Trophic Patterns of Freshwater Fish across the Balkan Biodiversity Hotspot

Abstract

Ecological information regarding trophic patterns and the inherent structure of freshwater aquatic communities is considered fundamental in ecological research. In this study, the trophic patterns of the Greek freshwater fish fauna, belonging to the Balkan biodiversity hotspot, were investigated, and comparisons of freshwater fish species feeding habits among the Greek freshwater ecoregions were performed. The analyses based on the widely used trophic level index and the available composition feeding datasets, along with the utilization of clustering methods, indicated seven major distinct trophic guilds in the Greek ecoregions. The trophic level of the studied fish species ranged from 2.0 to 4.5, thus being within the expected values for freshwater ecosystems, revealing the presence of both top predators and primary consumers. The highest trophic level values were found in the ecoregions of northern Greece. The results also exhibited considerable predominance of higher-trophic-level zooplanktivorous and insectivorous freshwater fish species in the mainland, compared to lower-trophic-level opportunist species in the island-isolated ecoregions. These results could be used for the application of ecosystem-based models and the formulation of conservation and fishery management schemes.

1. Introduction

Freshwaters are considered the most vulnerable aquatic ecosystems, facing threats such as water pollution and overexploitation, flow modification. and invasive species [1,2,3] that pose immense pressure to their biodiversity. The current status of these ecosystems and their provided ecosystem services is rather alarming [4,5], whereby nearly one-third of all freshwater fish species that have been assessed for their conservation status by the International Union for Conservation of Nature (i.e., IUCN) are threatened with extinction [6].

From a historical perspective, the antagonistic behavior among organisms has been recognized since the 19th century by Darwin in the fundamental work “On the Origin of Species”, referred to as the struggle of existence [7]. Observations related to resource competition and habitat diversity have stimulated nature scientists into exploring the underlying mechanisms and patterns of trophic dynamics. In addition to the intra/inter-taxa differences in behavior, the dynamics of energy and matter flow within the aquatic ecosystems are highly affected by the structure of the trophic interactions, which can impact the trophic state [8,9,10]. Trophodynamics can affect, among others, vegetation cover, diseases, erosion, and biochemical and hydrological cycles [11]. Hence, there is an urgent need for studies concerning the trophic structure of biological communities.

Trophic guilds stand out among the most widely used functional guilds, due to their ability to “describe” an ecosystem’s clear response to environmental changes, and the fact that changes in trophic interactions and food web disruptions can cause significant quantitative and qualitative ecosystem changes. For example, the shift in a freshwater fish community from piscivorous prevalence to omnivorous domination (mainly Cyprinidae species) may be indicative of the system’s eutrophication [12], as cyprinids prefer eutrophic waters [13,14]. Under the framework of operational research, information regarding trophic guilds and the interactions among them is deemed as fundamental for the exploration of aquatic food webs and their inherent structure, while it can be utilized for formulating restoration and conservation strategies [15,16,17], as well as for ecosystem-based modeling [18,19].

In terms of European freshwater ichthyofauna, numerous studies dealt with functional guilds in general, including trophic ones [20,21]. However, these studies mainly concerned northern and central Europe [22], with the Balkan subregion of the Mediterranean hotspot for freshwater fish biodiversity [23] being largely understudied. Specifically, studies on the trophic guilds of the freshwater ichthyofauna of Greece or the wider Balkan area are lacking, with previous studies focusing on the diet of individual species [24,25,26] or on fish species interactions in aquatic ecosystems [27,28,29]. Greek freshwaters have distinctive properties that both underpin and challenge ecologists, since they are characterized by rich, highly endemic biodiversity, as a result of vast geological processes in the Balkan Peninsula and the isolation of the highly fragmented hydrographic networks [23,30,31].

Since species traits are necessary to facilitate studies of ecology, conservation, and fishery management, we conducted an extensive analysis using the available information on the trophic patterns of freshwater fish species of Greece. In addition to constituting the first step toward the elucidation of traits of the Greek freshwater fish, the present work attempts to provide baseline information regarding species trophic guilds and ecology, as well as to interpret the underlying biogeographical mechanisms. Thus, the present study can function as a basis for further scientific and administrative initiatives, aimed at the development of freshwater resources sustainable management and conservation schemes.

