Species, populations, or individuals that co-exist locally potentially compete with each other for resources, notably when limited (Hutchinson, 1957; Chase, 2011). In this sense, resources can include any ecological features and their interactions such as diet and habitat (Chase, 2011). One way to limit intra- and interspecific competition is niche partitioning (Chase, 2011). Niche partitioning can occur spatially and/or temporally, following the variability of resource distribution through space and time. Specialization/generalization strategies theoretically respond to several environmental forces, including population size (Sargent and Otto, 2006, reviewed in Sexton et al., 2017), resource abundance, diversity, and accessibility (MacArthur and Pianka, 1966; Charnov, 1976; Svanbäck and Bolnick, 2007). In the context of interspecific competition and considering population size, models predict that specialization is expected to be promoted in locally rare populations/species, while generalization will be favoured when population/species are locally abundant (Sargent and Otto, 2006, reviewed in Sexton et al., 2017). Specialists are also expected to maximize their fitness in one environment, while generalist fitness will be maximized in multiple ones (Sexton et al., 2017). According to the niche theory, the smaller the local community (individuals and species), the greater the community will be impacted by stochastic ecological processes (Chase, 2003; Fukami, 2004). This is the case for instance when the type and abundance of preys vary in heterogeneous environments. In an environment where resources are patchily distributed, the optimal foraging theory predicts that the niche width of a consumer depends on habitat and prey availabilities as well as their distribution (MacArthur and Pianka, 1966; Charnov, 1976; Svanbäck and Bolnick, 2007). The range of preys is expected to be greater in generalist consumers than in specialist consumers. Therefore, determining niche width can reveal spatial and temporal patterns in resource use and interspecific resource competition (Bolnick et al., 2011), which is particularly important for species conservation. This conceptual framework is particularly relevant for diadromous species sharing common life history traits and resources, whose populations have declined worldwide for many reasons including the fragmentation of freshwater habitats (e.g., Costa-Dias et al., 2009; Lassalle et al., 2009; Duarte et al., 2021; Waldman and Quinn, 2022).

The diadromous fishes have peculiar life history traits as they migrate from the sea to freshwater to spawn (anadromous) or the reverse (catadromous) (Myers, 1949; Harden-Jones, 1968). Many temperate watersheds host multiple diadromous species with varied life histories both within and between species. In France, the potential richness in diadromous taxa is higher in the largest rivers, with, for instance, a maximum of seven taxa observed for the French Atlantic and the English Channel River systems (Seine, Loire, Adour, Dordogne, and Garonne) (Merg et al., 2020). They all share a life history that makes them particularly sensitive to river fragmentation induced by the construction of dams, weirs, and sluices (Pringle et al., 2000; Limburg and Waldman, 2009; Waldman and Quinn, 2022). In Europe, there are more than one million river barriers (Belletti et al., 2020). For diadromous fish, river barriers hamper connectivity and can lead to the confinement of populations and accumulation of individuals down to the barriers (Katano et al., 2006; Drouineau et al., 2018; Forget et al., 2018; Trancart et al., 2020; Verhelst et al., 2021; Teichert et al., 2022a). Over-density can increase intra- and interspecific competition, which in turn can hamper species diversity and individual growth, survival, and/or reproduction (Laffaille et al., 2003; Machut et al., 2007; Costa et al., 2008). Dams have been shown to decrease diadromous species diversity (Merg et al., 2020) and alter growth and resident/migrant strategies in the white-spotted char (Salvelinus leucomaenis) (Morita et al., 2000). It may also bias the sex determination of eels (Laffaille et al., 2006; Costa et al., 2008). Despite the large body of literature on diadromous fishes, very few studies have focused on sympatric diadromous fishes and their potential competition in a catchment obstructed by dams (but see Katano et al., 2006). Indeed, an accumulation of the different life stages of diadromous fish downstream to the dams is likely to enhance intra- and interspecific competition for food and habitats and promote generalization for abundant species and specialization for rare ones.

