INTRODUCTION
The alien mytilid species, Brachidontes pharaonis, which is a Lessepsian invader, is one of the pioneer species that were introduced from the Red Sea into the Mediterranean Sea via the Suez Canal, being reported near Port Said (Egypt) just 7 years after the opening of the Suez Canal in 1869 (Fuchs, Reference Fuchs1878). This small species (max. 4 cm long) was successively reported from Lebanon (Gruvel & Moazzo, Reference Gruvel and Moazzo1931), Israel (Haas, Reference Haas and Bodenheimer1937), Sicily (Di Geronimo, Reference Di Geronimo1971), and Syria, Greece and Turkey (Kinzelbach, Reference Kinzelbach1985). It was also found in the northern Adriatic, on the Slovenian coast, where it built up a stable population (De Min & Vio, Reference De Min and Vio1997).
It is one of the six species of the family Mytilidae that were considered as aliens for the Mediterranean Sea. The other species are Musculista perfragilis (Dunker, 1857), M. senhousia (Benson in Cantor, 1842), Modiolus auriculatus (Krauss, 1848), Xenostrobus securis (Lamarck, 1819) and Septifer forskali Dunker, 1855. Some of these species (S. forskali, M. auriculatus and M. perfragilis) could have been introduced to the Mediterranean via the Suez Canal; X. securis via aquaculture activities; and M. senhousia via multiple pathways (Suez Canal and aquaculture) (see Zenetos et al., Reference Zenetos, Gofas, Russo and Templado2004). Among these species, B. pharaonis is the most successful one, abundantly and frequently occurring in the eastern and the western Mediterranean (Sicily) and having already expanded its distributional range even to the north Aegean Sea (Zenetos et al., Reference Zenetos, Gofas, Russo and Templado2004; Doğan et al., Reference Doğan, Önen and Öztürk2007). It settles preferably on the rocky mediolittoral zone and rarely on the upper infralittoral zones of the region, and formed dense populations in Israel (max. 30,000 ind m−2) (Rilov et al., Reference Rilov, Benayahu and Gasith2004), Turkey (max. 40,000 ind m−2) (Doğan et al., Reference Doğan, Özcan, Bakir and Katağan2008), Sicily (25,000 ind m−2) (Sará et al., Reference Sará, Romano and Mazzola2006), and Malta (16,550 ind m−2) (Bonnici et al., Reference Bonnici, Evans, Borg and Schembri2012). This species totally or partly eliminated the native small mytilid species Mytilaster minimus (Poli, 1795) that was known to form continuous beds in the mediolittoral zone of the Levantine Sea (Gruvel & Moazzo, Reference Gruvel and Moazzo1931).
The faunal assemblages of this invasive species have not been investigated in detail in the Mediterranean Sea. However, some papers reported species of Serpulidae (Polychaeta) (Çinar, Reference Çinar2006), Decapoda (Crustacea) (Doğan et al., Reference Doğan, Özcan, Bakir and Katağan2008), Sipuncula (Açik, Reference Açik2011), Amphipoda (Crustacea) (Bakir & Katağan, Reference Bakir and Katağan2014), Eunicidae (Polychaeta) (Kurt-Sahin & Çinar, Reference Kurt-Sahin and Çinar2017) and Rissoidea (Gastropoda) (Bitlis-Bakir & Öztürk, Reference Bitlis-Bakir and Öztürk2016) associated with the B. pharaonis beds on the Levantine coast of Turkey. Ergen & Çinar (Reference Ergen and Çinar1997) gave a list of polychaetes along the Levantine coast of Turkey that were found on a variety of habitats including the B. pharaonis beds, but they did not specifically mention which polychaete species occurred on the mussel. Bonnici et al. (Reference Bonnici, Evans, Borg and Schembri2012) studied the population of B. pharaonis and its associated fauna at one locality in the Maltese waters and reported a total of 48 zoobenthic and nine phytobenthic species on the habitat.
The present study mainly aims to investigate macro-zoobenthic community structures of the alien mussel beds, to determine environmental variables mainly responsible for shaping the species assemblages of the habitat and to find out the importance of alien species associated with the mussel beds.
MATERIALS AND METHODS
Study area and sampling
Among 55 coastal stations selected for investigating macro-zoobenthic community structures of the shallow-water benthic habitats along the Turkish Levantine coast within the framework of a TUBITAK project (104Y065), seven stations, which had large, continuous beds of the mussel Brachidontes pharaonis, were selected for the study of faunistic assemblages of the mussel beds (Figure 1). Three random replicates were taken at the lower mediolittoral zone at each station using a quadrate of 20 × 20 cm in dimension and all materials within each quadrate were carefully scraped by a spatula and immediately put in jars with 10% seawater-formalin solution. In the laboratory, material was first washed with tap water on a 0.5 mm mesh and then sorted under a stereomicroscope. Specimens belonging to each taxonomic group were put into separate vials and preserved with 70% alcohol. Specimens were then identified to species level and counted. Biomass (wet weights) of the mussel and associated fauna were determined by a balance after resting them on absorbent paper to remove water from their bodies.
Fig. 1. Map of the investigated area with the location of sampling sites.
