Phytoplankton distribution and variation along a freshwater-marine transition zone (Kızılırmak River) in the Black Sea

Water samples were collected monthly from 5 sites at a depth of 0.5 meter between July 2007 and December 2008 using a Hydro-Bios Free Flow Water Sampler (2.5 liters) (Figure 1). A plankton net with a 20 μm mesh size was also used to collect qualitative samples for the determination of rare species.

Figure 1

Both temperature and pH in samples were measured using a Consort C534 model analyzer. Air temperatures were provided by the Samsun Meteorological Station. Water transparency was determined by a Secchi disk. Salinity and density were determined by Eutech Cyberscan Con 11. Nutrients; nitrite nitrogen (NO₂-N), nitrate nitrogen (NO₃-N), ammonium nitrogen (NH₄-N), orthophosphate phosphorus (PO₄-P) and silica (SiO₂) were measured spectrophotometrically according to the standard methods (APHA 1995). Chlorophyll a was measured after pheophytin a correction by acidification (APHA 1995).

Microscopic observations were conducted under a Prior phase-contrast inverted microscope, a Prior phase-contrast and Nikon E600 florescence microscope. Certain diatoms like Pseudonitzshia spp. and Skeletonema spp. were identified under a transmission electron microscope (JEOL-100SX). Thecate dinoflagellates were identified under a florescence microscope. The Lugol-fixed water samples were concentrated to 1-3 ml and then counted in the Sedgewick-Rafter chamber under an inverted microscope (Guillard 1978). Phytoplankton species were classified according to the taxonomic groups. Phytoplankton taxa were identified according to the following identification keys: Round et al. (1990), Krammer & Lange-Bertalot (1991a,b; 1999a,b), Sims (1996), Hasle & Syvertsen (1997), Steidinger & Tangen (1997), Throndsen (1997), Lange-Bertalot (2000), John et al. (2003). Potential HAB species were verified with the IOC HAB list and the Manual on Harmful Algae (Hallegraeff 2004). New records of taxa from the Turkish Algal Flora were checked according to Gönülol’s (2016) website. Furthermore, valid taxa names and arrangement of taxonomic groups were determined according to the website of Guiry & Guiry (2016).

The PRIMER-E statistical package was used for ecometric analyses. A similarity matrix was defined according to the Bray-Curtis similarity method after the square-root transformation was applied on the phytoplankton abundance matrix. Agglomerative Hierarchical Cluster Analysis and an nMDS algorithm was performed and the results were plotted on the two-dimensional MDS configuration. ANOSIM, SIMPER, BVSTEP and BIOENV routines of the PRIMER-E package (Clarke & Warwick 2001) and CCA (Ter Braak 1986) were applied on the data sets to determine relationships between clusters of samples and environmental variables.

Temperature fluctuated throughout the sampling period between 4.60°C and 27.50°C (Fig. 2A). The highest temperatures were observed during summer, while the lowest in winter. The pH varied from nearly neutral (7.26 in August 2008) to slightly alkaline (9.20 in March 2008) (Fig. 2B). Water density was at higher levels in the summer period, while the lowest density was recorded in the rainy period, in winter and spring. It varied from 1.00053 g cm^-3 to 1.0132 g cm^-3 (Fig. 2C). The lowest conductivity was observed in April 2008 (1.09 mmhos cm^-1) and the highest in July 2007 (34.00 mmhos cm^-1; Fig. 2D).

The Secchi disk depth varied between 0.30 meter (N1 and N2, January 2008) and 10.50 meter (K2, August 2007) (Fig. 2E). Figure 2F shows the inorganic N:P ratio ranging from 0.13 (site A1, December 2008) to 37 (site N1, March 2008). The highest value (29.63) of the inorganic N:Si ratio (Fig. 2G) was observed at the inner river station (N1) in October 2008 while the lowest (0.13 at 10 meter depth) was determined at the coastal site (K1) in December 2008. The concentrations of Chl-a (Fig. 2H) varied between 0.19 mg m^-3 (K1, in May 2008) and 4.90 mg m^-3, N1, in August 2008). The highest values were observed at the inner river and river mouth in late summer.