A diadromous niche can be studied through stable isotope analysis, as such analysis provides quantitative measures of both resource and habitat, generally represented in a multivariate space. This is particularly useful in the context of resource partitioning among sympatric species (Newsome et al., 2007); it informs on spatial and temporal variations of resources and consumers (Hobson, 1999). Due to isotope fractionation between prey and consumer, N stable isotope ratios (δ¹⁵N) of fish inform on their trophic position, while C stable isotope ratios (δ¹³C) indicate which type of C fuels their food chains, primary producers such as plant or algae, or decomposers. Both C and N stable isotope ratios are commonly used in combination to depict energy sources and transfers within food webs (e.g., Kostecki et al., 2010; Kostecki et al., 2012; Teichert et al., 2022a; Teichert et al., 2022b). In aquatic systems, δ¹³C indirectly informs on salinity variations, with marine ecosystems being ¹³C enriched compared to freshwater ones (Peterson and Fry, 1987; Hobson, 1999; Herzka, 2005; Reis-Santos et al., 2015). Thus, a diadromous fish entering a river after a significant feeding period in the ocean has the typical ¹³C enriched signature of the marine environment (Hobson, 1999). Comparatively, freshwater river resident fish can display a large range of ¹³C-depleted values compared to marine fish, depending on the origin of items they feed on, from the riparian ecosystem (leaf litter, terrestrial plants, or invertebrates) to aquatic primary producers (e.g., algae, plants, and microphytobenthos), invertebrates or fish. When studying the isotopic niche of a consumer, one can sample all the preys and alimentary items and then run an isotopic analysis to decipher which preys and items are more specifically consumed (e.g., Kostecki et al., 2010). This is nevertheless subjected to individual variations in life stage, size, and sexes and also to previously encountered habitats for migrating individuals (Saboret and Ingram, 2019). Therefore, instead of using the exact prey and alimentary items consumed by fishes, one can focus on the basal carbon and nitrogen sources of the isotopic ratio of the consumer (e.g., Lafage et al., 2021). This approach is particularly adapted to diadromous fish species with complex life histories, as they use different trophic strategies and habitats in terms of salinity.

In this study, we analysed the interspecific competition through niche partitioning for trophic and habitat resources between diadromous species in the sea–estuary–freshwater continuum on the Sélune River (Normandy, France). This continuum is obstructed by two large electric dams, which reduce the watercourse accessibility of diadromous fish to 12 km in the downstream part of the river. We hypothesize that an accumulation of different life stages of diadromous fish below a dam is likely to enhance intra- and interspecific competition for food and habitats, increasing niche overlaps and promoting generalization for abundant species and specialization for rare ones. We measured stable C and N isotope values of basal sources (particular organic matter, microphytobenthos and biofilm, and plants) along the continuum and compared the values with those of juveniles to adults of European eels (Anguilla anguilla), salmonids (Salmo salar and Salmo trutta), lampreys (Lampetra fluviatilis and Petromyzon marinus), European flounders (Platichthys flesus), and Mugilidae caught in the same study site. We ran Bayesian simulations (mixing models and ellipse areas) on isotopic data to infer feeding habitat, trophic position, niche width, and overlaps between the species by life stages.

The various diadromous species studied here are all simultaneously present in the Sélune River where they are confined to the 12-km downstream reach below the dam and estuary. Although they do not arrive, colonize, and reproduce at the same time, they all co-occur at some point for their development, growth, or migration for reproduction. The Sélune River and estuarine system is thus a peculiar aquatic ecosystem, where several diadromous species reside for some times: one is critically endangered (A. anguilla), another is vulnerable (S. salar), and others are least concern (S. trutta, P. flesus, Mugilidae, P. marinus, and L. fluviatilis), as classified by the International Union for Conservation of Nature (IUCN) red list of threatened species in Europe (IUCN, 2023). This study provides the first isotopic analysis by means of basal carbon sources to infer the trophic ecology and habitat use of several diadromous species living in the same river. Basal carbon sources are not directly assimilated by the diadromous species studied here, but they are transferred along the food chains from primary consumers to fish. Focusing on the basal carbon sources instead of the diet allowed us to infer and compare the niche partitioning of these species, although they feed on different sources of prey.