At each station, seawater samples were taken at 0.5 mm depth to determine some main environmental variables. Temperature and the dissolved oxygen concentration were encountered in the field. Water samples for analysing salinity, pH, nitrite, nitrate, ammonium, phosphate phosphorus and silicate were poured into the bottles, frozen and then transferred to the laboratory immediately. Salinity, pH and nutrients were analysed by using the Mohr–Knudsen method, a pH meter and a spectrophotometer, respectively (Parsons et al., Reference Parsons, Maita and Lalli1984).
The specimens identified were deposited at the Museum of the Faculty of Fisheries of Ege University (ESFM), Izmir (Turkey).
Statistical analyses
A species and samples matrix (abundance) was created and then analysed to assess the structures of species assemblages of B. pharaonis beds. The dominance and frequency of each species were estimated. In all faunistic analyses, the abundances of B. pharaonis were omitted to find out the structures of associated faunal communities. The Shannon–Weaver Diversity (H’) and Pielou's Evenness (J’) Indices were used to find out the species diversity and equilibrium on the mussel beds. Cluster analysis based on the Bray–Curtis similarity index was used to delineate species assemblages in the region. Principal coordinate analysis (PCO) based on the Bray–Curtis similarity was performed to determine the environmental variables and species principally responsible for structuring species assemblages in the mussel beds. Raw data were transformed by using the transformation of log (x + 1) before the multivariate analysis. SIMPROF analysis was used to assess significant species assemblages (P < 0.05) in the area. SIMPER analysis was applied to the species assemblages to assess which species contributed most to similarity and dissimilarity of species assemblages. One-way ANOVA was used to find out if there is a significant difference in community parameters (species numbers, diversity and evenness index) among stations. Before the analysis, the normality (Kolmogorov–Smirnov) and variance homogeneity (Cochran) tests were performed on the raw data. The Pearson correlation analysis was used to determine the correlation between the community parameters (diversity index, evenness index, species richness, number of species and number of individuals) and biotic indices, and environmental parameters. All analyses were performed using the PRIMER v7 and STATISTICA 7 packages.
The ecological status of each station was calculated using the biotic indices TUBI (Çinar et al., Reference Çinar, Bakır, Öztürk, Katağan, Dağli, Açik, Doğan and Bitlis-Bakir2015) and ALEX (Çinar & Bakir, Reference Çinar and Bakir2014). In the calculation of TUBI scores, each species was assigned to one of the three ecological groups [sensitive/indifferent species (G1), tolerant species (G2) and opportunistic species (G3)], based on their sensitivity to an increasing stress gradient (see Table 1). The national database was used to determine ecological groups of each species presented by Çinar et al. (Reference Çinar, Bakır, Öztürk, Katağan, Dağli, Açik, Doğan and Bitlis-Bakir2015). The ecological groups of the species that are absent in the national database were assessed by the expert team for zoobenthos in the laboratory. The boundary of each ecological status for TUBI that was presented by Çinar et al. (Reference Çinar, Bakır, Öztürk, Katağan, Dağli, Açik, Doğan and Bitlis-Bakir2015) was used in the present study. The alien biotic index ALEX (Çinar & Bakir, Reference Çinar and Bakir2014) was used to assess ecological status of stations based on the relative abundances of zoogeographic groups, which are native species (GI), casual alien species (GII), established alien species (GIII) and invasive alien species (GIV). The ecological status of stations was classified according to the ALEX scores from bad to high.
Table 1. List of species found in association with the invasive Red Sea mussel Brachidontes pharaonis, and their maximum densities (ind m−2) at stations.
The ecological (EG) and zoogeographic groups (ZG) to which each species was assigned is also given. Ecological Groups (EG) of TUBI: G1: sensitive and indifferent species, G2: tolerant species, G3: opportunistic species. Zoological Groups (ZG) of ALEX: GI: native species, GII: casual species, GIII: established species; GIV: invasive species. *denotes alien species.
RESULTS
Density and biomass of Brachidontes pharaonis
The density and biomass of the invasive mussel Brachidontes pharaonis significantly changed among stations (ANOVA, P < 0.01). The population density of the mussel ranged from 1575 ind m−2 (K19) to 27,525 ind m−2 (K28) in samples (Figure 2). The maximum mean density (22,250 ind m−2) of the mussel was calculated at station K28, the minimum mean density (1775 ind m−2) at station K19 (Figure 2). The mean shell length of the mussel was ~16 mm at stations 1 and 5 (max. 28 mm). The biomass (wet weight) of the mussel varied among samples, with scores between 454 g m−2 (K12) and 19,088 g m−2 (K28). The mean biomass of the mussel at stations ranged from 468 g m−2 (K12) to 17,711 g m−2 (K28) (Figure 2). At station K12, specimens were mainly composed of juveniles, which is why, although the density of the mussel was relatively high (mean: almost 10,000 ind m−2), the biomass of the mussel had the lowest score at this station. This could be mainly attributed to polluted freshwater discharges near the site.
Fig. 2. Mean and maximum density and biomass of Brachidontes pharaonis at stations with + standard error (SE).
