Copepod assemblage structure (Crustacea: Copepoda) along a longitudinal environmental gradient in a tropical river-floodplain system, Brazil

Aim: In this study, we analyzed the structure of the copepod community along a longitudinal axis of the last undammed stretch of the Upper Paraná River floodplain, downstream of Porto Primavera reservoir. We hypothesized that (i) copepod assemblage would show an increase in the abundance of organisms, species richness and specific diversity with the distance from the Porto Primavera reservoir, due to the arrival of species from adjacent lakes and tributaries; (ii) copepod abundance and species richness would be higher in tributaries and adjacent lakes located at the floodplain, which show higher environmental heterogeneity and function as a species source; and (iii) spatial patterns of copepod assemblage structure are related to the environmental gradient, since copepods show a rapid response to the variation of environmental conditions. Method: Sampling was performed quarterly from August 2013 to July 2014, at the subsurface of the pelagic region. Sampling sites were located along the Paraná River, in seven of its tributaries, and in eight lakes located in its floodplain. Results: We identified 29 species belonging to two families: Cyclopidae (21 species) and Diaptomidae (eight species). On contrary of our first hypothesis, ANOVA results showed a significant decrease in copepod abundance in the downstream direction. Similarly, the tributaries showed higher values of ecological attributes in the upstream stretch. RDA results showed that in axis 1, the Paraná River and the lakes were positively correlated with conductivity, pH, temperature and suspended organic matter, characterized by the most abundant copepod species. The tributaries, on the other hand, were negatively correlated with turbidity, total nitrogen, total phosphorus and depth. Axis 2 showed a positive correlation of the Paraná River with conductivity and suspended organic matter, and a higher abundance of Argyrodiaptomus azevedoi, Notodiaptomus iheringi, N. henseni, N. cearensis 2 Palhiarini, W.S. et al. Acta Limnologica Brasiliensia, 2017, vol. 29, e102 Serial Discontinuity Concept (Ward & Stanford, 1995), which complements the River Continuum Concept, highlights that river impoundments promote discontinuity in the longitudinal dynamics and alterations in biotic and abiotic components. However, aquatic communities and nutrients could show a gradient downstream of the dam, becoming gradually closer to the original conditions of the river. In some aquatic ecosystems, one can observe large reservoirs constructed in cascades, as is the case of the Paraná River (Agostinho et al., 2007). Despite having a 230 km undammed stretch, this river suffers the effects of 34 cascading reservoirs located upstream of Porto Primavera reservoir (Agostinho et al., 2008). The construction of Porto Primavera reservoir has drastically affected the Upper Paraná River floodplain, since discharge control altered the flood regime, nutrient concentration along the river


Introduction
Suitable environmental conditions are essential for the proper functioning and maintenance of watercourses, as described by the River Continuum Concept (Vannote et al., 1980).According to this theory the riparian vegetation, adjacent lakes and tributaries influence the patterns of environmental conditions in river systems.Therefore, aquatic communities are structured according to longitudinal changes in environmental gradients along the river.
Human activities such as reservoir construction have caused environmental disturbances and altered aquatic ecosystems, promoting changes in the hydrodynamics and chemical and physical characteristics of river systems, thus influencing the structure of aquatic communities (Simões et al., 2015;Agostinho et al., 2016).In this context, the and N. cf.spinuliferus.On the other hand, the lakes were positively correlated with temperature and pH and also with a higher abundance of Thermocyclops minutus and T. decipiens.Conclusion: We highlight the importance of the maintenance of the ecological integrity of the Upper Paraná River floodplain, especially of its tributaries, in order to support the high biodiversity registered in the last undammed stretch.Furthermore, we highlight the importance of manipulation of water levels in reservoirs located upstream of this stretch, a necessary condition to maintain the hydrological connectivity between floodplain habitats and the main river.
The Cascading Reservoir Continuum Concepts, in turn, attempts to predict the dynamics of these systems.This concept hypothesizes the existence of an interconnection between ecological processes in each type of reservoir, in the upstream-downstream direction, and the influence of lateral systems connected to the rivers, such as adjacent lakes and tributaries (Barbosa et al., 1999;Stanford & Ward, 2001), considering the differences in their communities (Barbosa et al., 1999).Thus, this concept has been used to evaluate the possibility of reestablishment of the original environmental conditions and the organization of aquatic communities in environments under the influence of reservoir construction.
In this sense, the study of the zooplankton community is relevant due to its high species diversity (Lansac-Tôha et al., 2009) and sensitivity to environmental alterations, making it a potential bioindicator of changes in these ecosystems (Maia- Barbosa et al., 2006).This community has been considered a key factor for the comprehension of alterations in aquatic environments (Eskinazi-Sant'anna et al., 2013), providing knowledge on ecological processes in hydrographic basins (Maia- Barbosa et al., 2006;Shah et al., 2013).
Among the organisms belonging to the zooplankton community, the copepods are generally filter-feeding organisms with a diet basically composed of phytoplankton (Perbiche-Neves et al., 2007), bacteria and detritus (Smith et al., 1979).Thus, copepods have a relevant ecological role in freshwater ecosystems, as they feed on primary producers and transfer the energy to higher levels of the food web (Sugumaran, 2016).
Alterations in physical and chemical factors caused by reservoir construction promote changes in food composition, altering the community structure of copepods (Perbiche-Neves et al., 2014).These changes can determine the type of feeding of copepods and favor the development of certain species, influencing on the community structure (Matsumura-Tundisi & Tundisi, 2003;Napiórkowski et al., 2006).In this context, alteration in the structure and dynamics of the copepod community can be an indicator and a reflex of changes in aquatic ecosystem functioning, as observed by Pedrozo & Rocha (2005).
Considering the sensitivity that some species show to environmental changes, in environments affected by reservoir construction, we aimed to verify the structure of the copepod assemblage along a longitudinal axis of the last undammed stretch of the Upper Paraná River floodplain, downstream of Porto Primavera reservoir.
We evaluated the hypotheses that (i) copepod assemblage would show an increase in the abundance of organisms, species richness and diversity with the distance from the Porto Primavera reservoir, due to the arrival of species from adjacent lakes and tributaries; (ii) copepod abundance and species richness would be higher in tributaries and adjacent lakes located at the floodplain, which show higher environmental heterogeneity and function as a species source; and (iii) spatial patterns of copepod assemblage structure are related to the environmental gradient, since copepods show a rapid response to the variation of environmental conditions.