Results confirm that sub-adults of A. anguilla (Cucherousset et al., 2011), juveniles of P. marinus (Quintella et al., 2003), and S. trutta (Michel and Oberdorff, 1995) exploit mainly the freshwater, or the marine part for adults of L. fluviatilis (Kelly and King, 2001), while some others, such as juveniles of P. flesus and Mugilidae (Lafage et al., 2021; Teichert et al., 2022b), use all habitats of the land–sea continuum. Several diadromous adults do not feed when they migrate in the Sélune aquatic system. This is the case of P. marinus and L. fluviatilis, which feed as external parasites of marine fishes and cetaceans, but also of S. salar 1SW and 2SW, spending respectively 1 and 2 years at sea to grow before migrating in the Sélune for reproduction. S. salar 1SW and 2SW were not included in the Bayesian models because they do not feed near the environment of the Sélune aquatic system (Keith et al., 2020), as opposed to L. fluviatilis, whose hosts are thought to live in marine habitats close to the Sélune River (Elliott et al., 2021). Indeed, S. salar 1SW feed and grow in Greenland seas, and S. salar 2SW in North Ireland and England seas (MacKenzie et al., 2012). In the same way, A. anguilla fast on their spawning migration from freshwater to the Sargasso Sea (Wright et al., 2022). However, the δ¹³C and δ¹⁵N isotopic signatures remain stable for approximately 1 month in their muscles (Buchheister and Latour, 2010) and therefore reflect the eel diet in the freshwater growth habitat they came from.

Niche partitioning between these diadromous fishes also occurs through the distinct use of basal carbon sources, but this depends on the habitat used. Indeed, sub-adults of A. anguilla and juveniles of S. trutta, P. flesus, and Mugilidae all use MPBB as their primary source of carbon in freshwater. In this habitat, juveniles of P. marinus equally relied on MPBB and POM as basal sources of carbon, but the small sample size for this species did not allow us to clearly infer their basal carbon source. In the marine habitat, diadromous basal carbon sources are more diverse, ranging from POM for adult L. fluviatilis, to POM, MPBB, and PLA for juveniles of P. flesus and Mugilidae even though MPBB is the main source of basal carbon for the later. Competition over basal carbon sources is thus more likely in the freshwater part of the land–sea continuum, and from our results, we identified three factors that could exacerbate this competition. Firstly, we found that diadromous species in the Sélune River watershed and Mont-Saint-Michel Bay depend more on the freshwater than on the marine habitats, and secondly, they all rely on the same basal carbon source (i.e., MPBB) in the freshwater habitat. Thirdly, these diadromous species, exploiting the freshwater habitat of the land–sea continuum, are physically coerced by the dams, particularly species that would normally colonize upper parts of rivers (i.e., A. anguilla, L. fluviatilis, P. marinus, S. salar, and S. trutta). These factors certainly play a role in increasing both intra- and inter-species competition between these diadromous when co-occurring in the same river. By contrast, we found that juveniles of P. flesus and Mugilidae exploit both the marine and freshwater parts of the land–sea continuum of the Sélune River watershed and Mont-Saint-Michel Bay, and they developed different strategies to exploit different sources of basal carbon. This could be a strategy to relax the likely competition both within and between these species. Niche partitioning through space is widely described in the literature, and strategies to minimize competition take place not only between but also within species (Sexton et al., 2017).

