Is it possible to simplify environmental monitoring ? Approaches with zooplankton in a hydroelectric reservoir É possível simplificar o monitoramento ambiental ?

Aim: In order to contribute to the knowledge about the simplification of biological surveys, this study evaluated the use of substitute groups, numeric and taxonomic resolution for the three main groups of zooplankton (cladocerans, copepods and rotifers) in a Hydropower Plant (UHE). The following issues were addressed: (i) the patterns of spatial and/or temporal ordering generated between each zooplankton group are in concordance? (ii) The concordance is maintained using presence/absence data instead of density data? (iii) The identification of organisms to the species level can be replaced by genus or family level? Methods: Samples were taken in seven sample units over five campaigns between 2009 and 2010 in the UHE Serra da Mesa (Goiás, Brazil). To evaluate the correlation between each pair of matrices was used the Mantel test. Results: The results demonstrate that the replacements should not be made among the zooplanktonic groups, requiring the monitoring of three groups (copepods, cladocerans and rotifers). Furthermore, the results suggest the use of density data of individuals rather than just presence/absence of species. Finally, the results of this study indicate the possibility to use data at species level instead of data at genus or family level. Conclusion: For zooplankton community monitoring purposes only the use of taxonomic resolution showed to be efficient for this area of study, not being recommended the use of surrogate groups or numerical resolution.

data (presence/absence) instead of organisms density (numerical resolution) would be sufficient (Carneiro et al., 2010;Heino, 2014).The use of simplifications in biomonitoring of aquatic environments becomes extremely relevant given the growing need for global supplies that add to the anthropogenic effects on the water quality of many reservoirs, such as: land use, inadequate water management and climate variations (Peters & Meybeck, 2000;Lee & Biggs, 2015).
Among the main groups of aquatic organisms, zooplankton stands out for being able to react quickly to environmental and toxicological changes (Moreira et al., 2014;Vieira et al., 2011).In this sense, the zooplankton can be used as variable for environmental monitoring because, in addition to responding human impacts, it plays an important role in nutrient cycling and energy flow of food webs (Gagneten & Paggi, 2009;Oberhaus et al., 2007;Vieira et al., 2011).
The aim of this study was to investigate the possibility of using only one of three zooplankton component groups: copepods, cladocerans and rotifers.It was also evaluated the use of numerical and taxonomic resolutions in this community.Thus, we sought to answer the following questions: (i) Are the spatial and/or temporal ordering patterns generated between each zooplankton group concordant?(ii) Is the concordance maintained using presence/absence data as a substitute for density data?(iii) Can the identification of organisms to the species level be replaced by genus or family level?Our expectations are: i) copepods and cladocerans might exhibit higher concordance because they are phylogenetically closer (both crustaceans) and they have more similar ecological niches ii) that it is possible to replace density data for presence/absence species data and iii) the substitution of data at species level may be also possible, both for genus and family level.

Introduction
The environmental impacts have been occurring faster than we can monitor and maintain biodiversity (Bini et al., 2007;Landeiro et al., 2012;Martinelli et al., 2010).The increase of impacted areas affects directly and indirectly the functioning of terrestrial and aquatic ecosystems, highly contributing to the extinction of species (Ceballos et al., 2015).In this way, the monitoring of the biological communities becomes an important practice to evaluate the level of environmental degradation, mainly through variations in richness, density, and functional traits (Harmon et al., 2009;Ribeiro et al., 2016).
Monitoring programs are most effective when they evaluate environmental and biological dimensions across space and time (Alahuhta & Aroviita, 2016;Alahuhta et al., 2016).However, the increase in sampling number brings cost growth, while evaluate biological dimension brings the need of specialists in different biological groups.To mitigate this problem is possible to use biological surrogate groups, which is the use of substitution between groups that present congruent patterns over time and space for biomonitoring purposes (Padial et al., 2012) or to work with presence/absence of species (rather than density) at lower taxonomic resolutions (Gomes et al., 2015;Machado et al., 2015).

Study area
The Serra da Mesa reservoir is located in Brazilian Midwest region (Figure 1) and its main drainage basin is constituted by the Tocantins River (Caramaschi et al., 2012).The dam's Hydropower Plant (UHE) of Serra da Mesa was built in 1996 in the upper Tocantins River and became the largest reservoir of the country in water volume, with 54.4 billion m 3 and an area of 1,784 km 2 (Caramaschi et al., 2012).The climate is tropical rainy with dry periods and average temperature of 20 °C (Sousa, 2003).
The samples were collected with a motorized pump at a depth of 0.5 m, being filtered 1000 of water per sample through a 68 μm mesh plankton net.The collected material was conditioned in polyethylene flasks and fixed in 4% formaldehyde solution.Samples were concentrated into a volume of 100 mL and the survey of density was conducted with 10 mL of downsampling in Sedgewick-Rafter chambers, taken with the Hensen-Stempel pippete and then analyzed using an optical microscope (Bottrell et al., 1976).After the counts, qualitative analyzes were performed, taking sub-samples until no new species were found.The final density was expressed in individuals * m -3 .

