David Mouillot, UMR CNRS-UMII 5119 Ecosystèmes Lagunaires, Université Montpellier II, CC 093, 34095 Montpellier cedex 5, France (fax: +33 467143 719; e-mail: mouillot@University-montp2.fr).
1Biodiversity is rapidly being lost in a world transformed increasingly by human activities. We need to determine urgently the factors which control the coexistence of species and thus allow local biodiversity to be maintained.
2Coexistence between interacting species can be explained by species-sorting, mass-effect, patch dynamic and neutral perspectives on metacommunities. According to the first two paradigms, community assembly rules are based on species’ ecological niches or functional roles, whereas the latter two emphasize the role of stochastic processes.
3Here I consider whether effective strategies for managing fish biodiversity in coastal lagoons depend on the predominant paradigm.
4The neutral perspective was not considered relevant because lagoon fishes belong to different trophic levels and have clear niche differences.
5Where fish coexistence is ruled by species-sorting, priority must be given to the preservation or restoration of habitat, whereas heterogeneity among lagoons and population densities become critical when mass effects predominate.
6According to the patch dynamics perspective, the key factors are individual turn-over and the fitness equivalence among species.
7Synthesis and applications. Coastal fish metacommunities are not ruled consistently by a single theoretical paradigm. However, consideration of such extreme cases sets clear boundaries for species assembly rules. Confronting the three main views is not only a controversial topic, but also has consequences for ecosystem and biodiversity management.
A key element in understanding biodiversity and how it changes is to understand the factors that determine how species coexist. The coexistence of species is explained classically by either a niche-assembly or a dispersal-controlled perspective on communities. The unified neutral theory of biodiversity and biogeography (Hubbell 2001) assumes the per capita ecological equivalence of individuals (whatever the species they belong to) and this dispersal-related view contrasts with proposals that coexistence is maintained by niche differentiation. In niche-assembled communities, interacting species are at equilibrium with, according to the limiting similarity theory (MacArthur & Levins 1967), the best competitor occupying each ecological niche. These two views of species coexistence are not mutually exclusive and are actually under scrutiny with regard to their ability to explain the relative abundance of species (McGill 2003; Volkov et al. 2003) and to predict species invasion (Fargione, Brown & Tilman 2003; Herben et al. 2004).
Much of the formal theory about community ecology focuses on a single scale, but local communities are not isolated assemblages and it is now generally accepted that community ecology has to be investigated within a metacommunity framework (Hubbell 1997; Loreau, Mouquet & Gonzalez 2003; Leibold et al. 2004). A metacommunity is defined as a set of local communities that are linked by dispersal of multiple, potentially interacting species (Wilson 1992). Of the four paradigms proposed in metacommunity theory (see review in Leibold et al. 2004), three assume niche differences among species and one proposes neutrality among individuals. Here I propose to fill the gap between these theoretical metacommunity paradigms and practical implications for management of biodiversity in brackish lagoons.
Coastal lagoons provide essential ecosystem services such as shoreline protection, water quality improvement and fisheries resources, as well as habitat and food for migratory and resident animals and recreational areas for human populations (Levin et al. 2001), and represent 13% of the world coastline. These highly productive ecosystems (300 g C m−2 year−1) (Knoppers 1994) are under severe stress: in addition to climatic change inducing sea-level rise, human impacts have led to problems such as anoxia, over-exploitation of resources, destruction of habitats, eutrophication and pollutant contamination from land use in the watershed (Joyeux & Ward 1998; Crooks & Turner 1999).
Many of the goods and services provided by coastal lagoon ecosystems rely on the species inhabiting these areas, with fish influencing ecosystem processes through trophic relationships with other compartments (e.g. Mancinelli, Costantini & Rossi 2002). Within lagoon fish communities we can distinguish freshwater (occasional visitors via freshwater discharges), brackish-water and marine species. Most of the marine species are transients (or at least non-residents), using lagoons as either feeding or breeding areas (Levin et al. 2001) and therefore exporting biomass and, indirectly, nutrients to adjacent marine ecosystems. Fish may also accumulate and export pollutants (Mouillot et al. 2000). The study of fish diversity is important for several reasons. First, the more fish species there are in the coastal lagoons, the more body shapes, colours or swimming abilities can be observed (after recreational fishing or not), increasing the aesthetic value of the lagoon ecosystem. Secondly, the more fish species there are in the coastal lagoons, the more various trophic levels can be expected to exist, increasing the complexity of the food web. Thirdly, the more fish species there are in the coastal lagoons, the more varied the biological responses can be expected to be during a perturbation or a crisis event, increasing the resistance of the system. Finally, the more fish species there are in the coastal lagoons, the more ecological niches can be occupied, increasing the resistance to invasion (Stachowicz et al. 2002).