Figure 2

A total of 447 taxa belonging to the divisions; Cyanobacteria (24), Bacillariophyta (209), Bigyra (1), Cercozoa (1), Charophyta (11), Chlorophyta (29), Cryptophyta (10), Miozoa (118), Euglenozoa (14), Haptophyta (13), Ochrophyta (10) and Protozoa Incertae Sedis (2) were identified in the study area. Phytoplankton composition, potential harmful species and new records for the algal flora of Turkey were given in Table 1. Seventy five of the total number of taxa were determined to be new records for the Algal Flora of Turkey and 41 were found to be HAB (Harmful Algal Bloom) organisms. Phytoplankton consisted of 52% freshwater and 48% marine species. However, 40% of the total phytoplankton was represented by euryhaline and brackish water species.

Figure 3 shows the phytoplankton variation among the sampling sites. The highest cell concentration was measured at the inner river site (N1) and three peaks were observed from May to October 2008.

Figure 3

The agglomerative hierarchical cluster analysis and the MDS plot from the surface water abundance data revealed assemblages with a stress value of 0.15 (Fig. 4 A-E). Accordingly, assemblages in the MDS plot were consistent with those resulting from hierarchical cluster analysis. Samples were divided into four groups: “Freshwater”, “Brackish”, “Marine” and “Earlyspring-Marine” at 44% similarity level. The assemblages from the MDS plots and cluster dendrograms were confirmed by the ANOSIM procedure (Global R values are 0.858 for surface groups and 0.746 for subsurface groups).

The BIOENV procedure applied to the surface phytoplankton data revealed the best correlation coefficient (0.68) not only for one environmental variable but also for the density, Secchi Disc depth, NH₃-N and silica. Figure 4(B-E) shows the effects of related parameters on the groups of samples in the MDS plot.

Figure 4

Temporal phytoplankton variation, distribution and interactions with environment were investigated at 5 sampling sites in the Kizilirmak River/Black Sea transition zone between July 2007 and December 2008.

During the sampling period, the sea and river water temperature varied between 4.60 and 27.50°C. Water temperature was significantly correlated with air temperature (4-30°C). Phytoplankton abundance increased due to the higher temperatures and peaked twice in August and September. However, some centric diatoms such as Chaetoceros socialis (5.6 × 10⁴ cells l^-1), Pontocsekiella kuetzingiana (1.28 × 10⁵ cells l^-1), C. meneghiniana (1.14 × 10⁵ cells l^-1) and P. calcar-avis (4.3 × 10⁴ cells l^-1) and coccolithophores such as Calyptosphaera globosa (6.6 × 10⁴ cells l^-), Holococcolithophora sphaeoridea (6.5 × 10⁴ cells l^-1) and Emiliana huxlei (5.5 × 10⁵ cells l^-1) occurred with their highest abundance in the cold period as compared to other studies located in the Black Sea (Türkoğlu & Koray 2002; Sorokin 2002; Baytut et al. 2010). Longterm data showed that centric taxa such as Aulacoseira and Cyclotella are dominant in the Danube river phytoplankton (Dokulil & Donabaum 2014).

Water samples in this study were consistent with the ratios declared by the European Commission and the inorganic N:P ratio varied during sampling period between 0.13 and 37.00 (European Commission 2002). Higher values were observed in the inner river (N1) and in the river-sea transition zone (N2). According to these findings, the phytoplankton production is limited by inorganic phosphorus in surface waters and by inorganic nitrogen in subsurface waters. The N:Si ratio was usually between 1 and 2 units in the healthy marine ecosystem and heterotrophic flagellates become dominant when it increases over 2 units (Roberts et al. 2003). Inorganic N:Si ratios ranged from 0.13 to 29.00 units. Diatom production was limited by higher inorganic N:Si values and the abundance of flagellates and cyanobacteria were dominant in the phytoplankton community.

