances (Ashton

et al. 2006), although populations can develop rapidly on navigational buoys far from shore (pers. comm., C. McKindsey, DFO Canada). The placement of stations inside and outside the mussel farms was thus intended to assess the importance of the proximity of source populations in the temporal dynamics of caprellid dispersal. The collectors were deployed for eight consecutive one-week periods from 11-Jul-09 to 29-Aug-09. On collection the mesh circles were removed and replaced. The total number of caprellids colonising the mesh circles was counted, and measurements were taken of all caprellids present, up to 50 randomly-selected individuals. As most individuals were juveniles, no attempt was made to distinguish between the two species. A two-way ANOVA was used to assess the significance of date and station location (inside vs. outside mussel farms) on caprellid recruitment.

5.5. Results

5.5.1. Observational relationship of the Caprella-Ciona interaction

Ciona recruitment was negatively correlated with caprellid abundance across the 72 stations (of the 88 deployed, 16 were lost) regardless of whether analysed for all stations (original and replaced: slope = - 0.70, r2 = 0.45, p<0.001 [Fig. 1]; n = 72), replaced stations only (slope = -0.73, r2 = 0.36, p<0.001; n = 33), or original stations only (slope = -0.6, r2 = 0.44, p<0.001; n = 39). Caprellid abundance ranged from 0 to 9329 ind.plate-1 (for all stations), and Ciona recruitment (ranging from 0 – 2655 ind.plate-1) generally only occurred where caprellid abundance was less that 150 individuals per plate (Fig. 1), a density of 0.37 ind.cm-2 based on the just the downward-facing surface area of the plates where most caprellids were observed. When caprellid abundance was higher than this threshold, we found little, if any, Ciona recruitment (mean = 1.47 ± 6.47 [SD]; max = 39).

Figure 5.2. Caprellid recruitment on PVC settlement plates treated with caprellids and control plates deployed on Sep 14 (Exp. 1) and Sep 29 (Exp. 2), 2010 in Boughton River, Prince Edward Island (standard error bars).

5.5.2. Experimental demonstration of the Caprella-Ciona interaction

As intended, caprellid densities were significantly higher on the caprellid-enhanced plates relative to the control plates (7x higher in Trial 1 [W = 25, p = 0.012] and 50x higher in Trial 2 [W = 25, p = 0.01]; Fig. 2), reaching average levels of 0.24 and 0.06 ind.cm-2 for the two trials, respectively. Thus, caprellids from the mesh circle colonized the plate, reaching levels much higher than the natural colonization of caprellids seen on the control plates. However, the level of natural colonization during the first trial was almost as high (0.04 ind.cm-2) as that seen on the caprellid-enhanced plates in the second trial (0.06 ind.cm-2).

In both trials, the recruitment of Ciona was markedly lower on caprellid-enhanced plates (Fig. 3; W = 0, p = 0.01 and W = 0, p < 0.01, for trials 1 and 2, respectively). In Trial 1, Ciona recruits were 20 × higher on control plates relative to those with enhanced levels of caprellids. In Trial 2, we found an even stronger effect with no Ciona recruiting on the caprellid-enhanced plates whereas control plates had recruitment approximately 50% of the level seen in the first trial (Fig. 3). Overall, the ratio of C. linearis to C. mutica for both trials was approximately 3:2 for the caprellid-enhanced plates.

Figure 5.3. Recruitment densities of Ciona intestinalis on settlement plates treated with caprellids (“caprellid- enhanced”) and control plates deployed September 14th (Exp. 1) and September 29th (Exp. 2) 2010 in Boughton River, Prince Edward Island (standard error bars).

5.5.3. Temporal and spatial variation of caprellid colonization

Although the colonization rates of caprellids varied six-fold over the eight weeks of the experiment (Fig. 5.4), there were no significant differences among the different weeks (Table 5.1), the low sample sizes notwithstanding. There was, however, significantly higher colonization within the mussel farm than outside, with colonization rates over five times higher within the mussel farm compared to locations outside the farm (mean ± SD: 238 ± 203 vs. 41 ± 46, respectively, per collector). There was no interaction detected between time and station location (inside vs. outside). The average size of recruits for all stations (within and outside the mussel leases) throughout the experiment was 0.58 ± 0.01 cm (SE), indicating that the majority of recruits were juveniles (Gierra-Garcia et al. 2002, Boos et al. 2009).

Figure 5.4. Time-series of weekly recruitment of caprellid amphipods (C. linearis and C. mutica) on plastic mesh collectors (25-cm diameter) in Boughton River, Prince Edward Island (standard errors bars)

5.6. Discussion

Our results show a dramatic and unexpected negative effect of caprellid amphipods on the recruitment of an invasive tunicate. Based on the initial and subsequent survey work, this effect appears to occur over both large spatial and temporal extents, and the experimental work strongly suggests that caprellids are indeed responsible for the observed impact. However, the exact mechanism behind this interaction remains unclear, due, in large part, to the poorly-known ecology of caprellids. We propose three possible non-mutually-exclusive explanations for the observed reduction in Ciona recruitment: (1) predation; (2) predator avoidance behaviour; and (3) interference during settlement.

With regard to predation, the flexible diet and diverse feeding methods (browsing, filter-feeding, predation, scavenging and scraping [Caine 1977]) of caprellids suggest they are capable of feeding on Ciona larvae and possibly the early juvenile stage, but consumption of later juvenile stages seems unlikely. In the laboratory, C. mutica have been found to be able to prey upon planktonic and recently settled (1-day recruits) Ciona larvae, but had no effect on 2-week old recruits (Rius et al. in press). This

Table 5.1. Two-way ANOVA results for caprellid recruitment in Boughton River as a function of date and location of the stations (inside vs. outside mussel farm) throughout July and August, 2009

Df Mean Sq F P

Date 7 22602 0.884 0.54

Inside/Outside 1 308309 12.064 0.003**

Date : In/Out 7 11924 0.467 0.84

Residuals 16 25557

result is consistent with laboratory work by Epelbaum et al. (2009), who found C. mutica did not prey on 7-day old Botrylloides violaceus recruits (the only stage investigated), which were relatively large (4.25 mm in diameter). The fast growth rate of Ciona (0.26 to 0.76 mm in diameter in 7 d; Bullard and Whitlatch 2004) suggests that the time period during which newly-settled tunicates are vulnerable to predation is limited (ca. 1-2 d) – once the larvae have metamorphosed and begin to grow, they probably become too large to be consumed (“ontogenetic vulnerability” sensu Osman and Whitlatch 2004). Caprellids, in particular C. mutica, are generally considered to be opportunistic feeders (Caine 1977, Sano et al. 2003, Guerra-García and Tierno de Figueroa 2009), capable of adapting their feeding habits to the food sources available to them (Cook et al. 2007). The rapid Ciona population growth in Boughton River, and the subsequent increase in larvae, has presented an abundant and readily-available food source. Our field results complement the laboratory results of Rius et al. (in press), and when combined, strongly suggest that C. mutica prey upon Ciona larvae in natural conditions.

Although predation seems thus the most likely cause for our observations of lower recruitment of Ciona in the presence of caprellids, we are, however, sceptical that it is the only factor involved. Based on our experiments, we conservatively estimate that a sustained feeding rate of 0.8 and 1.8 larvae.caprellid-1.hr-1 over the 4-d period (for trials 1 and 2, respectively, assuming conservatively that the final caprellid density was the average during the experiment) would be required to explain the differences observed between the two treatments. Even for a large adult C. mutica (2-3 cm in length), this seems an unrealistic quantity of

food and suggests other mechanisms are involved. The most likely alternative is avoidance behaviour by larvae during settlement (Rodriguez et al. 1993). In general, tunicate larvae respond to chemical cues (e.g., Davis et al. 1991; Roberts et al. 2007; Khalaman et al. 2008; Rius et al. 2010), with Ciona preferring substrata with developed biofilms (Szewzyk et al. 1991; Wieczorek and Todd 1997; but see Holmström et al. 1992) and avoiding substrata treated with an extract of the adult tunic (Rius et al. 2010). The only obvious difference between our treated and controlled plates was caprellid abundance, and we thus propose that a chemical cue produced by the caprellids inhibits settlement of Ciona larvae. This kind of response has been observed generally in invertebrate larvae (e.g., Johnson and Strathmann 1989), including tunicates (Grosberg 1981; Young and Chia 1981), and could explain the observed decrease in recruitment. An alternative possibility is competition for space where caprellids physically block or disturb larvae during the settlement process. All three factors described have the potential to impact Ciona recruitment on these settlement plates, but the identification of the relative importance of each, and at which level each species of caprellid contributes, will require further investigation (e.g., under more controlled laboratory conditions).

Despite uncertainty in the precise mechanism, the presence of caprellids can dramatically reduce Ciona recruitment. Moreover, this effect occurs at low densities of caprellids. Although we recorded very high densities in our broad-scale field survey (maximum of 22.5 ind.cm-2, similar to levels seen elsewhere, e.g., 31.9 ind.cm-2 [Ashton et al. 2010]), the threshold density above which recruitment was near zero was orders of magnitude smaller (0.37 ind.cm-2). Indeed, the enhanced densities in our shorter-term experiments were much lower (0.06 to 0.24 ind.cm-2), and the results were equally dramatic. Oddly, however, natural colonization by caprellids in Trial 1 led to an average density of 0.04 ind.cm-2 on the control plates where Ciona recruitment rates were still high (average density of 19.5 ind.cm-2), whereas the caprellid-enhanced treatment in Trial 2 only achieved an average caprellid density of 0.06 ind.cm-2, and Ciona recruitment was zero. Although caprellid density was higher in Trial 1, overall Ciona recruitment was also higher, suggesting that variations in the larval supply of Ciona alter the threshold value. Therefore, the effect of caprellids on Ciona recruitment may be more apparent during the early stages of invasion when larvae are less abundant, whereas their effect may decrease to undetectable levels as the Ciona population and the number of settling larvae increases.