Hence, the processes by which the number of coexisting fish species can be maintained is of major concern for coastal lagoons. My aim is to consider whether effective strategies for managing fish biodiversity in coastal lagoons depend on which metacommunity paradigm predominates and, if so, to propose appropriate management priorities.
Coastal lagoon fish and the metacommunity framework
Metacommunity models have either been based on a local community connected to a regional pool or on a set of spatially assembled local communities. In the former, the local community interacts with the regional pool (the metacommunity) through migration (Bell 2001; Hubbell 2001) while, in the latter, the metacommunity is partitioned into spatially explicit local assemblages that are connected by individual migrations (Chave, Muller-Landau & Levin 2002). A coastal lagoon comprises a set of localities (patches of macrophyte, macroalgal, sandy, muddy, shellfish or rocky habitat), which hold local communities and are connected to each other to constitute a lagoon metacommunity (Fig. 1). The fish metacommunity within a lagoon is connected to a larger-scale, regional metacommunity, embracing all coastal fish populations potentially able to colonize the lagoon. Metacommunity paradigms will be examined here at the regional level: spatial dynamics occurring among habitat patches within the lagoon metacommunity will not be discussed.
It is clear that fish species are not strictly equivalent, either in coastal lagoons (e.g. Dumay et al. 2004) or in coastal ecosystems in general (e.g. Bellwood et al. 2006 found a high functional versatility among coral reef fish belonging to the same trophic group). Thus the neutral theory is not relevant at the scale of biological or functional heterogeneity seen in coastal fish and the neutral metacommunity perspective will not be considered further.
Three paradigms might shape coexistence of fish species in coastal lagoons (Fig. 2). The species-sorting perspective postulates that environmental heterogeneity among and within localities is strong enough to promote spatial niche segregation among species. Strong gradients in abiotic factors or in habitat characteristics (three niches in Fig. 2a) lead to local environmental conditions acting as a niche filter on the regional pool of species (Zobel 1997; Statzner, Doledec & Hugueny 2004). Local community composition is thus fairly resilient to disturbance (Cottenie et al. 2003), and coexistence of species in the metacommunity is ruled by niche partitioning processes (e.g. Albrecht & Gotelli 2001). The mass-effect perspective (Fig. 2b) postulates that poor competitors can escape from local competitive exclusion by mass immigration. As in the species-sorting perspective, environmental differences among localities are strong enough to provide spatially segregated niches, but these localities are sufficiently connected by dispersal to establish source–sink relations and thus to allow species (such as species C in Fig. 2b) to colonize localities where they are poor competitors from adjacent localities where they are better competitors and are thus massively represented (Gonzalez et al. 1998; Mouquet & Loreau 2003). The third, patch dynamics perspective, postulates that localities are identical and that species coexistence is maintained by a trade-off between competitive ability and dispersal (Levin & Paine 1974). All species have the same probability of being present in a locality: even if species A is the strongest competitor, species B and C can also be present in the lagoon because of their better dispersal ability (Fig. 2c).
From metacommunity paradigms to a biodiversity management strategy
Here I consider the priorities for managing fish biodiversity in coastal brackish lagoons from the metacommunity viewpoint. Although these three pure theoretical rules are an oversimplification, their use has the merit of producing clear hypotheses and is potentially valuable for proposing management priorities.
According to both the two first perspectives (Fig. 2a,b), spatial niche differentiation occurs among localities within the lagoon (Fig. 1) but fish diversity within the lagoon is maintained by different processes (the presence of species C is promoted either by the presence of a niche where it is one of the strongest competitors or is dependent on a high regional density).