Study Area
Our study was performed in the last undammed stretch of the Upper Paraná River floodplain, between UHE Engenheiro Sérgio Motta (Porto Primavera reservoir) and Itaipu Reservoir (Figure 1).The Paraná River, which belong to the Prata basin, has an extension of 4,695 km, and has more than 145 great reservoirs along its course, all of them more than 10 meters tall, and of which 20% show an area greater than 100 km 2 (Agostinho et al., 2004b(Agostinho et al., , 2007)).
The mean declivity of the Upper Paraná River is 0.18m/km.In the studied stretch, its declivity is reduced (0.09m/km), and the river exhibits a broad anastomosing channel and an extensive alluvial plain, with accumulation of sediment in the riverbed, creating inlets and small islands (over 300), or large islands and a restricted floodplain (Agostinho, 1993).
The studied stretch of the Upper Paraná River showed an extensive floodable area of 480 km before the construction of Porto Primavera reservoir.However, after the dam construction, the floodplain was reduced to an extension of 230 km (Agostinho et al., 2004b).The river-floodplain system has several tributaries.On its left bank, Paranapanema, Ivaí and Piquiri are the main tributaries, and at the right bank Baía, Iguatemi, Ivinhema and Amambaí are the most important.This flood plain system can reach 20 km of width in its right bank, and show a series of adjacent lakes (Agostinho et al., 2004b).

Field surveys
Copepod sampling was performed in August (low water period) and November (high water period) 2013, and February (high water period) and May (low water period) 2014.Sampling was performed at the subsurface 30 cm of the central region, distributed in transects along a longitudinal axis of the Paraná River, from downstream Porto Primavera reservoir to the beginning of Itaipu Reservoir.We sampled in 10 sampling sites in the river channel with three replicates (left bank, center, and right bank), which were distributed upstream and downstream of the mouths of the seven largest tributaries, which were also sampled.Besides the lotic systems, we also sampled eight adjacent lakes, located in the floodplain itself.Thus, we obtained a total of 180 samples, 120 in the Paraná River, 28 in tributaries, and 32 in lakes.Plankton samplings were performed in the morning, with a suction pump, by filtering 600 liters of water per sample through a plankton net of 68 µm mesh size.Samples were further fixed in a 4% formaldehyde buffered with calcium carbonate.
Data on these physical and chemical variables were provided by the Laboratory of Basic Limnology of the Research Nucleus in Limnology, Ichthyology and Aquaculture (Nupelia) of the State University of Maringá.
Data on the water levels of the Paraná River were provided by the Itaipu Reservoir, of the monitoring station of São José (22° 43' 04"S, 53° 10' 36"O) (altitude above mean sea level -amsl).