To significantly affect recruitment and thus biofouling by Ciona, caprellids would need to be present throughout the reproduction season. Our time-series data suggests that caprellids were present and actively colonizing at a weekly-rate sufficient to achieve the observed threshold value (average of all 4 stations

[range] for the 9-wk period: 0.28 [0.09 - 0.59] ind.cm-2) during an important period of the Ciona recruitment season. If the caprellids had been allowed to accumulate, they would have certainly reached much higher densities by the end of the study (~ 2.6 ind.cm-2) if no negative intraspecific interactions occurred. However, these results did vary spatially, with caprellid colonization outside the mussel farms reaching only 17% of that observed inside the mussel farms. Thus, proximity to source populations may be a critical determinant of the ability of caprellids to affect Ciona recruitment, especially in situations where new substrata are involved (e.g., aquaculture equipment, seasonal docks). Regardless, the potential impact of caprellids appears to be both temporally and spatially widespread, at least within the domain of this study, and is likely to be more apparent when Ciona recruitment rates are low, i.e. during the early stages of invasion, or when caprellid densities are high.

At the time of this study, the caprellid fauna consisted of two species: Caprella linearis, an indigenous species, and C. mutica, a nonindigenous species. Although the two species were approximately equally abundant in the samples that we examined, we attribute more of the observed effect to C. mutica for several reasons. First, C. mutica has been observed to feed on a variety of foods (e.g., diatoms, algae, crustaceans, detritus, and other caprellids [Sano et al. 2003; Guerra-Garcia and Tierno de Figueroa 2009; Cook et al. 2007, 2010]), including Ciona larvae (Rius et al. in press), whereas C. linearis feeds predominantly on detritus (Guerra-Garcia and Tierno de Figueroa 2009). Second, there have been no previous studies that have demonstrated any predatory behaviour in C. linearis. Finally, the smaller size of C. linearis (1.1 cm for a mature male [Guerra-Garcia et al. 2002] vs. 2-3 cm for a mature male of C. mutica [Boos 2009]) suggests that even if it was predaceous, its capacity, both in terms of the size and number of prey, would be more limited than that of C. mutica. Therefore, if predation was a key contributor to our results, we consider C. mutica, rather than C. linearis, to be responsible for this result. Moreover, even if both species did contribute to the observed effect, we expect that C. mutica will become more dominant over time as it is more competitive and has displaced C. linearis and other native species of caprellids elsewhere (Shucksmith et al. 2009).

Surprisingly, negative NIS-NIS interactions as we have described here are poorly documented in other systems due either to a true rarity of this type of interaction or a lack of effort in documenting them. Those studies that have shown such results have generally focussed on the interactions of either taxonomically or functionally-similar species (e.g., the interaction between freshwater mussels [Ricciardi and Whoriskey 2004] or crabs [Griffen et al. 2008]) or species with a similar region of origin and presumably a shared evolutionary history (e.g., the zebra mussel, Dreissena polymorpha, and the round-nose goby, Neogobius

melanostomus, from the Ponto-Caspian; Ricciardi and MacIsaac 2000). In these cases, such interactions might be expected. More interestingly, the documented cases are usually where the most-recently introduced species negatively affects an NIS already present in the recipient community (e.g., the impacts of the Japanese shore crab, Hemigrapsus sanguineus, on the earlier-established green crab, Carcinus maenus; Griffen et al. 2008). Here we have seen the opposite in both regards – an unexpected negative interaction between species with no taxonomical, ecological, or historical association in which the already established invader adds to the biological resistance of the recipient community to the more-recently- arrived invader (an intriguing extension to this idea is the possibility that the presence of an established invader could even prevent the initial establishment of an inferior species, although this phenomenon would be difficult to detect). In our case, such findings might be more likely due to the strong association marine NIS have with artificial substrata (Glasby et al. 2007), where their concentration might favor negative NIS interactions, but we suspect that such interactions occur widely in other systems and will become more common as the composition of biological communities become more and more composed of NIS.

Overall, this kind of negative interaction between invaders has the potential to influence the invasion process and alter the overall impact on the native community. This outcome could prove fruitful for biocontrol, an approach that has been effective at times in terrestrial and lacustrine systems (e.g. see McFadyen 1998), but less so in marine systems due to the more open and widespread nature of dispersal of marine organisms (Lafferty and Kuris 1996). Traditionally, candidate species for use in biocontrol are sought from the native range of the NIS, which can lead to further NIS introductions (Lafferty and Kuris 1996). The identification of negative interactions among NIS that have already been introduced more easily allows for real-world assessments (i.e., field trials) of their use for biocontrol, but without the usual concerns of intentionally introducing additional NIS (Fagan et al. 2002). Certainly, the introduction of an additional NIS to control another runs the risk of exacerbating impacts on the native community, which is why a precautionary, multispecies-risk-assessment approach should always be applied. However, within the context of our study, the controlling NIS (C. mutica) had already been introduced accidentally prior to investigation, presenting a major advantage as its potential as an agent of biocontrol can be immediately explored without concern for irreversible consequences. Our findings should motivate other scientists to investigate other negative NIS interactions that may prove beneficial to biocontrol. With specific regard to Ciona management, if caprellids are already present, we encourage managers of aquaculture facilities to undertake measures that promote the development and maintenance of large caprellid populations (e.g., providing the physical structures that could harbor source populations) unless greater negative effects on the native community were expected. As with all biocontrol measures, eradication is not the goal, and

thus while we do not believe caprellids have the potential to eradicate Ciona from any particular system, we do think that enhancement of caprellid abundance similar to levels seen in this and other studies (e.g., Ashton et al. 2010) may have the potential to reduce the impact Ciona has on aquaculture and limit or slow the spread, especially in the context of an integrated pest management program.

Finally, we see the inclusion of negative NIS interactions as an important addition to the risk assessment process for invasive species (Kolar and Lodge 2002). Currently, risk analysis is done for individual species with little regard for the overall “invasion context” of the situation. As natural communities become more and more composed of NIS, there will be a need to conduct multispecies risk assessments (Leung et al. 2002) to account for the possible positive and negative interactions that can occur between NIS. Within the more general theoretical framework of invasion ecology, it will be essential to know that whereas two negatives might not make a positive, the whole can at least be less than the sum of the parts.

Chapitre 6

Conclusions

L'objectif principal de ma recherche était d'étudier en détail les mécanismes fondamentaux qui déterminent les schémas de recrutement du tunicier envahissant Ciona intestinalis à grandes (baies) et petites (plaques de fixation) échelles. De plus, j‟ai discuté la façon dont ces mécanismes peuvent améliorer les pratiques actuelles de surveillance et de gestion de la détection précoce de cette espèce tout en limitant la propagation et en contrôlant les impacts. Bien que l'accent ait été mis sur Ciona, l'approche générale de la conception de l'équipement de surveillance, de la répartition de l'effort de surveillance, et de l'enquête des possibilités de lutte biologique peut être largement appliquée à un grand nombre, sinon à la totalité, des espèces potentiellement envahissantes.

6.1. Résultats principaux

La détection d'une population naissante de Ciona à l'Île-du-Prince-Édouard a fourni une excellente occasion de recueillir des données de recrutement au cours des premières étapes de l'envahissement jusqu‟à ce qu‟elle devienne une population bien établie et problématique. Les populations isolées d‟ENI peuvent fournir de précieuses opportunités pour les écologistes afin de préciser les schémas de dispersion des adultes et des juvéniles et de recueillir des données sur le recrutement plutôt que par des mesures alternatives et indirectes de dispersion (ex. : les courants d'eau, Chapitre 2). La fenêtre d'opportunité pour la collecte de ces données est toutefois limitée par la capacité des espèces à se répandre et à établir une population plus grande. Pour Ciona, le délai pour la collecte de données est très court. En effet, seulement 1 an après la détection (été 2008), la population était déjà devenue trop grande et répandue pour la collecte des données (chapitre 3). Dans la majorité des études, les populations d‟ENI sont déjà largement répandues quand elles sont détectées (ou lorsque le monitorage débute), ce qui rend la collecte de données sur la dispersion difficile, voire impossible.

Modéliser la dispersion des populations naissantes a une grande valeur pour concevoir des stratégies de surveillance et de contrôle de la propagation. L'utilisation du tunicier Ciona comme modèle a démontré qu'il est possible d'identifier le potentiel de dispersion et la distance entre les pics de fixation et une population adulte (chapitre 2). Ces données peuvent être appliquées immédiatement à l'effort de gestion de Ciona à l'Île-du-Prince-Edouard. En effet, ces données nous informent que la propagation locale de Ciona

peut être contenue par le contrôle de l'activité humaine entre les baies (potentiel de dispersion <4 km) et que l'effort de surveillance devrait également être alloué à mi-distance (environ 1 km) des sites potentiels d'introduction. Ces données peuvent être collectées pour beaucoup d‟EAE, en particulier les espèces sessiles avec une portée de dispersion courte quand une population envahissante est détectée suffisamment tôt ou quand la répartition des adultes est limitée à un certain site, par exemple, un port (ex. : „‟invasion pinning‟‟).

La détection précoce des ENI permet également de surveiller les schémas de recrutement lorsque la population prend de l'ampleur, un phénomène rarement documenté en écologie marine puisque les populations sont souvent déjà bien établies ou renflouées par des propagules de populations voisines (ex. : la connectivité d‟une métapopulation). Les populations naissantes d‟ENI peuvent nous informer sur la façon dont certaines espèces surmontent les problèmes associés à de faibles densités, à la concurrence avec d'autres espèces, à l'autorecrutement, et à la propagation. De par les envahissements précédents à l'Île-du-Prince-Édouard, nous savons que Ciona est un colonisateur rapide, caractérisé par une courte portée de dispersion (chapitre 2), une fécondité élevée et une maturation rapide. En surveillant l'expansion de la population de Ciona dans la baie de la rivière Boughton, nous pouvons confirmer que, comme espèce pionnière, Ciona dispose d'un délai de latence court entre l'introduction et l‟établissement généralisé (environ 2 ans), mais, plus important encore, que les facteurs influençant le recrutement au cours de cette période de latence sont instables, mais deviennent de plus en plus stable lorsque la population augmente (chapitre 3). La dynamique des premiers schémas de recrutement en 2008 est causée par la dispersion et les changements dans la distribution des adultes, car la population augmente dans le milieu environnant, alors que les modèles plus cohérents associés à une plus grande population en 2009 sont dictés par les conditions plus stables de l'environnement (c.-à-d. le remplissage de la niche). Ceci a des implications importantes pour la répartition de l'effort de surveillance. L'utilisation actuelle des modèles de distribution des espèces (MDE) pour identifier les emplacements appropriés pour la survie de l'EAE (ex. : Therriault & Herborg 2008) et prédire sa distribution, une fois dans un état d'équilibre ont une certaine valeur, mais elle doit être considérée comme une étape préliminaire pour concevoir des stratégies de surveillance. Pour capter les patrons plus stochastiques et dynamiques du recrutement précoce, les gestionnaires ont besoin d‟intégrer le potentiel de dispersion dans la conception de la surveillance. En excluant la dispersion, le gestionnaire risque de concentrer ses efforts de surveillance à l'extérieur d'une région de dispersion de l'espèce cible, ce qui peut retarder la détection, et augmenter la probabilité de résultats faussement négatifs. Afin d'optimiser la répartition spatiale de l'effort de surveillance et d‟améliorer la probabilité de détection précoce, il est impératif que les gestionnaires intègrent à la fois la dispersion potentielle et l'adaptation à l'environnement dans la conception de la surveillance.

