Seasonal Di ﬀ erences in Taxonomic Diversity of Rotifer Communities in a Hungarian Lowland Oxbow Lake Exposed to Aquaculture E ﬄ uent

: With the intensiﬁcation of aquaculture technologies, the amount of feed input and waste material is increasing, creating potentially negative impacts on freshwater habitats receiving e ﬄ uent from such systems. Changes in biodiversity of zooplankton communities is often used to assess the e ﬀ ects of such impacts. Rotifers are suitable for bioindication of water quality due to their fast reaction to environmental changes. We examined seasonal changes in the diversity of rotifer communities along a 3.5 km section of the biggest oxbow lake in the Tisza River basin, Hungary, that received inﬂow from an intensive tank-based aquaculture farm. We detected a species-rich rotifer community with 26 species. Using the R é nyi one-parameter diversity index families, we found that biodiversity increased away from the point of inﬂow in spring, but after a summer transition period the situation became partially reversed during autumn. At the beginning of the study period, the nutrient-rich e ﬄ uent strengthened the dominance of common species, which decreased but did not disappear in summer. In autumn, the extra nutrient input delayed the decline of the rotifer community at the point of e ﬄ uent.


Introduction
With growing human population, protein production by aquaculture has greatly increased, becoming the fastest growing animal husbandry sector in the world [1]. As aquaculture production is intensifying, the amount of fish feed input and the resulting waste, including organic matter, nutrients, and suspended solids, is also increasing [2]. These create potentially negative impacts on neighbouring habitats, especially the freshwater habitats receiving outflow from such systems [3].
Wetlands in general have suffered a great reduction in Europe during the 20th century [4], and many of the remaining ones, often even in protected areas, are under encroachment from aquaculture [5]. The river regulation at our study site, the Hungarian lowland, was the biggest such enterprise in Europe during the 19th century and has profoundly modified the lowland environment and water regime [6]. The regulation decreased the meandering of the Tisza River to enhance flood management. After completion, the Tisza became 457 km shorter, but the originally ignored negative ecological impacts soon became evident [7]. An unexpected side benefit was that many of the isolated meanders became oxbow lakes and developed into valuable wetland habitats. The Kákafok oxbow lake near to the settlements of Szarvas and Békésszentandrás, formed after the regulation of River Körös (a tributary of the Tisza River), is the largest of the Tisza watershed, and also represents significant natural value. It has a significant role in the formation of the characteristic landscape of the Körös River basin, and provides different anthropogenic utilization opportunities [8], which influence the composition of the aquatic communities.
The assessment of changes in the biodiversity of impacted habitats is an indispensable part of the analysis of anthropogenic effects. A complete inventory of all elements of biodiversity is rarely feasible, thus the selection of a suitable indicator group is an important decision. In water quality assessment, several planktonic organisms have been suggested as indicator groups Rotifers have been used for water quality assessment from New Zealand [9] to Ireland [10]. Several authors have used rotifers to indicate eutrophication [11][12][13][14][15][16] and saprobity [17,18], and to develop trophic state indices [19,20]. The microscopic size of rotifers contributes to their cosmopolitan and ubiquity distribution [21,22], although at a much lower level than previously thought [23]. They also have several endemisms with narrow biogeographical distribution [24,25]. Because of this, the communities of rotifers are often included in faunistic and ecological investigations. Rotifers play an important role as the basis of the food web of aquatic ecosystems, transferring energy from lower to higher trophic levels [26]. Rotifers are amongst the fastest growing of planktonic indicator groups [27], and thus their communities react quickly to physical, chemical and biological changes in water [18,28,29].
Surveys of the zooplankton communities of the Tisza River began in the middle of the 20th century [30]. The research into the River Körös received some attention, [31,32] but the rotifers of oxbow lakes have been less investigated. A recent study [33] dealt with protected oxbow lakes nearby but left out the biggest, the Szarvas-Békésszentandrás (Kákafok) oxbow lake, because it is not under nature protection. A common feature of these studies is that they considered the individual water bodies to be homogeneous and only one or two samples per year were taken, providing information about the overall composition of the rotifer community. On the other hand, seasonal variations are also important, and these assessments did not explore the advantage that rotifers respond quickly to environmental changes.
The main goal of this article is to assess the cumulative effects of aquaculture effluent on the seasonal structure of rotifer communities in a spatially explicit manner and to describe the seasonal changes in rotifer communities along a 3.5 km long section of the Kákafok oxbow lake, at increasing distances from the point of effluent. For a synthetic analysis of diversity, we chose the Rényi one-parameter diversity index family [34]. We hypothesized that the effluent affects community composition but the effect is diluted as the distance from the inlet point increases. We expected that, under oligotrophic conditions, a generally species-rich community develops with low dominance, while there are fewer species with higher abundance under nutrient-rich conditions [35,36]. Consequently, our second hypothesis was that lower taxonomic diversity would be recorded near the inflow, while diversity would increase farther away from the point of inflow.