Faunal assemblages of the B. pharaonis bed
A total of 187 macro-benthic species belonging to 11 taxonomic groups were identified in B. pharaonis samples collected at seven stations (Table 1). Among the groups, Polychaeta was the most dominant group in terms of the number of species (87 species, 46% of total number of species) and the number of individuals (37% of the total number of individuals) in the assemblages, followed by Crustacea and Mollusca (Figure 3). Five species, namely Stenothoe gallensis, Spirobranchus kraussii, Mytilaster minimus, Elasmopus pocillimanus and Phascolosoma stephensoni, comprised 67% of the total number of individuals encountered on the mussel beds, with S. gallensis being the most dominant one in the area (Figure 4). Stenothoe gallensis was found only at two stations (K27 and K28) and reached a maximum density of 28,175 ind m−2 (see Table 1). Similary, P. kraussi was only found at two stations and highly dominated the associated fauna of the mussel beds at station K12 (max. density: 27,425 ind m−2). The species with the highest occurrence in samples were Pseudonereis anomala (100% of samples), P. stephensoni (81%), E. pocillimanus (76%), Syllis amica (76%) and Hyale crassipes (67%).
Fig. 3. A. Dominance of taxonomic groups in terms of the number of species, B. Dominance of taxonomic groups in terms of the number of individuals.
Fig. 4. Dominance of species associated with the Brachidontes pharaonis beds.
The mean number of species considerably varied among sampling stations, ranging from 17 (at station K19) to 39 (K28) (Figure 5). The macro-zoobenthos density ranked from 1800 ind m−2 (at station K19) to 42,550 ind m−2 (K12) among samples, with the highest mean density of 34,700 ind m−2. The biomass (wet weight) of the associated fauna varied between 4.04 g m−2 (K19) and 1503 g m−2 (K12), with the highest mean biomass of 1395 g m−2. The mean diversity index values (H’) were lower than 3 at two stations (K12 and K28), where the two most dominant species (S. kraussii and M. minimus at station K12, and S. gallensis and E. pocillimanus at station K28) comprised almost 70–75% of total number of individuals that considerably diminished the score of H’. The lowest mean evenness index value (J’ = 0.42) was calculated at station K28 and the highest (J’ = 0.82) at station K37. The differences of these community parameters among stations were statistically significant (P < 0.001).
Fig. 5. Mean number of species (S), density (N), biomass (B) of the associated fauna, and mean values of diversity (H’) and evenness (J’) indices, with + standard error (SE).
The similarity among replicates at each station was higher than 45% and the highest scores (>70%) were estimated at stations K1, K12 and K28. Two main species associations (A and B) were encountered in the area at the level of 40% and the similarity among them was statistically significant (SIMPROF analysis, P < 0.05) (Figure 6). These associations occurred at the neighbouring stations K1 and K5 (group A), and K27 and K28 (group B), and relatively different species associations were found among distant stations, K12, K19 and K37. The species responsible for the similarity in the groups and the dissimilarity between groups were indicated in Table 2. Two amphipods, Amphithoe ramondi (contribution: 14%) and Elasmopus pocillimanus (11%) contributed much to the similarity in the groups A and B, respectively. High differences in the abundance of A. ramondi, E. pocillimanus and Pseudonereis anomala resulted in the dissimilarity (67%) between the groups (see Table 2). A very low similarity (around 22%) was estimated between station K12 and the others, due to exceptionally high densities of Spirobranchus kraussii (max. 27,425 ind m−2) and Mytilaster minimus (17,250 ind m−2) at the station.
Fig. 6. Dendrogram indicating similarity and dissimilarity of replicated samples (upper graph) and stations (lower graph).
Table 2. The species (with per cent contributions) that contributed most to the species assemblages of the mussel Brachidontes pharaonis according to the SIMPER analysis.
AvA, Average Abundances.
The PCO analysis showed that the macro-zoobenthic assemblages of the mussel beds differed greatly among stations (Figure 7). The first two axes together explained almost 50% of the variability. The environmental variables which had the highest correlation with the axis PCO1 were salinity (r = 0.90), dissolved oxygen (r = 0.72) and total nitrogen (r = −0.70). Phosphate phosphorus (r = 0.44) and temperature (r = 0.41) showed high correlations with the axis PCO2.
Fig. 7. Principal coordinate analysis (PCO) ordination graph of stations and the correlation of environmental/habitat variables with PCO axes, represented by superimposed vectors. The similarity among stations was encountered using the Bray–Curtis similarity index and secondarily superimposed on the PCO plot.
Ecological status of stations
The importance of ecological groups (GI–GV) at stations is depicted in Figure 8. The sensitive (GI) and indifferent species (GII) comprised more than 50% at most stations, except for stations K5 and K12 where tolerant species (GIII) accounted for more than 50% of total specimens. The opportunist species (GIV and GV) had lowest percentages (<1%) at stations, but these species possessed 3.2% of total abundances at station K12 where Capitella teleta (max. 1350 ind m−2) and Paradella dianae (max. 1175 ind m−2) occurred in abundance. The values of TUBI ranged from 2.69 (K12) to 4.45 (K37) in samples. According to the mean values of TUBI, only K12 had a value lower than 3, which is a threshold between good and moderate ecological status. The ecological status of other stations (water bodies) was classified as good or high (K37).