Chlorophyll-a in lagoons of the Kızılırmak Delta ranged from 0.50 to18.00 mg m^-3 (Soylu & Gönülol 2006; Soylu et al. 2007; Gönülol et al. 2009). For the northwestern Black Sea, it was reported as 15.00 mg m^-3 in the 1980s and 4-5 mg m^-3 in the 1990s (Yunev et al. 2007). The results of this study revealed that chlorophyll-a concentrations were lower than those measured in the lagoons of the Kızılırmak delta but higher compared to hypereutrophic waters of the northwestern Black Sea. Chlorophyll-a values oscillated between 0.19 mg m^-3 (May 2008) and 6.90 mg m^-3 (August 2008).

Surface phytoplankton composition, especially at the inner river (N1) and at the river mouth (N2) site, consisted mainly of freshwater and euryhaline taxa. Remaining species were typically coastal marine species. The freshwater species belonged to the divisions: Cyanobacteria, Charophyta, Chlorophyta, Euglenozoa and their most abundant representatives were Pseudoanabaena catenata, Spirogyra fluviatilis, Oocycstis elliptica and Eutreptia lanowii, respectively. It was indicated that these species are common in nutrient-rich, eutrophic, polluted and shallow or slow flowing water systems (John et al. 2003).

The sampling sites in the study area showed major differences in the phytoplankton abundance. For instance, phytoplankton abundance at the inner river site (N1) exceeded 1 × 10⁶ cells l^-1 in October, May and August, while it reached only 0.5 × 10⁶ cells l^-1 at the river-sea transition zone (N2) in October, March and August. Bacillariophyta was the dominant taxonomic group in the inner river and in the transition zone. Cocconeis pediculus, Pontocsekiella kuetzingiana, C. meneghiniana, Gomphonema minutum, G. oliviaceum, Melosira varians, Navicula cryptocephala, Nitzschia spp., Pseudosolenia calcar-avis, Rhoicosphenia abbreviata, Thalassiosira antiqua and Ulnaria oxyrhyncus were the most abundant species of diatoms. Dinoflagellates, however, were absolutely dominant at the marine sites except for centric diatoms which peaked in March and May. The most abundant representatives of this group were Amphidinium crassum, Gymnodinium elongatum, Gymnodinium simplex, Gyrodinium estuariales, Heterocapsa rotundatum, Karlodinium micrum, Karenia brevis, Prorocentrum caudatum and P. micans.

Temporal changes in the abundance were usually consistent with chlorophyll-a. In certain months, however, various differences occurred during the sampling period. For instance, relatively lower chlorophyll-a values (1.20-1.36 mg m^-3) were observed at the inner river and river mouth sites (N1 and N2), while the increased values of abundance (0.6-2.3 × 10⁶ cells l^-1) were recorded in October 2007 and August 2008. The reason for this contradistinction could be the fact that benthic diatom cellswere included in the count (benthic cells may influence the phytoplankton through the water movement). Former studies in the Kizilirmak River (Dere & Sıvacı 2003; Hasbenli & Yıldız 1995; Yıldız & Özkıran 1991) reported that Amphora, Cocconeis, Cyclotella, Cymbella, Diatoma, Encyonema, Fragilaria, Gomphonema, Melosira, Navicula, Nitzschia and Rhoicosphenia were the most common taxa in the benthic community and they were frequently encountered in the inner river and in the riverine transition zone in the study area.

MDS and hierarchical cluster analyses revealed that there are four groups of samples. The results were tested and confirmed by ANOSIM in order to check the significance of differences between the groups of samples. These samples were divided into the following groups: “Freshwater”, “Brackish”, “Marine” and “Early spring-marine”. However, MDS and hierarchical cluster analyses of the subsurface phytoplankton abundance data, however, showed that the groups of samples varied according to the seasonal variation. Sample group A included the early spring samples (March and April), while group B1 comprised the late spring-early fall samples (from May to September). Group B2 included fall and winter samples (October and February) and group C comprised samples collected from July to August.

