Interactions Between Mussel Bed Area and Perimeter Predict the Response of Associated Communities and Ecosystem Functions Lauren

Interactions between mussel bed area and perimeter predict the response of associated communities and ecosystem functions

Laurence Paquette

Department of Biology McGill University Montréal, Québec, Canada April 2015

A thesis submitted to McGill University in partial fulfillment of the requirements of the degree of Master of Science

First, I sincerely thank Frédéric Guichard for his supervision, funding support and advices which allowed me to bring this project to completion. I also greatly thank Philippe Archambault for his support and his contagious enthusiasm which revived my pleasure in accomplishing this project, in the harder moments. Thanks to Anthoni Ricciardi, member of my supervisory committee, for his suggestions.

I would especially like to thank Myriam Lemelin for her very helpful field assistance. She was not only essential to the realization of this project, but it was also a pleasure to share a summer with her. I thank Julie Joseph, Nicolas LeCorre, Julie-Anne Dorval, Brian Boivin and Olivier Cloutier for field assistance. I am grateful to Lisa Treau De Coelli, Laure de Montety, Julien Massé Jodoin and César Largaespada for their contribution in the laboratory work. Thank to Gwenaёlle Chaillou and Steeven Ouellet for technical support in laboratory and to Alain Caron and Guillaume Larocque for statistical advising.

I wish to express my gratitude to the members (the list is too long) of the Guichard lab and the Benthic Ecology lab. Thank you for your advices, listening and support. It was nice to be in your company.

Finally, I want to express my deep gratitude to my partner in life and my family who always believed in me and supported me, even from a distance.

This work was supported by grants from the Natural Sciences and Engineering Research Council of Canada and the Fonds de recherche du Québec -Nature et technologies.

Contribution of authors

Laurence Paquette developed, planned, managed and participated to the collection and analyses of all data. Statistical analyses and writing of this thesis were also realized by Laurence Paquette. Frédéric Guichard and Philippe Archambault contributed by their suggestions and comments throughout the process and the writing.

Table of contents

Preface ...... ii Acknowledgements ...... ii Contribution of authors ...... ii Table of contents...... iii List of tables ...... v List of figures ...... v Abstract ...... vi Résumé ...... vii Chapter 1: General Introduction ...... 2 Introduction ...... 2 Area effect...... 3 Shape parameters are interrelated ...... 3 Perimeter effect ...... 5 Effect of the interaction between perimeter and area ...... 6 Mussel bed ecosystem in the St Lawrence estuary ...... 7 Functional groups ...... 9 Ecosystem processes ...... 10 Objective ...... 11 Chapter 2 ...... 12 From habitat geometry to ecosystem functions in marine mussel beds ...... 12 Introduction ...... 13 Method ...... 16 Data collection ...... 16 Study site ...... 16 Mussel transplants ...... 17 Community of associated taxa ...... 18 Ecosystem functions...... 19 Data analysis ...... 20 Community of associated taxa ...... 20 Ecosystem functions...... 22 Results ...... 23

iii

Community of associated taxa ...... 23 Community with Littorina sp...... 23 Community without Littorina sp...... 23 Functional groups ...... 24 Ecosystem functions ...... 24 Discussion...... 25 Community of associated taxa ...... 26 The spatial scale ...... 26 Ecosystem functions ...... 27 The effects of habitat area and perimeter on mussel bed ecosystem ...... 28 Area effect on community characteristics ...... 28 Effect of the interaction between area and perimeter on the community and on other ecosystem properties ...... 29 Conclusion ...... 31 Tables ...... 32 Figures ...... 37 Chapter 3: General conclusion ...... 43 Summary and future perspectives ...... 43 Shape effect in marine systems ...... 43 Community composition ...... 44 Interrelated shape effects and their mechanisms ...... 45 Spatial scale ...... 46 Development of tools for seascape studies ...... 47 Implications ...... 47 References ...... 49 Appendix 1 ...... 63 Appendix 2 ...... 64 Appendix 3 ...... 65 Appendix 4 ...... 66

List of tables Table 1. Linear mixed effects models showing the effect of A) area, perimeter and of B) P/A ratio on taxonomic characteristics of the community ...... 31

Table 2. Linear mixed effects models showing the effect of A) the presence of a central subunit and B) the position of subunits on taxonomic characteristics of the community ...... 32

Table 3. Linear mixed effects models showing the effect of A) area and perimeter and of B) P/A ratio on functional characteristics of the community ...... 33

Table 4. Linear mixed effects models showing the effect of A) the presence of a central subunit and B) the position of subunits on functional characteristics of the community ...... 34

Table 5. Generalized linear mixed effects models showing the effect of A) the presence of a central subunit and B) the position of subunits on the abundance of the groups 'swimmer', 'substrate' and 'sessile' ...... 35

Table 6. Mixed factorial ANCOVA showing the effect of area, perimeter and the number of mussels alive on oxygen uptake ...... 35

Table 7. Factorial ANOVA showing the effect of A) area and perimeter and of B) P/A ratio on ammonium release ...... 35

List of figures Figure 1. Schematic depiction of patch shape metrics...... 4

Figure 2. Representation of the experimental mussel transplants...... 18

Figure 3. Incubation chamber used to measure fluxes of oxygen and ammonium.. 20

Figure 4. Effects of area and perimeter on community taxonomic characteristics ...36

Figure 5. Effect of the presence of a central subunit on taxonomic characteristics. 37

Figure 6. Effect of area on Shannon and Evenness index for functional groups..... 38

Figure 7. Effect of area and perimeter on richness in functional groups...... 38

Figure 8. Effect of the presence of a central subunit on functional characteristics.. 39

Figure 9. Effect of the presence of a central subunit on the abundance of 'sessile' and 'substrate'...... 40

Figure 10. Effect of perimeter on oxygen uptake ...... 40

Figure 11. Effect of area and perimeter on ammonium release ...... 41

Abstract In terrestrial habitats, the Equilibrium Theory of Island Biogeography and Metapopulation Theory have incorporated the effect of area and isolation to explain patch-occupancy dynamics. More recent theories predict effects of fragmentation on communities from the more complex interplay between size of habitat area and edge. As patch area decreases, the relative proportion of edge increases, with corresponding increase in the perimeter-area ratio (P/A ratio). However, few studies have explicitly addressed the independent and interacting effects of perimeter and area on community and ecosystem dynamics. This study aimed to disentangle the relative effects of perimeter, area and their interaction on the structure of the macro- invertebrates community associated with mussel beds and on ecosystem functions measured by nutrient fluxes. We conducted our experiment along the coast of the St Lawrence estuary (Sainte-Flavie, Quebec). Live blue mussels were used to create artificial mussel transplants corresponding to nine combinations of area and perimeter, in a factorial design. Area and perimeter effects on biodiversity and on fluxes of oxygen and ammonium were assessed. The interaction between area and perimeter was found to affect community characteristics and ammonium release. This project also reveals scale-dependence in the effect of habitat shape metrics and suggests that communities and ecosystem processes can be decoupled in their response to changes in perimeter and area. This study underlines the importance of integrating the interactive effects of various metrics of landscape geometry to the study of the relationship between community dynamics and ecosystem functions in fragmented habitats.

Résumé

La théorie de la biogéographie insulaire et la théorie de la métapopulation intègrent l'effet de l'aire et de l'isolement pour expliquer la dynamique d'occupation des parcelles d'habitat en milieu terrestre. Des théories plus récentes prédisent que les effets d'un changement de la configuration d'un habitat sur la communauté découlent d'une interaction plus complexe entre la taille de la superficie totale de la parcelle et celle de sa bordure. Lorsque l'aire d'une parcelle diminue, la proportion relative de sa bordure augmente, correspondant à une augmentation du ratio périmètre-aire (ratio P/A). Pourtant, peu d'études ont explicitement abordées les effets de l'interaction entre l'aire et le périmètre d'une parcelle sur la communauté associée et la dynamique de l'écosystème. Cette étude a pour but de dissocier l'effet respectif de l'aire, du périmètre et de leur interaction sur la structure de la communauté de macro-invertébrés associée aux transplants de moules et sur les fonctions de l'écosystème mesurés par les flux d'oxygène et d'ammonium. L'expérience pris place le long de la côte de l'estuaire du Saint-Laurent (Sainte- Flavie, Québec). Des moules bleues vivantes ont été utilisées pour créer des transplants de moules expérimentaux de 9 formes correspondantes aux combinaisons d'aire et de périmètre d'un plan factoriel. Les résultats indiquent que l'interaction entre l'aire et le périmètre influence les caractéristiques de la communauté et les flux d'ammonium. De plus, ils suggèrent que l'effet des paramètres de forme dépendent de l'échelle spatiale et que la réponse de la communauté et des fonctions de l'écosystème peuvent divergées. Cette étude souligne le besoin d'intégrer l'effet de l'interaction entre plusieurs paramètres de la forme de l'habitat pour l'étude des relations entre la dynamique de la communauté et les fonctions de l'écosystème.

vii

Chapter 1: General Introduction

Introduction In the last decades, ecologists have shown a strong interest in trying to predict how ecosystems respond to habitat configuration. Landscape ecology is a field of study, developed in terrestrial environment, interested in relationships between spatial patterns and ecological processes at a wide range of spatial and temporal scales. It has been developed in the 1980's with an increase in the awareness of the broad- scale environmental issues that required a landscape perspective, the technological advances providing access to new tools, the recognition of the importance of scale and with the view of ecosystems as dynamic and open systems requiring an understanding of how mosaic of ecosystems interact to affect ecosystem processes (Turner et al. 2001). Knowledge on these relationships can assist managers in applying conservation plans and solving environmental issues (Pittman et al. 2011). For example, it is used to assess how habitat fragmentation affects population viability (Bender et al. 1998; Fahrig, 2003; Ewers et al. 2010; Bostrom et al. 2011; Mizerek et al. 2011; Arponen and Bostrom, 2012). A landscape is a heterogeneous area composed of interacting ecosystems (Forman and Godron, 1981). Habitat fragmentation not only affect the amount of habitat, but also the shape of patches. A patch is defined as a relatively homogeneous area, which differs from its surrounding (Forman 1995) and constitutes the basic unit of a landscape. The total area, the edge and core area, the geometric shape and the level of isolation are main attributes characterizing patches which have been studied (Ewers and Didham 2006). The term 'geometric shape' refers to only regular geometric shape such as circle, square and triangle (Airoldi 2003). The term shape has a broader meaning, similar to that used in landscape ecology (Farina 1998). This last term is referring to a wide range of characteristics of the dimensions and spatial configuration of a patch, such as those enumerated previously. Changes in these parameters, whether of natural or anthropogenic source, can influence the behaviour and distribution of organisms, modify the biotic and abiotic conditions in the patch and alter altogether the ecosystem functioning. Knowledge on those parameters effects, both individually and in interaction, is required in order to integrate them in landscape and seascape theories and when developing management plans.

Area effect In the field of landscape ecology, two tightly linked theories have been developed and well studied in terrestrial habitats: the Equilibrium Theory of Island Biogeography (MacArthur and Wilson 1967) and the Metapopulation Theory (Hanski and Gilpin 1991). Both incorporate the effect of area and isolation to explain patch-occupancy dynamics. According to these concepts, taxa richness would increase with island area size and decrease with island isolation. Species-Area relationships have been studied since the nineteenth century and several mechanisms have been proposed to explain the influence of area on community structure. Fletcher Jr. et al. (2007) reviewed two primary factors leading to area effects: the resource concentration hypothesis and the target effect. First, as patch area increases, the amount and diversity of resources available to organisms tend to increase (Root 1973), which has a positive effect on population size (MacArthur and Wilson 1967). The diversity of resources available also favours coexistence among a larger number of taxa (Williams 1943; Connor and McCoy 1979; Russell et al. 2006). Secondly, patch size can positively influence immigration rates (Gilpin and Diamond 1976). A larger area should sample more taxa from the available pool than a small patch, leading to greater taxa richness (Sousa 1984). Conspecific attraction can accentuate this positive effect of area on colonization rates (Fletcher 2006, 2009; Largaespada et al. 2012). Obviously, these direct effects can also lead to indirect changes in taxa interactions affecting community structure.

Shape parameters are interrelated Some patterns assumed to result from the effect of the patch size were derived from the effect of others shape parameters that have not been dissociated from patch area effect. A negative effect of patch size on emigration rate (e.g. Bowman et al. 2002), can be explained by the increase of perimeter associated with increasing patch area size (Mancke and Gavin 2000; Hamback and Englund 2005; Fletcher 2007).

We differentiate the notion of edge from perimeter, the later being the length of the interface between the patch and the surrounding environment (the patch circumference). The edge is the portion of an ecosystem near its perimeter, which

3 can be environmentally different from the core area, caused by the influence of the adjacent external environment (Forman 1995). In other words, the commonly called 'edge effect' can be calculated through 'lineal' (perimeter) and 'areal' (edge) measurements (Baskent and Jordan 1995). The perimeter is a lineal measure of the boundary length whereas the edge and core area are areal measures. Several authors have mentioned the importance of the edge length in addition to the penetration distance, or the edge width (Laurance and Yensen, 1991; Wu and Vankat, 1991). Variations in both, perimeter and total area, determine the percentage of core area and edge (Figure 1). As mentioned by Helzer and Jelinski (1999): ‘Because of the ambiguity about how far edge effects extend into patches (Faaborg et al. 1993), the use of a relative measure such as perimeter-area ratio seems appropriate’. Of all the various estimators of the edge, the P/A ratio is the one that has received the most consideration (Sousa 1984; Polis and Hurd, 1995; Helzer and Jelinski 1999).

Figure 1. Patch metrics

In natural landscapes, variations in perimeter and area are not independent, and their co-variation depends on the shape of habitats. For a patch of a given geometric shape, when patch area increases, the perimeter increases and the relative proportion of area exposed to borders decreases (namely the edge). This translates into a decrease in the perimeter-area ratio (P/A Ratio) (e.g. Fletcher et al. 2007; Banks-Leite et al. 2012).

Perimeter effect The access to adjacent resources and ecological flows on borders, referring to the movement of organisms, energy and matter (Wiens et al. 1985, Cadenasso et al. 2003, Ries et al. 2004), are factors that can lead to an effect of perimeter on the ecosystem (Reviewed by Ries and Sisk, 2004). A large perimeter promotes access to mobile organisms. Between patches of the same area, mobile fauna has more chance to enter patches presenting greater perimeter (Forman 1995; Haddad and Baum 1999; Fried et al. 2005). A higher immigration rate could lead to a greater abundance of the mobile fauna in patches with an important perimeter. Accordingly, Arponen and Boström (2012) have reported a positive effect of fragmentation on total epifauna abundance and support a positive edge effect in seagrass. Large perimeter could also lead to an increase in ecosystem functioning as a result of strong flows of resources.

Perimeter is defined as the outer limits of an area and hence the boundary between distinct habitats represented by different characteristics. If resources available across these discrete environments are complementary, a longer perimeter enhances the access to both resources (Dunning et al 1992; McCollin 1998; Fagan et al. 1999). Mobile organisms can reach resources more easily when the perimeter is longer and sessile species on the edge can also access resources from adjacent environments. A positive effect of perimeter, which leads to an increase of the density of individuals on the edge, is then expected. The response to perimeter length depends on taxa preferences and edge type (Forman 1995; Collinge and Palmer 2002; Fletcher and Koford 2003; Ries and Sisk 2004).

Taxa interactions are also likely to occur along the perimeter. Therefore, habitat shape can influence species interactions (MacArthur et al. 1972; Root 1973). This indirect factor is not restricted to edges but can have important outcomes on the distribution of taxa. For instance, scallops experienced a significantly higher predation risk along the borders of seagrass habitat patches (Bologna et al. 2000). Crabs also prey more on mussels along the edge of beds (Okamura 1986). However, taxa interactions are highly variable and depend on edge type, specific predators behaviours and landscape context (Forman 1995; Collinge and Palmer 2002; Fletcher and Koford 2003; Ries and Sisk 2004).