Study Site
The Kákafok oxbow lake located near Szarvas and Békésszentandrás (46 • 51 14.9" N, 20 • 30 44.6" E) is the biggest protected oxbow lake of the River Tisza watershed. This water body is 29 km long, spreading over 207 ha, with an average depth of 2.2 m, holding 4.5 million m 3 of water [8]. The lake is semi-paleopotamonic with a possibility of supplementing or replacing water, pumping over from the parent river [37]. The oxbow lake was filled up at the beginning of the year and the water level lowered in late autumn. Between these two events, there was no significant water movement or flow, so we considered it as stagnant water, with no species influx possible during the year. The oxbow serves for storing inland excess and irrigation water as well as for various recreational activities [8].
Along a 3.5 km section, we selected five sampling points (K1-K5) ( Figure 1) at different positions in relation to the combined inflow from a thermal water fish farm producing African catfish (Clarias gariepinus) and the experimental fish ponds of the Research Institute for Fisheries and Aquaculture, National Agricultural Research and Innovation Centre-(HAKI). No other source of contamination is known at this section of the oxbow lake. This inflow, located at site K1, was nutrient-rich, partly of geothermal origin, and before reaching the oxbow lake was treated by a wastewater-cleaning wetland system. During the treatment process, the thermal water from the catfish farm cooled down, so at the point of inflow no thermal pollution occurred. Further sampling points were selected based on the characteristics of the oxbow lake. Because there was no significant water flow, we only considered distance from inflow. The K2 point was at 500 m, K3 at 2.5 km (to avoid interference by a fish sampling operation), K4 at 3.0 km and K5 at 3.5 km from the influx point. The nutrient content of the water decreases away from the influx. In the year of the survey at the point of K1, the content of Total Nitrogen (TN) was 0.195-4.000 mg/L, the Total Phosphorous (TP) from 0.053 to 0.468 mg/L and the Total Suspended Solids (TSS) from 13.14 to 41.10 mg/L. These values at point K5 were the following: TN: 0.275-1.140 mg/L, TP: 0.062-0.210 mg/L and TSS: 7.1-28.6 mg/L [38]. The yearly amount of the effluent of the catfish farm was about 330,500 m 3 [39]. Along a 3.5 km section, we selected five sampling points (K1-K5) ( Figure 1) at different positions in relation to the combined inflow from a thermal water fish farm producing African catfish (Clarias gariepinus) and the experimental fish ponds of the Research Institute for Fisheries and Aquaculture, National Agricultural Research and Innovation Centre-(HAKI). No other source of contamination is known at this section of the oxbow lake. This inflow, located at site K1, was nutrient-rich, partly of geothermal origin, and before reaching the oxbow lake was treated by a wastewater-cleaning wetland system. During the treatment process, the thermal water from the catfish farm cooled down, so at the point of inflow no thermal pollution occurred. Further sampling points were selected based on the characteristics of the oxbow lake. Because there was no significant water flow, we only considered distance from inflow. The K2 point was at 500 m, K3 at 2.5 km (to avoid interference by a fish sampling operation), K4 at 3.0 km and K5 at 3.5 km from the influx point. The nutrient content of the water decreases away from the influx. In the year of the survey at the point of K1, the content of Total Nitrogen (TN) was 0.195-4.000 mg/l, the Total Phosphorous (TP) from 0.053 to 0.468 mg/l and the Total Suspended Solids (TSS) from 13.14 to 41.10 mg/l. These values at point K5 were the following: TN: 0.275-1.140 mg/l, TP: 0.062-0.210 mg/l and TSS: 7.1-28.6 mg/l [38]. The yearly amount of the effluent of the catfish farm was about 330,500 m 3 [39].