Fig. 8. The mean per cent abundances of ecological groups and the mean values of TUBI (with + standard error) at stations.
Relationships between biotic and abiotic data
The relationships between the environmental variables, and the community parameters of associated fauna and TUBI is depicted in Table 3. Three correlation values were found to be statistically significant; those between biomass, and salinity (r = −0.84) and total inorganic nitrogen (r = 0.93), and that (r = −0.79) between TUBI and phosphate phosphorus. The diversity index and TUBI values were negatively correlated with the concentrations of nutrients in ambient waters. Strong and negative correlations were estimated between TUBI, and the nitrogen and phosphate phosphorus. In contrast, correlations between TUBI, and salinity (r = 0.73) and dissolved oxygen (r = 0.61) were positive and relatively strong. The number of species was negatively correlated with silica (r = −0.71) and phosphate (r = −0.49), but was positively correlated with the mussel biomass (r = 0.62).
Table 3. Relationships between environmental variables, and the community parameters of associated fauna and TUBI.
S, Number of species; N, Number of individuals; B, Biomass; H’, Shannon-Weaver's diversity index; J’, Pielou's evenness index.
Bold values were statistically significant (P < 0.05).
Alien species
A total of 21 species belonging to four taxonomic groups (Hydrozoa, Polychaeta, Crustacea and Mollusca) were determined on the invasive alien mussel. The ratio between the number of alien species and native species is 0.13 and that between the number of individuals of alien species and native species is 0.77. The alien species comprised 43% of the total number of individuals of the associated fauna. The ratio between the number of native and alien species was almost the same at stations (almost 10%), but that between the number of individuals of natives and aliens considerably varied among stations, with the highest ratio (almost 70%) being estimated at K28 (Figure 9). The correlation between the abundances of alien species and native species in samples was positive (r = 0.74) and significant (P < 0.05). The total abundances of alien species in samples were positively and significantly correlated with the density of B. pharaonis (r = 0.67, P < 0.05). A weak and insignificant positive correlation (r = 0.35, P > 0.05) was estimated between abundances of native species and the density of B. pharaonis in samples.
Fig. 9. The ratios between the number of alien and native species (upper graph), and between the number of individuals of alien and native species (lower graph) at stations.
The number of alien species varied from 1 to 6 (K12 and 27) at stations, with the highest mean value being calculated at K12 (Figure 10). The density of alien species reached up to 30,300 ind m−2 in the area. The mean density of alien species ranged from 342 ind m−2 (station K19) to 23,600 ind m−2 (K28); the mean diversity index from 0.33 (station K5) to 1.28 (K1); the mean evenness index from 0.17 (station K28) to 0.64 (K1).
Fig. 10. Mean number of alien species (S) and the density of alien species (N), and the mean diversity (H’) and evenness (J’) indices based on solely alien species, with + standard error (SE).
Three species assemblages (groups A, B and C) were found in the area based on the abundance data of alien species (Figure 11). Station K12 showed a low similarity (25%) with the groups, mainly because of the high abundances of Spirobranchus kraussi, Hydroides operculatus and Paradella dianae at this station. The species mainly responsible for the formation of these assemblages were Pseudonereis anomala (contributions: 76, 83 and 53% in groups A, B and C, respectively) and Stenothoe gallensis (contribution: 47% only in group C). The dissimilarity among the groups were mainly due to the differences in abundances of the amphipods S. gallensis and Elasmopus pectenicrus, and the polychaetes Syllis ergeni, Spirobranchus kraussii and P. anomala.
Fig. 11. Principal coordinate analysis (PCO) ordination graph of stations and the correlation of species (upper graph) and environmental/habitat variables (lower graph) with PCO axes, represented by superimposed vectors. The similarity among stations was encountered using the Bray–Curtis similarity index and secondarily superimposed on the PCO plot. Cs: C. scabridum, Ds: D. similis, Gt: G. togoensis, Ho: H. operculata, La: L. antennata, Lc: L. canariensis, Lp: L. perkinsi, Pa: P. anomala, Pp: P. paucibranchiata, Pd: P. dianae, Se: S. ergeni, Sg: S. gallensis, Sk: S. kraussii.
The PCO analysis indicated that the first two axes explained a total of 64% variation. The species strongly correlated (r> ± 0.80) with the first axis (PCO1) were S. ergeni (r = −0.97), S. kraussi (r = −0.86), Gammaropsis togoensis (r = −0.82), Hydroides elegans (r = −0.82) and H. operculatus (r = −0.82). Stenothoe gallensis was only species that showed a strong correlation (r = 0.93) with the axis PCO2. The most important environmental variables affecting the community structures of alien species in the area are depicted in Figure 11. Salinity (r = 0.90) and total inorganic nitrogen (r = −0.74) showed strong correlations (r> ± 0.70) with the first axis PCO1, and the mussel biomass (r = 0.86) and silica (r = −0.71) with the second axis PCO2. The main reasons for the unique species assemblage occurring at station K12 might be lower salinity and higher nutrients values (especially TIN). Higher wet weights of the mussel estimated at stations K28 and K27 where S. gallensis only occurred resulted in the discrimination of the group A in the PCO graph.
