Contribution of different mesohabitats to the maintenance of fish richness and diversity in the lower Preto River

The presence of different mesohabitats in freshwater systems can support higher local and regional fish richness. Thus, their suppression by dam-building represents a real threat to aquatic biodiversity. Our aims were: (1) to survey the fish fauna in two different mesohabitats in lower Preto River: a riffle indirectly threatened by dam-building and a run with different physical structure; and (2) to analyze and compare the fish community structure in these different mesohabitats. Six samplings were made during one year in two reaches (R1 and R2). We conducted a ‘one way’ ANOSIM to assess the differences in fish community structure between R1 and R2. Fifty-three species were recorded, with the occurrence of Aphyocheirodon hemigrammus and Myleus tiete, two Brazilian threatened species. The highest richness was observed in R2. Nevertheless, diversity and evenness were significantly higher in R1. Rheophilic species were more common and restricted to R1 and species typical of lentic environments were predominant in R2. Fish community structure was different between R1 and R2 (R = 1, p = 0.02). Our results demonstrated that mesohabitats mosaic through rivers can contribute to the maintenance of a diverse fish assemblage.


Introduction
Brazil is the richest country in the world concerning freshwater fish, with approximately 3,000 species recorded to date (ICMBio, 2014).Despite this great richness and diversity, several Brazilian lotic systems are under strong anthropogenic pressure, especially related to dam-building, which modifies running (lotic) into standing (lentic) water ecosystems.This change has been associated with river homogenization, since it replaces the natural physical heterogeneity of the river by a large reservoir (AGOSTINHO et al., 2005).Different mesohabitats with particular characteristics (water velocity, depth, and substrate) in the same river can influence the fish community (YAN et al., 2011), supporting unique communities and contributing to species coexistence (TERESA; CASATTI, 2012).Thus, physical habitat can act as environmental filters by maximizing energy gain and reducing costs to different species (HALL et al., 1992).In this context, physical habitat homogenization caused by dam-building could be associated with deleterious consequences to aquatic biota (JOHNSON et al., 2008;WINEMILLER et al., 2008) and it may represent one of the most significant threats to diversity of freshwater fish (AGOSTINHO et al., 2005).
Despite these well-known negative impacts of dam-building, there are many hydroelectric projects in several Brazilian drainages, including in relative small rivers (AGOSTINHO et al., 2007).According to the hydroelectric sector, Small Hydroelectric Plants (SHPs) cause less impact due to small reservoir area.However, the hydropower potential of these systems is usually low in contrast with high fish richness and endemism (SCOTT; HELFMAN, 2001).The upper Paraná River basin comprises nearly 12% of the Brazilian fish diversity, and also contains an increasing number of new taxa, with more than 40 species described across the last five years (CARVALHO; LANGEANI, 2013).According to Liermann et al. (2012), the upper Paraná River basin is one of the most threatened ecoregions, due to heavy dam obstruction (more than 40% of the basin water surface) and above-average species richness and endemism.In the northwestern São Paulo State, two areas from the lotic drainage of the upper Paraná River basin, 'Cachoeira do Talhadão', in the Turvo River and 'Cachoeira de São Roberto', in the lower Preto River, are targeted by hydroelectric projects such as SHPs (HABTEC… 2010).However, the environmental cost associated with the impoundment in these areas could be very high, especially due to the lack of information about their importance concerning the fish richness, diversity, and conservation.
Therefore, in view of the dam-building pressure in this region and the physical and biological homogenization caused by the conversion of lotic to lentic habitats, the goals of this study were: (1) to survey the fish fauna in two different mesohabitats in lower Preto River, a riffle indirectly threatened by dam-building and a run with different physical structure; and (2) to analyze and compare the fish community structure in these different mesohabitats.