2. Materials and Methods

2.1. Data Collection

Data were collected through an extensive literature survey on the feeding habits and diet composition of fish species from the freshwater aquatic ecoregions of Greece. Initially, species trophic level (τ) was retrieved from the Fishbase database [32]. In addition, Scopus, Web of Science, and Google Scholar databases were explored using a Boolean search with the following combinations of keywords: Greece AND freshwater fish AND diet; Greece AND aquatic AND diet; Balkans AND fish AND diet; Balkans AND freshwater fish AND diet. Whenever trophic level or diet composition information was lacking, we used information for the same species in adjacent regions. Additionally, information was collected from gray literature (e.g., conference papers, academic theses, technical reports). Species and family names, along with their spatial allocation, were assigned according to previous works where the authors proceeded in an extensive coverage of all the types of aquatic ecosystems of Greece [33,34]. Each species was categorized to the Greek ecoregions on the basis of the biogeographical scheme proposed by Zogaris and Economou [35]. According to the aforementioned study, the Greek ichthyological ecoregions are (1) Thrace, (2) Macedonia-Thessaly, (3) Southeastern Adriatic, (4) Western Aegean, (5) Ionian, (6) Crete, (7) Eastern Aegean, and (8) Southern Anatolia. The Greek area of the latter ecoregion comprises Kastellorizo Island and its surrounding islets that are lacking permanent streams, wetlands, and native freshwater fish species; thus, it was excluded from the analyses.

2.2. Data Analysis

Under the framework of a consistent methodological protocol, the trophic level (τ) of the species was derived from the Fishbase database according to the following equation [36]:

$${\mathsf{\tau }}_i=1+{\displaystyle \sum }_{j=1}^{G}D{C}_{ij}{\mathsf{\tau }}_j,$$

where τ_j is the trophic level of prey j, DC_ij represents the proportion of j in the diet of species i, and G is the total number of preys [37]. Subsequently, trophic pattern analysis took place with the visualization of the mean τ-value (τ_mean) of each family connected to each ecoregion in the form of a visualized geometric rectangle (τ_mean step = 1). To achieve a better understanding of the species trophic patterns, we estimated the trophic guilds following the work of Pereira et al. [38].

A complete quantitative mass (w/w %) prey composition dataset was obtained for 60 species with data gathered in the bibliographic survey. Thus, the identified preys of the dataset were sorted into the following groups: detritus, algae, other plant material (e.g., macrophytes), aquatic insects (pupa and larvae), terrestrial insects, microcrustaceans (cladocerans, copepods, ostracods), other invertebrates (e.g., mollusks, worms), and nekton (fish). Initially, the data were scaled through square-root transformation, which significantly reduced skewness, and then ordinated through detrended correspondence analysis (DCA) [38,39,40]. This method has been proven to be one of the most efficient discriminating analyses in producing clusters with respect to habitats [39,41]. More specifically, the initial dataset was sorted in a matrix (n, p) with the included values of prey types j (p) in species i (n) being a_ij, and an arbitrarily chosen score (x_i) for each species. Hence, with the use of the following equations, scores for each taxon were calculated as weighted averages (y_i; Equation (1)), and then a new set of species scores (x_i; Equation (2)) was produced [42]:

$$y_j=\frac{{\sum }_{i=1}^n{a}_{ij}{x}_i}{a_{+j}},$$

(1)

$$x_i=\frac{{\sum }_{j=1}^p{a}_{ij}{y}_j}{a_{i+}}.$$

(2)

The x_i scores for each sample were centered and standardized so that their mean was equal to zero (Equation (3)) and their variance was equal to one (Equation (4)) with the procedure applied progressively until the scores were stabilized.

$${\displaystyle \sum }_{i=1}^n{a}_i+x_i=0.$$

(3)

$${\displaystyle \sum }_{i=1}^n{a}_i+x_i^2=1.$$

(4)

Moreover, a final detrending ordination was applied with the use of sliding moving-average window algorithm, while the rescaling of the axis was accomplished by equalizing the weighted variance of y_j scores along the axis segments [42]. Subsequently, the scores of the first two generated axes (DCA₁–DCA₂) created by the DCA were selected for clustering via the unsupervised k-means method, operated on 999 randomized iterations. The number of the optimal clusters was determined with the “average silhouette” and the “gap statistic” method operated through 999 Monte Carlo simulations. According to the silhouette routine, the optimal number of clusters during k-means clustering is estimated when the score of the average silhouette width is maximized [43]. The gap statistic method was used to validate the results of the silhouette routine, where the optimal number of clusters (k) is selected as the minimum point estimation ($\widehat{k}$) where the gap statistic is greater than or equal to the gap statistic of its higher by one (k + 1) cluster minus its standard error (S_k+1) deviation [44], where $\widehat{k}$ can be identified as the cluster point where the successively increase rate (slope) of the gap statistic is significantly minimized among the cluster populations.

