Structure and spatial distribution of the rotifer assemblages along a tropical reservoir

The construction of reservoirs is a common practice in the world. These systems modify the hydric landscape and alter the flow of rivers, becoming lotic environments in lentic. Here we investigated the structure and spatial distribution of rotifers along a tropical reservoir. We sampled four points in the Pedra do Cavalo Reservoir, Bahia, Brazil, bimonthly, between August 2014 and June 2015. We registered more than 70 taxa distributed in 17 families, with the majority of species belonging to the Lecanidae, Brachionidae and Trichocercidae families. The species rarefaction curve did not achieve a total asymptote, indicating that species richness in the reservoir is higher than what was registered. Based in the species frequency of occurrence, we identified 48 rare species, 16 common species, five constant species and one frequent species. The highest beta diversity values were registered in riverine P1 (0.513) and intermediate P2 (0.503), although there were no significant differences between the sampling points. Despite the high abundance values in P1, P2 and P3 no significant differences were found between the studied points. Thus, this study substantially increases the knowledge on the rotifer community in the Paraguaçu River and contributes to future studies that focus on biodiversity, ecology and conservation in the Brazilian reservoir ecosystems.


Introduction
The freshwater environments represent vital resources for human, also an ideals habitat for the maintenance of the biota and endemic species (Vörösmarty et al., 2010;Junk et al., 2014). These hydric systems are formed by lentic environments with slow or absence of flow (as lakes); lotic with flow (as rivers), and semilotic environments, with calm water that preserve characteristics of lotic environments (as reservoirs) (Esteves, 2011).
The construction of reservoirs is also a common practice in tropical and subtropical regions of South America, facilitated by high hydric availability, and their reasons are for electricity generation, as well urban, industrial and agricultural water supply, leisure and navigation (Rodgher et al., 2005). In Brazil, the increase in reservoir numbers started in the 1960's (Agostinho et al., 2002;Dudgeon et al., 2006), which have modified the hydric landscape and the flow of lotic environments, becoming lentic areas (Nogueira, 2006;Jorgensen et al., 2013;Schork and Zaniboni-Filho, 2018). Currently there are over 1,313 large reservoirs in operation and an enormous number of small all over the country (Dias et al., 2018).
The reduction of hydrologic flow provides favorable conditions to the development of species adapted to lentic conditions (Rocha et al., 1999). Such environments, as reservoirs, are suitable for the development of high planktonic populations, once that the plankton growing rates are elevated in these environments (Marzolf, 1990;Okulu et al., 2016). Studies have shown that high densities of zooplanktonic organisms in reservoirs may be due to the reduced water speed and low water residence time with the associated reduction in advective downstream losses (Takahashi et al., 2014;Beaver et al., 2015;Okulu et al., 2016). Besides that, the zooplankton abundance, generally, exhibited a trend similar that of the phytoplankton (Beaver et al., 2015).
Studies on the structure and functioning of the planktonic community in reservoirs provide opportunities to investigate patterns due to the cyclic variation and episodic disturbances (Nogueira, 2001;Xue et al., 2018). Reservoir shows marked changes in limnological conditions in relatively short periods (Geraldes and George, 2013). Thus, the understanding of the plankton dynamics in these systems can be used to evaluate the resilience of this type of ecosystem, once plankton are sensible to alterations in environmental characteristics and reflect the dynamics of the ecosystem (Margalef, 1983;Legendre et al., 1985;Dias et al., 2012;Silva Brito et al., 2018).
The zooplanktonic community has a fundamental role in aquatic ecosystems, due to its position in the food chain, where it regulates the structural heterogeneity of various biological communities (Lemke and Benke, 2009). Among the zooplankton, rotifers constitute the majority of the species richness, as they occupy diverse ecological niches, both in the limnetic and littoral zones (Lansac-Tôha et al., 2009;Azevêdo et al., 2015). Furthermore, because the environmental requirements their immediately respond to the environmental variations, associated with high rates of population renovation (Obertegger et al., 2007;Azevêdo et al., 2015).
In this way, characterizing the community structure of rotifers is useful both to the understanding of the growth of dependent populations and for studies aiming at conservation and biomonitoring in this region. This study aimed at investigating the structure and spatial distribution of the rotifer community along a tropical reservoir.