Among the zoogeographic groups, GI (native species) comprised more than 80% of the mean abundances at stations K1, K19, K27 and K37, which therefore had high ecological status according to the ALEX values (Figure 12). The invasive species accounted for more than 40% of total abundances of associated fauna at two stations (K12 and K28). These scores eventually increased the ALEX values that classified the ecological status of these stations as moderate or poor.
Fig. 12. The mean per cent abundances of zoogeographic groups and the mean values of ALEX (with + standard error) at stations.
The number of alien species and individuals, and diversity index and ALEX were negatively correlated with salinity but positively correlated with total inorganic nitrogen (TIN) (Table 4). The mussel biomass were positively correlated with total number of individuals of alien species and ALEX.
Table 4. Relationships between environmental variables, and the community parameters of associated alien fauna and ALEX.
S, Number of species; N, Number of individuals; B, Biomass; H’, Shannon–Weaver's diversity index; J’, Pielou's evenness index.
DISCUSSION
The invasive mussel Brachidontes pharaonis formed a dense population along the mediolittoral zone and upper infralittoral zone of the north Levantine Sea. The maximum density and biomass of the species were estimated as 22,250 ind m−2 and 19,088 wwt g m−2, respectively. Such high densities of this species were previously estimated in Israel (Rilov et al., Reference Rilov, Benayahu and Gasith2004) and Sicily (Sará et al., Reference Sará, Romano and Mazzola2006). Doğan et al. (Reference Doğan, Özcan, Bakir and Katağan2008) reported an exceptionally high density (40,000 ind m−2) of this species along the Levantine coast of Turkey, but the specimens of this species in the area were generally juveniles and densely attached themselves to thalli of Jania rubens (Linnaeus) J.V. Lamouroux, 1816 (pers. observation). The density of the mussel reached a maximum of 7000 ind m−2 in the Red Sea, a possible donor area of the species (Safriel et al., Reference Safriel, Gilboa and Felsenburg1980). It was observed in the field that this species mainly preferred horizontal rocky substrate and built up beds alone or together with other sessile species like Spirobranchus kraussi or Mytilaster minimus. Rilov et al. (Reference Rilov, Benayahu and Gasith2004) also found extensive beds of B. pharaonis on horizontal and relatively narrow rocky habitats where the Dendropoma rim was absent and algal coverage was low. The biomass (dry-weight) of B. pharaonis was previously estimated maximally as 50 g m−2 in Israel (Rilov et al., Reference Rilov, Benayahu and Gasith2004).
The invasive mussel was known to be in competition with the native mussel M. minimus. They were mixed together in a random spatial pattern, but in different abundances, depending on wave exposure (Safriel & Sasson-Frosting, Reference Safriel and Sasson-Frosting1988). Safriel & Sasson-Frosting (Reference Safriel and Sasson-Frosting1988) concluded that B. pharaonis was a typical species in sheltered areas, whereas M. minimus was common in wave-exposed sites. However, Rilov et al. (Reference Rilov, Benayahu and Gasith2004) proved this argument false and found dense settlements of B. pharaonis on wave-exposed rocks. This is true for our cases, in that continuous, large mussel beds were found at stations where strong hydro-dynamism occurred. A considerable increase was observed in the dominance of B. pharaonis along the coast of Israel (Rilov et al., Reference Rilov, Benayahu and Gasith2004). The ratio between the density of B. pharaonis and M. minimus on beach rocks changed 12-fold over a period of 25 years. The decline of M. minimus was thought to be due to the interference effect imposed by the larger B. pharaonis on M. minimus recruits (Safriel & Sasson-Frosting, Reference Safriel and Sasson-Frosting1988). In the present study, only six samples out of 21 included specimens of M. minimus, with a relatively high dominance at station K12 where the ratio between the density of B. pharaonis and M. minimus was estimated as 0.74. In other samples, densities of M. minimus can be negligible when compared with those of B. pharaonis. Why did M. minimus form dense populations at K12 and was almost absent at other stations? This question cannot be well explained by the present sampling design and the number of samples taken. However, the features that made station K12 unique among other stations were to have higher total inorganic nitrogen concentration (7.92 μg l−1, almost 10 times higher than those at other stations) and lower salinity value (38.4 PSU) due to domestic waste-water discharge near the site. It might indicate that polluted waters can act as a refuge for M. minimus and a barrier for the spread of B. pharaonis in the area. The salinity value measured at station K12 was at a manageable level both for M. minimus and B. pharaonis. However, the former species was observed to represent signs of stress only at salinities slightly above 37 PSU, whereas the latter species become stressed only at salinities over 45 PSU, which enables it to invade rocky shores even in hyper-saline environments (Sará & de Pirro, Reference Sará and de Pirro2011). The biomass of B. pharaonis was almost 20 times lower than the mean of biomass of the species in the area at the polluted station K12, where two sessile species S. kraussi and M. minimus were also dominant. It seems that B. pharaonis is not only in competition with M. minimus, but also the serpulid worm, S. kraussi in the Levantine Sea. In the SW Atlantic, it was proved in a sewage-impacted site that the massive settlement of a spionid polychaete Boccardia proboscidea Hartman, 1940 caused a drastic decline in the density of Brachidontes rodriguezii (d'Orbigny, 1842) due to the biogenic reef that suffocated the mytilid (Jaubet et al., Reference Jaubet, Garaffo, Sánchez and Elías2013).