Travailler à grande échelle permet d'étudier les facteurs de dispersion et de recrutement (chapitres 2 et 3), mais sans équipement conçu de manière appropriée, tout effort est vain. Pour les animaux ayant une phase planctonique comme Ciona, le comportement des larves pendant la dispersion et la fixation influe grandement sur les schémas de recrutement, la survie postcolonisation et la survie de la population adulte. Par conséquent, la conception du matériel doit être adaptée aux variations du comportement des larves en offrant des conditions appropriées pour le recrutement. Les larves de Ciona démontrent une variété de comportements géotactiques, en réponse aux variations de la lumière (chapitre 4), qui leur permettent de localiser les ombrages sur les surfaces inférieures. Or, il est important que l'équipement de surveillance de Ciona fournisse un endroit ombragé, sur sa surface inférieure, qui est facilement détectable par les larves. Une fois le contact avec la surface établit, le comportement des larves de Ciona est négativement phototactique, présentant des densités plus élevées vers le centre plus sombre de la plaque de fixation. Il a été observé une zone de basse densité, d'environ 2 cm de large, sur le pourtour de la plaque. Cela indique que des grandes plaques de fixation seraient plus efficaces pour détecter les larves de Ciona, car elles offrent une surface d'ombre plus grande qui est facilement détectée par les larves, et un pourcentage plus faible de la surface inférieure qui est affectée par l‟effet de la lumière (chapitre 4 et supporté par Delaney et al. en revision [Appendice B]).

Les facteurs abiotiques peuvent grandement influencer les schémas de recrutement (chapitres 2, 3 et 4), mais la résistance biotique d‟un système envers les ENI joue également un rôle important dans le recrutement et la survie. Dans certains cas, l‟établissement réussi d'une ENI crée une perturbation dans la communauté autochtone, qui agit comme une rétroaction positive et facilite l‟établissement d'autres ENI - une théorie populaire connue comme « invasional meltdown » (Simberloff 1999). Malgré le peu d‟exemples qui appuient cette théorie, elle est devenue un centre d'intérêt pour les écologistes spécialistes des envahissements. Étonnamment, peu d'études se sont penchées sur la possibilité d‟interactions négatives entre les ENI, où la présence d'une ENI peut contribuer à la résistance biotique du système envers un second envahisseur. L'interaction négative inattendue entre deux espèces de caprelles amphipodes, les Caprella linearis (native) et les Caprella mutica (invasive) et le recrutement larvaire de Ciona constitue une alternative intéressante à la théorie de l‟« invasion meltdown » (chapitre 5). Il y a peu d'exemples d'interactions négatives ENI-ENI et leur capacité de contrôle, pour prévenir l‟établissement d'une ENI, est inconnue et difficile à démontrer. Cependant, l'impact significatif que les caprelles ont sur le recrutement larvaire de Ciona suggère qu‟elles pourraient avoir un impact important sur les schémas de dispersion, à la fois temporellement et spatialement. En outre, ces interactions peuvent fournir des indications précieuses pour la lutte biologique. Les risques liés à l'introduction d'une espèce afin d‟en contrôler une autre en font une pratique controversée. En étudiant les interactions entre les espèces qui

existent déjà comme ENI, nous pouvons prendre des décisions éclairées sur les potentiels agents de lutte biologique et réduire les risques d'effets imprévus.

6.2 Contributions

Cette étude contribue à la masse croissante de travail sur les envahissements biologiques en fournissant une évaluation bien documentée et détaillée du processus de mise en place d‟un tunicier envahissant connu et problématique, Ciona intestinalis. Les résultats de ces études donnent un aperçu de la diversité des facteurs qui contribuent aux schémas de recrutement des larves de Ciona, qui intéressent les chercheurs et les gestionnaires de l'environnement. Bien que l'accent soit mis sur une seule espèce, les approches de la collecte et de l'analyse des données, ainsi que leur application en vue d'améliorer les stratégies de surveillance et de gestion, peuvent être adaptées à d'autres espèces envahissantes.

L'impact de Ciona et son potentiel de propagation dans l'est du Canada ont déjà été bien documentés, mais cette étude fournit la première description détaillée des premiers stades d'un envahissement de Ciona, en intégrant les données de recrutement à la fois spatiales et temporelles, dans le but spécifique d‟obtenir des données concrètes au début de la détection. L'importance d'intégrer le potentiel de dispersion, en plus de l‟importance de l'environnement, est indispensable pour optimiser l‟effort de surveillance - un élément clé de la détection précoce qui n'a pas été précédemment utilisé pour Ciona ou d'autres espèces envahissantes. Les études de populations naissantes pour recueillir des données fiables afin de modéliser la dispersion ont une grande valeur et devraient encourager des examens futurs plus approfondis de populations naissantes.

À plus petite échelle, l'importance des comportements phototactiques et géotactiques des larves de Ciona lors de la fixation souligne l'importance d'intégrer ces préférences environnementales dans la conception de l'équipement. Des études antérieures de ces comportements ont suggéré que la lumière est le facteur le plus déterminant à la fixation et que le comportement géotactique n'est influant que lorsque la lumière est absente (Ruiz et al. 2010). Les résultats présentés ici ne sont pas en accord avec ces résultats et suggèrent que le comportement géotactique est lié aux conditions ambiantes d'éclairage. Quand une diminution de la lumière est détectée, les larves de Ciona répondent par un comportement géotactique négatif. Cette réponse conditionnelle augmente la probabilité que des larves entrent en contact avec des surfaces orientées vers le bas. Discutée par Svane & Dolmer (1995) pour les larves d‟Ascidia mentula, sans manipulation de la direction de la lumière avec l'orientation de la surface, cette hypothèse n'avait pas pu être vérifiée. Le dispositif expérimental innovant mis en place ici (chapitre 4) a permis de séparer ces comportements, d‟identifier les rôles relatifs de chaque comportement, pour ensuite confirmer cet effet

conditionnel. Nous ne savons pas pourquoi ces résultats diffèrent des expériences de Ruiz et al. (2010), mais cette différence pourrait résulter des méthodes de laboratoire versus les méthodes utilisées sur le terrain. Les études de laboratoire fournissent des indications précieuses en ce qui concerne le comportement des espèces, mais il est important de reconnaitre que leurs résultats ne sont qu‟une approximation de ce qui se passe dans la nature. La nouvelle conception présentée dans le chapitre 4 atténue ces craintes en effectuant des expériences sur le terrain (manipulation de la lumière n'ayant jamais été précédemment réalisée dans des paramètres de laboratoire), qui ont ensuite fourni une représentation alternative plus précise du comportement des larves au cours de la fixation.

L'effet négatif que les amphipodes Caprellidae ont eu sur le taux de fixation de Ciona (chapitre 5) démontre le potentiel d'interactions négatives ENI-ENI. Comme le nombre d‟introductions d‟ENI augmente, les scientifiques seront confrontés à de nouvelles interactions entre des espèces qui n'ont aucun lien historique, taxonomique ou écologique. Ces interactions peuvent fournir des informations précieuses sur les capacités compétitives et adaptatives des ENI et aider à prédire les impacts écologiques (ex. : la perturbation de la communauté). L'interaction entre les caprelles et les larves de Ciona a été le résultat d'une observation fortuite (créé par une erreur de conception imprévue), qui a ensuite mené à la découverte d'une interaction potentiellement négative entre deux ENI. Une enquête plus poussée n‟a révélé aucun antécédent entre une interaction caprelle-Ciona (ou toutes autres espèces larvaires). Malgré la capacité d‟envahissement documentée des C. mutica, la quantité limitée de littérature disponible sur la biologie et le comportement des caprelles indiquent que leur potentiel d'impacts sur les communautés autochtones n'est pas été entièrement connu. Ceci est la première publication sur l‟impact potentiel des caprelles sur le taux de recrutement des espèces larvaires et il s‟agit aussi de l'une des rares études documentant le risque d'interactions négatives entre les ENI. Le faible nombre d'exemples d'interactions négatives ENI-ENI dans la littérature indique un faible niveau d'intérêt, mais l'omniprésence de l‟espèce suggère que ces nouvelles interactions pourraient être étudiées plus fréquemment. Ces interactions nouvelles ont le potentiel de fournir des indications précieuses, non seulement pour la gestion, mais aussi pour l'écologie et l'évolution.

6.3. Perspectives et considérations futures

L'étude de Ciona dans la baie de la rivière Boughton a fourni des informations et des idées utiles sur les schémas de recrutement et de comportement, mais, malgré cela, des questions restent encore en suspend en ce qui concerne les schémas de recrutement mis en évidence en 2009. Identifier quelles propriétés environnementales du système d‟étude favorisent la fixation était un objectif principal de cette étude.

Pourtant, bien que l'importance de la dispersion et de la variabilité environnementale aient été démontrées, nous n'avons pas pu identifier exactement pourquoi la fixation a augmenté de façon spectaculaire vers l'ouest de la baie.

Les données environnementales de base recueillies dans la rivière Boughton (température et salinité) ne permettent pas d‟expliquer les schémas de recrutement larvaire de Ciona qui pourraient être influencés par une variable environnementale encore non documentée (ex. : l'hydrodynamique locale, la vitesse des courants et la chlorophylle a [comme discuté dans le chapitre 5]). L'augmentation du recrutement est si importante qu‟un examen supplémentaire de la variabilité environnementale dans la rivière Boughton serait nécessaire, en combinant des expériences en laboratoire où les variables peuvent être isolées (ex. : la chlorophylle a) afin d'identifier les éléments conducteurs des schémas observés. Il est clair que les conditions dans la partie ouest de la rivière Boughton sont optimales pour le recrutement des larves, mais sans comprendre pourquoi, nous limitons notre capacité à identifier d'autres sites qui seraient également appropriés.