Study area
This study was made in the Pedra do Cavalo Reservoir, which supply a power plant in the dam (12°35'S and 38°59'W). This lake of 55 Km 2 area is in Bahia (Brazil) in the lower part of the Paraguaçu River, which large tributary is the Jacuípe River. This river rises in Chapada Diamantina Mountain, drain and crosses the semiarid region for 500 km until the confluence with the Paraguaçu River near the Pedra do Cavalo Reservoir (Fontes et al., 2011) (Figure 1).
We performed six samplings bimonthly between August 2014 and June 2015 in four sites distributed in the littoral zone of the Pedra do Cavalo Reservoir (Table 1).

Sampling design
We sampled rotifers in the morning period, at the sub-surface of the pellagic zone, using a graduated bucket and plankton net of 68 µm mesh size. We filtered 100 liters of water per sample totalizing 6 replicated. Samples were fixed with a 4% formaldehyde solution and buffered with calcium carbonate precipitated.

Laboratory analyses
Organisms were analyzed using stereoscope and microscope. We determined rotifer species richness in each sample until species curve was stabilized, that is, until no new species were registered in the sample. The individuals were estimated by analyzing a minimum of three subsamples (Bottrell et al., 1976) in a Sedgewick-Rafter chamber, using an optical microscope, at least 50 individuals being counted. Samples with few individuals were counted integrally. Species were identified using slides and coverslips and following basic literature as Koste (1972Koste ( , 1978, Koste and Roberton (1983), Paggi (1989), Nogrady et al. (1993), Nogrady and Pourriot (1995) and Segers (1995).
In order to determine the abundance of rotifers (individuals/m 3 ), we have taken subsamples from 0.5 to 4 ml with a graduated pipette after sample homogenization and counted in the Sedgwick-Rafter chamber under optic and stereoscopic microscope.

Data analyses
We have determined the frequency of occurrence of species considering the number of samples in which the species was present in relation to the total of samples. This was calculated according to the Formula 1: where C is the constancy index, p is the number of samples where the species was present and P is the total of samples.
Through species presence and absence data, we generated a rarefaction curve to verify if the samples were sufficient to estimate the incidence of rotifer species. We calculated
the mean percentage of richness extrapolation according to Heck Junior et al. (1975). We have used the Jackknife 2 estimator, which bases in the removal of a sample from the total of observed samples, recalculating the estimator from the remaining values. For this, we used the programs Estimate and Statistica 5.0.
Firstly we used an analysis of variance (ANOVA) in order to test for differences in the abundance and richness of species between sampled points (P1, P2, P3 and P4). However, as the homocedasticity assumptions for the ANOVA were not achieved, we realized a non-parametric Kruskal-Wallis test (Sokal and Rohlf, 1995).
We have evaluated the variability of the composition of the rotifer community (beta diversity) through a test for homogeneity of multivariate dispersions -PERMDISP (Anderson et al., 2006) using presence/absence data. This test is based on the distances between the sample units and the centroid of the group in an Analysis of Principal Coordinates (PCoA), built using a dissimilarity measure. For this we have used the Jaccard dissimilarity measure, so that the higher the Jaccard dissimilarity means, the higher the beta diversity. PERMDISP uses an ANOVA with 999 permutations to test for significant differences (p<0.05). For PERMDISP and PCoA, we used the software R 3.4 (R Development Core Team, 2017), and the packages vegan (Oksanen et al., 2013) and permute (Simpson, 2016).
We verified no significant differences in species richness between sampled points (p<0.05) (Figure 3). The frequency of occurrence indicated a higher number of rare species (48 species), followed by common species (16 species The rarefaction curve evidenced that our samplings were not enough to reach an asymptotic trend in P2, P3 and P4 (Figure 4). However more samples are needed to reach a P1 asymptote.
According to the mean percentage of richness extrapolation, rotifer data were capable of assessing 62.4% of species. The PERMDISP test, used to verify beta diversity, revealed higher values in P1 (0.513) and P2 (0.503), without significant differences between points ( Figure 5).  Koste, 1973 X Bdelloidea X X X X