Laboratory analysis
In the laboratory, we identified copepods using specialized literature (Santos- Silva, 2000;Lansac-Tôha et al., 2002;Silva, 2003;Perbiche-Neves et al., 2015).Individuals were counted under a binocular optic microscope, using Sedgewick-Rafter chambers.A minimum of 50 individuals were counted in three subsequent subsamples, obtained with a Hensen-Stempel pipette (2.5 mL) (method modified from Bottrell et al., 1976).Samples with a reduced number of individuals were counted entirely.Final density was expressed in individuals per cubic meter (ind.m -3 ).The subsample method, despite providing a reliable estimate of abundance, underestimates species richness, since rare species may not be registered.Considering that, in order to determine species richness, we analyzed subsequent subsamples of each sample until no new species were registered.

Statistical analyses
In order to verify the species diversity (H') of copepods in each sampling site, as described in hypotheses (i) and (ii), we calculated the Shannon-Wiener index (Pielou, 1975), which is described by -H' = -∑ (ni/N) x log 2 (ni/N), in which ni is the number of individuals of the ith species and N is the total number of individuals.
In order to test hypotheses (i) and (ii), we used Analysis of Variance (ANOVA) (Sokal & Rohlf, 1991) to verify if there are significant differences (p<0.05) in the abundance, richness and diversity of copepod species between sampling sites along the longitudinal axis of the river, lakes and tributaries.Moreover, we performed Fisher's LSD post-hoc test to identify significant differences between upstream and downstream of the sampled environments.The longitudinal axis was arbitrarily categorized in two stretches: upstream, represented by sampling sites P1 to P5 for the Paraná River, T1 to T4 for tributaries, and L1 to L4 for lakes; and downstream, represented by sites P6 to P10 for the Paraná River, T5 to T7 for tributaries and L5 to L8 for lakes.
Previously to all analyses, we logarithmized abundance data and checked the normality and homocedasticity assumptions.When assumptions were not met, we performed a Kruskal-Wallis test, using software Statistic 7.0 (Statsoft Inc., 2005).
In order to test hypothesis (iii) and evaluate the influence of the environmental heterogeneity on the structure of the planktonic copepods, we used a Redundancy Analysis (RDA) (Legendre & Legendre, 1998).Results were based on the values of total inertia and percentage of explanation of each retained axis (p<0.05).The abundance of organisms was logarithmized to reduce the influence of rare species and then transformed according to the Hellinger procedure (Legendre & Gallagher, 2001).Furthermore, the effect of multicolinearity between environmental variables was tested through the Variance Inflation Factor (VIF).These analyses were performed in software R. version 3.1.3(R Development Core Team, 2015), using the package 'vegan' (Oksanen et al., 2012).

Environmental characterization
In general, environmental variables showed low spatial variability during the study period in the different sampling sites.However, turbidity showed an increase in the upstream-downstream direction (P1 to P10).We also observed an increase in total phosphorus concentrations in the sampling sites located at the upstream stretch (P6 to P10), which are influenced by tributaries at the right bank of the Paraná River (Table 1).Water level showed little variation between sampling periods.The lowest value was observed in May 2014 (233.94amsl), and the highest in February 2014 (235.83amsl).Water levels varied only 1.89 m between those periods of low and high water levels (Figure 2).
Thermocyclops minutus and T. decipiens (Cyclopoida) showed the highest mean abundance in the lakes, whereas Notodiaptomus henseni and N. cearensis (Calanoida) were the most abundant in the river (Figure 3).
Significant differences in the copepod abundance (F = 4.33; p = 0.04) were found between stretches considering the sampling sites located in the Paraná River, with higher means in the upstream sites    (Kiefer, 1936) x Ectocyclops rubescens Brady, 1904 x Macrocyclops albidus (Jurine, 1820) x Mesocyclops sp.
x x Mesocyclops aspericornis (Daday, 1906) x x Mesocyclops ellipticus Kiefer, 1936 x x Mesocyclops longisetus (Thiébaud, 1912) x Mesocyclops meridianus (Kiefer, 1926) x     variables (p < 0.05), and the ordination model explained 57% of the variance in the data.Axis 1 explained 42% of the variance in the data and discriminated the sampling sites located in Paraná River and adjacent lakes, positively related to this axis and characterized by higher values of conductivity, pH, temperature and suspended organic matter, and also by the occurrence and abundance of the main copepod species, from the sampling sites located in the tributaries, negatively related to this axis, and which showed higher values of turbidity, total nitrogen, total phosphorus and depth.
Axis 2 explained 25% of the variance in the data and discriminated the sampling sites located in the Paraná River, with higher values of conductivity and suspended organic matter and characterized by the occurrence and higher abundances of diaptomids Notodiaptomus iheringi, N. cearenses, N. cf.spinuliferus and N. henseni, from the sampling sites located in the lakes, with higher pH and temperature, where cyclopids Thermocyclops minutus and T. decipiens predominated (Figure 8).Moreover, this analysis evidenced that higher nutrient concentrations (TN and TP) were positively related to axis 2, associated with sampling sites from the lakes and tributaries, which indicates a higher trophic state in these environments.