$$\widehat{k}=smallest k such that Ga{p}_{\left(k\right)}\ge Ga{p}_{\left(k+1\right)}-S_{k+1}.$$

Each trophic guild was named according to the food source type that appeared to exhibit the highest contribution to species diet. Finally, the guilds were aggregated on the basis of the food composition percentage and their distribution in the Greek ecoregions. All analyses were performed in the R programming environment [45].

3. Results

3.1. Trophic Level

Overall, the bibliographic survey yielded a dataset of 157 fish species of the Greek freshwater ichthyofauna, which belonged to 26 families. Cyprinidae was the most abundant family with the widest distribution. The species trophic level ± s.e. spanned values from 2.0 ± 0.00 to 4.5 ± 0.80 with an estimated mean trophic level (τ_mean) equal to 3.16. The most widespread species with the lowest trophic level were Ctenopharyngodon idella, Hypophthalmicthys molitrix, and Poecilia latipinna, while the species with the highest trophic level were Leuciscus aspius, Huso huso, and Perca fluviatilis. The greatest diversity regarding species richness was observed in the ecoregions of Thrace, Macedonia-Thessaly, and the Ionian. The τ_mean of the ecoregions ranged from 3.24 (Thrace) to 3.01 (Eastern Aegean) (

Table 1. Mean trophic level (τ_mean) values and the total number of species (N) of each Greek ecoregion, named according to Zogaris and Economou [35].

ID	Ecoregion	τmean	N
1	Thrace	3.24	67
2	Macedonia-Thessaly	3.23	58
3	Southeastern Adriatic	3.15	26
4	Western Aegean	3.1	38
5	Ionian	3.15	57
6	Crete	3.12	12
7	Eastern Aegean	3.01	17

), while the global mean trophic level for all Greek ecoregions was estimated at 3.14.

The hierarchical clustering of the species τ-value, as a function of the Euclidean and the Manhattan distance, resulted in the same pattern. As a result, the ecoregions of Thrace and Macedonia-Thessaly in northeastern and northern Greece formed a unique trophic–ecoregional cluster (C1, Figure 1), which actually hosted all the different families that constituted the aquatic freshwater fish diversity of Greece, except for Valenciidae, Coregonidae, and Cichlidae.

In addition, the Ionian ecoregion of Greece, located on the southwest side of the Balkan Peninsula, appeared to be quite species-distinct and isolated as a part of the second trophic cluster, while it appeared to function as an intermediary node between C1 and C2 (Figure 1). The hierarchical algorithm aggregated the ecoregions of Crete and Eastern Aegean, which are on the southeast Greek Island region, in one distinct subcluster of C2. Families Cyprinidae and Salmonidae demonstrated the largest heterogeneity regarding their species τ-values among the different ecoregions, while Cyprinidae and Anguillidae (i.e., Anguilla anguilla) appeared to be the two most widespread taxa (Figure 2).

3.2. Trophic Guilds

Prey composition was analyzed by DCA for the 60 species with diet data available. The scores of the DCA₁ (56.8%) and DCA₂ (34.9%) of the analysis that were imported into the k-means unsupervised algorithm resulted in the identification of seven distinct trophic guild groups in the Greek freshwater fish species (G1–G7, Figure 3a). The optimal number of clusters ($\widehat{k}$= 7;

Table 2. Cluster metrics obtained from the average silhouette and the gap statistic method via 999 Monte Carlo simulations for the determination of the optimal number of clusters ($\widehat{k}$ = 7) in the Greek freshwater icthyofauna based on their diet composition.

Number of Optimal Clusters ($\widehat{k}$)	Method
	Average Silhouette Width	Gap Statistic (GS)	Increase Rate of GS
1min	0	0.141	20.9%
2	0.192	0.178	34.8%
3	0.289	0.273	22.1%
4	0.334	0.350	6.3%
5	0.339	0.374	14.4%
6	0.390	0.436	11.8%
7	0.403	0.495	−2.4%
8	0.340	0.483	4.3%
9	0.352	0.505	4.9%
10max	0.351	0.531

) was estimated and validated with the use of the average silhouette width (width_max = 0.403) and the gap statistic estimation (GS; GS = 0.495, GS rate_min = −2.4%).