Chemical Water Characteristics
In the year of the survey, 1500 mL of effluent water from the African catfish farm was collected monthly and taken to the laboratory at the National Agricultural Research and Innovation Centre (NAIK) Research Institute of Irrigation and Water Management where TN [40] and TP [41] were measured by spectrophotometry, and TSS [42] by weight measurements. These water samples were collected from the outflow of the artificial wetland water treatment system.

Chemical Water Characteristics
In the year of the survey, 1500 mL of effluent water from the African catfish farm was collected monthly and taken to the laboratory at the National Agricultural Research and Innovation Centre (NAIK) Research Institute of Irrigation and Water Management where TN [40] and TP [41] were measured by spectrophotometry, and TSS [42] by weight measurements. These water samples were collected from the outflow of the artificial wetland water treatment system.

Zooplankton Sampling
We sampled the rotifer community nine times ( October 2016) at each sampling point, when we took 50 L of water from <50 cm below the surface, and filtered it using a 50 µm mesh plankton net. The filtered samples (100 mL were put into a 120 mL plastic bottle, preserved by adding formaldehyde (4% final concentration) and taken into the laboratory, where they were stored at 4 • C until identification. Zooplankton density was assessed using a 5-mL counting chamber. Rotifer identification was made under 40-125 × magnification, following standard keys [43,44].

Diversity Evaluation
The analysis of rank-abundance curves was performed with the vegan [45] package in R [46]. Rank-abundance graphs were visualised in EXCEL [47,48]. To compare rotifer diversity among sample sites, we used the diversity ordering with the Rényi generalised entropy function (HR; Equation (1)) [34]: where p i is the relative abundance of the i-th species, S is the total number of species in the sample, and α is the scale parameter which has no exact biological meaning. The Rényi-function is a generalised entropy function, because at special values of the scale parameter α, the value of the Rényi diversity will take the value of several frequently used entropy-based diversity indices (Table 1) [34,49]. In general, at low values of the scale parameter (close to 0) the Rényi index is sensitive to rare species, whereas with increasing α, the diversity value will gradually be influenced more by the more abundant species [34]. Diversity profiles were calculated using the vegan [45] package in R [46]. Table 1. Special values of scale parameters (α) of the Rényi diversity.

Scale Parameter (α)
Rényi Diversity (HR) logarithm of Berger-Parker index log 1 p max * Scale parameter cannot take the exact value of 1, but its limit as α tends to 1.

Nutrient Input
During the study period, an estimated 4200 kg·y −1 of nitrogen, 475 kg·y −1 of phosphorous and 5220 kg·y −1 of suspended solids were released into the oxbow lake from intensive aquaculture ( Table 2).

Composition of the Rotifer Assemblage
During the sampling period, 28 rotifer taxa were collected: 26 were identified to species level, and 2 additional ones to genus level (Table S1). The highest species richness (14 species) and density (163,272 ind·m −3 ) were registered at the inlet point (K1) in the summer and the lowest values (2 species and 864 ind·m −3 ) were found at the most distant point (K5) in autumn (Table S1). The most frequent and widespread species were Brachionus calyciflorus and B. leydigi; these were absent from only 2 sites. There were 5 singleton species, observed in only one sample. The most species-rich genus, Brachionus was represented by 10 species.
In spring, the most species-rich site was K5 (11 species), but this site had the lowest density (7560 ind·m −3 ). The highest abundance (140,472 ind·m −3 ) was found at K2. B. nilsoni and B. leydigi were most abundant at the point of inflow and at a distance of 0.5 km (K2).
In summer, the highest species richness and density were observed at the inflow (K1 site). In summer, the abundance varied between 71,928 and 163,488 ind·m −3 and was dominated by B.angularis and B. calyciflorus. The number of species varied between 7 (K5) and 14 (K1).
In autumn, the highest number of species (12) as well as individuals (12,096 ind·m −3 ) were detected at K1, and the lowest (2 species and 864 ind·m −3 ) at the most distant site, K5.