Nonetheless, the wider picture of potential competition within and between these diadromous fishes is not easily retrieved from their use of habitats (freshwater, estuarine, and marine) and carbon sources (POM, MPBB, and plants). Niche overlap assessment is thus essential to estimate the competition between diadromous species. Our results showed that niche overlaps occurred between diadromous, despite the use of indirect sources of carbon (we concentrated on basal primary producers as basal sources). This means that any basal source compartment alteration could impact diadromous fish and possibly the competition between each other. Some diadromous species such as A. anguilla (sub-adults) and P. flesus (juveniles) overlapped other diadromous species’ niches, more than they were overlapped by P. marinus and Mugilidae, both as juveniles. This is mainly because they have a wider niche than the impacted species. The dominance of A. anguilla in this system, in terms of both niche use and width, is congruent with other studies (Cucherousset et al., 2011; Denis et al., 2022) and explains why this euryecious and opportunistic species was classified, in France until 1984, as a pest in salmonid producing rivers before their decline (CSP, 1998). In A. anguilla, some individuals are more specialized in fish and others in invertebrates (Cucherousset et al., 2011). This translates not only in morphological differences but also in distinct habitat niche use in a river, with individuals of the former being located further from the riverbank (Cucherousset et al., 2011). A similar pattern of specialization is observed for P. flesus where some individuals exploit different habitats of the land–sea continuum (Teichert et al., 2022b) and different sources of basal carbon, albeit no morphological differences have been described in this species. Globally at the scale of the studied area, A. anguilla and P. flesus appeared to be more generalist than other diadromous species, as individuals strategically exploit different sources of carbon and/or habitats within the same watershed. The other diadromous fishes; juveniles of S. trutta, P. marinus, and Mugilidae; and adults of L. fluviatilis appeared more specialized in their habitat use and basal source of carbon.

Intraspecific competition is one of the main drivers of inter-individual variation (individual specialization; Bolnick et al., 2003), with theory predicting an increase in phenotypic variation leading to population niche expansion as competition increases (Van Valen, 1965; Bolnick, 2001; Svanbäck and Bolnick, 2005; Sexton et al., 2017). Phenotypic variation in the head morphology of A. anguilla has been demonstrated according to their specialization degree in fishes or invertebrates (Cucherousset et al., 2011). This suggests that A. anguilla presents some level of intraspecific competition in the Sélune watershed, as demonstrated by their wide niche. This idea is reinforced by the presence of dams obstructing the Sélune watershed over more than a century; the La-Roche-qui-Boit and Vezins dams were built in 1914 and 1927, respectively. The concentration of different diadromous species on an obstructed catchment is expected to lead to co-occurrence of species that do not co-occur in an open watershed and vice versa because of alteration/delay of the species-specific patterns of migration. Patterns of migration depend on the local environmental conditions experienced by individuals, and intraspecific competition will decrease the availability of individual resources. This will in turn increase interspecific competition over the same resources exploited by different species.

In the presence of interspecific competition, specialization and therefore niche width reduction are favoured when species are locally rare, while generalization is favoured when species are locally abundant (relative to other species) (Sargent and Otto, 2006; Sexton et al., 2017). How niches are partitioned, how intraspecific and interspecific competitions impact them, and whether abundant and rare species differ in their niche width are critical questions, as these will determine their ability to cope with changes (global change and/or anthropogenic pressures) (Colles et al., 2009; Sexton et al., 2017). There are concerns and signals of the global decline of specialist species (Clavel et al., 2011), and some generalist species such as A. anguilla are already critically endangered. In the Sélune River, dam removal is expected to relax the likely competition within and between diadromous species as densities downstream of the dams will decrease, rendering novel habitats upstream of the dams accessible (Merg et al., 2020). This may increase the niche width of diadromous species relaxing their local densities and potentially altering their generalist/specialist feeding strategy. Thus, restoration actions of the watersheds used by diadromous fish, such as dam removal, improved connectivity, and promotion of habitat diversity, are evidently needed for their conservation.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding authors.

Ethical approval was not required for the study involving animals in accordance with the local legislation and institutional requirements because tissue samples were retrieved from collections.

AL: conceptualization, data curation, validation, formal analysis, methodology, and writing—original draft, review and editing. AC: conceptualization, methodology, funding acquisition, supervision, and writing—review and editing. J-MR: conceptualization, validation, formal analysis, methodology, and writing—review and editing. NT: methodology, formal analysis, and writing—review and editing. EF: conceptualization, funding acquisition, and writing—review and editing. AA: conceptualization and writing—review and editing. All authors contributed to the article and approved the submitted version.