Discussion
The results obtained through the faunistic inventory of the rotifer community showed that the number of taxa may be considered relevant when compared to the values found in other studies. Sampaio et al. (2002), for example, High abundances of K. cochlearis and B. calyciflorus have been commonly registered in reservoirs and are possibly related to the great tolerance of these species to eutrophic environments (Aoyagui et al., 2004;Almeida et al., 2006;Bezerra et al., 2015), especially in the studied region (Almeida et al., 2012). Some species Keratella are good explorers of a high range of feeding resources and are resistant to cyanobacteria toxins, which favors the development of large populations in eutrophic environments (Bezerra et al., 2015). Brachionus calyciflorus is tolerable to eutrophic environments with high conductivity (Serafim-Junior et al., 2010). Thus, water quality changes in reservoirs resulting from the eutrophication process may have contributed to the higher abundance of these species.
Brachionidae, Lecanidae, and Trichocercidae have been frequently registered in various studies in Brazil and others countries as the families with the highest species richness, and may correspond to 50% of the present taxa Mantovano et al., 2015;Augustin et al., 2018;Nandini et al. 2019). The predominance of these families is considered typical of Neotropical regions (Serafim Júnior et al., 2003;Aoyagui et al., 2004).
Lecanidae presented the highest species richness and was mainly registered in P2, P3 and P4, which possess macrophyte patches. Taxa of these families are    more frequent in the littoral zone, and in the benthonic and periphytic communities. They usually occur in the plankton as occasional migrants, mainly in areas with marginal vegetation and extensive macrophyte patches (Almeida et al., 2006;Melo-Junior et al., 2007;Bezerra et al., 2015). On the other hand, Brachionidae, highly endemic to South America, have in general planktonic habit and are associated with environments with great accumulation of organic matter (Almeida et al., 2006;Dantas-Silva and Dantas, 2012). Thus, the higher richness of this family in the reservoir is expected, since three sampled points present some kind of activities that could generate nutrient accumulation (see Table 1).
The highest frequency of occurrence of Keratella tropica, Lecane bulla, Keratella cochlearis, Brachionus calyciflorus, Keratella americana and Bdelloidea is commonly observed in other continental water bodies of the tropical region. As in our study, Eler et al. (2003) observed an elevated frequency of occurrence for K. cochlearis, K. tropica and Bdelloidea in fee fishing lakes, in the Mogi-Guaçu River Basin (SP). Serafim-Junior et al. (2006) have also registered highest frequency of occurrence for K. cochlearis and L. bulla, besides more Bdelloidea, with 100% of occurrence in the Itajaí-Açu River at Santa Catarina State. Keratella tropica and K. americana were the dominant species in the zooplankton community of some semi-arid reservoirs of the Rio Grande do Norte State (Eskinazi-Sant'Anna et al., 2007). High frequencies of B. calyciflorus and K. cochlearis have also been commonly observed in reservoirs (Zaganini et al., 2011;Dantas-Silva and Dantas, 2012;Pedrozo et al., 2012). Serafim-Junior et al. (2016), studying the effects of a reservoir construction in the Iguaçu River (PR) on zooplankton also verified an elevated frequency for K. americana and K. cochlearis.
Although the number of rotifer species is considerably high, the species rarefaction curve did not reach a total asymptote, indicating that species richness in the reservoir is higher than has been registered. According to the mean percentage of richness extrapolation, our rotifer data were capable of assessing 62.4% of species. However, Heck Junior et al. (1975) suggested that inventories that obtain 50% to 75% of the total number of species that occur in a determinate area might be considered satisfactory without loss of information, provided that common species are registered in the study.
Our results evidenced no significant differences between sampled points for beta diversity. This may be a result of the high water residence time (180 days) that the Pedra do Cavalo Reservoir presents when compared to other small and medium sized reservoirs, such as Rosana Reservoir (18.6 days) and Salto Segredo (47 days) (Julio Junior et al., 2005). Environments with low flow velocity and long residence times, such as reservoirs, favor rotifers, as they are organisms with a short life cycle (Takahashi et al., 2014).
The Paraguaçu River, with its 500 Km of extension, still remains unknown as to its planktonic species, as rotifers. In this sense, we highlight the importance of realizing a spatially integrated analysis, rich in information on local species. This study substantially amplifies the knowledge of the rotifer communities in the Paraguaçu River Basin and shows us that the trends found in the study are similar to those of other eutrophic reservoirs. It also aid in future studies that focus on biodiversity, ecology and conservation and, more broadly, contribute with studies on Brazilian reservoirs.