Effect of the interaction between perimeter and area Strong interactions are expected between area and perimeter. Edge and core area effects could explain part of the influence of area-perimeter interactions on community structure. Variation in the amount of habitat edge has been used to explain the effect of the interaction between perimeter and area through simple mechanisms. It has been suggested that a habitat with a large area and a small perimeter, and consequently a small P/A ratio, is more subject to present a diversity of environmental conditions, which supports taxa richness. If species perceived the habitat on the edge differently than the habitat from the core area, then patches with small P/A ratio have the advantage of providing a diversity of habitats (Helzer and Jelinski 1999; Ewers et al. 2007; Fletcher et al. 2007; Ewers and Didham, 2007; Bostrom et al. 2011; Didham and Ewers, 2012; Arponen and Bostrom 2012). As an example, the richness of grassland birds as been found to be inversely correlated with perimeter-area ratio. Patches with a small P/A ratio provided wet and dry sites with short and tall vegetation (Helzer and Jelinski 1999).

Nevertheless, the P/A ratio may not capture the various facets of the influence of the interaction area-perimeter. Note that, patches of different areas and perimeters, and therefore of different shapes, can have the same P/A ratio. Consequently, the effect of an interaction between perimeter and area can be unrelated to the P/A ratio. The joint effects of area and perimeter are expected to display strong interactions, even if there are no differences in the percentage of edge or core area. Patches of different shapes with the same amount of core area can present different conditions. For example, a patch can have a simple core area while others can present more complex shapes with multiple disjointed core areas that are separated by bands of habitat influenced by the matrix (Ewers 2004).

The level of complexity of the perimeter line (Nams 2014) and the patch geometric shape (Laurance and Yensen 1991; Baskent and Jordan, 1995; Yamaura et al. 2008) also depend on area-perimeter interactions and are likely to influence biotic displacements and abiotic conditions in the habitat. The presence of concave or convex angles and curves, determines the level of complexity of the line. In a marine system, these characteristics can modify the hydrodynamics surrounding a patch of habitat and thus influence abiotic and biotic components within the patch. Perimeter- area ratio, often used as 'shape index', automatically classifies an elongated, narrow

6 patch as highly complex even if it is geometrically quite simple (Hulshoff 1995; Moser et al. 2002). Therefore, P/A ratio is thought to be a better index for compactness and for the percentage of edge, than for geometric shape complexity (Baskent and Jordan 1995; Haines-Yang and Chopping 1996). Other indexes can be used to test shape complexity effect (e.g. Forman and Godron 1986; Ripple et al. 1991).

Area and perimeter (and edge by extension) are main parameters of the habitat shape, which can influence resources mapping, species distribution and ecosystem functions. A review on faunal responses to patch shape mentioned that results in marine system differ from terrestrial studies to a certain extent and are still highly inconsistent, with positive, negative and neutral effects reported (Bostrom et al. 2011). Area and edge effects, and thus perimeter effect, have often been confounded in studies (Reviewed in Parker et al. 2005; Fletcher and al. 2007). The importance of distinguishing the contribution of each shape parameter has now been recognized (Jelbart et al. 2006; Ewers et al. 2007; Barbaro et al. 2012; Soga et al. 2013; Carpintero and Reyes-Lopez 2014), but it has rarely been applied to marine systems (Airoldi 2003). Airoldi (2003) used a design to investigate separately the relative importance of area, perimeter and distance from edge on colonisation of algae on a rocky coast. The patches sizes (150, 280 and 320 cm2) were small and covered a narrow range of spatial scales. This study revealed shape parameters to influence algae colonisation. The goal of my thesis was to test the influence of shape parameters on the community structure associated to a wider range of large habitat patches in marine systems. Major theories developed in terrestrial systems are based on studies using patches of one or more square kilometer. Some mechanisms observed on large patches may also be relevant at smaller sized patches and could explain ecosystem response. My thesis also reports results from a field experiment that explicitly tested the interacting effects of area and perimeter, which are not captured by indices used in previous studies, such as the P/A ratio (Sousa 1984) or the distance between the perimeter line and an estimated core area (Airoldi 2003).

Mussel bed ecosystem in the St Lawrence estuary Mussels beds are a suitable model for experimental studies in the field of coastal seascape ecology. Blue mussels (composed of Mytilus edulis, M. trossulus and hybrids, hereafter named Mytilus Spp.) are abundant on the rocky coast of the St.

Lawrence estuary, Canada, and their aggregations form a mosaic of mussel beds. The spatial structure of those beds provides resistance to wave exposure (Hunt and Scheibling 2001, 2002), reduces current velocities, dampens waves, traps sediments and influences water quality (Widdows et al. 1998; Dolmer 2000; Folkard and Gascoigne 2009; Van Leeuwen et al. 2010, Newell 2004). Mussel leads to the production of faeces and pseudofaeces which are deposited in mussel aggregations and are a nutrient source for many organisms (e.g. Tsuchiya and Nishihira 1985; Richard et al. 2007). Therefore, mussel beds modify physical and biological habitat conditions, control the availability of resources to other organisms, promotes the establishment of a wide variety of associated taxa (Commito et al. 2005; Crain and Bertness 2006; Largaespada et al. 2012) and form ecosystems. Hence, mussels are considered as bioengineers (Jones et al. 1994, 1997, Cole 2010).

Recent studies have provided evidence for the importance of small-scale connectivity, related to proximity (<1m) between patches, for mussel engineering (Brazeau 2009; Largaespada et al. 2012). They have revealed that high connectivity, favoring population colonization, can lead to a homogenization in community structure and enhance within-patch ecosystem processes. Mussel beds are dynamic. Indeed, the wave force of seasonal storms and active crawling behavior of mussels (Underwood 1999; Schneider et al. 2005) influence mussel bed configuration due to mussel dislodgement and redistribution (e.g., Paine and Levin 1981; Denny 1995; Hunt and Scheibling 1998, 2001; Petrovic and Guichard 2008 ). It can lead to changes in patch connectivity, but also to changes in its shape. Mussels can form irregular patches of various size (cm2 up to over 100 square meters) or more regular elongated beds (e.g. Van Leeuwen et al. 2010). An aggregate can be composed of as few as 9 mussels and still have an effect of bioengineering (Cole 2010). The small size of a mussel bed ecosystem and the small scales of spatial and temporal variability make experimental manipulation feasible and appropriated at small scale (Connell 1972). It is also facilitated by the great number of macroscopic taxa, often sessile or slow-moving, which interact over small spatial scales (Underwood and Chapman, 1996).

Functional groups The shape of mussel beds could influence taxonomic and functional diversity, and ecosystems functions. Community structure and dynamics can influence ecosystem processes, but the mechanisms underlying the relationship between taxa and ecosystem functions is subject to debate (Loreau et al. 2001; Giller et al. 2004). Among the diverse points of view, it was suggested that the functional characteristics of the taxa could be more important than the number of taxa (Bengtsson, 1998; Bolam et al. 2002; Biles et al. 2003; Raffaelli et al. 2003). Several taxa can have similar traits and functions. The absence of some taxa performing a particular role in an ecosystem or a reduction in their frequency may be compensated by the presence of other taxa with similar characters (Frost et al. 1995). Thereby, the use of index based on functional traits can be more relevant than taxonomic diversity measures to assess ecosystem response to environmental conditions (e.g. Vandewalle et al. 2010). In any case, it is understood that the presence of many taxa helps to support the functioning of ecosystems, since more taxa will most probably represent a greater diversity of traits (sampling effect, Cornwell et al. 2006; Villeger et al. 2008). A great richness in taxa can lead to a complementarity in the resources usage (Loreau and Hector 2001), to a facilitation (Cardinale et al. 2002), or increase the probability of including a dominant taxon (Huston et al. 2000). To assess the effect of the habitat shape on the community structure, both taxonomic diversity, and the diversity of functional groups should be considered.

Groups of taxa with different life history traits can respond differently to the parameters of habitat configuration (Ewers and Didham, 2006). Organisms with different degrees of mobility can show different dynamics of colonization (Antoniadou et al. 2011). Depending on the type of locomotion of organisms, the habitat shape may present barriers or, at the opposite, favor displacements, and thus influence colonization as well as the community structure. In addition, since the diet is often associated to a mode of locomotion in benthic system, an important diversity in functional groups based on the type of locomotion may be related to complementarity of resources usage. In mussel bed ecosystems, there is a great richness of taxa characterized by small body size, low trophic level and moderate mobility. Functional groups, based on the mobility, have contrasting biology and are easily identified (Borthagaray et al. 2009). For these reasons, locomotion seems to

9 be an adequate trait to subdivide the taxa into functional groups in rocky intertidal communities.

Ecosystem processes Accounting for biotic and abiotic components allow a better understanding of the overall functioning of the ecosystem (e.g. Loreau et al. 2002; Daufresne and Hedin 2005). Most studies have focused on the effect of habitat parameters on biodiversity. Biodiversity draws a lot of interest, partly because its potential relationship with functions of the ecosystem. However, we have no evidence that the effect of patch shape on biodiversity and on ecosystem functions are mediated by such relationship, or by similar underlying processes. Therefore, it is of interest to study not only the effect of habitat shape on community structure, but also its influence on ecosystem functions directly. In addition to hosting a high biodiversity of associated species, mussels beds are important for exchanges between the water column and the benthic system and for nutrient turnover in shallow coastal environments (Pfister, 2007). Invertebrates (mussels and associated taxa included) are involved in the regulation of processes such as the carbon and nutrient cycling and the decomposition of organic matter (Richard 2006, 2007a, b; Norling and Kautsky, 2008). Those ecosystem functions can be approximated through the measure of fluxes. Fluxes are a more accurate measure than concentration as they outline a tendency over a time period. An influence of patch shape on hydrodynamics and thus deposition of sediment and matter in mussel beds (van Leeuwen et al. 2010) can potentially influence ecosystem functions such as nutrient cycling and decomposition of organic matter. These effects on ecosystem functions have been observed under Mytilus culture (Richard et al., 2007a; Stenton-Dozey et al., 2011; McKindsey et al. 2011; Robert et al. 2013).

Mussels enable a transfer of energy from the water column to the benthic compartment through their activity of filtration. The benthic invertebrates can excrete faeces and pseudofaeces in substantial quantity, which have a high content of ammonium (Jensen and Muller-Parker 1994). Faeces and pseudofaeces are a valuable source of nutrients for invertebrates of the intertidal zone. The decomposition of organic matters and the bioturbation from invertebrates are

10 important for ecosystem functioning. Furthermore, microorganisms can convert the organic matter into ammonium or the ammonium into nitrate. These processes, respectively named mineralization and nitrification, as well as their rates, depend on the availability of organic matter, ammonium and oxygen (Kemp et al. 1990). Nitrate is the most available form of nitrogen for plants. Primary production can thus be enhanced by the combined action of mussels and microorganisms (Pfister, 2007). An influence of the habitat shape on the conditions of the habitat and on the community could then affect the input of energy in the ecosystem, and the ecosystem functions.

Objective The knowledge of the influence of habitat shape and of the interaction among its metrics on communities and ecosystems is essential for a better understanding of fragmented habitats and of the impacts of habitat fragmentation in marine systems. To my knowledge, no study has explicitly tested the interacting effects of perimeter and area on both community structure and on ecosystem functions. The aim of this study is to test the effect of area, perimeter and their interaction on both the community and the ecosystem. This study proposes to further the issue of habitat configuration beyond the level of macro-community structure and look at the effect of patch shape on rate of ecosystem functions measured as rate of oxygen uptake and ammonium release. Results reported in this thesis also extend the issue of fragmentation beyond the use of simplifying metrics such as the P/A ratio, by resolving the explicit interactions between perimeter and area. My results reveal important scale-dependence in the effect of these components of habitat geometry and suggest that communities and ecosystem processes can be decoupled in their response to changes in perimeter and area.

Chapter 2

From habitat geometry to ecosystem functions in marine mussel beds

Laurence Paquette1, Frédéric Guichard1 and Philippe Archambault2

1Department of Biology, McGill University, 1205 Dr. Penfield, Montréal (Québec), Canada H3A 1B1

2Institut des Sciences de la Mer (ISMER), Université du Québec à Rimouski, 310 allée des Ursulines, Rimouski (Québec), Canada G5L 3A1

Introduction Landscape ecology concepts and techniques are increasingly applied to marine environments in order to gain a better understanding of the consequences of habitat configuration and fragmentation (Bostrom et al. 2011). Knowledge of ecosystem responses to seascape configuration has important implications for the management of coastal resources (e.g. Mizerek et al. 2011; Green et al. 2012). However, studies interested in the effects of habitat configuration parameters in marine system are scarce compared to those in terrestrial studies and there is a need to tease apart their effects, often confounded in studies, in order to identify shape metrics and indexes mainly influencing marine coastal ecosystems. This study is interested in the influence of shape metrics of a habitat patch on its ecosystem.

The patch-matrix (or island) model of landscape cover is based on the concept of island biogeogarphy theory (MacArthur and Wilson 1967; Franklin and Lindenmayer, 2009; Wedding et al. 2011). According to this model, a patch is defined as an area of suitable habitat, which differs from its surrounding matrix of unsuitable habitats. This model focuses on the study of biotic responses to the various attributes of the patch, such as its area, perimeter, geometric shape complexity and its isolation. Area and perimeter are main parameters characterizing the shape of habitats and can influence resources and taxa distribution and ecosystem functions. We can further distinguish the notion of perimeter, which is a linear measure of the boundary length, from the notion 'edge' which is an 'areal' measure (Baskent and Jordan, 1995) of the portion of the area influenced by the adjacent environment. Variations in both, perimeter and total area, determine the percentage of core area and edge. In marine systems, results of studies that have investigated the effect of these parameters at the patch scale have been highly variable (Bostrom et al. 2006). In natural landscapes, a change in area also involves a change in perimeter and edges. Each parameter involves mechanisms, which can influence the ecosystem. The inconsistency in the effects reported for any specific shape parameter (e.g. area), can be in part attributed to the fact that its effect is confounded with others parameters such as perimeter. The importance of considering the distinct effect of each habitat shape parameter is now recognized (Ewers et al. 2007; Barbaro et al. 2012; Soga et al. 2013; Carpintero and Reyes-Lopez 2014), but it has rarely been

13 applied in marine systems (Airoldi 2003; Jelbart et al. 2006; Arponen and Bostrom 2012).

The majority of seascape studies were carried out in seagrass ecosystems (49%) and tidal marshes (32%). Fewer studies were interested in coral reefs (11%) and mangroves (6%), and almost none were conducted in rocky shores with algae mosaics and beds of oysters or mussels (Airoldi, 2003; Bostrom et al. 2011). Yet, Mytilus spp. are bioengineers present in great abundance on marine coasts, and playing a key role on the dynamics of shallow coastal ecosystems. Aggregates of mussels dampen waves, clarify the water, trap sediments, and are important for pelago-benthic transfers of carbon and nutrients (Widdows et al. 1998; Richard 2007a, b; Folkard and Gascoigne 2009; Van Leeuwen et al. 2010). They form a mosaic of mussel beds of various sizes and shapes along coastal habitats. Both biotic (Menge 1976; Okamura 1986; Hunt and Scheibling 2001; Cusson and Bourget 2005; Petrovic and Guichard 2008) and abiotic factors (Levin and Paine 1974; Paine and Levin 1981; Hunt and Scheibling 2001) were showed to affect patch dynamics, and spatial configuration. Changes in habitat configuration, both at the landscape and at the patch scales, may affect their associated community and ecosystem functions. Brazeau (2009), showed that small scale connectivity (<1m) between mussel patches enhances population colonization and can lead to a homogenization of the communities structure. Some studies assessed the influence of mussel patch size on taxa diversity within associated communities (Tsuchiya and Nishihira, 1985; Peake and Quinn, 1993; Norling and Kautsky, 2008; Koivisto and Westerbom, 2012). In these studies the effect of area might have been overestimated. The effects of other shape metrics were confound with that of the area. Variability in community characteristics might be better explained by another shape parameter or the interaction between some. Area and perimeter are main attributes of patch shape. Knowledge on those parameters effects, both individually and in interaction, are necessary to a better understanding of the dynamics of populations and ecological processes.