En raison de la nature unique et imprévisible des évènements d‟envahissements, il est difficile et irréaliste de s'attendre à pouvoir répéter des études sur les premiers schémas de dispersion et de recrutement des populations naissantes, ce qui a été l‟une des principales critiques sur l'écologie des envahissements. Sans la surveillance constante des sites à risque d‟envahissement et d'une équipe en veille permanente, prête à commencer la collecte des données si une ENI est détectée, nous sommes limités à des études sur seulement un site spécifique. Par conséquent, la généralisation de ces données devrait être effectuée avec prudence, car tous les systèmes diffèrent, ce qui peut influencer les schémas d'envahissements. Même à l‟IPE, où les baies sont similaires au niveau de l'environnement, les données recueillies et analysées à la rivière Boughton devraient être utilisées comme un guide pour le monitorage, plutôt que comme un protocole strict.

En adoptant une approche proactive à l'égard de l‟étude des ENI dès qu'elles sont détectées, nous pouvons commencer à surmonter les contraintes liées à la spécificité du site et commencer à identifier des généralités qui peuvent être appliquées à plusieurs systèmes. À présent, notre attention est naturellement dirigée vers le contrôle et la réduction des impacts des ENI les plus problématiques, mais cette réponse réactive fournit peu d‟informations sur la manière de mener des études de détection précoce et ne nous renseigne pas sur pourquoi certaines ENI ne parviennent pas à devenir envahissantes. Pour améliorer notre capacité à contrôler et, éventuellement éradiquer les espèces envahissantes, nous avons besoin d'une meilleure compréhension des causes de l'échec de l'envahissement, ce qui nécessite l'étude de toutes les

ENI, qu'elles soient problématiques ou non. Lorsque c'est possible, nous devrions examiner les espèces qui ont un taux de réussite d‟envahissement variable dans les différents systèmes afin de déterminer pourquoi certains sites sont plus sensibles que d'autres. De telles études pourraient améliorer notre compréhension des échecs d‟envahissements et de la résistance biotique envers les ENI. Par ailleurs, pour améliorer la détection et la gestion des ENI, nous devrions augmenter le nombre d'études concernant le succès ou l‟échec des envahissements sur plusieurs sites. Avec ces données, nous pourrions établir un bassin plus diversifié de données à partir duquel il serait possible de générer des solutions plus largement applicables pour la gestion des ENI.

Comme pour la surveillance des espèces rares, un élément clé d‟une surveillance efficace et, par conséquent, d‟une détection précoce des populations naissantes, est notre capacité à surveiller avec précision et fiabilité le milieu marin. Notre capacité à déterminer avec certitude la présence, l'absence et l'abondance des espèces marines est grandement améliorée en augmentant l'effort. Dans cette étude, une stratégie d'échantillonnage intense (~ 90 stations répétées huit fois) a été déployée pour documenter en détails un envahissement par Ciona et échantillonner efficacement la rivière Boughton afin d'examiner l‟importance de l'effort d‟échantillonnage par rapport à la probabilité de détection en utilisant du sous- échantillonnage (Edwards PhD thèse 2013). Il s'agissait d'une étude de cas extrême et ciblée sur les ENI et il serait inutile et irréaliste d'attendre des gestionnaires un effort d‟échantillonnage similaire dans les programmes de surveillance. Cependant, un équilibre entre l'effort, le coût et la probabilité de détection doit être davantage étudié. Mon étude a montré qu'une augmentation seule de l'effort n'augmente pas nécessairement la probabilité de détection et qu'en intégrant les aspects tels que la dispersion et l‟adéquation de l'environnement, la probabilité de détection peut être accrue sans augmentation de l'effort.

Une des préoccupations majeures des programmes de gestions actuels concerne le manque de standardisation des équipements utilisés entre les sites et/ou les organismes chargés de la surveillance des ENI. Actuellement, les stratégies de contrôle ont tendance à être conçues selon le matériel qui est le plus facilement disponible plutôt que spécifiquement conçu. La surveillance actuelle de Ciona au Canada est un bon exemple de cas où l'équipement de surveillance varie d'un site à l'autre (par exemple, des plaques de PVC de 10x10 à l'IPE et une assiette de céramique de 25cm de diamètre associée à des boîtes de pétries en Colombie-Britannique). Cela suggère un manque de communication entre les responsables des sites en plus du fait qu‟ils n‟utilisent pas les équipements disponibles les plus efficaces. Les résultats du chapitre 4 et un projet connexe sur l'effet de la taille de la plaque sur la fixation des larves (Delaney et al.. en revision [Appendice B]) montrent comment de petits changements dans la conception des équipements peut grandement influencer les taux de fixation et par la suite, la détection. Plus important encore, les résultats

de ces études indiquent que l'équipement de surveillance actuellement en place à l‟IPE, où Ciona est le plus problématique, est en fait la méthode actuellement utilisée au Canada la moins efficace (plaques de fixation en PVC de 10x10 cm). S‟assurer d‟un bon choix de l'équipement devrait être prioritaire dans tous les programmes de surveillance, mais la variété des équipements actuellement utilisés suggère que l'importance du design a souvent été négligée. Pour les espèces globalement problématiques comme Ciona, la communication et le partage de l'information sont indispensables pour optimiser la gestion. De plus, nous devrions nous assurer que ces données sont facilement accessibles aux gestionnaires dans un format pratique et compréhensible.

Un aspect essentiel du travail expérimental (chapitres 4 et 5) a été effectué sur le terrain plutôt qu‟en laboratoire. Les taux de recrutement élevés dans la rivière Boughton ont créé une rare opportunité de déployer des expériences à court terme (aussi courte que 6 h) pour étudier le comportement des larves avec peu de compétition d‟autres espèces. Toutefois, en raison des difficultés de travailler dans des conditions moins contrôlées, des observations directes du comportement animal n'ont pas été possibles et le taux de recrutement des larves de Ciona ou d'autres espèces n‟a pas pu être contrôlé. Cela était particulièrement évident avec l'expérience d'interaction caprelle-Ciona (chapitre 5), où la manipulation de l'abondance de chaque espèce de caprelles n'était pas possible et l‟observation directe du mécanisme exact qui a créé la diminution du recrutement larvaire de Ciona n‟a pas été effectuée. L'analyse a également été compliquée par les similitudes anatomiques existant entre les deux espèces de caprelles présentes, en particulier entre les juvéniles. Cela a conduit à une interprétation spéculative, des résultats qui, bien qu‟informée n'a pas pu confirmer définitivement une interaction négative ENI-ENI. Des analyses en laboratoires complémentaires n'ont pas été possibles dans ce projet en raison de contraintes de temps et de coûts, mais il s‟agirait logiquement de la prochaine étape afin de déterminer la relation exacte entre ces espèces et évaluer si ce genre de caprelles ont eu un impact sur l'envahissement de Ciona. En outre, une étude temporelle continue de la taille de la population de caprelles pourrait nous informer de tous les impacts que la présence de Ciona a peut avoir sur les espèces déjà présentes dans la rivière Boughton. Si les larves de Ciona fournissent une nouvelle source de nourriture pour les caprelles, cela pourrait-il se traduire par une augmentation de l'abondance des caprelles et, par conséquent, pourrait-il être un impact négatif accru sur le recrutement du Ciona?

Bien que cette thèse ait porté sur l'importance du dépistage précoce de l‟ENI, il est important de reconnaitre que la détection précoce seule n'est pas une solution pour contrer les espèces envahissantes. Elle doit être considérée comme un précurseur à la gestion par des réponses rapides. Cependant, sans protocole d'intervention rapide adéquat en place, le dépistage précoce est peu utile pour les gestionnaires

de l'environnement. Par exemple, Ciona a été une espèce envahissante très problématique et couteuse à l'IPE depuis 2004 et beaucoup de mesures ont porté sur les moyens de maitriser son impact sur l'industrie de la mytiliculture. Quand Ciona a été détecté dans la rivière Boughton en 2007, alors qu‟il n‟y avait que quelques individus, aucune tentative n‟a été faite afin de contrôler ou d‟éliminer la population avant qu'elle ne devienne problématique. Ceci n'est pas une critique du gouvernement de l'IPE, mais plutôt un exemple de la façon dont nous sommes mal préparés pour faire face aux ENI, si elles ne sont pas détectées assez tôt, ou pas détectées du tout. Cela est probablement le résultat d‟études se concentrant essentiellement à documenter la présence et l'impact des espèces envahissantes plutôt que de fournir des solutions pratiques pour les gestionnaires. Pour y remédier, la recherche sur la détection précoce doit être couplée avec des protocoles d'intervention rapide de confinement et, si possible, d'éradication.

Les recherches présentées dans cette thèse ne couvrent qu'une petite partie de ce qui est devenu un domaine d'étude vaste et diversifié. La rareté des études quantitatives et expérimentales sur les espèces envahissantes a laissé des lacunes dans nos connaissances sur le processus d'envahissement, en particulier pendant les premiers stades, ce qui a limité le nombre de directives à la disposition des gestionnaires chargés de la conception et de la mise en œuvre des programmes de surveillance. Les études présentées dans cette thèse ont démontré l'utilité des enquêtes quantitatives approfondies en ce qui concerne les populations d‟ENI naissantes en matière de gestion et devraient inspirer des études plus rigoureuses et détaillées des autres ENI. En augmentant le nombre d'études quantitatives sur les mécanismes fondamentaux qui conduisent à la survie de l‟espèce, à l'établissement de la population, et à la propagation, plutôt que de simplement documenter les impacts de l'envahissement, nous pouvons continuer à améliorer notre compréhension des envahissements ainsi que des aspects plus fondamentaux de l'écologie.