The ratio of the between-cluster sum of squares and the total sum of squares was estimated at 91.7%, indicating increased internal cohesion and external separation of the allocated clustered guilds. The group related to the centroid of the scaled ordination analysis (Figure 3a; red circle) was G1, while the most isolated groups were G3 and G6.

More specifically, the “piscivorous” trophic guild group (G3, N = 7) appeared to be the one with species that demonstrated almost exclusive prey preference in nekton (fish) sources of food (94.1%,

Table 3. Mean value of quantitative mass prey contribution (w/w %) to the diet of each functional trophic guild group of the Greek freshwater fish species (G1: zooplanktivorous; G2: opportunistic piscivorous; G3: piscivorous G4: insectivorous; G5: opportunistic zooplanktivorous; G6: herbivorous; G7: opportunistic detritivorous). The asterisk (*) denotes the most abundant preys of food in each group.

Trophic Guild	Algae	Detritus	Aquatic Insects	Terrestrial Insects	Micro-Crustaceans	Other Invertebrates	Nekton	Plant Matter
G1	1.09	5.45	12.59	2.71	57.10 *	17.71	0.78	2.54
G2	0.00	0.00	15.48	7.62	24.29 *	5.23	36.70 *	10.65
G3	0.00	0.00	1.25	3.17	0.18	0.75	94.06 *	0.02
G4	2.86	0.00	65.09 *	0.00	7.34	22.71	0.66	1.30
G5	9.09	35.15 *	0.30	0.70	38.63 *	5.18	0.00	10.10
G6	79.15 *	20.85	0.00	0.00	0.00	0.00	0.00	0.00
G7	0.22	29.85 *	3.02	27.10 *	15.76	1.19	3.17	19.66

). The sixth trophic guild group (G6, N = 4), which was the subsequent most isolated according to the DCA ordination, included species that demonstrated a high feeding preference in algae (79.1%) and, thus, was named the “herbivorous” group. Another two quite distinct trophic guild groups were the fourth (G4, N = 12), which showed main preference for aquatic insects (65.1%), and the seventh (G7, N = 7), which exhibited preference in detritus (29.8%) and terrestrial insects (27.1%). Thus, the G4 and G7 trophic guild groups were named “insectivorous” and “opportunistic detritivorous”, respectively. The second “opportunistic piscivorous” group (G2, N = 6) included species which fed mainly with nekton (36.7%) and microcrustaceans (24.3%). Lastly, the fifth trophic guild group G5 (N = 11) included species that fed with microcrustaceans (38.6%) and detritus (35.1%), while G1 (N = 13) exhibited a specialized preference in microcrustaceans (57.1%). Therefore, G5 was described as “opportunistic zooplanktivorous” and G1 was described as “zooplanktivorous”.

According to the ordination map of the prey resources (Figure 3b), it appears that the only preys that were grouped together were the generic invertebrates (e.g., mollusks and worms) and the aquatic insects. Moreover, according to their location, the analysis raised the narrow linking between the prey types of plant matter and terrestrial insects, along with algae and detritus (Figure 3b). The nekton food resources point of ordination was isolated as a cluster of distinct preference, while microcrustacean prey was characterized as the centroid point of the ordination analysis.

The relation of the Greek ecoregions to each functional trophic guild revealed that the ecoregion of Thrace was dominated by the zooplanktivorous species (G1 and G5), with the other trophic guilds following a descending order (Figure 4). Similarly, the ecoregion of Macedonia-Thessaly displayed a similar pattern where zooplanktivorous species (G1 and G5) were dominant, while insectivorous species (G4) also exhibited a considerable presence. The southeastern Adriatic showed a slightly different composition, as it mainly hosted opportunistic zooplanktivorous (G5) and opportunistic detritivorous species (G7).

A common pattern can be observed in the ecoregions of Western Aegean and Ionian, where zooplanktivorous species were the majority, while insectivorous and opportunistic piscivorous species showed an increase in the Ionian ecoregion compared to the rest. Regarding the exclusively island ecoregions of Greece (Crete and Eastern Aegean), a common trophic pattern existed as opportunistic zooplanktivorous species were the most abundant.