A study conducted on a small scale like this one, over habitat patches with dimensions of the order of one square meter, allows to manipulate the shape parameters and thus test their respective effects and their interaction. It provides valuable fundamental information that might orient future studies. It would be

14 interesting to test their effect at different size scales. It is pertinent to relate the mechanisms observed in studies using larger plots (about one square kilometer) as some of these mechanisms may also be involved in a small patch, even in other types of ecosystem.

An interaction between area and perimeter of mussel patches could be explained by the perimeter-area ratio. The P/A ratio is a relative measure of the percentage of edge. In terrestrial habitats, the edge and core area effects are recognized to have a greater influence on the community structure than the total patch area (Ewers et al. 2007; Ewers and Didham, 2007; Fletcher et al. 2007; Banks-Leite et al. 2010; Didham and Ewers, 2012). Edge effect has also been reported in marine system (Sousa 1984; Murphy et al. 2010; Bostrom et al. 2011; Arponen and Bostrom 2012), but are taxon-specific and can be positive or negative. Patches with a small P/A ratio have a high percentage of core area and are more likely to present a diversity of environmental conditions that support high taxonomic richness. The roughness of the bed created by mussel shells reduces current velocity and favors deposition of sediments and pseudofaeces in core areas of patches. Such faeces, pseudofaeces and sediments accumulation in core areas can in turn benefit some taxa (e.g. Tsuchiya and Nishihira, 1985). Oxygen depletion in the boundary layer over large beds can also occur (Wright et al. 1982; Fréchette et al. 1989). In contrast, other taxa benefit from lower accumulation on the edges that are exposed to erosion. Mussels and associated species could benefit from such improved environment through a better access to resources (Ries et al, 2004). A large patch perimeter could enhance ecosystem processes. Area can influence immigration rates and have a positive effect on taxonomic richness and abundance. The greater concentration and diversity of resources on large areas may be actively target by some taxa (Root, 1973; Gipin and Diamond, 1976) and might lead to a positive area effect. Thus a positive effect of area and perimeter might be expected on taxonomic diversity associated to marine mussel patches. We might also expected an effect of the interaction between these shape metrics, explained by a negative effect of the P/A ratio on taxa diversity. In addition to the response of taxonomic diversity to patch geometry, taxa response to patch shape can largely depend on their traits (Ewers and Didham, 2006), suggesting the importance of functional diversity for predicting community and ecosystem response to habitat shape.

One of the main challenges for seascape ecology is to identify structural attributes, at the patch scale, which drive the biotic assemblages and key ecological processes (Pittman et al. 2011). Few studies have considered the effects the habitat shape attributes on ecosystem functions. Most have focused on their effects on taxa distribution (Bostrom et al. 2011, Wedding et al. 2011). Decoupling the geometric properties in the study of landscape would allow to determine what attribute of the structure drive biotic assemblages, distribution of biodiversity, ecological functions and processes. This would orient research towards these attributes to understand mechanisms involved and how community and ecosystem responses are linked. It would be necessary in order to reach the same level of practical benefits for coastal management than what was achieved in terrestrial environments (Botequilha Leitao and Ahern, 2002; Pittman et al. 2011).

Mussel beds form dynamic habitats across a broad range of spatial and temporal scales, and hosting a great diversity of macro-invertebrates (Underwood and Chapman, 1996). These characteristics, along with their compatibility with experimental manipulation make them an ideal model system for seascape experimental studies (Connell, 1972). In order to test for the interacting effects of multiple metrics of habitat geometry on both communities and ecosystems, we conducted field experiments to identify structural attributes, at the patch scale, that drive community characteristics and ecosystem functions within mussel beds. We more specifically tested the effect of area, perimeter and of their interaction on macro benthic community structure (taxonomic and functional diversity, abundance) and on ecosystem functions (oxygen and ammonium fluxes).

Method

Data collection

Study site We conducted our experiment in a rocky intertidal area near Sainte-Flavie, on the south of the St. Lawrence Estuary, (Quebec, Canada; see also Guichard et al., 2001; Brazeau, 2009; Largaespada et al., 2012; Lemieux and Cusson, 2014). The weak inclination and the absence of fresh water tributaries, large boulders and deep tidal

16 pool limit potential confounding factors. The regular shoreline is exposed to wave action, semi diurnal tide cycles and ice scouring. The average tidal range was between 1.7 and 3.7 meters at the time of the study, between June 24 and August 22 2013 (Canadian Hydrographic Service 2013). Macroalgae (Fucus sp.) and blue mussels (composed of Mytilus edulis, M. trossulus and hybrids, hereafter named Mytilus spp.) are dominant species (Archambault and Bourget, 1996). A 50m stretch was chosen as our study site between 0.7 and 1.2 m of shore elevation, where natural mussel beds covered 30 to 40% of the site. It was cleared of macroalgae and all natural mussel aggregates prior to the experiment. The site was selected because it was clear from topographic irregularities >30cm and had <20% tidal pool cover (see also Guichard et al., 2001; Brazeau, 2009).

Mussel transplants Experimental mussel transplants (Figure 2) were used as habitat to control for habitat area and perimeter. Galvanized steel grids with a mesh size of 1 cm2 were used as the bottom substratum of transplants of 3 controlled areas (800, 1000 and 1200 cm2) of specific shapes that were filled with mussels to form experimental mussel beds. A thin rubber mat in the bottom of the transplants facilitated mussel byssal attachment (Largaespada et al. 2012). A thin plastic net with a mesh size of 1 cm2 covered each transplant to limit mussel dislodgement. It was attached to each transplant so as to define 10x10 cm subunits limiting changes in mussel aggregation size and shape through passive and active movement within the transplant (Fig. 2). Live mussels between 3 and 4 cm were collected from natural mussel beds located 1.2 kilometers downstream from the study site, and rinsed with seawater to remove associated species and sediments. Mussels were then added to each transplant at a fixed density of approximately 300 ind./m2 (30 ind./10cm2). Mussel transplants were left for 4 days in flow tanks supplied with unfiltered water to allow for byssi attachment. Transplants were randomly distributed among selected adequate spots of fixation 1m apart from each other to avoid between-transplant interactions (Brazeau 2009). All transplants were fixed on the bare rock of the study site between 24 and 26 June using stainless metal screws for a period of 56 days corresponding to the peak of the colonization period for most associated taxa (Brazeau, 2009; Largaespada et al., 2012). A total of 36 transplants were used, each with a controlled area (800, 1000 and 1200cm2) and perimeter (160, 180 and 200cm) according to a

17 factorial design (Fig. 2). Each combination of area and perimeter included 4 replicate transplants with identical shape (Fig. 2).

Figure 2. Shapes of the experimental mussel transplants corresponding to the nine combinations (a, b, c, d, e, f, g, h and i) of area (800, 1000 and 1200 cm2) and perimeter (160, 180, 200 cm) in a factorial design. Each transplant is composed of subunits (10x10cm) delimited by the net covering the transplant. Value of perimeter-area ratio (cm-1) associated to each transplant is indicated in the bottom-right corner. Subunits with a central or border position are identified in black and grey respectively. Transplants c, d, f and i have a central subunit.

Community of associated taxa Transplants were removed from the study site between August 19-22 and brought to the laboratory for analysis. Macro invertebrates associated to each subunit were sampled by rinsing the mussels and the subunit structure over a sieve with a mesh size of 500 um. They were preserved in a buffered solution of formaldehyde 4%. In each subunit, live and dead mussels were counted and weighed. Macro invertebrates were transferred in ethanol 70%. We randomly selected four subunits from each transplant. All sampled organisms from each subunit were identified under a dissecting microscope to the lowest taxonomic level possible, and the density of each taxon was evaluated. Number of individuals for each taxon was used to determine taxonomic richness, total abundance (density), Shannon diversity index and Pielou's evenness index (Pielou, 1966).

Density and richness of functional groups were also calculated. Functional groups were assigned to each taxon based on mode of locomotion of their adult stage: 'swimmers' were those that are highly mobile and can swim in the water column, 'Substrate' was used for mobile taxa that crawl on the substrate, and taxa with a sessile adult stage were simply referred to as 'sessile'.

Ecosystem functions Ecosystem functions were assessed by measuring ammonium and oxygen fluxes. 4 replicate transplants in each of the 4 extreme combinations of area and perimeter (Treatment a, c, g and i identified in Figure 2, in 4 replicates for a total of 16 transplants) were used to measure oxygen consumption and ammonium release. Five subunits were randomly selected within transplants and transported to the laboratory. They were incubated in sterile dark chambers (27 cm in diameter and 20 liters in volume) made of high-density opaque polyethylene (Figure 3). Absence of light allowed to control for photosynthesis. Thereby, only the processes associated with respiration activity of the overall community were studied (Hargrave 1969; Boucher and Clavier 1990; Plante-Cuny et al. 1998). Mussels and their associated community were incubated together and a single chamber was used per transplant. The chambers were previously washed with acid water (HCl 10%), rinsed abundantly and filled with unfiltered water pumped off the St. Lawrence estuary (hereinafter referred as the reference water). Each chamber, hermetically closed, contained a submersible pump that ensured water mixing. An increase in the temperature of the water is known to induce an increase oxygen consumption from benthic communities and influence biogeochemical fluxes (e.g. Upton et al., 1993; Banta et al., 1995; Cowan et al., 1996). To avoid water warming within chambers, they were installed in a network of basins supplied with circulating seawater at approximately 10°C. Ten milliliter of water were sampled at the beginning of the incubation, followed by four other sampling at intervals of 60 minutes, for a total incubation time of 4 hours. When a sample was taken, the same volume of reference water was injected to keep the water volume constant in the chamber and prevent contact with air. Water samples were filtered on Watman GF/F filters and directly preserved at -20 °C. Oxygen concentration was measured with an YSI Pro2030 probe within the chambers and the reference water at each sampling interval. By the end of the incubations, oxygen concentration was not depleted by more than 30%. This has been planned to

19 prevent hypoxic conditions which could modify bivalve metabolism (Mazouni et al. 1996). The chambers were emptied after incubation, washed and rinsed a few times, and filled with reference water the next day for a new series of incubations. The ammonium concentration of the samples was measured with the method of ionic chromatography (Hall and Aller, 1992). Fluxes of oxygen and ammonium were determined from the slopes of the linear regressions established between incubation time and concentration.

Figure 3. Incubation chamber used to measure fluxes of oxygen and ammonium. Each chamber contains 5 subunits of transplant and a water pump. They are placed in a basin filled with circulating water to keep the temperature constant.

Data analysis

Community of associated taxa We constructed linear mixed models and generated p-values to assess the variability of community characteristics explained by habitat shape variables (using lmer4 and lmerTest in R). The models included the fixed variables 'area' (3 levels), 'perimeter' (3 levels) and their interaction. Random effects for transplant (36 levels) were included in the model to account for non-independency between subunits from a same experimental mussel transplant. Richness, total abundance, Shannon diversity index and Evenness index, calculated for taxa and functional groups, were used as dependent variable. Practical constraints and random losses of subunits after collection of transplants have limited the number of subunits available for community analysis on some transplants and it results in unequal sample sizes. Assumptions of

20 normality and linearity were checked by visual inspections of plots of residuals against fitted values. Data of abundance (density) and richness were log+1 transformed to meet the normality assumption. The Bartlett's test was used to test homogeneity of variances (Bartlett, 1937; Snedecor and Cochran, 1989). The presence of influential data points was assessed using Cook's distance (Cook, 1977).

Inspection of model residuals showed that negative binomial errors provided the best fit for models of functional groups abundances (sessile, substrate, swimmer). The dataset included zero count-data that could not be normalized by transformation, and there was overdispersion under a Poisson distribution. Consequently, we used generalized linear mixed models for these analysis. Models were built to test the influence of the fixed effect area (3 levels), perimeter (3 levels) and their interaction on the total abundance (density) of the functional groups 'swimmer', 'substrate' and 'sessile' separately. Transplant was included as a random variable.

Pairwise comparisons were calculated on least-squares means and with Tukey method for adjustment on p-values for family of three tests. When an effect of the interaction between area and perimeter was significant, we test the hypothesis of an effect of the fixed variable P/A ratio (8 levels) with the same method determined to be appropriate for the dependant variable. Transplant was included as random variable.

These analyzes were performed on the data for the complete community of macro- invertebrates associated to mussels transplants and on the data with Littorina sp. excluded. We were interested in the response of the community excluding Littorina sp. because it accounted for 78% of the total density of the community in mussel transplants and its own distribution was not influenced by shape treatments.

We were interested to know if there was variability in the taxonomic and functional characteristics of the community between the border and the central position within transplants. Thus, we built a linear mixed effects models with the fixed variable position (two levels: position of the subunit in the center or on the border of a transplant) as main source of variation (Figure 2). Random effects for transplant were included in the models. Richness, total abundance (log(density+1)), Shannon diversity index and Evenness index, calculated for taxa and functional groups, were

21 the dependant variables. It reveals no significant effects of position on taxonomic and functional characteristics (Table 2 and 4). Subsequent analyzes aimed to test whether the presence of a central subunit (without a border along the perimeter line) influences the community throughout the transplant. We built similar models which included the fixed variable central (two levels: with or without a central subunit in the transplant; Figure 2) instead of position. The same fixed and random structures described were used in General Linear Mixed Models (GLMMs) with a binomial negative distribution for analysis on data of abundance (density) for each functional group (sessile, substrate and swimmer) as dependant variables.

Two permutational multivariate analyses of variance (9999 permutations) (Anderson 2001) were performed to test the effect of area (fixed with 3 levels), perimeter (fixed with 3 levels) and their interaction on 1) assemblies of taxa and on 2) assemblies of functional groups, which were fourth root transformed. We added transplant as random effect. We used the Gower Index (S15) to include double-zeros (Legendre and Legendre, 2012). This index increases the similarity value between two transplants that include only a few taxa. Littorina sp. was not included in the data set. The sum of the fixed effects for mixed terms was zero and the sums of squares was of type III. It reveals no significant effect (Appendix 2, Table 9).

Ecosystem functions We tested the effect of area (fixed with 2 levels), perimeter (fixed with 2 levels) and their interaction on oxygen uptake using analysis of covariance (ANCOVA) controlling for the number of live mussels (continuous) at the incubation time.

Mussel mortality in transplants during the colonization period introduced variability in the density of live mussels, with potential effect on ecosystem processes. This covariable presented a significant relationship with the response variable at one level of the treatment perimeter (P= 0.081, R2=0.422 for group 160 cm and P= 0.148 R2=0.315 for group 200 cm) and no relationship for treatments of area (P= 0.329, R2=0.158 for group 800 cm2 and P= 0.268 R2=0.199 for group 1200 cm2). We fixed the level of significance at 0.1 to make sure to consider the potential effect of mortality. Homogeneity of regression slopes was tested and showed that treatment regression lines have similar slope.