the species-sorting perspective
According to the species-sorting perspective, the number of species is limited by the number of niches available within the lagoon. Consequently, a large spectrum of alpha (resource-related) or beta (habitat-related) niches is needed to maintain or increase fish richness (Wilson 1999). The number of beta niches depends on both diversity and heterogeneity of habitats (Ziv 1998; Guidetti et al. 2002) and, in coastal lagoons, habitat diversity is often related to macrophyte diversity and cover (Sfriso, Birkemeyer & Ghetti 2001), with macrophyte beds and meadows providing microhabitats and protection from predators (Dolbeth et al. 2003). Transplanted macrophyte beds can provide the structural and functional attributes of natural beds in previously degraded sites (Dolbeth et al. 2003) and mobile macrofauna often colonize the new habitat within a couple of weeks (Sogard 1989). Fish and shrimp richness was therefore higher in transplanted eelgrass beds than on unplanted substrates after less than 1 year (Fonseca et al. 1990). Shellfish aquaculture can also provide new habitats, with oyster reefs, for instance, creating corridors between shelter and foraging grounds (Peterson & Lipcius 2003; Peterson, Grabowski & Powers 2003). Sea urchins also provide biogenic structures and thus new microhabitats that allow more fish species to become established (Hartney & Grorud 2002).
In order to increase the number of alpha niches available in coastal lagoons we can increase the diversity of resources and the diversity of prey. Resource partitioning is a key factor in promoting fish coexistence because competition is reduced when diet overlap is low (DePirro, Marchetti & Chelazzi 1999; Sibbing & Nagelkerke 2001; Carrasson & Cartes 2002). The amount and diversity of food sources are again dependent on the heterogeneity and diversity of habitats, with oyster reefs, for instance, providing structured habitat for finfish, crabs and other organisms (Breitburg et al. 2000) and other man-made structure providing suitable habitats for amphipods in brackish lagoons (Aikins & Kikuchi 2001). Nevertheless, Parker, Duffy & Orth (2001) demonstrated experimentally that most indices of animal diversity in marine environments were more strongly related to total vegetal surface area than to vegetal diversity, indicating that increasing vegetation cover is another strategy for maintaining a large alpha niche spectrum in brackish lagoons where fish communities are niche-assembled.
The diversity of both beta and alpha niches depends on water quality. Eutrophication and its associated oxygen imbalance (Justic 1991) and increasing organic particulate matter ratio (Dell’Anno et al. 2002) has become a major problem in numerous coastal areas (Flindt et al. 1999). Environmental constraints such as oxygen depletion (Ishitobi et al. 2000) are likely to reduce the number of fish species because fewer species are able to tolerate eutrophic conditions (e.g. Ludsin et al. 2001). Increased primary production may translate into additional production for pelagic fishes but not for demersal ones feeding on benthic resources. Also, periodic dystrophic crises observed in coastal lagoons with high eutrophication levels (Bachelet et al. 2000) lead to high mortality of both fish and invertebrates and to a degradation of fish habitats (Karim, Sekine & Ukita 2003; Powers et al. 2005; Bishop et al. 2006) and, inevitably, to a decrease in species richness. Watershed management must therefore aim to decrease nutrient and pollutant inputs (Ludsin et al. 2001). Dispersion is also critical for maintaining local diversity; because species must be able to reach the niches they are able to occupy, sea-lagoon exchanges must be carefully maintained.
In summary, when richness of the fish community within a lagoon depends on species-sorting, management should mainly focus on habitat protection, habitat diversification within the lagoons, quality improvement of run-off waters and biotic exchanges (Fig. 2).
the mass-effect perspective
If mass-effects operate, a species can be present in a lagoon even if it is not a good competitor in any of the available niches, providing the species (C in Fig. 2b) is abundant somewhere that can act as a source of individuals for the sink lagoon. Management priorities are therefore rather different: they need to preserve a high regional density for all fish populations, to maximize dispersal and thus maintain connectivity between habitats, and to promote regional habitat diversity to ensure that each species has a habitat where it is among the best competitors.
Strong heterogeneity among coastal habitats will prevent any biotic homogenization at the regional scale (Olden et al. 2004) and thus preserve the source–sink relationships and the mass effects (Mouquet & Loreau 2003). In highly eutrophic lagoons, planktivorous and detritivorous fish (low trophic levels) are likely to maintain dense populations, whereas carnivorous demersal fishes will benefit from a high secondary production of invertebrates in more oligotrophic lagoons where they will reach higher densities. Different lagoons may act as complementary sources of various populations, each one providing optimal conditions for a given group of species (Fig. 2b).