References

Abelson A, Miloh T, Loya Y (1993) Flow patterns induced by substrata and body morphologies of benthic organisms, and their roles in determining availability of food particles. Limnolology and Oceanography 38:1116-1124

Abelson A, Denny M (1997) Settlement of marine organisms in flow. Annual Review of Ecology and Systematics 28: 317-339

Albins MA, Hixon MA (2008) Invasive Indo-Pacific lionfish Pterois volitans reduce recruitment of Atlantic coral-reef fises. Marine Ecology Progress Series 367: 233-238

Allendorf, Lundquist (2003) Introduction: Population biology, evolution and control of invasive species. Conservation Biology 17: 24-30

Altman S, Whitlatch RB (2007) Effects of small-scale disturbance on invasion success in marine communities. Journal of Experimental Marine Biology and Ecology 342: 15-29

Ashton GV (2006) Distribution and dispersal of the non-native caprellid amphipod, Caprella mutica Schurin 1935. PhD Thesis, University of Aberdeen, Scotland

Ashton GV, Burrows MT, Willis KJ, Cook EJ (2010) Seasonal population dynamics of the non-native Caprella mutica (Crustacea, Amphipoda) on the west coast of Scotland. Marine Freshwater Research 61: 549-559

Bakus GJ, Targett NM, Schulte B (1986) Chemical ecology of marine organisms: an overview. Journal of Chemical Ecology 12: 951-987

Bax N, Carlton JT, Matthews-Amos A, Haedrich RL, Howarth FG, Purcell JE, Rieser A, GraySource A (2001) The control of biological invasions in the world‟s oceans. Conservation Biology 15: 1234-1246

Bax N, Wlliamson A, Aguero M, Gonzalez E, Geeves W (2003) Marine invasive alien species: a threat to global biodiversity. Marine Policy 27: 313-323

Baxter PWJ, Possingham HP (2010) Optimizing search strategies for invasive pests: learn before you leap. Journal of Applied Ecology 48; 86-95

Berrill NJ (1947) The development and growth of Ciona. Journal of the Marine Biological Association UK 26: 616-625

Bingham BL, Young CM (1993) Larval phototaxis in barnacles and snails associated with bathyal sea urchins. Deep-Sea Research 40: 1-12

Bone Q (1992) On the locomotion of ascidian tadpole larvae. Journal of the Marine Biological Association of the United Kingdom 72: 161-186

Boots B. (2002) Local measures of spatial association. Ecoscience 9: 168-176

Boos K (2009) Mechanisms of a successful immigration from north-east Asia: population dynamics, life history traits and interspecific interactions in the caprellid amphipod Caprella mutica Schurin, 1935 (Crustacea: Amphipoda) in European coastal waters. PhD Thesis, Freie University, Berlin

Brown JH, Sax DF (2004) An essay on some topics concerning invasive species. Austral Ecology 29: 530- 536

Bruijs MCM, Kelleher B, van der Velde G, de Vaate AB (2001) Oxygen consumption, temperature, and salinity tolerance of the invasive amphipod Dikerogammarus villosus: indicators of further dispersal via ballast water transport. Archive fur Hydrobiologie, 152: 633-646

Buhle ER, Margolis M, Ruesink JL (2005) Bang for buck: cost-effective control of invasive species with different life histories. Ecological Economics 52: 355-366

Bullard SG, Whitlatch RB (2004) A guide to the larval and juvenile stages of Long Island Sound ascidians and bryozoans. Connecticut Seagrant Publication, CTSG-04-07

Buric M, Koci L, Petrusek A, Kouba A, Kozak P (2009) Invaders eating invaders: potential trophic interactions between the amphipod Dikerogammarus villosus and juvenile crayfish Orconectes limosus. Knowledge and Management of Aquatic Systems 5: 394-395

Buschbaum C, Gutow L (2005) Mass occurrence of an introduced crustacean (Caprella cf. mutica) in the south-eastern North Sea. Helgoland Marine Research 59: 252-253

Butman CA (1987) Larval settlement of soft-sediment invertebrates: the spatial scales of pattern explained by active habitat selection and the emerging role of hydrodynamical processes. Oceanography and Marine Biology Annual Review 25:113-165

Caine EA (1977) Feeding mechanisms and possible resource partitioning of the Caprellidae (Crustacea: Amphipoda) from Puget Sound, USA. Marine Biology 42: 331-336

Campbell ML, Gould B, Hewitt CL (2007) Survey evaluations to assess marine bioinvasions. Marine Pollution Bulletin 55: 360-378

Carlton JT (1985) Transoceanic and interoceanic dispersal of coastal marine organisms: the biology of ballast water. Oceanography and Marine Biology Annual Review 23: 313-371

Carlton JT (1987) Patterns of transoceanic marine biological invasions in the Pacific Ocean. Bulletin of Marine Science 41: 452-465

Carlton JT (1989) Man‟s role in changing the face of the ocean: biological invasions and implications for conservation of near-shore environments. Conservation Biology 3: 265-273

Carlton JT, Geller JB (1993) Ecological Roulette: The global transportation of nonindigenous marine organisms. Science 261: 78-82

Carr MH, Neigel JE, Estes JA, et al. (2003) Comparing marine and terrestrial ecosystems: implications for the design of coastal marine reserves. Ecological Applications 13: 90-107

Carver CE, Chisholm A, Mallet AL (2003) Strategies to mitigate the impact of Ciona intestinalis (L.) biofouling on shellfish production. Journal of Shellfish Research 22: 621-631

Cirino P, Toscano A, Caramiello D, Macina A, Miragliaet V, Monte A (2002) Laboratory culture of the ascidian Ciona intestinalis (L.): a model system for molecular developmental biology research. Marine Models Electronic Record, http://hermes.mbl.edu/BiologicalBulletin/MMER/cirino/CirCon.html

Clarke Murray C, Pakhomov EA, Therriault TW (2011) Recreational boating: a large unregulated vector transporting marine invasive species. Diversity and Distributions 17: 1161-1172

Colautti RL, Grigorovich IA, MacIsaac HJ (2003) Propagule pressure: a null model for biological invasions. Biological Invasions 8: 1023-1037

Colautti RI, MacIsaac HJ (2004) A neutral terminology to define „invasive‟ species. Diversity and Distributions 10: 135-141

Collin SB, Oakley JA, Sewell J & Bishop JDD (2010) Widespread occurrence of the non-indigenous ascidian Corella eumyota Traustedt, 1882 on the shores of Plymouth Sound and Estuaries Special Area of Conservation, UK. Aquatic Invasions 5: 175-179

Connell JH, Slatyer RO (1977) Mechanisms of succession in natural communities and their role in community stability and organization. The American Naturalist 111, 1119-1144

Connell JH, Keough MJ (1985) Disturbance and patch dynamics of subtidal marine animals on hard substrata. The ecology of natural disturbance and patch dynamics, Academic Press, Orlando Florida, pp 125-151

Cook EJ, Willis KJ, Lozano-Fernandez M (2007) Survivorship, growth and reproduction of the non-native Caprella mutica Schurin, 1935 (Crustacea: Amphipoda). Hydrobiologia 590: 55-64

Cook EJ, Shucksmith R, Orr H, Ashton GV, Berge J (2010) Fatty acid composition as a dietary indicator of the invasive caprellid, Caprella mutica (Crustacea: Amphipoda). Marine Biology 157: 19-27

Courchamp F, Clutton-Brock T, Grenfell B (1999) Inverse density dependence and the Allee effect. Trends Ecology and Evolution 14: 405-410

Cowen RK, Sponaugle S (2009) Larval dispersal and marine population connectivity. Annual Review of Marine Science 1: 443-466

Crisp DJ (1961) Territorial behaviour in barnacle settlement. Experimental Biology 38: 429-446

Crisp DJ, Ghobashy AFAA (1971) Responses of the larvae of Diplosoma listerianum to light and gravity. Fourth European Marine Biology Symposium, Cambridge University Press, NY. 443-465

Crooks JA (2002) Characterizing ecosystem-level consequences of biological invasions: the role of ecosystem engineers. Oikos 97: 153-166

Cuddington K, Hastings A (2004) Invasive engineers. Ecological Modelling 178: 335-347

Darbyson E, Locke A, Hanson JM, Willison JHM (2009) Marine boating habits and the potential for spread of invasive species in the Gulf of St. Lawrence. Aquatic Invasions 4: 87-94

Davidson JD, Landry T, Johnson G, et al. (2012) Evaluating the productivity and cost benefit of high pressure water treatment regimes on Ciona intestinalis infested mussel socks on Prince Edward Island. Journal of Shellfish Research 31: 274

Davis AR (1987) Variation in the recruitment of the subtidal colonial ascidian Podoclavella cylindrica (Quoy and Gaimard) – the role of substratum choice and early survival. Journal of Experimental Marine Biology and Ecology 106: 57-71

Davis AR, Butler AJ, van Altena I (1991) Settlement behaviour of ascidian larvae: preliminary evidence for inhibition by sponge allelochemicals. Marine Ecology Progress Series 72: 117-123

Delaney DG, Collin SB, Descôteaux R, Johnson LE (in prep) Optimizing early detection and monitoring strategies of non-indigenous species: A recruitment study of the invasive tunicate Ciona intestinalis. deRivera CE, Ruiz GM, Hines AH, Jivoff P (2005) Biotic resistance to invasion: Native predator limits abundance and distribution of an introduced crab. Ecology 86: 3364-3376

Dick JTA, Platvoet D (2000) Invading predatory crustacean Dikerogammarus villosus eliminates both native and exotic species. Proc R Soc B 267: 977-983

Durante K.M (1991) Larval behaviour, settlement preference, and induction of metamorphosis in the temperate solitary ascidian Mogula citrina Alder and Hancock. Journal of Experimental Marine Biology and Ecology 145: 175-187

Dybern BI (1965) The life cycle of Ciona intestinalis (L.) f. typical in relation to the environmental temperature. Oikos 16: 109-131

Dybern BI (1967) The distribution and salinity tolerance of Ciona intestinalis (L.) f. typica with special reference to the waters around southern Scandinavia. Ophelia 4: 207-226

Edwards PK, Collin SB, Johnson LE, Leung B (in review) Capturing dispersal patterns and tolerance of early invaders with enhanced species distribution models: applications for design of detection programs. Diversity and Distributions

Edwards PK, Leung B (2009) Re-evaluating eradication of nuisance species: invasion of the tunicate, Ciona intestinalis. Frontiers in Ecology and the Environment 7: 326-332

Elkin C, Marshall DJ (2007) Desperate larvae: influence of deferred costs and habitat requirements on habitat selection. Marine Ecology Progress Series 335: 143-153

Elton CS (1958) The ecology of invasions by animals and plants. The University of Chicago Press, Chicago 60637, Illinois, USA

Epelbaum A, Pearce CM, Barker DJ, Paulson A, Therriault TW (2009) Susceptibility of non-indigenous ascidian species in British Columbia (Canada) to invertebrate behaviour. Marine Biology 156: 1311-1320

Fagan WF, Lewis MA, Neubert MG and van den Driessche P (2002) Invasion theory and biological control. Ecology Letters 5: 148-157