When there is no significant linear relationship between the response variable and covariable for all treatments, then the analysis of covariance offers no improvement over the analysis of variance in detecting differences between the group means. Since there was no significant relationship between the number of live mussels and ammonium release for treatments of perimeter (P= 0.748, R2=0.087 for group 160 cm and P= 0.803 R2=0.011 for group 200 cm) and area (P= 0.740, R2=0.020 for group 800 cm2 and P= 0.288 R2=0.185 for group 1200 cm2) analysis of variance were conducted instead of ANCOVAs to test the effect of area (fixed with 2 levels) and perimeter (fixed with two levels) of mussel transplants on rate of ammonium release, and the effect of perimeter-area ratio (fixed with 4 levels) on the same dependant variable. Assumption of normality was assessed by visual inspections of plots of residuals against fitted values. Bartlett's test was used to check homogeneity of variances (Bartlett, 1937). The presence of influential data points was assessed using Cook's distance (Cook, 1977). When an effect of the interaction between area and perimeter was significant, we tested the hypothesis of an effect of the fixed variable P/A ratio (4 levels) with the same method.

Results

Community of associated taxa

Community with Littorina sp. The interaction between perimeter and area had a significant effect on taxonomic richness (Appendix 1, Table 8), when the full community of macro invertebrates was considered, including Littorina sp. The gastropoda Littorina sp. constituted 78% of the total density of macro-invertebrates associated to mussels transplants and was also abundant on the rocky substratum surrounding experimental mussel transplants (personal observation). There was no effect of the area and perimeter on Littorina sp. density (Appendix 1, Table 8).

Community without Littorina sp. Analysis conducted on the community data without Littorina sp. similarly revealed an effect of the perimeter-area interaction on taxonomic richness because Littorina sp.

23 was present in all transplants and did not affect the mean number of taxa. However, the community with Littorina sp. excluded also revealed an effect of the perimeter- area interaction on Shannon diversity index and Equitability index (Table 1). When the perimeter was small (160cm), area had a negative effect on taxonomic diversity (richness, Shannon index and Equitability index; Figure 4). Transplants with a small area also had a lower diversity (richness, Shannon and Equitability indexes) at intermediate (180 cm) compared to small (160 cm) perimeter (Figure 4). P/A ratio did not explain a significant part of the variation, for any of our dependent variables (Table 1). However, we observed that the presence of a central subunit of transplant (subunit) without a border along the perimeter line was revealed as a significant factor influencing the community characteristics (Table 2, Figure 5). Though, linear mixed models incorporating the fixed variable position revealed there was no difference in community characteristics between subunits with different positions in the transplant (Table 2). The presence of a central subunit negatively impacts the taxonomic diversity (richness, Shannon diversity index and Equitability index) through the whole transplant (Table 2, Figure 5).

Functional groups We found that diversity (Shannon and Equitability indexes) in locomotion groups was lower on transplants with a large area (Table 3, Figure 6). We also observed a significant effect of the perimeter-area interaction on functional groups. There was a negative effect of area on the richness of functional groups (modes of locomotion) in small perimeter transplants (Table 3, Figure 7). Once again, the effect of the area- perimeter interaction was not explained by an influence of the P/A ratio (Table 3). This was rather linked to a significant negative effect of the presence of a central subunit in the transplants (Table 4, Figure 8). Sessile taxa and those with a reduced mobility (the functional group named 'substrate') were less abundant in transplants with a central subunit (Table 5, Figure 9).

Ecosystem functions We found significant effects of area and perimeter on ecosystem processes (Table 6 and 7). The oxygen uptake by the community (mussels and associated taxa) was highest in large perimeter transplants (Figure 10). It was also in large perimeter transplants that we found a positive effect of area on ammonium release (Figure

11). This interaction between area and perimeter was not explained by the P/A ratio (Table 7).

Discussion

Several studies have examined the effect of mussel bed size on diversity of organisms associated to it (Tsuchiya and Nishihira, 1985; Peake and Quinn, 1993; Lintas and Seed, 1994; Svane and Setyobudiandi, 1996; Norling and Kautsky, 2008; Koivisto and Westerbom, 2012). Effects of patch size may have been overestimated because it can be confounded with effects of other patch shape attributes. Dissociating effects of shape parameters and their interactions would allow identifying shape parameters influencing the ecosystem and would give direction to research investigating the mechanisms involved. Airoldi (2003) investigated the relative importance of area, perimeter and distance of the central zone from the edge in rocky intertidal algae mosaic. The 'distance from the edge' (corresponding to the distance between an estimated core area and the perimeter line) is a linear measure of the edge effect, which depends on the interaction between perimeter and area. Airoldi's study revealed taxa specific effects of the different shape metrics on algae colonisation and that shape is an important property of patches in a rocky intertidal system. Still, the shape effects could be even more complex and imply interactions between shape parameters. The interaction between area and perimeter has not been directly tested. Other shape parameters than the distance from the edge, which depend on the interaction between area and perimeter, could affect the community structure.

We have experimentally controlled the geometry of mussel beds and used a design that allowed distinguishing the effect of area, perimeter and their interaction on the ecosystem. Together, our study reveal 1) the non-additive effect of area and perimeter on community of associated taxa, 2) the important scale-dependence in the effect of parameters of habitat geometry, 3) the possibility to decouple ecosystem functions and community responses and 4) it draw main lines of some mechanisms that were likely involved in area and perimeter effects.

Community of associated taxa Our results revealed that both area and perimeter have influenced the structure of the macro invertebrates community associated with experimental mussels transplants on the intertidal rocky coast and that their effects were not additive. They were highly interrelated. Ecosystem response depended on the interaction of area and perimeter at many levels. Sousa (1984, 1985) suggested that habitat geometry effect in marine system might be attributed to variations in the ratio between perimeter and area. The P/A ratio is an estimator of the edge effect, which could explain the influence of the area-perimeter interaction on community structure. A habitat of large area with a small perimeter, and consequently a small P/A ratio, is expected to present a higher diversity of environmental conditions that increase taxa richness (Helzer and Jelinski 1999; Arponen and Bostrom 2012). In contrast, we found that the percentage of edge, estimated with the P/A ratio, did not significantly explain variability in community structure between experimental mussel transplants. The lowest diversity of invertebrates was associated to transplants with the smaller P/A ratio (160cm/1200cm2) and transplants with a more complex shape and thus a relatively high P/A ratio (180cm/ 800cm2). These results suggest that mechanisms linked to the P/A ratio could operate on patches with small perimeter, but other mechanisms driven by the explicit interaction between area and perimeter also affect taxa distribution. This result supports the idea that to assess patch shape influence on associated community, it is essential to consider the effects of the interaction between area and perimeter and it cannot be based on a single estimator of edge effects such the P/A ratio.

The spatial scale We found no difference in diversity and density of taxa and functional groups between subunits with different levels of exposition to the perimeter within transplants (Tables 2, 4 and 5). In other words there was no core area effect detected from the community characteristics, even if we found a significant interaction between area and perimeter. It has been suggested that the patch size may control edge effect (Jelbart et al. 2006). An edge effect on fish richness in large seagrass habitats was observed by Jelbart, Ross and Connolly (2006), but when the patches were small (< 6500m2), it was rather an effect of the area. They found higher taxa richness in small patches. This is congruent with the absence of P/A ratio effect

26 and the negative effect of area we observed in our experimental mussel transplants (< 0.12 m2). The absence of relationship between P/A ratio and every components of the ecosystem studied suggests that habitat patches were perhaps too small to have a central section of the area not influenced by the external environment (core area). It is also possible that the experimental mussel transplants had not enough layers to trap a good amount of sediments, faeces and pseudofaeces in their center (Tsuchiya 1980), and thus did not provide singular conditions in the core area. It is likely that there is a general tendency over an interaction between patch size and edge effect. The threshold habitat size from which P/A ratio effect could be detected might depend on the taxa and the type of habitat. P/A ratio and thus edge and core area effect might not be a universal, general explanation for shape effect at any spatial scale. As long as its effect is not demonstrated at different spatial scales, we cannot rely on a single parameter to estimate shape effect.

Ecosystem functions The influence of habitat geometry on ecosystem functions seems to have received little attention compared to its influence on communities. Accounting for biotic and abiotic components allows a better understanding of the overall functioning of the ecosystem (Loreau et al. 2002; Daufresne and Hedin 2005). We tested the effect of area, perimeter and their interaction, at the mussel transplant scale, on oxygen fluxes and ammonium release per unit of area. It revealed a significant effect of the interaction area-perimeter on ammonium release (influencing nitrogen cycling), and an influence of the perimeter on oxygen uptake. Importantly, the characteristics of mussel transplants shape which significantly influenced community characteristics were not the same as those that influenced oxygen and nitrogen fluxes. This suggests that linking communities and ecosystem functions in natural landscapes depends on resolving the differential responses of community structure and ecosystem fluxes to perimeter and area. Our results also strengthen the idea that, the total area and the length of the border are both of interest for issues concerning community dynamics and ecosystem processes. Our study provides evidence that the habitat characteristics at the patch-scale, even for patches of small size (800, 1000 and 1200cm2), are important determinants of the structure and functions of natural communities within an open marine system. Recent studies using experimental mussel transplants in intertidal system also provided evidence for the

27 small-scale dynamics in meta-ecosystems (Brazeau 2009; Largaespada et al. 2012). In addition to connectivity, patch shape points out as a local property influencing community structure and ecosystem dynamics.

The effects of habitat area and perimeter on mussel bed ecosystem

Area effect on community characteristics The species-area relationship, developed in terrestrial environment, predicted an increase in the number of taxa with the area (MacArthur and Wilson 1967; Root 1973; Gilpin and Diamond 1976). In opposition, we found a negative influence of mussel beds area. In terrestrial system such as in marine, patch size effect is highly variable and differences in taxa life history, functional traits and habitat can explain the divergence in response (e.g. Eggleston et al. 1998, 1999; Johnson and Heck, 2006; Bostrom et al. 2011). Bender et al. (1998) found a negligible influence of area on generalists, a positive one on interior taxa and a negative area effect on edge taxa. Accordingly, there was a lower functional diversity (Shannon and Equitability index) in large transplants, indicating that functional groups are not all influence the same way by the habitat shape. This prevents the possibility to establish general trend on community response and stress to take into account the characteristics of the taxa composing the community in order to predict response to habitat shape. We observed that mussel transplants were mostly inhabit by a diversity of taxa with a low mobility. They can face a restriction in the access to resources. Highly mobile taxa have the advantage of reaching more easily complementary resources outside of the patch. Sessile taxa and taxa with a low mobility found in mussel patches likely were edge taxa and could have been disadvantaged on large patches. We observed no significant general effect of area and perimeter on abundance of each functional groups (Appendix 3, Table 10), but variability in Evenness index for functional groups indicates that they were not equally represented between patches of different area. Diversity in functional groups was also lower on large patches. This suggests that some mechanisms linked to patch shape might have influenced functional groups but their effects are not captured by the area, perimeter and P/A ratio metrics.

Effect of the interaction between area and perimeter on the community and on other ecosystem properties

The effect of area on functional richness and taxonomic community characteristics depended on the patch perimeter length (Tables 1 and 3). This indicates the importance to consider the effect of the interaction between shape parameters since one can control the effect of another. Ries et al. 2004 have reviewed mechanisms of edge effects in forest systems. Perimeter length could affect the strength of these mechanisms, namely access to resources and fluxes of nutrients, organic matter and organisms. A short perimeter might have limited the access to resources for sessile and low-mobility taxa, in habitats of large area where the access to the border is less direct. This is supported by the positive effect of perimeter on oxygen uptake per unit of area we found. Mussels on transplants with a long perimeter may have benefited from enhanced conditions, such as strong flows of seston. This could explain their greater metabolic activity and oxygen consumption during the incubation. The active feeding of a large amount of mussels (1200 cm2 of area), over a limited flow of seston (160 cm of perimeter), might have reduced the amount of food available for every organism and increased competition with other filter feeders. These are mechanisms associated to a P/A ratio effect but this shape parameter failed to explain variability in community characteristics. It might be due to its combined effect with mechanisms not related to the P/A ratio which depend on the combination of area and perimeter sizes. On small transplants, richness in taxa and functional groups was lower when the transplants had an intermediate perimeter (Figures 4 and 7). The complexity of the perimeter line (the geometric shape) or other indexes of the shape could explain this observation (e.g. McGarigal and Marks, 1995; Botequilha Leitao and Ahern, 2002; Ries et al. 2004; Fletcher, 2005; Nams, 2014). The irregularity of shape could have nullified the effect of the P/A ratio (e.g. by hydrodynamics). We observed that the presence of a central subunit was strongly related to variability in community characteristics between transplants (Tables 2 and 4). The abundance of sessile taxa and taxa with a low mobility was lower in the presence of a central subunit within the transplant (Table 5, Figure 9). This indicates that the edge effect can be addressed differently than with the P/A ratio or the distance from the edge. The generalists with a high mobility have been shown to be less sensitive to patch area (Franzen et al. 2012). Our results suggested that they

29 might be also less sensitive to other patch characteristics, as they were not significantly influenced by the presence of a central subunit (Table 5). The rationale behind the effect of a shape parameter such as this one is to find. It would be interesting to study the mechanisms implies and specify the nature of the parameter that can best explain the influence of the area-perimeter interaction on the associated community at this spatial scale. The observed positive effect of area on ammonium release when the perimeter was large suggests that the influence of the interaction between area and perimeter on the ecosystem is even more complex as it is influencing many compartment of the ecosystem on different ways. This highlights the need to acquire more knowledge on the mechanisms leading to an effect of habitat shape on taxa composition, community characteristics, abiotic conditions and ecosystem functions.

Conclusion

Our results demonstrate that there are effects of the habitat area and perimeter on the ecosystem and that they are not additive. The interaction between area and perimeter was found to be a strong determinant of the community characteristics and could influence ecosystem functions, even on patches of small size. It highlights the importance to integrate the interactive effect of various metrics in the study of ecosystem response to habitat shape. The present study suggests that the different components of the ecosystem, community structure and ecosystem fluxes, are influenced differently by the parameters of habitat shape. Area, perimeter and their interaction all have an effect on the whole ecosystem that can influence the links between community structure and ecosystem functions. Further studies on those links could resolve the mechanisms underlying the effect of habitat geometry reported here on an intertidal mussel bed ecosystem. In this study conducted over small spatial scales, the perimeter-area ratio did not explain the variability in community characteristics, which challenges the relevance of single metrics to predict the influence of habitat shape across scales. Understanding the non-additive impacts of multiple patch attributes on communities and ecosystems is relevant not only for the conservation and management of fragmented habitats, but also for models of community distribution and ecosystem dynamics over natural landscapes.

Tables

TABLE 1. P-values extracted from linear mixed effects models testing the effects of (A) area and perimeter of mussel transplants and (B) the effect of the perimeter-area ratio, on Shannon diversity index, richness in taxa (log+1 transformed), abundance (log+1 transformed) and Evenness index. Transplant was included as random variable. Littorina sp. is not included to the data set.

Source of variation Num df Den df Mean square F P Shannon diversity index A Area 2 36 0.532 3.491 0.041 * Perimeter 2 36 0.193 1.269 0.293 Area*Perimeter 4 36 0.426 2.796 0.041 * B Ratio P/A 1 37 0.088 0.586 0.449

Richness in taxa (log(richness+1)) A Area 2 37 0.081 1.269 0.293 Perimeter 2 37 0.087 1.361 0.269 Area*Perimeter 4 36 0.196 3.050 0.029 * B Ratio P/A 1 37 0.008 0.125 0.725

Abundance in taxa (log(density+1)) Area 2 37 0.127 0.299 0.744 Perimeter 2 37 0.598 1.410 0.257 Area*Perimeter 4 37 0.958 2.257 0.081

Evenness index A Area 2 36 0.193 3.491 0.041 * Perimeter 2 36 0.070 1.269 0.293 Area*Perimeter 4 36 0.154 2.796 0.041 * B Ratio P/A 1 37 0.032 0.586 0.449

Notes: Calculation using alpha = 0.05 Signif. codes: 0 '***' 0,001 '**' 0,01 '*' 0,05 '.' 0.1

TABLE 2. P-values extracted from linear mixed effects models models with (A) the presence of a central subunit (2 levels) or (B) the position of the subunits (2 levels) in mussel transplants as fixed variable. Transplant (36 levels) is included in models as random variable.The dependant variables are the Shannon diversity index, the richness in taxa (log+1 transformed), the abundance (log+1 transformed) and the Evenness index. Littorina sp. is not included to the data set.