the patch dynamics perspective
In the patch dynamics view, local habitats within the lagoon are considered to be equivalent (Fig. 2c). Fish species coexistence is permitted by trade-offs between functional attributes and is ruled by the dispersal-assembly hypothesis, i.e. communities are non-equilibrium assemblages of species, their presence or absence being dictated by random dispersal. A fish species that is a poor competitor can be present in the lagoon if it compensates with high dispersal abilities (such as species C in Fig. 2). In this case we need to preserve the dispersal of fish individuals, for instance by managing channels to allow fish movements throughout the year and reducing or removing passive nets and other physical barriers (Crespi 2002). Fish dispersal and migration are also strongly influenced by physical oceanographical features such as shifts in hydrological conditions in coastal lagoons and estuaries (Garcia et al. 2004). Fish can detect a wide range of external stimuli (chemicals, temperature and atmospheric pressure) (Bullen & Carlson 2003) and rapid shifts in hydrological conditions are likely to modify routes and intensity of fish migration (Brehmer et al. 2006) and, thus, richness under this perspective.
Variability of abiotic factors within the lagoon will promote turn-over in species composition, suggesting that lagoons must be protected from increasing marine influence (and thus more stable temperature, salinity and pH). A hydrological shift from hypohaline (late 1950s) to hyperhaline conditions in the Terminos lagoon (Gulf of Mexico) restricted the richness, the taxonomic diversity and density of estuarine species, as well as the presence of freshwater species following freshwater discharge from the watershed (Ramos Miranda et al. 2005; Sosa-Lopez et al. 2007).
Another critical point is to preserve the fitness equivalence among species within the regional pool. Functional differentiation of the mosquitofish Gambusia affinis (Baird & Girard) (Dumay et al. 2004), its higher aggressiveness than native species (Mills, Rader & Belk 2004) and its greater dispersal abilities than some congeners (Rehage & Sih 2004) all contribute to the relative invasiveness of this species. Particular attention must therefore be paid to invasive species which may break the trade-off rule and thus prevent coexistence and richness of fish species within coastal lagoons.
The current management programmes for biodiversity are largely pattern-based and focus on hotspots (Turpie, Beckley & Katua 2000; Roberts et al. 2002). They should become process-based and include functional mechanisms. Thus, preserving goods and services provided by ecosystems requires not only the investigation of biodiversity patterns but also the identification of processes that generate and maintain local biodiversity: different metacommunity paradigms lead to different management strategies to maintain lagoon fish biodiversity. Priorities for increasing local richness in coastal lagoon fish communities and the number of functions provided by the species depend upon the communities’ assembly rules.
Although it is unlikely that any coastal fish metacommunity is ruled solely by one of the three hypotheses, one assembly rule may be predominant. Indeed, Leibold & Norberg (2004) suggested that plankton metacommunities may be ruled by species-sorting paradigm. Regional fish metacommunities embracing lagoon and coastal habitats may be consistent with the mass-effect paradigm because the many different patches enable niche segregation (Mouillot, Dumay & Tomasini 2007), connectivity among coastal sites is very high (Sanvicente-Anorve, Flores-Coto & Chiappa-Carrara 2000) and because regional habitat diversity (e.g. Mouillot et al. 2005) may promote source–sink relationships between localities. In the open sea, where differences among localities are much less pronounced, fish metacommunities are more likely to follow the patch dynamics paradigm.
The challenge now is to move from conjecture and speculation to confront metacommunity paradigms with empirical data. Assembly rules in lagoon fish can be investigated using functional differences to test whether coexisting species in local patches are more similar than expected by chance (i.e. than random samples from the regional pool). If so, local habitat characteristics act as environmental filters, allowing only a narrow spectrum of species to survive (Peres-Neto 2004; Mouillot et al. 2007) and species-sorting and mass-effect are better explanations than patch-dynamics. If coexisting species are less similar than expected by chance (i.e. compared to random samples from the pool of species able to survive in these conditions) we may conclude that interspecific competition limits species similarity (Jensen 1997) even in connected habitats, and that species-sorting is the best candidate. Innovative approaches, using acoustic methods (Gaudreau & Boisclair 2000), might allow estimation of fish fluxes between coastal lagoons and the sea (dispersal) and evaluation of fish density within the lagoon (carrying capacity). In any case, before proposing an integrated management strategy for local biodiversity in limnetic ecosystems, it is necessary to take into account the processes that rule local assemblages in metacommunities.
I wish to thank J. C. Joyeux, T. Bouvier and J. A. Tomasini for comments on the manuscript. J. B. Wilson and R. de Wit corrected the English and provided useful suggestions.