Τhe piscivorous and herbivorous groups had the highest (G3; τ_mean = 4.15, τ_median = 4.1) and the lowest (G6; τ_mean =2.25, τ_median = 2.00) trophic level values (Figure 5). The groups of insectivorous (G4; τ_mean = 3.2, τ_median = 3.2) and the zooplanktivorous (G1; τ_mean = 3.33, τ_median = 3.3) species demonstrated a more balanced variation in their τ-values. Lastly, the species that belonged to the opportunistic detritivorous (G7; τ_mean = 2.94, τ_median = 2.7), opportunistic piscivorous (G2; τ_mean = 3.48, τ_median = 3.35), and opportunistic zooplanktivorous (G5; τ_mean = 2.78, τ_median = 2.8) groups were characterized by an even more centroid-alike allocation of their τ-values due to their wider spectrum of food preference.

4. Discussion

This work represents the first attempt at the analysis of the trophic patterns of the freshwater fish species of Greece, in the southern Balkan biodiversity hotspot. The range of the trophic level of the studied fish species (i.e., from 2.0 to 4.5) was within the expected values for freshwater ecosystems, revealing the presence of both top predators and primary consumers [29]. In general, at the family level, the species belonging to Siluridae, Salmonidae, Percidae, Clupeidae, and Esocidae could be characterized as the top predators, while the species of Cichlidae, Poeciliidae, Mugilidae, and Cyprinidae appeared to be the trophic baseline of the freshwater ichthyofauna.

The fact that the most abundant family with the widest distribution in Greece is Cyprinidae is in accordance with regional works that also highlighted the family’s domination in the Balkan freshwater ichthyofauna [27,46,47,48]. This predominance could be attributed to their omnivorous behavior, coupled with their competitive advantage of adapting to eutrophic and turbid waters [49]. The complex trophic structure of the Greek freshwater ichthyofauna described under the theoretical scheme of trophic guilds [50,51] resulted in the identification of seven guilds, each including species with similar food resources.

The main representative trophic guilds at a national level were the zooplanktivorous (G1) and opportunistic zooplanktivorous (G5) species, with the insectivorous (G4) guilds being dominant mainly in continental Greece (Figure 4). These results are in agreement with previous studies that confirmed the considerable abundance of omnivorous trophic guilds of species inhabiting the Mediterranean shallow freshwater ecosystems, in contrast to those in northern climates [52,53,54]. Piscivorous (G3) and herbivorous (G6) species demonstrated a distinct, specialized trophic cluster in their food preferences (Table 3), an observation which is in accordance with previous studies in subtropical regions [55,56].

With reference to the biogeographical scheme, the ecoregions of Thrace, Macedonia-Thessaly, and the Ionian exhibited the highest species richness; a finding supported by the fact that they are the largest regions that include the majority of the aquatic ecosystems in Greece [33,46]. Our analysis revealed that the ecoregions of Thrace and Macedonia-Thessaly (northern Greece) were both dominated by zooplanktivorous, insectivorous, and opportunistic zooplanktivorous species (G1, G4, and G5; Figure 4). The latter guild, which exhibited main preference in microcrustacean and detritus species, demonstrated a considerable dominance in the continental and island Greek regions (Figure 4). Accordingly, especially in meso/eutrophic basins (e.g., Greece: Lake Volvi [27]), detritus is considered as a food source that can be consumed accidentally, while it is available during all seasons [57,58]. Its consumption by opportunistic species (e.g., G7; τ_mean = 2.94; Table 3) can cause negative ecosystem impacts, since it can potentially enhance the decrease in zooplankton communities, trigger the development of algal populations, and eventually result in the emergence of eutrophication events [54].

The Ionian ecoregion (i.e., Western Greece) did not share the same features with the ecoregions mentioned above. This ecoregion is characterized by promontories, karstic features, canyons, and numerous small lentic and small-sized lotic basins, offering the ability of isolation and the capacity to accumulate and host a unique biodiversity [46] and, thus, a variety of feeding habits and strategies. The relative predominance of opportunistic piscivorous (G2), opportunistic zooplanktivorous (G5), and insectivorous (G4) species was an interesting observation in Western Greece (Figure 4). In conjunction with the first published data on the Albanian (Adriatic basin) freshwater fish species [59], there are species similarities with the Ionian ecoregion, indicating the considerable presence of zooplanktivorous (G1 and G5) and insectivorous (G4) species in the mountainous freshwaters of the southwestern Balkans. The major west–east split of Greece has been well documented in many studies and is able to describe the vast compositional dissimilarity of the Balkan biota [31,60], thus providing further support to our findings on the distribution and the grouping of species based on their trophic level indices.