Source of variation Num df Den df Mean square F P

Shannon diversity index A

Presence of a central subunit 1 34 1.981 13.071 0.001 *** B Position of subunits 1 120 0.004 0.027 0.871

Richness in taxa (log(richness+1)) A Presence of a central subunit 1 36 0.858 7.007 0.012 * B Position of subunits 1 115 0.006 0.050 0.824

Abundance in taxa (log(density+1)) A Presence of a central subunit 1 36 0.459 0.082 0.305 B Position of subunits 1 110 0.034 0.080 0.777 Evenness index A Presence of a central subunit 1 34 0.719 13.071 0.001 *** B Position of subunits 1 120 0.001 0.027 0.871

Notes: Calculation using alpha = 0.05 Signif. codes: 0 '***' 0,001 '**' 0,01 '*' 0,05 '.' 0.1

TABLE 3. P-values extracted from linear mixed effects models testing (A) the effects of area and perimeter of mussel patches and (B) the effect of the perimeter-area ratio, on Shannon diversity index, richness in functional groups, abundance (density log+1 transformed) and equitability index. The random effects of transplant were included in models. Littorina sp. is not in the data set.

Source of variation Num df Den df Mean square F P Shannon diversity index A Area 2 37 0.546 4.986 0.012 * Perimeter 2 38 0.084 0.770 0.470 Area*Perimeter 4 37 0.276 2.521 0.057 . B Ratio P/A 1 38 0.192 1.759 0.193

Richness in functional groups A Area 2 38 1.023 1.950 0.157 Perimeter 2 38 0.446 0.850 0.435 Area*Perimeter 4 37 1.400 2.664 0.047 * B Ratio P/A 1 38 0.222 0.424 0.519

Abundance in functional groups (log(density+1)) A Area 2 37 0.127 0.299 0.744 Perimeter 2 37 0.598 1.409 0.257 Area*Perimeter 4 37 0.958 2.257 0.081 .

Evenness index A Area 2 37 0.546 4.986 0.012 * Perimeter 2 38 0.084 0.770 0.470 Area*Perimeter 4 37 0.276 2.521 0.057 . B Ratio P/A 1 38 0.159 1.759 0.193

Notes: Calculation using alpha = 0.05, Signif. codes: 0 '***' 0,001 '**' 0,01 '*' 0,05 '.' 0.1

TABLE 4. P-values extracted from linear mixed effects models with (A) the presence of a central subunit (2 levels) or (B) the position of the subunits (2 levels) in mussel transplants as fixed variable Transplant (36 levels) is included in models as random variable. The dependant variables are the richness in functional groups, the abundance (log+1 transformed), the Evenness index and Shannon diversity index calculated for functional groups. Littorina sp. is not included to the data set.

Source of variation Num df Den df Mean square F P

Shannon diversity index for functional groups A

Presence of a central subunit 1 36 1.718 15.719 0.000 *** B Position of subunits 1 120 0.060 0.554 0.458

Richness in functional groups A Presence of a central subunit 1 36 4.687 8.930 0.005 ** B Position of subunits 1 116 0.117 0.224 0.637

Abundance in functional groups (log(density+1)) A Presence of a central subunit 1 36 0.459 1.082 0.305 B Position of subunits 1 110 0.034 0.080 0.777

Evenness index for functional groups A Presence of a central subunit 1 36 1.424 15.719 0.000 *** B Position of subunits 1 120 0.049 0.554 0.458

Notes: Calculation using alpha = 0.05 Signif. codes: 0 '***' 0,001 '**' 0,01 '*' 0,05 '.' 0.1

TABLE 5. P-values extracted from Generalized linear mixed effects models with (A) the presence of a central subunit (2 levels) or (B) the position of the subunits (2 levels) in mussel transplants as fixed variable. Transplant (36 levels) is included in models as random variable. The dependant variables are the total abundance (density) for the functional groups 'sessile', 'substrate' and 'swimmer'. Littorina sp. is not included to the data set.

Source of variation Num df Den df F P

Abundance for the group 'swimmer' A

Presence of a central subunit 1 32 0.23 0.637 B Position of subunits 1 106 0.09 0.761

Abundance for the group 'substrate' A Presence of a central subunit 1 42 5.39 0.025 * B Position of subunits 1 132 0.89 0.347

Abundance for the group 'sessile' A Presence of a central subunit 1 32 4.44 0.043 * B Position of subunits 1 132 0.24 0.628

Notes: Calculation using alpha = 0.05 Signif. codes: 0 '***' 0,001 '**' 0,01 '*' 0,05 '.' 0.1

TABLE 6. Results from ANCOVA testing the effect of area and perimeter of mussel beds on the rate of oxygen uptake, with the number of mussels alive as covariable.

Source of variation df Mean square F P

Number of mussels alive 1 1076.7 4.064 0.069 . Area 1 3.8 0.014 0.907 Perimeter 1 1457.5 5.502 0.038 * Area*Perimeter 1 73.2 0.276 0.610 Residuals 11 264.9

Notes: Calculation using alpha = 0.05 Signif. codes: 0 '***' 0,001 '**' 0,01 '*' 0,05 '.' 0.1

TABLE 7. Results from ANOVAs testing the effect of (A) area and perimeter of mussel beds on rate of ammonium release and (B) the effect of perimeter-area ratio.

Source of variation df Mean square F P A Area 1 343.1 4.325 0.060 . Perimeter 1 29.2 0.368 0.555 Area*Perimeter 1 414.7 5.228 0.041 * Residuals 12 79.3 B

Ratio P/A 1 239.0 2.231 0.157 Residuals 14 107.1

Notes: Calculation using alpha = 0.05 Signif. codes: 0 '***' 0,001 '**' 0,01 '*' 0,05 '.' 0.1

Figures a 4 A

3 AB

B 2

Richness in taxa (Mean ± SE) 0 160 180 200 b 1,2 A 1,0 AB 0,8

0,6 B Area 800 cm2 2 0,4 1000 cm 1200 cm2

for taxa (Mean ± SE) 0,2

Shannon index diversity 0,0 160 180 200 c 0,7 A

0,6 AB 0,5

0,4 B 0,3

(Mean ± SE) 0,2

Evenness index for taxa 0,1

0,0 160 180 200

Perimeter (cm2)

Figure 4. Effects of transplant area and perimeter on (a) mean associated taxa richness, (b) mean Shannon Diversity index and (c) mean Evenness index. Black, light gray and dark gray bars illustrate area values of 800 cm2, 1000 cm2 and 1200 cm2 respectively. Least-squares means with Tukey method for adjustment on p-values for family of three tests, were calculated for factors area and perimeter of linear mixed effects models (P< 0.05). Horizontal black lines with different levels, above the bars, illustrate a significant difference within groups (P< 0.05). An unbroken line mean that there is no significant difference with the adjacent bar. Different uppercase letters indicate significant differences between transplants of different perimeters for a same value of area.

a 3,5

3,0 *

2,5

2,0

1,5

1,0

0,5

Richness in taxa (Mean ± SE) 0,0 With Without b 1,0

0,8 * 0,6

0,4

0,2

for taxa (Mean ± SE)

Shannon diversity index 0,0 With Without

c 0,6

0,5 * 0,4

0,3

0,2

(Mean ± SE) 0,1

Evenness index for taxa 0,0 With Without Presence of a central subunit (Mean ± SE)

Figure 5. Effects of the presence in the transplants of a central subunit on (a) mean associated taxa richness, (b) mean Shannon Diversity index and (c) mean Evenness index. A star above the bars illustrate a significant difference between groups (P< 0.05).

38 a b

0,7 0,7 A A 0,6 A 0,6 A

0,5 0,5 B 0,4 0,4 B

0,3 0,3

(Mean ± SE) 0,2 0,2

groups (Mean ± SE) 0,1 0,1

Evenness Index for functional 0,0 0,0 Shannon indiversity functional groups 800 1000 1200 800 1000 1200

Area (cm2) Area (cm2)

Figure 6. Effects of the transplant area on (a) mean Shannon Diversity index and (b) Evenness index for functional groups. Different uppercase letters indicate significant differences between groups.

3,0 A

2,5 AB

2,0 B Area 800 cm2 1,5 1000 cm2 1200 cm2

(Mean ± SE) 1,0

0,5

Richness in functional groups 0,0 160 180 200 Perimeter (cm)

Figure 7. Effects of transplant area and perimeter on mean richness in functional groups. Black, light gray and dark gray bars illustrate area values of 800 cm2, 1000 cm2 and 1200 cm2 respectively. Least-squares means with Tukey method for adjustment on p-values for family of three tests, were calculated for factors area and perimeter of linear mixed effects models (P< 0.05). Horizontal black lines with different levels, above the bars, illustrate a significant difference within groups (P< 0.05). An unbroken line mean that there is no significant difference with the adjacent bars. Different uppercase letters indicate significant differences between transplants of different perimeters for a same value of area.

a 2,5 * 2,0

1,5

1,0

0,5

groups (Mean ± SE)

Richness in functional 0,0 b With Without 0,7 0,6 * 0,5 0,4 0,3 0,2 0,1

Shannon diversity Index for 0,0 functional groups (Mean ± SE) With Without

c 0,7 0,6 * 0,5 0,4 0,3 0,2

groups (Mean ± SE) 0,1 0,0 Evenness Index for functional With Without

Presence of a central subunit

Figure 8. Effects of the presence in the transplants of a central subunit on (a) mean richness in functional groups, (b) mean Shannon Diversity index for functional groups and (c) mean Evenness index for functional groups. A star above the bars illustrate a significant difference between groups (P< 0.05).

40 a b 3,0 1,2 * * 2,5 1,0

2,0 0,8

1,5 0,6

1,0 (Mean ± SE) 0,4

(Mean ± SE)

0,5 0,2

Abundance of group 'sessile' 0,0 Abundance of group 'substrate' 0,0 With Without With Without

Presence of a central subunit Presence of a central subunit

Figure 9. Effects of the presence in transplants of a central subunit on (a) mean abundance for the functional group 'sessile' and (b) mean abundance for the functional group 'substrate'. A star above the bars illustrate a significant difference between groups (P< 0.05).

Perimeter (cm)

160 200 0

-120

-130

-140

-150

-160

-170

-180 *

Rate uptakeof oxygen (Mean ± SE) -190

Figure 10. Effects of the transplant perimeter on mean rate of oxygen uptake. A star between bars illustrate a significant difference between groups (P< 0.05).

50 B 40

AB AB 30

A 20

(Mean ± SE)

Rate of ammonium release

0 160 200

Perimeter (cm)

Figure 11. Effects of transplant area and perimeter on mean rate of ammonium release. Black and light gray bars illustrate area values of 800 cm2 and 1200 cm2 respectively. Different uppercase letters indicate significant differences between groups (P< 0.05).

Chapter 3: General conclusion

Summary and future perspectives Our results highlight the importance to consider the interactive effect of various shape parameters on both the community and the ecosystem functions in the study of the impacts of habitat configuration on the ecosystems. Developing knowledge on the mechanisms involved in shape effect would be necessary in order to choose wittingly the explicative variables and dependant variables to include in models. These can be used by managers and decision makers as baseline for management and conservation decisions in marine environment.

Shape effect in marine systems In numerous terrestrial habitats, the patch shape has been shown to be a main parameter influencing the abundance and distribution of species (Gutzwiller and Anderson, 1992; Farina, 1998; Fletcher et al. 2007). Studies in marine environments exploring the influence of the shape metrics of habitat patches on the ecosystem has been limited by the complexity of achieving non-confounded experiments. Airoldi (2003) showed that area, perimeter and the distance from the edge (as a measure of edge effect) are patch shape attributes influencing algae colonization on rocky substrate, even at small spatial scale (150, 280 and 320cm2). Our experiment pushed further by testing the presence of an interaction between patch area and perimeter. Our results confirm the influence of patch shape metrics in marine coastal system at a slightly larger scale (800, 1000 and 1200 cm2). Our experimental design also shed light on the importance of the synergistic interaction between area and perimeter, affecting ecosystem functions and community characteristics. Our study highlights the need to consider the effect of the interactions between various shape metrics which have been often confounded in seascape studies. Area effect has been usually confounded not only with perimeter and shape, but also with the effect of the habitat heterogeneity (e.g. Peake and Quinn, 1993). The size of mussels and the presence of dead mussels have been shown to influence the community structure (Matias et al. 2010; Fausto et al. 2011). Experimental mussel beds allow a simplification of the ecosystem in order to understand the relationships between its components.

We showed that the significant effect of the interaction between area and perimeter on community diversity and ammonium release was not explained by an index such as the P/A ratio. The P/A ratio is widely used as an index of shape complexity or of edge effect both in terrestrial and marine studies. We suggest being cautious with its use because it could be neglecting important effects of the interaction between area and perimeter. Our results suggest that the interaction might be better represented by another index than the P/A ratio, which remains to be identified more precisely. Presence of a central subunit was shared as a common characteristic on mussel patches with lower diversity. Finding the mechanisms underlying shape effects could allow the identification of a reliable index, possibly valid throughout scales of patch sizes.

There are important differences between marine and terrestrial system such as the influence of hydrodynamics, tides and presence of a high connectivity. Different shape parameters could influence the habitat depending on the system. Moreover, the sense of their effect can also diverge. We observed that response to shape parameters in mussel bed diverge from terrestrial studies (MacArthur and Wilson, 1967; Helzer and Jelinski, 1999), although they were variable. As in terrestrial studies, the variability in responses was related to taxa traits (Forman 1995; Bender et al. 1998; Collinge and Palmer 2002; Fletcher and Koford 2003; Ries and Sisk 2004; Ewers and Didham, 2006). The variation in taxa responses prevents the identification of a general trend. The identification of key traits of taxa related to shape effects would help to select appropriate response variables. The choice of response variables and factors to include in a model and the spatial scale of the study both influence the outcomes. Landscape ecology concepts and techniques are increasingly applied to marine environments in order to gain a better understanding of the consequences of habitat configuration and the mechanisms involved. However, the number of papers applying these concepts to marine systems remains relatively low (Garrabou et al., 1998; Bostrom et al. 2011).

Community composition It is likely that the influences of intrinsic properties of the habitats (e.g. differences in shape and colonization rates) are not only perceivable through differences in diversity indices, but also through composition dissimilarities between communities.

A next step would be to investigate shape effect on community composition and to evaluate if area and perimeter influence populations abundance. The abundance of the various taxa composing the community associated to mussel beds of different shapes could be represented with pie charts. Information on the community composition can reveal mechanisms of patch shape effect on ecosystem functions. Some are known to be associated with the presence of specific taxa. For example, even if the functional group 'swimmer' regrouping highly mobile taxa was not significantly more abundant, knowing that Hediste diversicolor and/or Alitta virens were only present on patches of large area combined with a long perimeter (present in greater abundance) would provide information on the mechanisms that might have led to a high ammonium release. Hediste diversicolor is known to affect ammonium availability by the production of excretions and mucus, by bioturbation and by influencing the oxidizers community structure affecting rates of microbial mineralization (Papaspyrou et al., 2007; Gilbertson et al., 2012). Microbial mineralization produces ammonium from organic matter, which increases ammonium availability and influences nitrogen cycle. Allita virens influences O2 flux at the water-sediment interface, bacterial abundances and biogeochemical fluxes (Nielsen et al. 1995; Piot, 2014). Their presence could have led to the observed positive effect of area on ammonium release when the perimeter is large. Piot (2014) have reported no effect of taxa richness on ecosystem properties, but an effect of taxa composition. Information on taxa composition could also help to identify unique taxa. These sensitive taxa which respond strongly to patch metrics could be of greater conservation concern in systems subject to fragmentation.