Species Distributions
The spring assemblages from K1 to K5 had a similar number of species (8)(9)(10)(11), but the evenness increased with increasing distance from the inflow point ( Figure 2). In summer, the species richness was the highest at the inflow. Considering the most frequent and the rarest species, K1 had the highest number of individuals but, in the case of the moderately common species (species rank: 3-10), K2, K4, K5 sites showed higher diversity. In autumn, the assemblage at the inlet point (K1) was the most species-rich with a correspondingly high number of individuals. At K3-K5, the number of species and individuals were lower ( Figure 2).

Composition of the Rotifer Assemblage
During the sampling period, 28 rotifer taxa were collected: 26 were identified to species level, and 2 additional ones to genus level (Table S1). The highest species richness (14 species) and density (163,272 ind·m −3 ) were registered at the inlet point (K1) in the summer and the lowest values (2 species and 864 ind·m −3 ) were found at the most distant point (K5) in autumn (Table S1). The most frequent and widespread species were Brachionus calyciflorus and B. leydigi; these were absent from only 2 sites. There were 5 singleton species, observed in only one sample. The most species-rich genus, Brachionus was represented by 10 species.
In spring, the most species-rich site was K5 (11 species), but this site had the lowest density (7560 ind·m −3 ). The highest abundance (140,472 ind·m −3 ) was found at K2. B. nilsoni and B. leydigi were most abundant at the point of inflow and at a distance of 0.5 km (K2).
In summer, the highest species richness and density were observed at the inflow (K1 site). In summer, the abundance varied between 71,928 and 163,488 ind·m −3 and was dominated by B.angularis and B. calyciflorus. The number of species varied between 7 (K5) and 14 (K1).
In autumn, the highest number of species (12) as well as individuals (12,096 ind·m −3 ) were detected at K1, and the lowest (2 species and 864 ind·m −3 ) at the most distant site, K5.

Species Distributions
The spring assemblages from K1 to K5 had a similar number of species (8)(9)(10)(11), but the evenness increased with increasing distance from the inflow point ( Figure 2). In summer, the species richness was the highest at the inflow. Considering the most frequent and the rarest species, K1 had the highest number of individuals but, in the case of the moderately common species (species rank: 3-10), K2, K4, K5 sites showed higher diversity. In autumn, the assemblage at the inlet point (K1) was the most species-rich with a correspondingly high number of individuals. At K3-K5, the number of species and individuals were lower ( Figure 2).

Rényi Diversity Profiles
In spring, according to the Rényi diversity ordering, the unequivocally most diverse site was K5 (Figure 3). The other four assemblages could not be unequivocally ordered. At small values of the scale parameter (emphasis on rare species), the order was K2 > K3 >K1 > K4. Over most of the range,