Discussion
We observed a longitudinal gradient of nutrients along the sampled stretch of the Paraná River.This tendency of lower values of total phosphorous concentrations and turbidity in sampling sites closer to Porto Primavera reservoir is certainly related to the effects of upstream reservoir cascades causing several alterations downstream, such as the reduction in the transport of suspended matter transport, decrease in nutrient concentrations causing oligotrophication, increase in water transparency, and discontinuity of physical and chemical characteristics (Barbosa et al., 1999).These effects have been widely reported as possible causes of changes in the aquatic biota (Istvánovics et al., 2010;Bortolini et al., 2016) and have been observed for the Upper Paraná River floodplain (Agostinho et al., 2008;Mantovano et al., 2015;Simões et al., 2015) and other reservoirs in the world (Ney, 1996).
Water levels showed little variation between sampling periods, which did not allow for a characterization of low water and high water periods.Inundations of important impact over habitats adjacent to the Paraná River occur when water levels reach at least 4.6 meters (Souza-Filho, 2009).In this way, we evaluated only the spatial scale, while the temporal scale was not taken into account, considering that the effect of flood pulse on the copepod assemblage could not be tested.
Copepod abundance decreased with the distance from Porto Primavera reservoir, in the upstream-downstream direction.This pattern of decreasing abundance could also be observed in other studies performed in environments located downstream of reservoirs (Mitsuka & Henry, 2002;Chang et al., 2008;Perbiche-Neves et al., 2012;Grabowska et al., 2013).
This decrease in abundance is probably caused by the influence of flow velocity in the stretch downstream of the reservoir (Mitsuka & Henry, 2002;Chang et al., 2008;Portinho et al., 2016).Thus, there is an expressive contribution of the reservoir fauna for the community structure immediately downstream of the reservoir, which frequently suffers a decrease along the dowstream stretch, which exhibits with lotic characteristics.The abundance of these organisms is influenced by hydrodynamic conditions, since the current velocity affects the availability of food resources through the removal of nutritive compounds of the biotic community (Mitsuka & Henry, 2002).Furthermore, the reproduction of planktonic organisms is also limited by the high current velocity (Ward, 1994), considering that these organisms are exported in a downstream direction in a rate that is frequently higher than their reproductive capacity.In addition, the large volume of water coming from downstream, promotes a dilution effect in the abundance of copepods (Nadai & Henry, 2009;Neto et al., 2014).
Moreover, lentic environments such as floodplain lakes favor the development of planktonic populations, including copepods (Viroux, 1997;Velho et al., 2001).Since the upstream stretch is more extensive and characterized by an elevated number of lakes and secondary channels, we suggest that the higher values of copepod abundance observed in sampling sites located upstream of the Paraná River could also be the result of the contribution of fauna from environments adjacent to the floodplain.However, the absence of flood during the studied period leads us to disregard this idea.In this sense, Casanova & Henry (2004) verified that low water levels determined a small contribution of lakes, both in terms of abundance and diversity, to the increment of copepod species in the Paranapanema River.
Thus, it is more plausible to assume that the higher abundance of copepods in the stretch closer to Porto Primavera reservoir was due to the arrival of copepods coming from the reservoir.The higher water retention time in the reservoir frequently favors an increase in the abundance of these microcrustaceans (Sartori et al., 2009).
The higher diversity of copepod species in the main river channel was probably determined by the influence of tributaries, mainly those located at the right bank (Baía, Ivinhema and Ivinheminha), in the floodplain.This result suggests an importance of these tributaries for species dispersal, which maintains a higher diversity of copepods in the main river channel despite the low abundance found in this environment.Velho et al. (2004) suggested that the Paraná River functions as a collector of the fauna from all the floodplain.
The redundancy analysis showed that environmental conditions influenced the structure of the copepods assemblage.In this way, the highest copepod abundances and occurrence in lakes corroborate the pattern frequently registered for zooplankton, owing to the lentic characteristic of these environments, which allow the successful reproduction and establishment of populations, especially of copepods (Paggi & Paggi, 1990).In our study, cyclopoid species, especially T. minutus and T. decipiens, predominated in these environments, characterized by the higher values of water temperature and pH.Due to the higher water residence time and reduced flow velocity of these lakes, compared to the river, one could expect that calanoids would be better adapted and predominate in these environments (Ferrareze & Nogueira, 2011), and that cyclopoids would predominate in sampling sites located at the Paraná River.However, we observed an inverse pattern, which could be explained by the fact that the studied lakes were mainly mesotrophic, which favours cyclopoid species (Landa et al., 2007), to the detriment of calanoid species, which prevail in oligotrophic environments.
In the Paraná River, as previously highlighted, calanoids dominated, in terms of abundance and occurrence.This suggests that this predominance is determined by the influence of the upstream reservoir (Porto Primavera), which retain both nutrients and sediments (Agostinho et al., 2004a), Copepod assemblage structure (Crustacea: Copepoda)... Acta Limnologica Brasiliensia, 2017, vol. 29, e102 promoting oligotrophication, considering that this is the last of a long series of reservoirs located at the Paraná River basin.Other authors also registered a similar pattern of calanoid abundance in lotic environments impacted by upstream reservoirs (Lopes et al., 1997;Portinho et al., 2016).
In turn, tributaries were characterized by the low occurrence and abundance of copepods, likely due to the high current velocity of these environments.Therefore, although tributaries show considerably high values of nutrient concentration, which was also verified in the lakes, the water residence time, extremely reduced in these environments, limits the development of populations of copepods (Pourriot et al., 1997).Moreover, higher turbidity values found in these tributaries can limit the occurrence of these organisms, by obstructing the respiratory tract (Mitsuka & Henry, 2002).
Therefore, although the hydrodynamics was considered the main factor influencing the occurrence and dominance of copepod groups, the results obtained in our study also indicate the relevance of the trophic degree as a structuring factor for this assemblage, when comparing the Paraná River (oligotrophic) with the adjacent mesotrophic lakes.In summary, the results obtained in our study evidenced a decrease in the copepod abundance along the longitudinal axis of the Paraná River, from upstream to downstream, refuting our first hypothesis.This was likely due to the influence of fauna coming from the reservoir in the organization of the copepod assemblage in the upstream stretch of the study area.
Our second hypothesis was partially refuted considering that the highest copepod diversity, contrary to our expectations, was registered in the Paraná River, although the highest abundance was verified in the floodplain lakes.Finally, the third hypothesis was corroborated, since the abundance and composition patterns of copepods were related to the environmental gradient.
Thus, we highlight the importance of the maintenance of the ecological integrity of the Upper Paraná River floodplain, in order to maintain the high biodiversity registered in the last undammed stretch of this important tropical river.We emphasize the importance of the regulation of outflow from reservoirs located upstream this stretch, which is a necessary condition to the existence of hydrological connectivity between the floodplain habitats and the main river, fundamental for the biodiversity of river-floodplain systems.