Interrelated shape effects and their mechanisms We found that ecosystem functions and community characteristics were not influenced by the same combinations of area and perimeter size. Investigations on the links between community and ecosystem functions and on the mechanisms leading to an effect of patch shape on both ecosystem components will be necessary to understand mussel bed shape effect on ecosystem dynamics. Further experimental studies should investigate separately the links between 1) shape parameters of mussel patch and taxa colonization (in benthocosm), 2) shape parameters, hydrodynamics and biodeposition (flume experiment), and 3) shape parameters and other abiotic conditions throughout the mussel bed. An experiment

45 testing the influence of patch shape on current speed and flows regime around and over a mussel bed would provide significant information on its potential effect on mussel's filtration activity and ecosystem dynamics. The flow speed is known to influence filtration rate (Wildish and Kristmanson, 1979; Fréchette et al. 1989). Flume experiments have reported enhanced turbulence production on the borders of a patch and deceleration in flow speed in the center of the mussel patch due to mussel bed surface complexity (Fréchette et al., 1989; Butman et al., 1994; Green et al., 1998). These conditions can significantly influence rates of sedimentation, resuspension, food supply and regulate ecological conditions, taxa composition and abundance in mussel beds (Fréchette and Bourget, 1985 a,b; Fréchette et al., 1989; Ricciardi et al., 1997; Widdows et al., 1998). It would be meaningful to know if patch shape metrics can influence hydrodynamics and their effects on the ecosystem. Understanding the mechanisms implied could help to identify shape parameters and other environmental or biological parameters to include in models for predicting ecosystem response to seascape configuration.

Spatial scale We also suggested that the effect of shape metrics on the community might depend on the spatial scale. For example, area effect can depend on perimeter size while P/A ratio could be an important factor only over a threshold patch size. Therefore, the choice of the factors can depend on the scale of the studies. Efforts should be focused on finding a shape metric that encompasses all spatial scales from cm2 to km2. This metric should be a representative index of the interaction between area and perimeter. Another perspective would be to find the threshold points in spatial scales where the different shape metrics have a significant effect. These information should be consciously included in the model in order to obtain strong results at multiple scales (Banks-Leite et al., 2011, 2012). Incorporating various variables allow to account for differences in scales and habitat type. For example, to allow response to vary among scales, an interaction term can be added in the model. One of the great challenges with the use of these complex models is finding ways to compare results between studies. Ewers et al. (2010) proposed to integrate statistical models to geographic information systems in a way to generate a comparable estimate of the total impact of change in habitat configuration on different taxa.

Development of tools for seascape studies Knowledge developed in landscape studies are of interest for marine studies as they have started to investigate those possibilities and develop tools which could be slightly modified to serve seascape studies. For example, to measure patch metrics at large scale and with fine precision, high-resolution satellite imagery can be used (Mathieu et al. 2007; Ewers et al. 2010; Hepcan, 2013). It allows calculating simultaneously dozens of landscape metrics facilitating the use of more complex models. This method might be applicable to coastal seascape at low tide, but the use of satellite is limited for imagery at important depths (Stumpf and Holderied, 2003). Thus, multibeam mapping systems set up on research vessels can be used to map benthic seascape. It can be combined with backscatter data allowing the classification of seafloor bottom characteristics and mapping of benthic ecosystems (Lundblad et al., 2006). This is a promising and valuable tool for seascape studies. A great number of seascape metrics could be quickly calculated using geographic information system (GIS) and they could be used in complex models to consider scale effects and confounding factors (Banks-Leite et al., 2012).

Implications Knowledge on the ecosystem responses to seascape configuration parameters has important implications for the management of coastal resources. This study gives insights to the importance of considering the distinct effect of main geometry parameters and their interaction to understand ecosystem responses. The shape characteristics of habitat patches in fragmented seascape could significantly alter biodiversity and processes in shallow coastal ecosystems. We found an influence of habitat shape on ammonium release which could influence nitrogen cycle. The perimeter was found to have positively influenced oxygen consumption by the community. This presumes an increase in mussel filtration rate or microbial activity which can significantly influence the importance of pelago-benthic transfers of organic matter and nutrients turnover (Pfister, 2007; Richard et al., 2007; Stenton- Dozey et al., 2011; McKindsey et al. 2011; Robert et al. 2013).

Understanding habitat configuration effects on hydrodynamics, abiotic habitat conditions, ecosystem functions and taxa distribution is essential to ecology and conservation biology in marine systems. Knowledge could be incorporated in ecological theories and help to build statistical models. In the last decade, predictive models of species distribution have become a tool increasingly important in addressing various issues (Guisan and Thuiller, 2005). They can be used to evaluate potential management actions, help maximize biodiversity, and interpret the effects of disturbances such as habitat fragmentation. An understanding of ecosystem responses to seascape configuration, and the mechanisms implied, would help to improve model predictions and management of resources. At a larger extent, this knowledge would also be baseline for the design of marine protected area and decisions relative to their number, size and shape. In this perspective, georeferenced maps of habitat-species and habitat configuration assessments would be valuable tools to marine management (Lundblad et al. 2006). In a context of severe issues of habitat fragmentation and loss of biodiversity, understanding how habitat configuration influences ecosystems is critical in order to improve model predictions and management of resources. This project reveals scale-dependence in the effect of habitat shape metrics and underlines the importance of integrating the interactive effects of various metrics of landscape geometry to the study of the relationship between community dynamics and ecosystem functions in fragmented landscapes.

References

Airoldi, L. 2003. Effects of patch shape in intertidal algal mosaics: roles of area, perimeter and distance from edge. Marine Biology 143: 639-650

Anderson, M. 2001. Permutation tests for univariate or multivariate analysis of variance and regression. Journal candien des sciences halieutiques et aquatiques 58(3): 626-639

Antoniadou, C., E. Voultsiadou and C. Chintiroglou. 2011. Seasonal patterns of colonization and early succession on sublittoral rocky cliffs. Journal of Experimental Marine Biology and Ecology 403: 21-30

Archambault, P., and E. Bourget. 1996. Scales of coastal heterogeneity and benthic intertidal species richness, diversity and abundance. Mar. Ecol. Prog. Ser. 136: 111- 121

Arponen, H., and C. Bostrom. 2012. Responses of mobile epifauna to small-scale seagrass patchiness: is fragmentation important? Hydrobiologia 680: 1-10

Banks-Leite, C., R.M. Ewers, and J.P. Metzger. 2010. Edge effects as the principal cause of area effects on birds in fragmented secondary forest. Oikos119: 918–926

Banks-Leite, C., R.M. Ewers, V. Kapos, A.C. Martensen and J.P. Metzger. 2011. Comparing species and measures of landscape structure as indicators of conservation importance. J. Appl. Ecol. 48: 706–714

Banks-Leite, C., R.M. Ewers and J.P. Metzger. 2012. The confounded effects of habitat disturbance at the local, patch and landscape scale on understorey birds of the Atlantic Forest: implications for the development of landscape-based indicators. Ecological Indicators 31: 82-88

Banta, G. T., A. E. Giblin, J. E. Hobbie and J. Tucker. 1995. Benthic respiration and nitrogen release in Buzzards Bay, Massachusetts. Journal of Marine Research 53: 107–135

Barbaro, L., E.G. Brockerhoff, B. Giffard and I.V. Halder. 2012. Edge and area effects on avian assemblages and insectivory in fragmented native forests. Landscape Ecol 27: 1451-1463

Bartlett, M. S. 1937. Properties of sufficiency and statistical tests. Proceedings of the Royal Statistical Society 160 (901): 268–282

Baskent, E.Z., and G.A. Jordan. 1995. Characterizing spatial structure of forest landscapes. Can. J. For. Res. 25: 1830-1849

Bates, D., M. Maechler, B. Bolker and S. Walker. 2014. lme4: Linear mixed-effects models using Eigen and S4. R package version 1.1-6

Bender, D.J., T.A. Contretas and L. Fahrig. 1998. Habitat loss and population decline: a meta-analysis of the patch size effect. Ecology 79(2): 517-533

Bengtsson, J. 1998. Which species? What kind of diversity? Which ecosystem function? Some problems in studies of relations between biodiversity and ecosystem function. Applied Soil Ecology 10: 191–199

Biles, C.L., M. Solan, I. Isaksson, D.M. Paterson, C. Emes, D.G. Raffaelli. 2003. Flow modifies the effect of biodiversity on ecosystem functioning: an in situ study of estuarine sediments. Journal of Experimental Marine Biology and Ecology 285–286: 165–177

Bolam, S.G., T.F. Fernandes, M. Huxham. 2002. Diversity, biomass and ecosystem processes in the marine benthos. Ecological Monographs 72 (4): 599–615

Bologna, P.A.X., L.H. Kenneth and Jr. 2000. Impacts of seagrass habitat architecture on bivalve settlement. Estuaries 23(4):449-457

Borthagaray, A.I., A. Brazeiro and L. Gimenez. 2009. Connectivity and patch area in a coastal marine landscape: disentangling their influence on local species richness and composition. Austral Ecology 34: 641-652

Bostrom, C., E.L. Jackson and C.A. Simenstad. 2006. Seagrass landscapes and their effects on associated fauna: a review. Estuarine, Coastal and Shelf Science 68: 383-403

Bostrom, C., S.J. Pittman, C. Simenstad, R.T. Kneib. 2011. Seascape ecology of coastal biogenic habitats: advances, gaps, and challenges. Mar. Ecol. Prog. Ser. 427: 191-217

Botequilha Leitao, A. and J. Ahern. 2002. Applying landscape ecological concepts and metrics in sustainable landscape planning. Landscape and Urban Planning 59: 65-93

Boucher, G. and J. Clavier. 1990. Contribution of benthic biomass to overall metabolism in New Caledonia lagoon sediments. Mar. Ecol. Prog. Ser. 64: 271-280

Bowman, J., N. Cappuccino and L. Fahrig. 2002. Patch size and population density: effect of immigration behavior. Conservation Ecology 6(1): 9. [online] URL: http://www.consecol.org/vol6/iss1/art9

Brazeau, C. 2009. In healing the ills of fragmented populations: Meaningful level of connectivity that affects community composition within intertidal mussel beds. Dissertation. McGill University, Montreal, Canada

Butman, C.A., M. Fréchette, W.R. Geyer and V.R. Starczak. 1994. Flume experiments on food supply to the blue mussel Mytilus edulis L. as a function of Boundary-layer flow. Limnology and Oceanography 39(7): 1755-1768

Cadenasso M.L., S.T.A. Pickett, K.C. Weathers and C.G. Jones. 2003. A framework for a theory of ecological boundaries. BioScience 53:750–758

Canadian Hydrographic Service. 2013. Tides, currents, and water levels, Web: http://www.marees-tides.gc.ca

Cardinale, B.J., M.A. Palmer and S.L. Collins. 2002. Species diversity enhances ecosystem functioning through interspecific facilitation. Nature 415(6870): 426-429

Carpintero S. and J. Reyes-Lopez. 2014. Effect of park age, size, shape and isolation on ant assemblages in two cities of Southern Spain. Entomological Science 17:41-51

Cole, V.J. 2010. Alteration of the configuration of bioengineers affects associated taxa. Mar. Ecol. Prog. Ser. 416: 127-136

Collinge, S.K. and T.M. Palmer. 2002. The influences of patch shape and boundary contrast on insect response to fragmentation in California grasslands. Landscape Ecology 17:647-656

Commito, J.A., E.A. Celano, H.J. Celico, S.Como and C.P. Johnson. 2005. Mussels matter: Postlarval dispersal dynamics altered by a spatially complex ecosystem engineer. J. Exp. Mar. Biol. Ecol. 316:133-147

Connell, J.H. 1972. Community interactions on marine rocky intertidal shores. Annual Review of ecology and Systematics 3: 169-192

Connor, E.F. and E.D. McCoy. 1979. The statistics and biology of the species-area relationship. The American Naturalist 113(6): 791-833

Cook, R. D. 1977. Detection of Influential Observations in Linear Regression. Technometrics (American Statistical Association) 19 (1): 15–18

Cornwell, W.K., D.W. Schwilk and D.D. Ackerly. 2006. A trait-based test for habitat filtering: convex hull volume. Ecology 87: 1465–1471

Cowan, J. L. W., J. R. Pennock and W. R. Boynton. 1996. Seasonal and interannual patterns of sediment-water nutrient and oxygen fluxes in Mobile Bay, Alabama (USA): regulating factors and ecological significance. Mar. Ecol. Prog. Ser. 141: 229- 245

Crain, C.M. ad M.D. Bertness. 2006. Ecosystem engineering across environmental gradients: implications for conservation and management. Bioscience 56(3): 211-218

Cusson, M., and E. Bourget. 2005. Small-scale variations in mussel (Mytilus spp.) dynamics and local production. J. Sea. Res. 53: 255-268

Daufresne, T., L.O. Hedin. 2005. Plant coexistence depends on ecosystem nutrient cycles: Extension of the resource-ratio theory. Proceedings of the National Academy of Sciences of the United States of America 102(26): 9212-9217

Denny, M. 1995. Predicting physical disturbance: mechanistic approaches to the study of survivorship on wave-swept shores. Ecol. Mono. 65:371-418

Didham, R.K., and R.M. Ewers. 2012. Predicting the impacts of edges in fragmented habitats: Laurance and Yensen's core area model revisited. Biological Conservation 155: 104-110

Dolmer, P. 2000. Algal concentration profiles above mussel beds. J. Sea Res. 43: 113–119

Dunning, J. B., B. J. Danielson, and H. R. Pulliam. 1992. Ecological processes that affect populations in complex landscapes. Oikos 65:169–175

Eggleston, D.B., L.L. Etherington, W.E. Elis. 1998. Organism response to habitat patchiness: Species and habitat-dependent recruitment of decapod crustaceans. J. Exp. Mar. Biol. Ecol. 223: 111–132

Eggleston, D.B., W.E. Elis, L.L. Etherington, C.P. Dahlgren and M.H. Posey. 1999. Organism responses to habitat fragmentation and diversity: habitat colonization by estuarine macrofauna. J. Exp. Mar. Biol. Ecol. 236: 107-132

Ewers, R.M. 2004. The extent of forest fragmentation in New Zealand and its effects on arthropod biodiversity. Ph.D. Thesis, University of Canterbury, Christchurch Online : URL http://hdl.handle.net/10092/1519

Ewers, R.M. and R.K. Didham. 2006. Confounding factors in the detection of species responses to habitat fragmentation. Biol. Rev. 81: 117-142

Ewers, R.M., and R.K. Didham. 2007. Habitat fragmentation: panchreston or paradigm? Trends in Ecology and Evolution 22(10): 511

Ewers, R.M., S. Thorpe and R.K. Didham. 2007. Synergistic interactions between edge and area effects in a heavily fragmented landscape. Ecology 88(1): 96-106

Faaborg, J., M. Brittingham, T. Donovan, and J. Blake. 1993. Habitat fragmentation in the temperate zone: a perspective for managers. Pages 331-338 in D. Finch and P. S. Stangel, editors. Status and management of neotropical migratory birds. U.S. Forest Service General Technical Report RM- 229, Rocky Mountain Forest and Range Experiment Station, Fort Collins, Colorado, USA.