Rényi Diversity Profiles
In spring, according to the Rényi diversity ordering, the unequivocally most diverse site was K5 (Figure 3). The other four assemblages could not be unequivocally ordered. At small values of the scale parameter (emphasis on rare species), the order was K2 > K3 >K1 > K4. Over most of the range, diversity increased with distance from the inflow (K4 > K3). There was little difference between the first two sites, K1 and K2. In summer, the five assemblages did not show any easily interpretable differences: K4 was unequivocally more diverse than K2, K4 or K5, but mutual relationships with emphasis on rare species were chaotically different from values under higher influence of moderately common or dominant species. In autumn, both K1 and K2 were unequivocally more diverse than the communities at the other three sites, which could also be unequivocally ordered among themselves (K4 > K3 > K5). K1 was more diverse than K2 only when rare species were influential (0 < α < 2). At α ≥ 2 and thereafter, site K2 was the most diverse, and K5 the lowest (Figure 3).
Water 2019, 11, x FOR PEER REVIEW 6 of 11 common or dominant species. In autumn, both K1 and K2 were unequivocally more diverse than the communities at the other three sites, which could also be unequivocally ordered among themselves (K4 > K3 > K5). K1 was more diverse than K2 only when rare species were influential (0< α < 2). At α > = 2 and thereafter, site K2 was the most diverse, and K5 the lowest (Figure 3). Regarding the seasonal biodiversity of the sampling points, a clear ordering can only be identified for K5. In this case, the spring > summer > autumn order can be seen (Figure 4). At the point of inflow (K1), the initial, rare-species-influenced order of summer > autumn > spring changed to autumn > summer > spring at higher values of α. Diversity in autumn was clearly higher than in spring. For K2, the order of summer > spring > autumn at α = 0 changed to autumn > summer > spring at α = 1 and at α = 2: rotifer diversity in summer was clearly higher than in spring. At the third point (K3), the highest diversity was found in spring, but the order of summer>autumn at α = 0 scale parameter reversed at α = 1. At the fourth point (K4), rotifer diversity was clearly the lowest in autumn. The other two profiles crossed each other twice. At α = 0 rotifer diversity was greater in summer than in spring. This was reversed at α = 1 and turned back at α = 2, meaning that while there were more species in summer, these were rare, and the assemblage was dominated more by the common species.  Regarding the seasonal biodiversity of the sampling points, a clear ordering can only be identified for K5. In this case, the spring > summer > autumn order can be seen (Figure 4). At the point of inflow (K1), the initial, rare-species-influenced order of summer > autumn > spring changed to autumn > summer > spring at higher values of α. Diversity in autumn was clearly higher than in spring. For K2, the order of summer > spring > autumn at α = 0 changed to autumn > summer > spring at α = 1 and at α = 2: rotifer diversity in summer was clearly higher than in spring. At the third point (K3), the highest diversity was found in spring, but the order of summer>autumn at α = 0 scale parameter reversed at α = 1. At the fourth point (K4), rotifer diversity was clearly the lowest in autumn. The other two profiles crossed each other twice. At α = 0 rotifer diversity was greater in summer than in spring. This was reversed at α = 1 and turned back at α = 2, meaning that while there were more species in summer, these were rare, and the assemblage was dominated more by the common species.

Discussion
The survey indicated a species-rich rotifer fauna with 26 species (2 identified only to genus level). An earlier survey [33] found 13 and 8 species in the adjacent, protected Aranyos and Borza oxbow lakes, respectively, and 14 species in the similar oxbow lake of Dán-zug. In the Hármas-Körös, the parent river of the Kakafok oxbow, a survey conducted 25 years ago found 65 species [50], but 10 species present in the Kákafok oxbow lake were missing from that list. An extensive, 2 years survey in the oxbows near the Danube at Gemenc (between 1503 km and 1469 km), found 71 rotifer taxa [51], 16 of which were in common with our survey. However, 8 found in ours (30%) were not present in the Danube, indicating the taxonomic distinctness between the Danube and the Tisza Rivers.
Most of the species found in the Kákafok oxbow lake were characteristic of oligo-and oligo/β mesosaprobic waters [17] except the Brachionus species that prefers contaminated, β and β/α mesosaprobic water bodies [17]. The species we detected (Brachionus angularis, B. calyciflorus, B.  diversicornis, B. falcatus, B. forficula, B. leydigi, B. nilsoni, B. quadridentatus, B. urceolaris and B. variabilis) typically occurred in summer, mainly at K1-K2 points, indicating the impact of the effluent water combined with favourable thermal conditions. We found more Brachionus spp. than in the two previous studies [33,50], which mostly belonged to the β mesosaprobic indicator group [17,18]. Species previously detected in the mother river are Brachionus angularis, B. bennini, B. falcatus, B.  budapestiensis, B. calyciflorus, B. urceolaris and B. quadridentatus [50], and in the other three oxbow lakes Brachionus bidentata, B. quadridentatus and B. calyciflorus were found [33]. In our samples, the most common species and in most places (Brachionus calyciflorus) also belonged to this genus. Trophic conditions are related to saprobic ones [52,53]. These species are also indicators of trophic conditions [11][12][13][14]16]. A similar species is Trichocerca pusilla in the work of Berzins [11], Jarenefelt [13], Thunmark [14] and Pejler [16] which also appeared only sporadically at the point of influx in summer and autumn. Sladacek [18] suggested that we can deduce the trophic level from the quotient of Brachionus vs. Trichocerca species. However, we could not apply this parameter because Trichocerca spp. could