The macro-zoobenthic fauna associated with the B. pharaonis mussel beds proved to be diversified, with a total of 187 species belonging to 11 higher taxa. Most of them belonged to polychaetes and crustaceans. Several papers (Çinar, Reference Çinar2006; Doğan et al., Reference Doğan, Özcan, Bakir and Katağan2008; Açik, Reference Açik2011; Bakir & Katağan, Reference Bakir and Katağan2014; Bitlis-Bakir & Öztürk, Reference Bitlis-Bakir and Öztürk2016; Kurt-Sahin & Çinar, Reference Kurt-Sahin and Çinar2017) indicated the species of some taxonomic groups associated with this mussel along the coasts of Turkey, but all of these papers were produced completely or partly based on the dataset presented in the present study. The associated fauna of B. pharaonis were previously mentioned in the Mediterranean Sea by Bonnici et al. (Reference Bonnici, Evans, Borg and Schembri2012) on the coast of Malta, who identified 48 invertebrate species on the mussel bed. The species richness of the mussel beds on the Turkish coast is almost four times higher than that found on the Maltese coast. The mean species number at stations ranged from 17 to 30 species in the present study, whereas Bonnici et al. (Reference Bonnici, Evans, Borg and Schembri2012) reported it between four and 15 in the Maltese waters. This big difference in the species richness between distant mussel beds might be complicated, but the structures of the beds, the sampling placement in the littoral zone, the hydrographic features of ambient waters, species interactions in beds, the differences in quadrate sizes (4 times larger in the present study) might lead to this difference. However, Bonnici et al. (Reference Bonnici, Evans, Borg and Schembri2012) stated that the faunal assemblages of the mussel beds were considerably affected by the age of the mussel bed. They also suggested that the B. pharaonis beds in Malta, which had intermediate-sized individuals, was recently settled and possessed a low proportion of dead mussel shells, resulting in lower abundance and richness of the associated fauna. Mohammed (Reference Mohammed1997) carried out a faunistic study on the B. pharaonis beds in the Great Bitter Lake in Suez Canal and reported a total of 47 species belonging to eight taxonomic groups, of which Crustacea and Mollusca comprised 46% and 41% of total abundances, respectively. He postulated that the age of mussel beds was the main factor affecting the species assemblages of the mussel. In older mussel beds, the complexity of the habitat increases as a result of larger shell lengths of the mussel that could provide more available spaces for the associated fauna and accumulate more sediments and detritus from ambient waters (Tsuchiya & Nishihira, Reference Tsuchiya and Nishihira1986; Tsuchiya & Bellan-Santini, Reference Tsuchiya and Bellan-Santini1989). The difference in the diversity of faunal assemblages of mussel beds in different areas could also be linked to the availability of taxonomic expertise on different taxonomic groups, the number of samples, the geo-morphology of shores and structural complexity of the bed (Hammond & Griffiths, Reference Hammond and Griffiths2004).
The dominant species of the B. pharaonis beds on the Turkish coast were Stenothoe gallensis, Spirobranchus kraussii, Mytilaster minimus, Elasmopus pocillimanus and Phascolosoma stephensoni that accounted for 67% of total abundances of the associated fauna. The top two dominant species (comprising 35% of total abundances), namely S. gallensis and S. kraussii, were alien species possibly introduced to the Mediterranean from the Red Sea via the Suez Canal (Çinar et al., Reference Çinar, Bilecenoğlu, Öztürk, Katağan, Yokes, Aysel, Dagli, Açik, Özcan and Erdoğan2011). Stenothoe gallensis is a small amphipod (~3 mm) that mainly occurred in the shallow-water benthic habitats (1–25 m) of the western and eastern Mediterranean Sea such as algae beds (Krapp-Schickel, Reference Krapp-Schickel1976; Kocatas & Katağan, Reference Kocatas and Katağan1978), mussel beds (Çinar et al., Reference Çinar, Katağan, Koçak, Öztürk, Ergen, Kocatas, Önen, Kirkim, Bakir, Kurt, Dagli, Açik, Doğan and Özcan2008) and Posidonia oceanica meadows (Kocatas & Katağan, Reference Kocatas and Katağan1978; Zakhama-Sraieb & Charfi-Cheikhrouha, Reference Zakhama-Sraieb and Charfi-Cheikhrouha2010). The maximum density of these species was previously estimated as 463 ind m−2 on a P. oceanica bed in Tunisian waters (Zakhama-Sraieb & Charfi-Cheikhrouha, Reference Zakhama-Sraieb and Charfi-Cheikhrouha2010) and as 400 ind m−2 on Mytilus galloprovincialis in the harbour environment of the Aegean Sea (Çinar et al., Reference Çinar, Katağan, Koçak, Öztürk, Ergen, Kocatas, Önen, Kirkim, Bakir, Kurt, Dagli, Açik, Doğan