Fagan, W. E., R. S. Cantrell, and C. Cosner. 1999. How habitat edges change species interactions. American Naturalist 153:165-182

Fahrig, L. 2003. Effects of habitat fragmentation on biodiversity. Annu. Rev. Ecol. Syst. 34:487-515

Farina, A. 1998. Principles and methods in landscape ecology: towards a science of the landscape. Second edition. Chapman and Hall, London. 412p.

Fausto, N.F., F.J. Hidalgo, B.J. Lomovasky, E. Ramos, P. Gamero and O.O. Iribarne. 2011. Habitat structure is more important than nutrient supply in modifying mussel bed assemblage in an upwelling area of the Peruvian coast. Helgol. Mar. Res. 65: 187-196

Fletcher, R.J.Jr. 2005. Multiple edge effects and their implications in fragmented landscapes. Journal of Animal Ecology 74: 342-352

Fletcher, R.J.Jr. 2006. Emergent properties of conspecific attraction in fragmented landscapes. The American Naturalist 168(2): 207-219

Fletcher, R.J. 2009. Does attraction to conspecifics explain the patch-size effect? An experimental test. Oikos 118:1139-1147

Fletcher, R.J.Jr. and R.R. Koford. 2003. Spatial responses of Bobolinks (Dolichonyx oryzivorus) near different types of edges in northern Iowa. The Auk 120(3):799-810

Fletcher, R.J. Jr., L. Ries, J. Battin and A.D. Chalfoun. 2007. The role of habitat area and edge in fragmented landscapes: definitively distinct or inevitably intertwined? Can. J. Zool. 85: 1017-1030

Folkard, A.M. and J.C. Gascoigne. 2009. Hydrodynamics of discontinuous mussel beds: Laboratory flume simulations. Journal of Sea Research 62: 250-257

Forman, R. T. T. and M. Godron. 1981. Patches and structural components for a landscape ecology. BioScience 31:733-740.

Forman, R.T.T. and M. Godron. 1986. Landscape ecology. John Wiley, New York, 619 pp.

Forman, RT. 1995. Some general principles of landscape and regional ecology. Landscape ecology 10(3): 133-142

Franklin, J.F. and D.B. Lindenmayer. 2009. Importance of matrix habitats in maintaining biological diversity. PNAS 106(2): 349-350

Franzen, M., O. Schweiger and P-E. Betzholtz. 2012. Species-area relationship are controlled by species traits. PLoS ONE 7(5): e37359. doi:10.1371/ journal.pone.0037359

Fréchette, M. and E. Bourget. 1985a. Energy flow between the pelagic and benthic zones: factors controlling particulate organic matter available to an intertidal mussel bed. Can J. Fish. Aquat. Sci. 42: 1158-1165

Fréchette, M. and E. Bourget. 1985b. Food-limited growth of Mytilus eduis L. in relation to the benthic boundary layer. Can J. Fish. Aquat. Sci. 42: 1166-1170

Fréchette, M., C.A Butman and W.R. Geyer. 1989. The importance of boundary- layer flows in supplying phytoplankton to the benthic suspension feeders, Mytilus edulis L. Limnology and Oceanography 34(1): 19-36

Fried J.H., D.J. Levey and J.A. Hogsette.2005. Habitat corridors function as both drift fences and movement conduits for dispersing flies. Oecologia 143: 645–651

Frost, T.M., S.R. Carpenter, A.R. Ives, T.K. Kratz. 1995. Species compensation and complementarity in ecosystem function. In: Jones, C., Lawton, J. (Eds.), Linking Species and Ecosystems. Chapman and Hall, New York, 224–239p.

Garrabou, J., J. Riera and M. Zabala. 1998. Landscape pattern indices applied to Mediterranean subtidal rocky benthic communities. Landscape Ecology 13: 225-247

Gilbertson, W.W., M. solan and J.I. Prosser. 2012. Differential effects of microorganism-invertebrate interactions on benthic nitrogen cycling. FEMS Microbiol Ecol 82: 11-22

Giller, P.S., H. Hillebrand, U-G. Berninger, M.O. Gessner, S. Hawkins, P. Inchausti, C. Inglis, H. Leslie, B. Malmqvist, M.T. Monaghan, P.J. Morin, and G. O‟Mullan. 2004. Biodiversity effects on ecosystem functioning: emerging issues and their experimental test in aquatic environments. Oikos 104:423-436

Gilpin, M.E. and J.M. Diamond. 1976. Calculation of immigration and extinction curves from the species area distance relation. Proc. Natl. Acad. Sci. U.S.A., 73(11): 4130-4134

Green, B.C., D.J. Smith and G.J.C. Underwood. 2012. Habitat connectivity and spatia complexity differentially affect mangrove and salt marsh fish assemblages. Mar. Ecol. Prog. Ser. 466:177-192

Green, M.O., J.E. Hewitt and S.F. Thrush. 1998. Seabed drag coefficient over natural beds of horse mussels (Atrina zelandica). Journal of Marine Research 56: 613-637

Guichard, F., E. Bourget and J-L. Robert. 2001. Scaling the influence of topographic heterogeneity on intertidal benthic communities: alternate trajectories mediated by hydrodynamics and shading. Mar. Ecol. Prog. Ser. 217: 27-41

Guisan, A. and W. Thuiller. 2005. Predicting species distribution: offering more than simple habitat models. Ecology Letters 8: 993-1009

Gutzwiller, K.J. and S.H. Anderson. 1992. Interception of moving organisms: influences of patch shape, size and orientation on community structure. Landscape Ecology 6(4): 293-303

Haddad N.M. and K.A. Baum.1999. An experimental test of corridor effects on butterfly densities. Ecol. Appl. 9:623–633

Haines-Young, R., and M. Chopping. 1996. Quantifying landscape structure: a review of landscape indices and their application to forested landscapes. Progress in Physical Geography 20(4):418-445

Hall, P.O.J. and R.C. Aller. 1992. Rapid, small-volume, flow injection analysis for SCO2, and NH4+ in marine and freshwaters. Limnology and Oceanography 37(5): 1113-1119

Hamback, P.A. and G. Englund. 2005. Patch area, population density and the scaling of migration rates: the resource concentration hypothesis revisited. Ecology Letters 8:1057-1065

Hanski, I., M. Gilpin. 1991. Metapopulation dynamics: brief history and conceptual domain. Biological Journal of the Linnean Society 42(1-2): 3-16

Hargrave, B.T. 1969. Similarity of oxygen uptake by benthic communities. Limnology and Oceanography 14(5): 801-805

Helzer, C.J., and D.E. Jelinski. 1999. The relative importance of patch area and perimeter-area ratio to grassland breeding birds. Ecological applications 9(4): 1448- 1458

Hepcan, C.C. 2013. Quantifying landscape pattern and connectivity in a Mediterranean coastal settlement: the case of the Urla district, Turkey. Environ. Monit. Assess. 185: 143-155

Hulshoff, R.M. 1995. Landscape indices describing a Dutch landscape. Landscape Ecology 10: 101–111

Hunt, H. L., and R. E. Scheibling. 1998. Spatial and temporal variability of patterns of colonization by mussels (Mytilus trossulus, M. edulis) on a wave-exposed rocky shore. Marine Ecology Progress Series 167:155–169.

Hunt, H.L. and R.E. Scheibling. 2002. Movement and wave dislodgment of mussels on a wave-exposed rocky shore. Veliger 45(4):273-277

Hunt, H.L., and R.E. Scheibling. 2001. Patch dynamics of mussels on rocky shores: integrating process to understand pattern. Ecol. 82 (11): 3213-3231

Huston, M.A., L.W. Aarssen, M.P. Austin, B.S. Cade, J.D. Fridley, E. Garnier, J.P. Grime, J. Hodgson, W.K. Lauenroth, K. Thompson, J.H. Vandermeer, D.A. Wardle. 2000. No consistent effect of plant diversity on productivity. Science 289(5483): 1255

Jelbart, J.E., P.M. Ross and R.M. Connolly. 2006. Edge effects and patch size in landscapes: an experimental test using fish. Mar. Ecol. Prog. Ser. 319: 93-102

Jensen, S. L., and G. Muller-Parker. 1994. Inorganic nutrient fluxes in anemone dominated tide pools. Pacific Science 48: 32–43

Johnson, M.W. and K.L. Heck Jr. 2006. Effects of habitat fragmentation per se on decapods and fishes inhabiting seagrass meadows in the northern Gulf of Mexico. Mar. Ecol. Prog. Ser. 306: 233-246

Jones, C. G., J. H. Lawton, and M. Shachak. 1994. Organisms as ecosystem engineers. Oikos 69:373-386

Jones, C. G., J. H. Lawton, and M. Shachak. 1997. Positive and negative effects of organisms as physical ecosystem engineers. Ecology 78:1946-1957

Kemp, W.M., P. Sampou, J. Caffrey, M. Mayer, K. Henriksen and W.R. Boynton. 1990. Ammonium recycling versus denitrification in Chesapeake Bay sediments. Limnology and Oceanography 35(7): 1545-1563

Koivisto, M. and M. Westerbom. 2012. Invertebrate community associated with blue mussel beds in a patchy environment: a landscape ecology approach. Mar. Ecol. Prog. Ser. 471: 101-110

Kuznetsova, A., P.B. Brockhoff and R.H.B. Christensen. 2014. lmerTest: Tests in linear mixed effects models. R package version 2.0-20

Largaespada, C., F. Guichard and P. Archambault. 2012. Meta-ecosystem engineering: nutrient fluxes reveal intraspecific and interspecific feedbacks in fragmented mussel beds. Ecology 93(2): 324-333

Laurance, S.G.W. 2008. Responses of understory rain forest birds to road edges in central Amazonia. Ecological Applications 14(5): 1344-1357

Laurance, W.F., and E. Yensen. 1991. Predicting the impacts of edge effects in fragmented habitats. Biological Conservation 55: 77-92

Laurance, WF., L.V. Ferreira, J.M. Rankin-de Merona, S.G. Laurance, R.W. Hutchings and T.E. Lovejoy. 1998. Effects of forest fragmentation on recruitment patterns in Amazonian tree communities. Conservation Biology 12(2): 460-464.

Legendre, P. and L.F.J. Legendre. 2012. Numerical Ecology. Third english edition, Elsevier, Oxford, UK. 1006 p.

Lemieux, J. and M. Cusson. 2014. Effects of habitat-forming species richness, evenness, identity, and abundance on benthic intertidal community establishment and productivity. PLoS ONE 9(10): e109261. doi:10.1371/journal.pone.0109261

Levin, S.A., and R.T. Paine. 1974. Disturbance, patch formation, and community structure. Proc. Nat. Acad. Sci. USA. 71: 2744-2747

Lintas, C. and R. Seed. 1994. Spatial variation in the fauna associated with Mytilus edulis on a wave-exposed rocky shore. J. Moll. Stud. 60: 165-174

Loreau, M. and A. Hector. 2001. Partitioning selection and complementarity in biodiversity experiments. Nature 412: 72-76

Loreau, M., S. Naeem and P. Inchausti. 2002. Biodiversity and ecosystem functioning: synthesis and perspectives. Oxford University Press, Oxford, UK. 294 p.

Loreau, M., S. Naeem, P. Inchausti, J. Bengtsson, J. P. Crime, A. Hector, D. U. Hooper, M. A. Huston, D. Raffaelli, B. Schmid, D. Tilman and D. A. Wardle. 2001. Biodiversity and ecosystem functioning: current knowledge and future challenges. Science 294: 804-808

Lundblad, E.R., D.J. Wright, J. Miller, E.M. Larkin, R. Rinehart, D.F. Naar, B.T. Donahue, S.M. Anderson and T. Battista. 2006. A benthic terrain classification scheme for American Samoa. Marine Geodesy 29(2): 89-111

MacArthur, R. H., E. O. Wilson. 1967. The theory of island biogeography. Monographs in Population Biology, vol. 1.

MacArthur, R.H., J.M. Diamond and J.R. Karr. 1972. Density compensation in island faunas. Ecology 53: 330–342

Mancke, R.G., and T.A. Gavin. 2000. Breeding bird density in woodlots: effects of depth and buildings at the edges. Ecol. Appl. 10: 598–611

Mathieu, R., C. Freeman and J. Aryal. 2007. Mapping private gardens in urban areas using object-oriented techniques and very high-resolution satellite imagery. Landscape and Urban Planning 81: 179-192

Matias, G.M., A.J. Underwood, D.F. Hochuli and R.A. Coleman. 2010. Independent effects of patch size and structural complexity on diversity of benthic macroinvertebrates. Ecology 91(7): 1908-1915

Mazouni, N., J-C. Gaertner, J-M. Deslous-Paoli, S. Landrein and M.G. d'Oedenberg. 1996. Nutrient and oxygen exchanges at the water-sediment interface in a shellfish farming lagoon (Thau, France). J. Exp. Mar. Biol. Ecol. 205: 91-113

McCollin, D. 1998. Forest edges and habitat selection in birds: a functional approach. Ecography 21:247–260

McGarigal, K., and B.J. Marks. 1995. Spatial pattern analysis program for quantifying landscape structure. Gen. Tech. Rep. PNW-GTR-351. US Department of Agriculture, Forest Service, Pacific Northwest Research Station.

McKindsey, C.W., P. Archambault, M.D. Callier and F. Olivier. 2011. Influence of suspended and off-bottom mussel culture on the sea bottom and benthic habitats: a review. Canadian Journal of Zoology 89(7): 622-646

Menge, B.A. 1976. Organization of the New England rocky intertidal community: role of predation, competition, and environmental heterogeneity. Ecological Monographs 46: 355-393

Mizerek, T., H.M. Regan and K.A. Hovel. 2011. Seagrass habitat loss and fragmentation influence management strategies for a blue crab Cllinectes sapidus fishery. Mar. Ecol. Prog. Ser. 427: 247-257

Moser D., H.G. Zechmeister, C. Plutzar, T. Wrbka and G. Grabherr. 2002. Landscape patch shape complexity as an effective measure for plant species richness in rural landscapes. Landscape Ecology 17: 657-669

Moser, D., H.G. Zechmeister, C. Plutzar, N. Sauberer, T. Wrbka, and G. Grabherr. 2002. Landscape patch shape complexity as an effective measure for plant species richness in rural landscapes. Landscape Ecology 17: 657-669

Murphy, H.M., G.P. Jenkins, J.S. Hindell and R.M. Connolly. 2010. Response of fauna in seadrass to habitat edges, patch attributes and hydrodynamics. Austral Ecology 35: 535-543

Nams, V.O. 2014. Tortuosity of habitat edges affects animal movement. Landscape Ecology 29:655-663

Newell, R.I.E. 2004. Ecosystem influences of natural and cultivated populations of Nielsen, A.M., N.T. Eriksen, J.J. Lonsmann Iversen and H.U. Riisgard. 1995. Feeding, growth and respiration in the polychaetes Nereis diversicolor (facultative filter-feeder) and N. virens (omnivorous) a comparative study. Mar. Ecol. Progr. Ser. 125: 149-158

Norling, P. and N. Kautsky. 2008. Patches of the mussel Mytilus spp. are islands of high biodiversity in subtidal sediment habitats in the Baltic Sea. Aquat. Biol. 4: 75-87

Okamura, B. 1986. Group living and the edge effects of spatial position in aggregations of Mytilus edulis. Oecologia 69(3): 341-347

Paine, R. T. and S. A. Levin. 1981. Intertidal landscapes - Disturbance and the dynamics of pattern. Ecological Monographs 51:145-178

Papaspyrou S, E. Kristensen and B. Christensen. 2007. Arenicola marina (Polychaeta) and organic matter mineralisation in sandy sediments: in situ and microcosm comparison. Estuar. Coast. Shelf. Sci. 72: 213–222