Discussion
The survey indicated a species-rich rotifer fauna with 26 species (2 identified only to genus level). An earlier survey [33] found 13 and 8 species in the adjacent, protected Aranyos and Borza oxbow lakes, respectively, and 14 species in the similar oxbow lake of Dán-zug. In the Hármas-Körös, the parent river of the Kakafok oxbow, a survey conducted 25 years ago found 65 species [50], but 10 species present in the Kákafok oxbow lake were missing from that list. An extensive, 2 years survey in the oxbows near the Danube at Gemenc (between 1503 km and 1469 km), found 71 rotifer taxa [51], 16 of which were in common with our survey. However, 8 found in ours (30%) were not present in the Danube, indicating the taxonomic distinctness between the Danube and the Tisza Rivers.
However, we could not apply this parameter because Trichocerca spp. could only be detected twice, at the point of influence in summer and autumn. However, most Brachionus species can also be detected here.
Our initial hypotheses were only partially supported by the data. The diversity profiles provide more sophisticated and articulated results than the usual one-dimensional diversity indices [54,55]. In spring, as we expected, the biodiversity increased with increasing distance from the point of inflow, but after a summer transition period, the situation became partially reversed during autumn. There was no clear seasonal difference between the inflow point and the closest one, indicating that the effect of the effluent did not decrease at a distance of 500 m. In most cases, (spring and autumn), and for rare species in summer, point K3 showed medium diversity, conforming to our hypothesis. By summer, this site became the one of the lowest diversities as a result of a few species reaching high abundance and dominating the assemblage. The third most distant point (K4) gave surprising results. While by the end of the study period it could be clearly ranked, in spring it became the second most diverse community when the influence of rare species diminished. Unexpectedly, even in summer, diversity at K4 point was significantly higher than nearer and farther from the influx point. The probable reason is the presence of several species at higher abundance, which possibly competed with each other, and none of them managed to dominate the assemblage, creating high evenness in the assemblage. The most distant point (K5) was the most diverse in spring and the least diverse in autumn. In summer, its rotifer diversity was not clearly different from those at the other points. Comparing the different sampling points, the decreasing abundance with increasing distance from the inflow was most often observed in Brachionus species, which are indicators of more saprobic waters [17]. At the point of aquaculture inlet (K1) and nearby (K2), Brachionus spp. showed greater dominance in the same species set, leading to a decrease of diversity in spring. The species pool reached its maximum in summer; however, due to the predominance of more saprobic species, it did not result in a clearly identifiable increase in diversity. In autumn, the number of species and individuals decreased at the more distant points, while at the point of inflow a relatively rich and abundant community emerged. The monthly sampling may capture seasonal trends; however, given that rotifer generation time (from hatching to maturity) is typically about 24 h, monthly sampling likely missed changes in rotifer communities that occur on a shorter time scale. Nonetheless, our survey indicated that rotifer species richness was higher than previously documented.
Summarizing, we detected a seasonally different effect of the aquaculture effluent in the composition of the rotifer community of the Kákafok oxbow lake. This effect was strongest during spring and autumn. In summer, the differences in taxonomic diversity disappeared. According to our investigations at the beginning of the study period, around the nutrient-rich effluent the dominance of common species increased. The differences in dominance decreased but did not disappear in summer. In autumn, however, the extra nutrient inflow enabled the maintenance of a diverse community at the point of effluent inlet. The question rightly arises as to how other factors (biotic interactions, phylogenetic connections, etc.) may have influenced the results. In addition, other environmental factors that had not yet been examined may play a role, including competition with copepods and cladocerans and predation or food availability. Due to the presence of closely related and different saprobiotic indicator species, functional and phylogenetic analysis may provide additional information.