and Özcan2008), whereas it reached up to 28,175 ind m−2 on the B. pharaonis bed (this study). This species comprised almost 30% of total amphipod abundances found on Cystoseira spp. on the coast of Sicily (Krapp-Schickel, Reference Krapp-Schickel1976). The density of the invasive serpulid polychaete species S. kraussi reached 27,425 ind m−2 on the mussel bed. It occurred both on mussel shells and spaces among mussels as an erect form. This species was previously known from the area and reported to build up a belt like biogenic habitat in the low mediolittoral zone where its density and biomass was calculated as 31,375 ind m−2 and 80 wwt g m−2, respectively (Çinar, Reference Çinar2006). Its distribution within the Mediterranean is confined to the eastern Levantine Sea (Mersin and İskenderun Bays, and Israeli coast). This species was also known to be a component of fouling communities and built up a dense aggregation in the Suez Canal (35,000 ind m−2) (Emara & Belal, Reference Emara and Belal2004) and Hong Kong (30,000 ind m−2) (Jianjun & Zongguo, Reference Jianjun and Zongguo1993). In the B. pharaonis beds on the coast of Malta, Bittium spp. were the most abundant species, followed by Tanaidacea (sp.) and Rissoa sp (Bonnici et al., Reference Bonnici, Evans, Borg and Schembri2012). However, their population densities on the mussel beds from the Levantine Sea were too low (i.e. 40 ind m−2 for Bittium spp.). The most dominant species on the B. pharaonis beds in the Suez Canal were Pirenella conica (Blainville, 1829), Gammarus sp. and Amphibalanus amphitrite (Darwin, 1854), all comprising 63% of total abundance (Mohammed, Reference Mohammed1997). The similarity between these two studies performed in different regions was that the dominant species belonged to the superfamily of Gastropoda Cerithoidea. However, abundances of this superfamily that were represented by three species (Cerithidium scabridium and Bittium spp.) were negligible (<0.1%) in the present study.
The PCO analysis indicated two main species associations in the area. The environmental variables mainly responsible for these groupings were dissolved oxygen, total inorganic nitrogen and salinity. Çinar et al. (Reference Çinar, Katağan, Koçak, Öztürk, Ergen, Kocatas, Önen, Kirkim, Bakir, Kurt, Dagli, Açik, Doğan and Özcan2008) found a similar finding that nutrient concentrations in ambient waters were main drivers in structuring faunal assemblages of M. galloprovincialis in the Aegean Sea. The species contributed most to the similarity of assemblages were Amphithoe ramondi and Elasmopus pocillimanus. However, abundances of Pseudonereis anomala, Syllis amica and Phascolosoma stephensoni enabled stations to be grouped together at relatively high similarity levels. Çinar et al. (Reference Çinar, Katağan, Koçak, Öztürk, Ergen, Kocatas, Önen, Kirkim, Bakir, Kurt, Dagli, Açik, Doğan and Özcan2008) also indicated the importance of amphipods and syllids in forming different assemblages on the Mediterranean mussel. Bonnici et al. (Reference Bonnici, Evans, Borg and Schembri2012) determined three different species assemblages on the B. pharaonis beds in Malta and encountered the importance of Bittium spp., Tanaidae (sp.), Rissoa sp. in constituting these assemblages. In the present study, no Rissoa species was found on the B. pharaonis beds, but three tanaid species (Apseudes latreilli, Chondrochelia savignyi and Tanais dulongii) were encountered in the area. Among these species, A. latreilli formed a relatively high population (3625 ind m−2) at station K27 and was one of the main species responsible for the dissimilarity between the species assemblages of the B. pharaonis beds.
According to TUBI scores, the ecological status of stations was good or high, except for K12 whose ecological status was found to be moderate. As pointed out earlier, this station is located near a waste-water outfall and was mainly characterized by higher nitrogen concentration. The nutrients (especially phosphate and nitrogen) and TUBI were negatively and significantly correlated, indicating that TUBI enabled detection of deterioration in ecological functioning. This index was newly developed, based on macro-zoobenthos data from soft sediments along the Aegean and Levantine coasts of Turkey and has proved to be more effective in assessing the ecological status of water bodies than other biotic indices widely used in the Mediterranean Sea such as AMBI, M-AMBI, BENTIX and MEDOCC (Çinar et al., Reference Çinar, Bakır, Öztürk, Katağan, Dağli, Açik, Doğan and Bitlis-Bakir2015). The present study indicated that it is also a useful tool in classifying benthic quality status of hard bottom substrates.