The southeastern Adriatic ecoregion stands at the junction of the two abovementioned major groupings of the regions (western and eastern) and shows the highest level of distinctiveness, hosting endemic ichthyofauna with strong affinities to the Dessaretes lake group fauna (i.e., Ohrid, Prespa, and Mikri Prespa Lakes [61]). This compositional dissimilarity is linked to the dispersal limitation of species, the complex geological history of the area, and the high rates of extinction [60]. It is the smallest ecoregion in our study, and it is part of the transboundary system between North Macedonia and Albania [62], with two drainages in the Greek territory (Aoos River and Prespa Lakes). According to the results of this study, the two main trophic guilds of the southeastern Adriatic were the opportunistic zooplanktivorous (G5) and the opportunistic detritivorous (G7) species. The preference in nekton exhibited by the opportunistic piscivorous species in the Greek region of this ecoregion can be explained since the Prespa Lakes represent a large lentic ecosystem that can adequately host fish predators in terms of spatial and temporal prey abundance.

The Western Aegean ecoregion is severely impacted by high anthropogenic pressure and presents a depauperate freshwater ichthyofauna, compared to the other continental ecoregions, despite showing affinities to the Macedonia-Thessaly freshwater ecoregion. One important observation is that dominant species belong to the opportunistic zooplanktivorous guild (G5), which feeds mainly on microcrustacean and detritus prey, exhibited a proportional increased presence in regions with an increased level of isolation, similar to the (partially) Western Aegean and to the (exclusively) Crete–Eastern Aegean Island Greek ecoregions (Figure 4). An explanation could be that the predominance of generalist opportunistic species, which is greater in hydrologically variable environments [63], occurs due to the exploitation of the high nutritional detritus, in aquatic regions where resource input is limited [64].

The island freshwater ecoregions (i.e., Crete and Eastern Aegean) host very few native fish species, mainly secondary (of marine origin) or diadromous species (e.g., Anguilla anguilla). It is possible, however, that the aquatic biodiversity of Crete) was determined by the extirpation of several aquatic species, due to anthropogenic activities [35]. In both island ecoregions, opportunistic zooplanktivorous (G5) species form the most distributed trophic guild (Figure 4). A noteworthy observation is that the herbivorous (G6) species guild was absent from both ecoregions, while, in the Eastern Aegean ecoregion, piscivorous (G3) species were also absent. This observation can be indirectly related to the analogous relationship of species richness and total ecosystem size [65,66]. In particular, smaller, isolated areas (i.e., Eastern Aegean ecoregion) tend to provide less diverse aquatic basins (temporary and shallow habitats), which are associated with increased environmental stochasticity, resulting in spatiotemporal limitation of energy resources. Thus, the presence of species that utilize low-trophic-level prey is expected (e.g., τ_mean: Crete = 3.12; Eastern Aegean = 3.01), such as species that feed on microcrustaceans and detritus prey (G5; Figure 5), a finding confirmed in the present work.

Mediterranean aquatic ecosystems (e.g., Greek lakes) are characterized by higher insolation and increased temperature, compared to northern temperate ecosystems, which can activate the development of phytoplankton populations [67]. Additionally, an aquatic community that hosts a high density of zooplanktivorous fish can cause a decline in the zooplankton communities and, thus, an increase in phytoplankton organisms, due to the absence of predators [68]. On the basis of the assumption that an omnivorous strategy can mitigate the propagation of trophic cascades to lower trophic levels [69], the considerable dominance of opportunist species in the ecoregions of this study, especially in the isolated and urbanized/industrialized ones, may constitute a potential response to the aforementioned climatic conditions in Mediterranean basins, as well as the anthropogenic pressures.

5. Conclusions

In conclusion, the Greek freshwater ichthyofauna is characterized by a strong heterogeneity in its species composition and trophic structure, which is principally a result of biogeographical history. In general, there is a considerably higher trophic level in the freshwater fish located in northern Greece compared to the rest of the country. There are some basic trophic patterns that are consistent in the Mediterranean ecoregions and that were confirmed in the present study, such as the predominance of zooplanktivorous and insectivorous species and the lower contribution of specialists in the most geographically isolated regions. The study was limited by the lack of complete food resource datasets for all the species and by the seasonal discrepancies which shape interspecific differences in fish species as a condition factor [70,71]. Nevertheless, it can be considered as providing useful baseline information and, thus, it can be utilized to resolve further research questions, as well as for the application of ecosystem-based models and the formulation of conservation and fishery management schemes.