Feng D, Caihuan K, Lu C, Li S (2010) The influence of temperature and light pre-settlement metamorphosis: a study of the effects of environmental factors on pre-settlement metamorphosis of the solitary ascidian Styela canopus. Marine and Freshwater Behaviour and Physiology 43: 11-24

Finnoff D, Shogren JF, Leung B, Lodge D (2007) Take a risk: preferring prevention over control of biological invaders. Ecological Economics 62: 216-222

Fitt BDL, Gregory PH, Todd AD (1987) Spore dispersal and plant disease gradients; a comparison between two empirical models. Journal of Phytopathology 118: 227-242

Flores AR, Faulkes Z (2008) Texture preferences of ascidian tadpole larvae during settlement. Marine and Freshwater Behaviour and Physiology 41: 155-159

Forsyth DM, Duncan RP (2001) Propagule size and the relative stress of exotic ungulate and bird introductions to New Zealand. The American Naturalist 157: 583-595

Gertzen E, Leung B (2011) Predicting the spread of invasive species in an uncertain world: accommodating multiple vectors and gaps in temporal and spatial data for Bythotrephes longimanus. Biological Invasions 13: 2433-2444

Glasby TM, Connell SD (2001) Orientation and position of substrata have large effects on epibiotic assemblages. Marine Ecology Progress Series 214: 127-135

Glasby TM, Connell SD, Holloway MG, Hewitt CL (2007) Nonindigenous biota on artificial structures: could habitat creation facilitate biological invasions? Marine Biology 151: 887-895

Gollasch S (2002) The importance of ship hull fouling as a vector of species introductions into the North Sea. Biofouling 18: 105-121

Gollasch S (2007) International collaboration on marine bioinvasions – the ICES response. Marine Pollution Bulletin 55: 353-359

Green PT, O‟Dowd DJ, Abbott KL, Jeffery M, Retalick K, MacNally R (2011) Invasional meltdown: invader-invader mutualism facilitates a secondary invasion. Ecology 92: 1758-1768

Griffen BD, Guyt T, Buck JC (2008) Inhibition between invasives: a newly introduced predator moderates the impacts of a previously-established invasive predator. Journal of Animal Ecology 77: 32-40

Grosberg RK (1981) Competitive ability influences habitat choice in marine invertebrates. Nature 290: 700-702

Guerra-Garcia JM (2002) Re-descriptions of Caprella linearis (Linnaeus, 1767) and C. septentrionalis Kröyer, 1838 (Crustacea : Amphipoda : Caprellidae) from Scotland, with an ontogenetic comparison between the species and a study of the clinging behaviour. Sarsia 87: 216-235

Guerra-Garcia JM, Tierno de Figueroa JM (2009) What do caprellids (Crustacea : Amphipoda) feed on? Marine Biology 156: 1881-1890

Havenhand JN, Svane I (1991) Roles of hydrodynamics and larval behaviour in determining spatial aggregation in the tunicate Ciona intestinalis. Marine Ecology Progress Series 68:271-276

Hayes KR, Cannon R, Neil K, Inglis G (2005) Sensitivity and cost considerations for the detection and eradication of marine pests in ports. Marine Pollution Bulletin 50: 823-834

Heimpel GE, Frelich LE, Landis DA, Hopper KR, Hoelmer KA, Sezen Z, Asplen MK, Wu K (2010) European buckthorn and Asian soybean aphid as components of an extensive invasional meltdown in North America. Biological Invasions 12: 2913-2931

Helms KR, Hayden CP, Vinson SB (2011) Plant-based food resources, trophic interactions among alien species, and the abundance of an invasive ant. Biological Invasions 13: 67-79

Herborg ML, Jerde CL, Lodge DM, Ruiz GM, MacIsacc HJ (2007) Predicting invasion risk using measures of introduction effort and environmental niche models. Ecological Applications 17: 663-674

Hewitt CL, Campbell ML (2007) Mechanisms for the prevention of marine bioinvasions for better biosecurity. Marine Pollution Bulletin 55: 395-401

Hilbe JM (2007) Negative Binomial Regression. Cambridge University Press

Holmström C, Rittschof D, Kjelleberg S (1992) Inhibition of settlement by larvae of Balanus amphitrite and Ciona intestinalis by a surface-colonizing marine bacterium. Applied Environmental Microbiology 58: 2111-2115

Holomuzki JR (1986) Predator avoidance and diel patterns of microhabitat use by larval tiger salamanders. Oecologia 71: 142-148

Hooper DU, Chapin III FS, Ewel JJ, Hector A, Inchausti P, Lavorel S, Lawton JH, Lodge DM, Loreau M, Naeem S, Schmid B, Setala H, Symstad AJ, Vandermeer J, Wardle DA (2005) Effects of biodiversity on ecosystem functioning: a consensus of current knowledge. Ecology Monographs 75: 3-35

Howes S, Herbinger CM, Darnell P, Vercaemer B (2007) Spatial and temporal patterns of recruitment of the tunicate Ciona intestinalis on a mussel farm in Nova Scotia, Canada. Journal of Experimental Marine Biology and Ecology 342: 85-92

Hulme PE (2006) Beyond control: wider implications for the management of biological invasions. Journal of Applied Ecology 43: 835-847

Hulme PE (2009) Trade, transport, and trouble: managing invasive species pathways in an era of globalization. Journal of Applied Ecology 46: 10-18

Hurlbut CJ (1993) The adaptive value of larval behaviour of a colonial ascidian. Marine Biology 115: 253-262

Inglis GJ, Hurren H, Oldman J, Haskew R (2006) Using habitat suitability index and particle dispersion models for early detection of marine invaders. Ecological Applications 16: 1377-1390

Johnson LE, Strathmann RR (1989) Settling barnacle larvae avoid substrata previously occupied by a mobile predator. Journal of Experimental Marine Biology and Ecology 128: 87-103

Kanary L, Locke A, Watmough J, Chassé J, Bourque D, Nadeau A (2011) Predicting larval dispersal of the vase tunicate Ciona intestinalis in a Prince Edward Island estuary using a matrix population model. Aquatic Invasions 6: 491-506

Kang PA, Bae PA, Pyen CK (1978) Studies on the suspended culture of oyster, Crassostrea gigas in Korean coastal waters. 5. On the fouling organisms associated with culturing oysters at the oyster culture farms in Chungmu. Bulletin of Fisheries Research Agency 20: 121-127

Keitt TH, Lewis MA, Holt RD (2001) Allee effects, invasion pinning, and species borders. American Naturalist 157: 203-216

Kelly NE, Wantola K, Weisz E, Yan ND (2013) Recreational boats as a vector of secondary spread for aquatic invasive species and native crustacean zooplankton. Biological Invasions 15: 509-519

Keough MJ, Downes BJ (1982) Recruitment of marine invertebrates: the role of active larval choices and early mortality. Oecologia 54: 348-352

Khalaman VV, Belyaeva DV, Flyachinskaya LP (2008) Effect of excretory-secretory products of some fouling organisms on settling and metamorphosis of the larvae of Styela rustica (Ascidiae). Russian Journal of Marine Biology 34: 170-173

Kimbro DL, Cheng BS, Grosholz ED (2013) Biotic resistance in marine environments. Ecology Letters 16: 821-833

Kingsford MJ, Leis JM, Shanks A, Lindeman KC, Morgan SG, Pineda J (2002) Sensory environments, larval abilities and local self-recruitment. Bulletin of Marine Science 70: 309-340

Kinlan BP, Gaines SD (2003) Propagule dispersal in marine and terrestrial environments: a community perspective. Ecology 84: 2007-2020

Kinlan BP, Hastings A (2005) Rates of population spread and geographic range expansion – what exotic species tell us. Species Invasions: Insights into Ecology, Evolution, and Biogeography, Sinauer Associates Inc. pp. 381-419

Kolar CS, Lodge DM (2001) Progress in invasion biology: predicting invaders. Trends in Ecology and Evolution 16: 199-204

Kolar CS, Lodge DM (2002) Ecological predictions and risk assessment for alien fishes in North America. Science 298: 1233-1236

Kot M, Lewis MA, van den Driessche P (1996) Dispersal data and the spread of invading organisms. Ecology 77: 2027-2042

Kraft CE, Johnson LE (2000) Regional difference in rates and patterns of North American inland lake invasions by zebra mussels (Dreissena polymorpha). Canadian Journal of Fisheries and Aquatic Sciences 57: 993-1001

Kunze HB, Morgan SB, Lwiza KM (2013) Field test of the behavioral regulation of larval transport. Marine Ecology Progress Series 487: 71-87

Kusabe T, Tsuda M (2007) Photoreceptive systems in ascidians. Photochemistry and Photobiology 83: 248-252

La Pierre KJ, Harpole WS, Suding KN (2010) Strong feeding preference of an exotic generalist herbivore for an exotic forb: a case of invasional antagonism. Biological Invasions 12: 3025-3031

Lafferty KD, Kuris AM (1996) Biological control of marine pests. Ecology 77: 1989-2000

Lambert CC, Brandt CL (1967) The effect of light on the spawning of Ciona intestinalis. Biological Bulletin 132: 222-228

Lambert CC, Lambert G (1998) Non-indigenous ascidians in southern California harbors and marinas. Marine Biology 130: 675-688

Lemire M, Bourget E (1996) Substratum heterogeneity and complexity influence micro-habitat selection of Balanus sp and Tubularia crocea larvae. Marine Ecology Progress Series 135: 77-87

Leung B, Roura-Pascual N, Bacher S, Heikkila J, Brotons L, Burgman MA, Dehnen-Schmutz K, Essl F, Hulme PE, Richardson DM, Dol D, Vila M (2012) TEASIng apart alien species risk assessments: a framework for best practices. Ecology Letters 15: 1475-1493

Locke A, Hanson JM, Ellis KM, Thompson J, Rochette R (2007) Invasion of the southern Gulf of St. Lawrence by the clubbed tunicate (Styela clava Herdman): potential mechanisms for invasions of Prince Edward Island estuaries. Journal of Experimental Marine Biology and Ecology 342: 69-77

Lockwood JL, Cassey P, Blackburn T (2005) The role of propagule pressure in explaining species invasions. Trends Ecology and Evolution 20: 223-228

Lockwood JL, Hoopes MF, Marchetti MP (2008) Invasion Ecology. Blackwell Publishing Ltd.