Data analysis
Firstly, all data were separated by species, genus, families and zooplanktonic groups (copepods, cladocerans and rotifers) and transformed into Acta Limnologica Brasiliensia, 2017, vol.29, e8 logarithmic scale (x + 1).Subsequently, the data set was converted into distance matrices using Bray-Curtis for density data and Jaccard index for presence/absence species (Legendre & Legendre, 2012).
To evaluate the correlation between each pair of matrices was used the Mantel test (Legendre & Legendre, 2012), which calculates the correlation between two distance matrices.The correlation strength was measured using the value of r (ranging from -1 to +1) and its significance was measured using 999 randomizations.
All analyzes were performed using vegdist and mantel functions of vegan package (Oksanen et al., 2013) in the statistical software R (R Core Team, 2016).

Results
A total of 45 taxa were sampled, being distributed as follows: 27 taxa of rotifers, 12 of cladocerans and six of copepods (Table 1).Regarding density, 354,320 individuals were sampled, with 210,359 rotifers, 97,289 copepods and 46,672 cladocerans, comprising respectively: 59.4%, 27.4% and 13.2% of all occurrences.The rotifers had the highest species richness in all campaigns and higher densities in Aug/09 and Nov/09 (Figure 2).
Only copepods and rotifers presented spatial distribution patterns concordant being the high level of concordance in Nov/09 and medium in Aug/09 and Sep/10 (Table 2).
The results of the numerical resolution ranged between the campaigns in a way that significant values were observed for all groups in Nov/09 and only for copepods in March/10 and cladocerans in Sep/10 (Table 3).On the other hand, the taxonomic resolution showed high values and homogeneous in all campaigns, indicating that the use of data on genus or family level is equivalent to the use of species (Table 3).The copepods were not included in this analysis because it presented a small number of species and genera.

Discussion
The results showed that the three zooplanktonic groups (copepods, cladocerans and rotifers) should not be used as substitutes for each other, due to high variability over time with relatively low levels of concordance (<0.7) (Heino, 2010).Therefore, our expectations were not corroborated, since (i) the correlation was higher among copepods and rotifers and (ii) the few significant concordance levels were low.
Even among taxonomically different biological groups, such as macrophytes and macroinvertebrates (Traversetti et al., 2015), vascular plants, bryophytes and birds (Rooney & Azeria, 2015), concordant distribution patterns has been found.However, Our results do not indicate concordance between the zooplanktonic community groups.This absence of concordance was also observed in other studies related to zooplankton community (Bessa et al., 2011;Bini et al., 2008;Vieira et al., 2015).Although the species of each group may respond differently to environmental gradients (Adamczuk et al., 2015) the absence of concordance among copepods, cladocerans and rotifers is interesting, once they are part of the same assembly.This result reinforces the need to incorporate all zooplanktonic groups (copepods, cladoceran e rotifers) in environmental monitoring programs.
The numerical resolution also showed that the use of presence/absence values of species as substitute of density data is not indicated, which is similar to results found by other authors      (Bessa et al., 2011;Giehl et al., 2014;Heino, 2008;Valente-Neto et al., 2016).However, some studies indicate the use of numerical resolution for zooplanktonic community (Gomes et al., 2015) and other groups (Carneiro et al., 2010;Ribas & Padial, 2015).
Our results indicate that the use of genus level for zooplankton groups of Serra da Mesa hydroelectric reservoir would be equivalent to the use of the species.Similar results were found for different groups (Giehl et al., 2014;Souza et al., 2016;Valente-Neto et al., 2016) including zooplankton (Carneiro et al., 2013).The concordance with higher taxonomic levels can be justified by the low number of species found by genus and family, which creates a higher similarity between the more and less specific data sets (Giehl et al., 2014).This approach is acceptable when there is not a major loss of information between the data sets (high concordance level), thus resulting in lower effort for taxonomic identification of taxa (Carneiro et al., 2010;Heino & Soininen, 2007;Khan, 2006).A study using ants, for example, revealed that the reduction of actual costs for the use of taxonomic resolution at genus level could reach 40% (Souza et al., 2016).
Finally, for zooplankton community monitoring purposes, only the use of taxonomic resolution (Dens = density; PA = presence/absence of species) and taxonomic (Sp = species, Gn = genus, Fa = family); P<0.05; Copepods were not included in the second part of the analysis due to their low species richness.
Acta Limnologica Brasiliensia, 2017, vol. 29, e8 showed to be efficient for this area of study, not being recommended the use of substitutes or numerical resolution.It is clear that this strategy (identifying organisms only at the genus level) should not be used indiscriminately.Even with the high values found for Mantel r, only in extreme situations, such as the absence of taxonomists, the need for immediate assessments and/or a significant financial shortfall, the taxonomic resolution at the genus level would be an interesting alternative.

Figure 2 .
Figure 2. Species richness and density of rotifers, cladocerans and copepods in Serra da Mesa Reservoir (Brazil).