Figure 1 .
Figure 1.Sampling sites in the Upper Paraná River floodplain, between Porto Primavera reservoir and Itaipu Reservoir.Upstream: stretch directly influenced by Porto Primavera reservoir and the large tributaries of the left river bank; Downstream: stretch influenced by the small tributaries of the right river bank and Itaipu reservoir.

Figure 2 .
Figure 2. Variation in water level in relation to sea level (amsl) from August 2013 to July 2014.Areas signed with arrows represent sampled months.The value zero in the y-axis is represented by 231.5 (level zero corrected by altitude above mean sea level).

Figure 3 .
Figure 3. Copepod abundance in the three types of environment sampled.Only the most abundant species were considered.

Figure 4 .
Figure 4. Copepod abundance (a), species richness (b) and diversity (c) in the sampled stretches (upstream and downstream) of the main channel of the Paraná River.Symbols represent mean values and vertical bars represent standard error.

Figure 5 .
Figure 5. Copepod abundance of organisms (a), species richness (b) and diversity (c) recorded in the tributaries sampled in the different stretches (upstream and downstream) of the Paraná River.Symbols represent mean values and vertical bars represent standard error.

Figure 6 .
Figure 6.Copepod abundance (a), species richness (b) and diversity (c) recorded in the lakes sampled in the different stretches (upstream and downstream) of the Paraná River.Symbols represent mean values and vertical bars represent standard error.

Figure 7 .
Figure 7. Copepod abundance (a), species richness (b), and diversity (c) in each type of environment sampled.Symbols represent mean values and vertical bars represent standard errors.

Table 1 .
Palhiarini, W.S. et al.Mean values and standard deviation of environmental variables in sampling sites of the Upper Paraná River floodplain.River: upstream-downstream (P1 to P10).

Table 2 .
Species list and spatial occurrence of copepods collected in this study in the Upper Paraná River floodplain.