Parker, T.H., B.M. Stansberry, C.D. Becker and P.S. Gipson. 2005. Edge and area effects on the occurrence of migrant forest songbirds. Conservation Biology 19(4): 1157-1167

Peake, A.J. and G.P. Quinn. 1993. Temporal variation in species-area curves for invertebrates o clumps of an intertidal mussel. Ecography 16: 269-277

Petrovic, F., and F. Guichard. 2008. Scales of Mytilus spp. Population dynamics: importance of adult displacement and aggregation. Mar. Ecol. Prog. Ser. 356: 203- 214

Pfister. C.A. 2007. Intertidal invertebrates locally enhance primary production. Ecology 88(7): 1647-1653

Pielou, E.C. 1966. The measurement of diversity in different types of biological collections. J. Theor. Biol., 13: 131-144

Piot, A., C. Nozais and P. Archambault. 2014. Meiofauna affect the macrobenthic biodiversity-ecosystem functioning relationship. Oikos 123: 203-213

Pittman, S., R. Kneib, C. Simenstad and I. Nagelkerken. 2011. Seascape ecology: application of landscape ecology to the marine environment. Mar. Ecol. Prog. Ser. 427: 187-190

Plante-Cuny, M-R., R. Plante, N. Mazouni, M-F. Fontaine, P. Souchu, J-M. Deslous- Paoli and C. Grenz. 1998. Oxygen fluxes involving the benthic micro- and macrophytic components in the Thau Lagoon under pre-anoxic conditions. Oceanologica Acta 21(6): 819-829

Polis, G.A., and S.D. Hurd. 1995. Extraordinarily high spider densities on islands: flow of energy from the marine to terrestrial food webs and the absence of predation. Proc. Natl. Acad. Sci. USA. 92: 4382-4386

Raffaelli, D.G., M.C. Emmerson, M. Solan, C.L. Biles, D.M. Paterson. 2003. Biodiversity and ecosystem processes in shallow coastal waters: an experimental approach. Journal of Sea Research 49: 133–141

Ricciardi, A., F.G. Whoriskey and J.B. Rasmussen. 1997. The role of the zebra mussel (Dreissena polymorpha) in structuring macroinvertebrate communities on hard substrata. Can. J. Fish. Aquat. Sci. 54: 2596-2608

Richard, M., P. Archambault, G. Thouzeau and G. Desrosiers. 2006. Influence of suspended mussel lines on the biogeochemical fluxes in adjacent water in the îles- de-la-Madeleine (Quebec, Canada). Can. J. Fich. Aquat. Sci. 63: 1198-1213

Richard, M., P. Archambault, G. Thouzeau, C.W. McKindsey and G. Desrosiers. 2007a. Influence of suspended scallop cages and mussel lines on pelagic and benthic biogeochemical fluxes in Havre-aux-Maisons Lagoon, îles-de-la-Madeleine (Quebec, Canada). Can. J. Fish. Aquat. Sci. 64(11):1491-1505

Richard, M., P. Archambault, G. Thouzeau and G. Desrosiers. 2007b. Summer influence of 1 and 2 yr old mussel cultures on benthic fluxes in Grande-Entrée lagoon, Iles-de-la-Madeleine (Québec, Canada). Mar. Ecol. Prog. Ser. 338: 131-143

Ries, L., and T.D. Sisk. 2004. A predictive model of edge effects. Ecology 85(11): 2917-2926

Ries, L., R.J. Fletcher Jr., J. Battin, and T.D. Sisk. 2004. Ecological responses to habitat edges: mechanisms, models, and variability explained. Annu. Rev. Ecol. Evol. Syst. 35: 491-522

Ripple, W.J., G.A. Bradshaw and T.A. Spies. 1991. Measuring forest landscape patterns in the cascade range of Oregon, USA. Biological Conservation 57(1): 73-88

Robert, P., C.W. Mckindsey, G. Chaillou and P. Archambault. 2013. Dose-dependent response of a benthic system to biodeposition from suspendedblue mussel (Mytilus edulis) culture. Mar. Pol. bul. 66(1-2): 92-104.

Root, R.B. 1973. Organization of a plant-arthropod association in simple and diverse habitats: the fauna of collards (Brassica Oleracea). Ecological Monographs 43(1): 95-124

Russel, G.J., J.M. Diamond, T.M. Reed and S.L. Pimm. 2006. Breeding birds on small island biogeography or optimal foraging? Journal of Animal Ecology 75: 324- 339

Schneider K.R., Wethey D.S., Helmuth B.S.T. 2005. Implications of movement behavior on mussel dislodgement: exogenous selection in a Mytilus spp. hybrid zone. Mar. Biol. 146:333–343

Snedecor, G. W. and W.G. Cochran. 1989. Statistical Methods, Eighth Edition, Iowa State University Press

Soga, M., N. Kanno, Y. Yamaura, and S.Koike. 2013. Patch size determines the strength of edge effects on carabid beetle assemblages in urban remnant forests. J. Insect. Conserv. 17: 421-428

Sousa, P.W. 1984. Intertidal mosaics: patch size, propagule availability, and spatially variable patterns of succession. Ecology 65(6):1918-1935

Sousa, W. P. 1985. Disturbance and patch dynamics on rocky intertidal shores. In, The ecorosy of natural disturbance andpatch dynamics, edited by ST. A. Pickett and P.S. White, Academic Press, New York, New York, 101-124p.

Stenton-Dozey, J., T. Probyn and A. Busby. 2011. Impact of mussel (Mytilus galloprovincialis) raft-culture on benthic macrofauna, in situ oxygen uptake, and nutrient fluxes in Saldanha Bay, South Africa. Can. J. Fish. Aquat. Sci. 58(5): 1021- 1031

Stumpf, R.P. and K. Holderied. 2003. Determination of water depth with high- resolution satellite imagery over variable bottom types. Limnol. Oceanogr. 48(1-2): 547-556

Svane, I. and I. Setyobudiandi. 1996. Diversity of associated fauna in beds of the blue mussel Mytilus edulis L.: effects of location, patch size, and position within a patch. Ophelia 45: 39–53

Tsuchiya, M. 1980. Biodeposit production by the mussel Mytilus edulis L. on rocky shores. J. exp. mar. Biol. Ecol. 47: 203-222

Tsuchiya, M. and M. Nishihira. 1985. Islands of Mytilus as a habitat for smal intertidal animals: effect of island size on community structure. Mar. Ecol. Prog. Ser. 25: 71-81

Turner M.G., R.H. Gardner and R.V. O’Neill. 2001. Landscape Ecology in Theory and Practice: Pattern and Process. Springer, New York, 401p.

Underwood, A.J. 1999. Physical disturbances and their direct effect on an indirect effect: responses of an intertidal assemblage to a severe storm. J. Exp. Mar. Biol. Ecol. 232:125–140

Underwood, A.J. and M.G. Chapman. 1996. Scales of spatial patterns of distribution of intertidal invertebrates. Oecologia 107(2): 212-224

Upton, A. C., D. B. Nedwell, R. J. Parkes and S. M. Harvey. 1993. Seasonal benthic microbial activity in the southern North Sea; oxygen uptake and sulphate reduction. Mar. Ecol. Prog. Ser. 101: 273–281

Van Leeuwen, B., D.C.M. Augustijn, B.K. van Wesenbeeck, S.J.M.H. Hulscher and M.B. de Vries. 2010. Modeling the influence of a young mussel bed on fine sediment dynamics on an intertidal flat in the Wadden Sea. Ecological Engineering 36:145-153

Vandewalle, M., F. de Bello, M. P. Berg, T. Bolger, S. Dole´dec, F. Dubs, C.K. Feld, R. Harrington, P. A. Harrison, S. Lavorel, P. M. da Silva, M. Moretti, J. Niemela, P. Santos, T. Sattler, J. P. Sousa, M. T. Sykes, A. J. Vanbergen, B. A. Woodcock.2010. Functional traits as indicators of biodiversity response to land use changes across ecosystems and organisms. Biodivers Conserv 19: 2921-2947

Villeger, S., N.W.H. Mason and D. Mouillot. 2008. New multidimensional functional diversity indices for a multifaceted framework in functional ecology. Ecology 89: 2290–2301

Wedding, L.M., C.A. Lepczyk, S.J. Pittman, A.M. Friedlander and S. Jorgensen. 2011. Quantifying seascape structure: extending terrestrial spatial pattern metrics to the marine realm. Mar. Ecol. Prog. Ser. 427: 219-232

Widdows, J., M.D. Brinsley, P.N. Salkeld and M. Elliott. 1998. Use of annular flumes to determine the influence of current velocity and bivalves on material flux at the sediment-water interface. Estuaries 21(4):552-559

Wiens J.A., C.S. Crawford and J.R. Gosz. 1985. Boundary dynamics: a conceptual framework for studying landscape ecosystems. Oikos 45:421–27

Wildish, D.J., and D.D. Kristmanson. 1979. Tidal energy and sublittoral macrobenthic animals in estuaries. J. Fish. Res. Bd. Can. 36: 1197-1206

Williams, C.B. 1943. Area and number of species. Nature 152: 264-267

Wright, R.T., R.B. Coffin, C.P. Ersing and D. Pearson. 1982. Field and laboratory measurements of bivalve filtration of natural marine bacterioplankton suspension- feeding bivalve mollusks: a review. J. Shellfish Res. 23: 51–61

Wu, J. and Vankat, J.L. 1991. An area-based model of species richness dynamics of forest islands. Ecol. Modelling 58: 249-271

Yamaura, Y., T. Kawahara, S. Iida, and K. Ozaki. 2008. Relative importance of the area and shape of patches to the diversity of multiple taxa. Conservation Biology 22(6): 1513-1522

Appendix 1

TABLE 8. Results from linear mixed effects models, with the fixed factors area (3 levels) perimeter (3 levels) and the interaction between area and perimeter as main sources of variation. The model included the random variable transplant (36 levels).The dependent variables are Shannon diversity index, richness in taxa (log+1 transformed), abundance (log+1 transformed) and Evenness index and abundance in Littorina sp.

Source of variation Num df Den df Mean square F P Shannon diversity index

Area 2 37 0.012 0.245 0.784 Perimeter 2 37 0.015 0.304 0.740 Area*Perimeter 4 37 0.104 2.144 0.095 .

Richness (log(richness+1))

Area 2 37 0.081 1.269 0.293 Perimeter 2 37 0.087 1.360 0.269 Area*Perimeter 4 36 0.196 3.050 0.029 *

Abundance in taxa (log(density+1))

Area 2 36 0.028 0.319 0.729 Perimeter 2 36 0.203 2.274 0.117 Area*Perimeter 4 36 0.058 0.648 0.632

Evenness index

Area 2 37 0.009 0.245 0.784 Perimeter 2 37 0.011 0.304 0.740 Area*Perimeter 4 37 0.077 2.144 0.095 . Abundance in Littorina sp.

Area 2 36.254 102.415 0.940 0.400 Perimeter 2 36.282 118.731 1.090 0.347 Area*Perimeter 4 36.243 93.705 0.860 0.497

Notes: Calculation using alpha = 0.05 Signif. codes: 0 '***' 0,001 '**' 0,01 '*' 0,05 '.' 0.1

Appendix 2

TABLE 9. Permutational multivariate analysis of variance (9999 permutations) with Gower dissimilarity index on fourth root transformed data, testing the effect of area (fixed with 3 levels) and perimeter (fixed with 3 levels) on assemblies of taxa and locomotion groups, with transplant as random factor. Fixed effects sum to zero for mixed terms and sums of squares is of type III. Littorina sp. is not in the data set.

Source of variation df Mean square F P Taxa

Area 2 24.72 0.696 0.674 Perimeter 2 26.43 0.746 0.639 Area*Perimeter 4 26.18 0.737 0.713 Transplant (Area*Perimeter) 27 35.97 1.479 0.004 * Residuals 90 24.32

Groups of locomotion

Area 2 1311.50 1.053 0.414 Perimeter 2 339.24 0.273 0.829 Area*Perimeter 4 2098.90 1.676 0.153 Transplant (Area*Perimeter) 28 1284.50 2.827 0.000 *** Residuals 97 454.34

Notes: Calculation using alpha = 0.05 Signif. codes: 0 '***' 0,001 '**' 0,01 '*' 0,05 '.' 0.1

Appendix 3

Table 10. P-values extracted from Generalized linear mixed effects models relating (A) the fixed variables area and perimeter and (B) the fixed variable perimeter-area ratio (P/A ratio), to abundance of the functional groups sessile, substrate and swimmer. Transplant is included as random variable in models. Littorina sp. is not in the data set.

Source of variation Num df Den df F P

Abundance for the group 'swimmer' Area 2 25 1.11 0.900 Perimeter 2 25 0.53 0.595 Area*Perimeter 4 25 0.96 0.445

Abundance for the group 'substrate' Area 2 37 0.81 0.452 Perimeter 2 38 0.82 0.449 Area*Perimeter 4 34 2.07 0.106

Abundance for the group 'sessile' Area 2 23 1.45 0.254 Perimeter 2 24 2.35 0.117 Area*Perimeter 4 23 1.27 0.309

Notes: Calculation using alpha = 0.05, Signif. codes: 0 '***' 0,001 '**' 0,01 '*' 0,05 '.' 0.1

Appendix 4

Table 11: Taxa identified within our mussel transplants and functional group based on locomotion type associated to each taxa. Swimmers (Swi), substrate (Sub) and sessile (Ses) are the three functional groups.

Sub- Fonct. Phylum Class Order Family Genus genus Species Group Annelida Clitellata Sub

Polychaeta Sub

Polychaeta Phyllodocida Nereididae Nereis Alitta virens Swi Nereis Hediste diversicolor Swi

Nereis pelagica Swi

Nereis sp Swi

Pholoidae Pholoe longa Sub

Pholoe minuta Sub

Pholoe sp Sub

Sabellida Fabriciidae Fabricia sabella Sub

Spionida Spionidae Sub

Spionidae Polydora cornuta Sub

Polydora websteri Sub

Polydora sp Sub

Spiophanes kroyeri Sub

Arthropoda Malacostraca Amphipoda Caprellidae Caprella linearis Sub

Gammaridae Gammarus lawrencianus Swi

Gammaridae Gammarus oceanicus Swi

Gammaridae Gammarus setosus Swi

Gammaridae Gammarus sp Swi

Gammaridae Swi

Hyalidae Apohyale prevostii Swi

Hyalidae Hyale sp Swi

Lysianassidae Orchomenella pinguis Swi

Amphipoda Swi

Decapoda Cancridae Cancer sp Sub

Isopoda Janiridae Jaera Jaera albifrons Sub

Maxillopoda Swi

Ostracoda Swi

Cnidaria Anthozoa Actiniaria Ses Campanulariida Ses Hydrozoa Leptothecata e Verticillina sp

Campanulariidae Ses

Campanulinidae Ses

Sertulariidae Hydrallmania sp Ses

Sertulariidae Ses

Hydrozoa Ses

Table 11 (continued): Taxa identified within our mussel transplants and functional group based on locomotion type associated to each taxa. Swimmers (Swi), substrate (Sub) and sessile (Ses) are the three functional groups.

Sub- Fonct. Phylum Class Order Family Genus genus Species Group Mollusca Bivalvia Myoida Myidae Mya arenaria Ses

Mytiloida Mytilidae Mytilus Ses

Veneroida Tellinidae Macoma balthica Ses Tellinidae Ses Bivalvia Ses Gastropoda Littorinimorpha Littorinidae Lacuna vincta Sub

Littorina saxatilis/ Sub obtusata

Littorina sp Sub Skeneopsidae Skeneopsis planorbis Sub Mollusca Gastropoda Littorinimorpha Sub Gastropoda Sub Nematoda Sub Nemertea Sub Platyhelminthes Turbellaria Sub Platyhelminthes Swi