The invasive alien mussel B. pharaonis hosted a number of alien species (a total of 21 species) with high abundances. Most of these alien species were the Lessepsian invaders, but five species, namely, Linopherus canariensis, Hydroides elegans, Pseudopolydora paucibranchiata, Paradella dianae and Conomurex persicus could have been introduced to the eastern Mediterranean by shipping. The alien species reported in the present study were known from the Turkish coasts (Çinar et al., Reference Çinar, Bilecenoğlu, Öztürk, Katağan, Yokes, Aysel, Dagli, Açik, Özcan and Erdoğan2011). However, it is the first time Syllis ergeni has been classified as alien species in the Mediterranean Sea. This species was originally described from the Aegean Sea (Çinar, Reference Çinar2005) and then subsequently reported from the Mediterranean coast of Egypt (Abd-Elnaby & San Martín, Reference Abd-Elnaby and San Martín2011) and the Red Sea (Faiza Abd-Elnaby, pers. comm.). As its distribution is only confined to the Levantine and its neighbouring sea (Aegean Sea), and it dominated shallow-water benthic habitats in the areas, this species could be an alien species introduced to the Mediterranean from the Red Sea via the Suez Canal. The alien species occurring in all samples with relatively high abundance was the nereidid polychaete worm Pseudonereis anomala. This species is known to be an invasive species dominating shallow-water benthic habitats of the Levantine (Çinar & Altun, Reference Çinar and Altun2007) as well as Aegean Seas (Çinar & Ergen, Reference Çinar and Ergen2005). This species greatly expanded its distributional range to the coast of Sicily (D'Alessandro et al., Reference D'Alessandro, Castriota, Consoli, Romeo and Andalora2016). Limited information regarding the population density and biomass of P. anomala is available. This species was reported to attain a highest density on the coast of Turkey [2475 ind m−2 (on the alga Jania rubens), 7.95 g m−2 (on B. pharaonis)] (Çinar & Altun, Reference Çinar and Altun2007). In the present study, the maximum density of it was estimated as 8925 ind m−2. This species is known to be in competition with the native nereidid species such as Perinereis cultrifera and Platynereis dumerilii (Ben-Eliahu, Reference Ben-Eliahu, Spanier, Steinberger and Luria1989).
Does high alien diversity (=xenodiversity) found on the invasive mussel bed support the invasion meltdown hypothesis? This hypothesis points out that the successful establishment of one invasive species in a new environment facilitates other alien species to invade (Simberloff & Von Holle, Reference Simberloff and Von Holle1999). The alien species found in the mussel beds were also found in different coastal habitats, but the habitat selectivity of alien species should be investigated if the mussel bed habitat favours or facilitates the establishment of any alien species in the area. Almost half of the individuals of the associated fauna belonged to alien species, indicating the importance of alien species in the benthic habitats of the Levantine Sea. In Mersin Bay, alien species comprised 31% of the total number of individuals in soft substrata (Çinar et al., Reference Çinar, Katağan, Öztürk, Dagli, Açik, Bitlis, Bakir and Doğan2012b). However, in polluted soft substrata in İzmir Bay, alien species comprised more than 75% of total individuals (Çinar et al., Reference Çinar, Katağan, Öztürk, Bakir, Dagli, Açik, Doğan and Bitlis2012a). On the Mediterranean mussel beds in the Aegean Sea, Çinar et al. (Reference Çinar, Katağan, Koçak, Öztürk, Ergen, Kocatas, Önen, Kirkim, Bakir, Kurt, Dagli, Açik, Doğan and Özcan2008) reported 10 alien species, of which Hydroides elegans and H. dianthus (Verrill, 1873) were the most dominant ones. These species accounted for up to 80% of total individuals at some stations. Three different alien species assemblages were developed in the area, mainly based on the presence/absence and abundances of Pseudonereis anomala, Spirobranchus kraussi, Paradella dianae, Syllis ergeni and Stenothoe gallensis. The nutrient concentrations, salinity and mussel biomass were the factors mainly controlling the structures of these assemblages. High dominance of alien species at samples increased the ALEX values which classified two stations as moderate (K12) or poor (K28). This index also showed a positive correlation with total inorganic nitrogen in ambient waters. Çinar et al. (Reference Çinar, Katağan, Öztürk, Bakir, Dagli, Açik, Doğan and Bitlis2012a, Reference Çinar, Katağan, Öztürk, Dagli, Açik, Bitlis, Bakir and Doğanb) also indicated the importance of nutrient concentrations in ambient waters, depth and sediment structure in the distribution of alien species in the eastern Mediterranean Sea. The abundances of alien species increased with increasing abundance of native species. The present study did not detect any competition between native and alien species. Such a tendency was also noted by Çinar et al. (Reference Çinar, Katağan, Öztürk, Bakir, Dagli, Açik, Doğan and Bitlis2012a, Reference Çinar, Katağan, Öztürk, Dagli, Açik, Bitlis, Bakir and Doğanb). However, some alien species were reported to eliminate some common opportunistic species in the polluted environments (Çinar et al., Reference Çinar, Katağan, Öztürk, Egemen, Ergen, Kocatas, Önen, Kirkim, Bakir, Kurt, Dagli, Kaymakçi, Açik, Doğan and Özcan2006).
The present study indicated that the invasive mytilid species, B. pharaonis provided shelter and food for a diversified faunal assemblage in the region. The interactions between the structures, orientations of the mussel beds and the associated fauna still remain largely unexplored. In addition, spatio-temporal variations of the faunal assemblages of the beds in larger areas merit detailed investigations to shed more light on the assessment of community structures of the invasive mussel beds.
ACKNOWLEDGEMENTS
We are much indebted to two anonymous referees for their constructive comments on the manuscript.
FINANCIAL SUPPORT
This work was financially supported by TUBITAK (Project Number: 104Y065).