Lodge DM (1993) Biological invasions: lessons for ecology. Trends Ecology and Evolution 8: 133-137

Lodge DM, Williams S, MacIsaac HT, Hayes KR, Leung B, Richard S, Mack RN, Moyle PB, Smith M, Andow DA, Carlton JT, McMichael A (2006) Biological invasions: Recommendations for U.S. policy and management. Ecological Applications 16: 2035-2054

Lohrer AM, Whitlatch RB (2002) Interactions among aliens: apparent replacement of one exotic species by another. Ecology 83: 719-732

Lonsdale WM (1999) Global patterns of plant invasions and the concept of invisibility. Ecology 80: 1522- 1536

MacIsaac HJ, Robbins TC, Lewis MA (2002) “Biological invasions of aquatic habitats in Europe and the Great Lakes”. Modelling ships‟ ballast water as invasion threats to the Great Lakes. Canadian Journal of Fisheries and Aquatic Sciences 59: 1245-1256

Mack RN, Simberloff D, Lonsdale WM, Evans H, Clout M, Bazzaz FA (2000) Biotic invasions: causes, epidemiology, global consequences, and control. Ecological Applications 10: 689-710

Marsico TD, Burt JW, Espeland EK, et al. (2010) Underutilized resources for studying the evolution of invasive species during their introduction, establishment, and lag phases. Evolutionary Applications 3: 203-219

McDonald J (2004) The invasive pest species Ciona intestinalis (Linnaeus, 1767) reported in a harbour in southern Western Australia. Marine Pollution Bulletin 49: 9-10

McFadyen REC (1998) Biological control of weeds. Annual Review of Entomology 43: 369-393

McHenry MJ, Strother JA (2003) The kinematics of phototaxis in larvae of the ascidian Aplidium constellatum. Marine Biology 142: 173-184

Mileikovsky SA (1973) Speed of active movement of pelagic larvae of marine bottom invertebrates and their ability to regulate their vertical position. Marine Biology 23: 11-17

Millar RH (1958) Some ascidians from Brazil. Journal of Natural History Series 13: 1, 497-514

Miller SE, Hadfield MG (1986) Ontogeny of phototaxis and metamorphic competence in larvae of the nudibranch Phestilla sibogae Bergh (Gastropoda: Opisthobranchia). Journal of Experimental Marine Biology and Ecology 97: 95-112

Minchin D (2001) Introduction of exotic species. Encyclopedia of Ocean Sciences. Elsevier Science Ltd. pp. 877-889

Minchin D, Gollasch S, Cohen AN, Hewitt CL, Olenin S (2009) Characterizing vectors of marine invasion. Biological Invasions in Marine Ecosystems and Ecological Studies 204: 109-116

Mogami Y, Oobayashi C, Yamaguchi T, Ogiso Y, Baba SA (1988) Negative geotaxis in sea urchin larvae – a possible role of mechanoreception in the late stages of development. Journal of Experimental Biology 137: 141-156

Montgomery WI, Lundy MG, Reid N (2012) „Invasional meltdown‟: evidence for unexpected consequences and cumulative impacts of multispecies invasions. Biological Invasions 14: 1111-1125

100

Moody ME, Mack RN (1988) Controlling the spread of plant invasions: the importance of nascent foci. Journal of Applied Ecology 25: 1009-1021

Morales JM, Carlo TA (2006) The effects of plant distribution and frugivore density on the scale and shape of dispersal kernels. Ecology 87: 1489-1496

Morgan SB, Fisher JL, Miller SH, McAfee ST, Largier JL (2009) Nearshore retention in a region of strong upwelling and recruitment limitation. Ecology 90: 3489-3502

Mullineaux LS, Butman CA (1991) Initial contact, exploration and attachment of barnacle (Balanus amphitrite) cyprids settling in flow. Marine Biology 110: 93-103

Murray CC, Therriault TW & Martone PT (2012) Adapted for invasion? Comparing attachment, drag and dislodgement of native and nonindigenous hull fouling species. Biological Invasions 14: 1651-1663

Myers JH, Simberloff D, Kuris AM, Carey JR (2000) Eradication revisited: dealing with exotic species. Trends in Ecology and Evolution 15: 316-320

Nakagawa M, Miyamoto T, Ohkuma, M, Tsuda M (1999) Action spectrum for the photophobic response of Ciona intestinalis (Ascidieacea, Urochordata) larvae implicates retinal protein. Photochemistry and Photobiology 70: 359-362

Nathan R (2006) Long-distance dispersal of plants. Science 313: 786-788

O‟Dowd DJ, Green PT, Lake PS (2003) Invasional „meltdown‟ on an oceanic island. Ecology Letters 6: 812-817

Occhipinti-Ambrogi A, Savini D (2003) Biological invasions as a component of global change in stressed marine ecosystems. Marine Pollution Bulletin 46: 542-551

Occipinti-Ambrogi A (2007) Global change and marine communities: Alien species and climate change. Marine Pollution Bulletin 55: 342-352

Olenin S, Elliot M, Bysveen I et al. (2011) Recommendations on methods for the detection and control of biological pollution in marine coastal waters. Marine Pollution Bulletin 62: 2598-2604

Olson RR (1983) Ascidian-Prochloron Symbiosis: The role of larval photoadaptations in midday larval release and settlement. Biological Bulletin 165: 221-240

Osman RW, Whitlatch RB (2004) The control of the development of a marine benthic community by predation on recruits. Journal of Experimental Marine Biology and Ecology 311: 117-145

Paetzold SC, Davidson J (2011) Aquaculture fouling: efficacy of potassium monopersulphonate triple salt based disinfectant (Virkon Aquatic) against Ciona intestinalis. Biofouling 27: 655-665

Park S, Epifanio CE, Grey EK (2004) Behavior of larval Hemigrapsus sanguineus (de Haan) in response to gravity and pressure. Journal of Experimental Marine Biology and Ecology 307: 197-206

Pawlik J (1992) Chemical ecology of the settlement of benthic marine invertebrates. Oceanography and Marine Biology 30: 273-335

101

Pech D, Ardisson PL, Bourget E (2002) Settlement of a tropical marine epibenthic assemblage on artificial panels: influence of substratum heterogeneity and complexity scales. Estuarine Coastal and Shelf Science 55: 743-750

Pechenik JA (1999) On the advantages and disadvantages of larval stages in benthic marine invertebrate life cycles. Marine Ecology Progress Series 177: 269-297

Petersen KS, Rasmussen KL, Heinemeier J, Rud N (1992) Clams before Columbus? Nature 359: 679

Petersen JK, Svane I (1995) Larval dispersal in the ascidian Ciona intestinalis (L.). Evidence for a closed population. Journal of Experimental Marine Biology and Ecology 186: 89-102

Peterson AT, Vieglais DA (2001) Predicting species invasions using ecological niche modelling: New approaches from bioinformatics attack a pressing problem. Bioscience 51: 363-371

Pimentel D, Zuniga R, Morrison D (2005) Update on the environmental and economic costs associated with alien-invasive species in the United States. Ecological Economics 52: 273-288

Platvoet D, Dick JTA, MacNeil C, van Riel MC, van der Velde G (2009) Invader-invader interactions in relation to environmental heterogeneity leads to zonation of two invasive amphipods, Dikerogammarus villosus (Sowinsky) and Gammarus tigrinus Sexton: amphipod pilot species project (AMPIS) report 6. Biological Invasions 11: 2085-2093

Ramsay A, Davidson J, Bourque D, Stryhn H (2009) Recruitment patterns and population development of the invasive ascidian Ciona intestinalis in Prince Edward Island, Canada. Aquatic Invasions 4: 169-176

Reed DC, Raimondi PT, Carr MH, Goldwasser L (2000) The role of dispersal and disturbance in determining spatial heterogeneity in sedentary organisms. Ecology 81: 2011-2026

Relva MA, Nunez MA, Simberloff D (2010) Introduced deer reduce native plant cover and facilitate invasion of non-native tree species: evidence for invasional meltdown. Biological Invasions 12: 303-311

Ricciardi A (2001) Facilitative interactions among aquatic invaders: is an “invasional meltdown” occurring in the Great Lakes? Canadian Journal of Fisheries and Aquatic Sciences 58: 2513-2525

Ricciardi A, MacIsaac HJ (2000) Recent mass invasion of the North American Great Lakes by Ponto- Caspian species. Trends in Ecology and Evolution 15: 62-65

Ricciardi A, Rasmussen JB (1998) Predicting the identity and impact of future biological invaders: a priority for aquatic resource management. Canadian Journal of Fisheries and Aquatic Sciences 55: 1759- 1765

Ricciardi A, Whoriskey FG (2004) Exotic species replacement: shifting dominance of dreissenid mussels in the Soulanges canal, upper St. Lawrence River, Canada. Journal of North American Benthological Society 23: 507-514

Richardson DM, Pysek P, Carlton JT (2011) Chapter 30. A compendium of essential concepts and terminology in invasion ecology. Fifty years of invasion ecology: the legacy of Charles Elton, 1st Edition. Blackwell Publishing Ltd.