Reynolds (2006) reported that only 20-30 species succeed in the whole community even if resident taxa are represented by large numbers. Likewise, dissimilarities between groups of samples determined by the SIMPER routine and Spearman rank correlations performed on the data subsets (BVSTEP) revealed that 35 successful species were found in the phytoplankton community in spite of the 430 taxa identified in the Kizilirmak River/Black Sea transition zone. Some of them were typical eutrophic freshwater and marine diatom species in the surface water phytoplankton, except for a few eutrophic dinoflagellates. For instance, the diversity of Nitzschia species in polluted and eutrophic zones was usually at the highest levels (Petrov et al. 2010). Similarly, the Nitzchia genus was represented by 20 species in this study. In the eutrophic Varna Bay, which receives a great part of the Danube’s freshwaters, the most abundant species of the phytoplankton community were similar to those occuring in the current study area (C. socialis, E. huxleyi, P. delicatissima and P. cordatum) (Petrova et al. 2006).

The BIOENV procedure revealed the Spearman rank correlations between the Bray-Curtis similarity matrix of the abundance data and Euclidean distance matrix of environmental parameters. Thus, the surface phytoplankton assemblages varied along the salinity gradient and the Secchi disc depth. Subsurface assemblages differed due to the water temperature and the N:P ratio. Flagellates (including Dinoflagellata) were reported to be able to bloom in eutrophic coastal waters in the Black Sea (Nesterova et al. 2008). There are, however, some differences in the phytoplankton dynamics between the findings from the study area and the studies from the whole Black Sea basin. The abundance of the coccolithophore Emiliana huxleyi reached 9 × 10⁶ cells l^-1 in spring and summer. It was, however, noted in the former studies that the coccolithophores were also able to increase their abundance in winter (Krupatkina et al. 1991). Another discrepancy related to the phytoplankton community from the Kizilirmak River/Black Sea transition zone is the low abundance of a hypereutrophic water species from the genus Skeletonema which bloomed several times and caused massive fish kills due to the hypoxia in the northwestern Black Sea (Nesterova & Terenko 2007). The highest abundance of Skeletonema determined in the water samples was only 15 × 10³ cells l^-1 and this may be a result of allelopathic stress caused by the neritic toxic dinoflagellate Prorocentrum cordatum. Tameishi et al. (2009) made an abundance model between Skeletonema and P. cordatum and reported that P.cordatum allelopathically suppressed the abundance of Skeletonema. It was also reported that P. cordatum is able to form harmful toxic blooms in temperate or subtropical waters (Steidinger & Tangen 1997) and is gradually able to form dense blooms in eutrophic coastal waters (Heil et al. 2005).

A total of 41 potentially harmful algal species were identified in the phytoplankton of the study area. Some of them were among succeeding species of the phytoplankton community. The abundance of harmful taxa in the study area becomes even more important in the context of the fact that only 300 taxa among the thousands of algal species form blooms able to change the color of seawater and 80 of them are toxic (Hallegraef 2004). It was noted that these events have recently increased worldwide due to anthropogenic eutrophication and the global climate change (Hallegraeff 2004). Some taxa may be harmful even when present in small numbers in the seawater. For instance; shellfish farms and fisheries activities must be closed in many parts of the world when the abundance of Dinophysis and Phalachroma spp. exceeds 500 cells l-1 in the seawater (European Commission 2002). In this study, the abundance of the potentially harmful Dinophysis caudata reached 5200 cells l^-1 (July 2007) at the coastal sites (K1 and K2) at a depth of 10 meter. Phalachroma rotundatum reached 8000 cells l^-1 at a depth of 5 and 10 meter at the open water site (A1) in September 2008. Not only brackish and marine harmful species, but also potentially toxic freshwater species were observed in the phytoplankton community of the Kizilirmak River/Black Sea transition zone. Among them, Aphanizomenon flos-aquae, Microcystis aeruginosa, M. flos-aquae, Phormidium formosum, Planktothrix agardhii, Pseudanabaena catenata, Snowella lacustris and Trichormus variabilis were found to be toxic in the community of polluted, eutrophic freshwaters (Cronberg et al. 2004).

This study provides new contributions to the algal flora of the Turkish Seas together with additional taxonomic and ecological investigations. A total of 71 new taxa and 41 potentially harmful taxa have been identified in this study for the first time. Taxonomic, physicochemical and statistical analyses have revealed that the predominance of heterotrophic and mixotrophic species instead of the autotrophic ones, as well as the increase in number of potentially HAB species and severe eutrophication contribute to an unstable system and pose a threat to the ecosystem and human health.