Rittschof D, Forward RB, Cannon G, Welch JM, McClary M, Holm ER, Clare AS, Conova S, McKelvey LM, Bryan P, van Dover CL (1998) Cues and context: larval responses to physical and chemical cues. Biofouling 12: 31-44

102

Rius M, Branch GM, Griffiths CL, Turon X (2010) Larval settlement behaviour in six gregarious ascidians in relation to adult distribution. Marine Ecology Progress Series 418: 151-163

Roberts B, Davidson B, MacMaster G, Lockhart V, Ma E, Smith Wallace S, Swalla BJ (2007) A complement response may activate metamorphosis in the ascidian Boltenia villosa. Development of Genes and Evolution 217: 449-458

Robinson TB, Griffiths CL, McQuaid C, Rius M (2005) Marine alien species of South Africa – status and impacts. African Journal of Marine Science 27: 297-306

Roques L, Auger-Rozenberg M, Roques A (2008) Modelling the impact of an invasive insect via reaction- diffusion. Mathematical Bioscience 216: 47-55

Rodriguez SR, Ojeda FP, Inestrosa NC (1993) Settlement of benthic marine invertebrates. Marine Ecology Progress Series 97: 193-207

Rouget M, Richardson DM (2003) Inferring process from pattern in plant invasion: a semimechanistic model incorporating propagule pressure and environmental factors. American Naturalist 162: 713-724

Roughgarden J, Gaines S, Possingham H (1988) Recruitment dynamics in complex life cycles. Science 241: 1460-1466

Ruiz GM, Fofonoff PW, Carlton JT, Wonham MJ, Hines AH (2000) Invasion of coastal marine communities in North America: apparent patterns, processes, and biases. Annual Review Ecological Systems 31: 481-531

Sakai AK, Allendorf FW, Holt JS, Lodge DM, Molofsky J, With KA, Baughman S, Cabin RJ, Cohen JE, Ellstrand NC, McCauley DE, O‟Neil P, Parker IM, Thompson JN, Weller SG (2001) The population biology of invasive species. Annual Review Ecological Systems 32: 305-332

Sanford E, Holzman SB, Haney RA, Rand DM, Bertness MD (2006) Larval tolerance, gene flow, and the northern geographic range limit of fiddler crabs. Ecology 87: 2882-2894

Sano M, Omori M, Taniguchi K (2003) Predator-prey systems of drifting seaweed communities off the Tohoku coast, northern Japan, as determined by feeding habitat analysis of phytal animals. Fisheries Science 6: 260-268

Sargent PS, Wells T, Matheson K, McKenzie CH, Deibel D (2013) First record of vase tunicate, Ciona intestinalis (Linnaeus, 1767) in coastal Newfoundland waters. Bioinvasions Records 2: 89-98

Sax DF, Kinlan BP, Smith KF (2005) A conceptual framework for comparing species assemblages in native and exotic habitats. Oikos 108: 457-464

Sebens KP (1981) Recruitment in a sea anemone population: juvenile substrate becomes adult prey. Science 213: 785-787

Secord D (2003) Biological control of marine invasive species: cautionary tales and land-based lessons. Biological Invasions 5: 117-131

Sephton D, Vercaemer B, Nicolas JM, Keays J (2011) Monitoring for invasive tunicates in Nova Scotia, Canada (2006-2009). Aquatic Invasions 6: 391-403

103

Sharov AA, Liebhold AM, Roberts AE (1998) Optimizing the use of barrier zones to slow the spread of gypsy moths (Lepidoptera: Lymantriidae) in North America. Journal of Economical Entomology 91: 165- 174

Shucksmith R, Cook EJ, Hughes DJ, Burrows MT (2009) Competition between the non-native amphipod Caprella mutica and two native species of caprellids Pseudoprotella phasma and Caprella linearis. Journal of the Marine Biological Association UK 89: 1125-1132

Siegel DA, Kinlan BP, Gaylord B, Gaines SD (2003) Lagrangian descriptions of marine larval dispersion. Marine Ecology Progress Series 260: 83-96

Sih A (1982) Foraging strategies and the avoidance of predation by an aquatic insect, Notonecta hoffmanni Ecology 63: 786-796

Simberloff D (2000) Nonindigenous species : a global threat to biodiversity and stability. Nature and human society: the quest for a sustainable world. National Academy Press, Washington, DC

Simberloff D (2006) Invasional meltdown 6 years later: important phenomenon, unfortunate metaphor, or both? Ecology Letters 9: 912-919

Simberloff D, Von Holle B (1999) Positive interactions of nonindigenous species: invasional meltdown? Biological Invasions 1: 21-32

Simberloff D, Parker IM, Windle PN (2005) Introduced species policy, management, and future research needs. Frontiers in Ecology and Environment 3: 12-20

Sponaugle S, Cowen RK, Shanks A, et al. (2002) Predicting self-recruitment in marine populations: biophysical correlates and mechanisms. Bulletin of Marine Science 70: 341-375

Stachowicz JJ, Whitlatch RB, Osman RW (1999) Species diversity and invasion resistance in a marine ecosystem. Science 286: 1577-1579

Stachowicz JJ, Fried H, Osman RW, Whitlatch RB (2002) Biodiversity, invasion resistance, and marine ecosystem function: reconciling pattern and process. Ecology 83: 2575-2590

Stoner DS (1992) Vertical distribution of a colonial ascidian on a coral reef: the roles of larval dispersal and life-history variation. American Naturalist 139: 802-824

Stoner DS (1994) Larvae of a colonial ascidian use a non-contact mode of substratum selection on a coral reef. Marine Biology 121: 319-326

Svane I, Dolmer P (1995) Perception of light at settlement: a comparative stdy of two invertebrate larvae, a scyphozoan planula and a simple ascidian tadpole. Journal of Experimental Marine Biology and Ecology 187: 51-61

Svane I, Havenhand JN (1993) Spawning and dispersal in Ciona intestinalis (L.). Marine Ecology 14: 53- 66

Svane I, Young CM (1989) The ecology and behaviour of ascidian larvae. Oceanography and Marine Biology 27: 45-90

Swearer SE, Shima JS, Hellberg ME, et al. (2002) Evidence of self-recruitment in demersal marine populations. Bulletin of Marine Science 70: 251-271

104

Szewzyk U, Holmström C, Wrangstadh M, Samuelsson MO, Maki JS, Kjelleberg S (1991) Relevance of exopolysaccharide of marine Pseudomonas sp. strain S9 for the attachment of Ciona intestinalis larvae. Marine Ecology Progress Series 75: 259-265

Tan CKF, Nowak BF, Hodson SL (2002) Biofouling as a reservoir of Neoparamoeba pemaquidensis (Page, 1970), the causative agent of amoebic gill disease in Atlantic salmon. Aquaculture 210: 49-58

Taylor CM, Hastings A (2005) Allee effects in biological invasions. Ecology Letters 8: 895-908

Therriault TW, Herborg L (2008) Predicting the potential distribution of the vase tunicate Ciona intestinalis in Canadian waters: informing a risk assessment. ICES Journal of Marine Science 65: 788-794

Thompson B, MacNair N (2004) An overview of the clubbed tunicate (Styela clava) in Prince Edward Island. PEI Agriculture and Fisheries Aquaculture Technical Report 234: viii + 29

Thorson G (1964) Light as an ecological factor in the dispersal and settlement of larvae of marine bottom invertebrates. Ophelia 1: 167-208

Thuiller W, Richardson DM, Pyseks P, Midgley GF, Hughes GO, Rouget M (2005) Niche-based modelling as a tool for predicting the risk of alien plant invasions at a global scale. Global Change Biology 11: 2234-2250

Trebitz AS, Kelly JR, Hoffman JC, Peterson GS, West CW (2009) Exploiting habitat and gear patterns for efficient detection of rare and non-native benthos and fish in Great Lakes coastal ecosystems. Aquatic Invasions 4: 651-667

Uribe E, Etchepare I (2002) Effects of biofouling by Ciona intestinalis on suspended culture of Argopecten purpuratus in Bahia Inglesa, Chile. Bulletin of Aquaculture Association Canada 102: 93-95

Uygun N, Ulusoy MR, Sekeroglu E, Ohnesorge B, Gozel U (1994) Interactions between two introduced species of whiteflies in the Mediterranean area of Turkey: Dialeurodes citri (Ashmead) and Parabemisia myricae (Kuwana) (Hom., Aleyrodidae). Journal of Applied Entomology 118: 365-369

Vermeij MJA, Fogarty ND, Miller MW (2006) Pelagic conditions affect larval behavior, survival, and settlement patterns in the Caribbean coral Montastraea faveolata. Marine Ecology Progress Series 310: 119-128

Vander Zanden MJ, Olden, JD (2008) A management framework for preventing the secondary spread of aquatic invasive species. Canadian Journal of Fisheries and Aquatic Sciences 65: 1512-1522

Walker B, Steffen W (1997) An overview of the implications of global change for natural and managed terrestrial ecosystems. Conservation Ecology 1: http://www.consecol.org/vol1/iss2/art2/

Wieczorek SB, Todd CD (1997) Inhibition of Bugula flabellata and facilitation of Ciona intestinalis settlement by natural multi-species biofilms: effects of film age and the roles of active and passive larval attachment. Marine Biology 128: 463-473

Williamson M, Fitter A (1996) The varying success of invaders. Ecology 77: 1661-1666

Willis KJ, Cook EJ, Lozano-Fernandez M, Takeuchi I (2004) First record of the alien caprellid amphipod, Caprella mutica, for the U.K. Journal of the Marine Biological Association UK 84: 1027-1028

Woodbridge H (1924) Botryllus schlosseri (Pallas) – the behaviour of teh larva with special reference to the habitat. Biological Bulletin 47: 223-230 105

Young CM, Chia FS (1981) Laboratory evidence for delay of larval settlement in response to a dominant competitor. International Journal of Invertebrate Reproduction 3: 221-226

Young CM, Chia FS (1984) Microhabitat-associated variability in survival and growth of subtidal solitary ascidians during the first 21 days after settlement. Marine Biology 81: 61-68

Young CM, Chia FS (1985) An experimental test of shadow response function in ascidian tadpoles. Journal of Experimental Marine Biology and Ecology 85: 165-175

Zenetos A, Cinar ME, Pancucci-Papadopoulou MA, Harmelin JG, Furnari G, Andaloro F, Bellou N, Streftaris N, Zibrowius (2005) Annotated list of marine alien species in the Maditerranean with records of the worst invasive species. Mediterranean Marine Science 6: 63-118

106

APPENDICE A

107

Figure 1. Ciona intestinalis recruitment 08-Aug-2008 – Grid 1

Figure 2. Ciona intestinalis recruitment 22-Aug-2008 – Grid 2

108

Figure 3. Ciona intestinalis recruitment 18-Sep-2008 – Grid 3

Figure 4. Ciona intestinalis recruitment 27-Sep-2008 – Grid 4

109

Figure 5. Ciona intestinalis recruitment 10-Aug-2009 – Grid 5

Figure 6. Ciona intestinalis recruitment 24-Aug-2009 – Grid 6

110

Figure 7. Ciona intestinalis recruitment 07-Sep-2009 – Grid 7

Figure 8. Ciona intestinalis recruitment 21-Sep-2009 – Grid 8

111

APPENDICE B

112

0.8 )

m 0.7

q s

0.6

s 0.5

r c

e 0.4

(

n 0.3

o 0.2

i s

n 0.1

e D 0.0

Plate size (cm) 5 10 20 40 5 10 20 40 5 10 20 40 Depth (m) 1 3 5

Figure 1. Ciona intestinalis larval settlement densities on PVC settlement plates of various size (5x5, 10x10, 20x20, and 40x40 cm) at various depths (1, 3, and 5 m) – Delaney et al. in prep.

113