Study area
This study was performed in the Preto River, one of the most important rivers of the northwestern São Paulo State.The Preto River runs approximately 230 km from the headwaters (Cedral, São Paulo State) to its confluence with the Turvo River, in the upper Paraná River basin (COMITETG, 2009).Along its continuum, the upper Preto River is highly altered within an urban area at São José do Rio Preto city (São Paulo State), by stream flow channeling and impoundments.Near to its mouth, in the Turvo River, Preto River has a series of riffles, a tourist and recreation area targeted by dam-building (CETESB), popularly known as 'Cachoeira de São Roberto' (Pontes Gestal, São Paulo State).

Sampling and data analysis
Six samplings were conducted during one year (April 2013 to February 2014) in both sections.Fish were sampled using a small net (2.0 x 0.95 m, 2 mm mesh size) and a dip net (0.8 x 0.4 m, 2 mm mesh size).Standardized sampling effort (involving three collectors) was performed for two hours over a distance of approximately 150 m at each site.Fish were fixed in 10% formalin for 72 hours and then transferred to 70% alcohol for final preservation.Specimens were identified, counted, weighed, and vouchered in the fish collection of the Department of Zoology and Botany at Universidade Estadual Paulista "Júlio de Mesquita Filho", campus of São José do Rio Preto, São Paulo State (DZSJRP).
Species abundance in R1 and R2 was grouped and randomized as a function of the sampling effort, i.e., the number of sampling events (12 events, 6 from R1 plus 6 from R2), to generate an accumulation curve for the area (lower Preto River).The Chao 1 index (with 100 permutations), a function of the ratio of the number of species with only one specimen in a sample ('singletons') to the number of species with two specimens ('doubletons') was used to estimate the fish richness of the area.According to Magurran (2011), richness estimators based on Chao 1 are robust and are useful for revealing the true richness of an area.All analyses described above were performed in R statistical software.The Shannon index (based on natural logarithms) (alpha diversity) and Pielou Acta Scientiarum.Biological Sciences Maringá, v. 37, n. 3, p. 301-308, July-Sept., 2015 evenness were calculated for each site (R1 and R2) in the software Primer 6 (CLARKE; GORLEY, 2006), and then compared by a t-test in the software Statistica 7.0.The raw data of species abundance was squareroot transformed to reduce the contribution of highly abundant species and used to create a Bray-Curtis similarity matrix.To assess the differences in fish community structure between R1 and R2, this Bray-Curtis matrix was then compared using a 'one way' ANOSIM with 999 permutations, at significance level of 0.05.A SIMPER routine was used to identify which species contributed most to the dissimilarity between sites.Analyses referring to fish community structure were performed in Primer 6.0 (CLARKE; GORLEY, 2006).

Discussion
According to our results, habitat characteristics can influence several community aspects, like fish richness, diversity, abundance, and biomass.The section R2, with lower water velocity, sand substrate and marginal grasses, showed higher fish richness and abundance, while R1 (Cachoeira de São Roberto) presented greater fish biomass, diversity, and evenness.Additionally, the community structure was different between R1 and R2.Species sampled in the lower Preto River represent 15% of the total species recorded in the upper Paraná River basin (CARVALHO; LANGEANI, 2013), and this number can be higher in view of the richness estimate (65 spp.) for the area.The high species richness found in this study (53 spp.) showed the importance of this river portion to the maintenance and conservation of the ichthyofauna in this basin.
The lower alpha diversity and evenness registered in R2 were due to the dominance of the characids Hyphessobrycon eques and Serrapinnus notomelas.Both species are associated with grasses (Poaceae) and macrophytes on the riverbanks, similarly to other studies (i.e.CASATTI et al., 2003;CETRA;PETRERE JR., 2007).Marginal grasses associated with deforestation are usually responsible for physical, biological, and functional stream homogenization (BUNN et al., 1997;CORTELEZZI et al., 2013;CASATTI et al., 2015).Nevertheless, in large rivers less connected to riparian vegetation and, consequently with lower input of wood debris (important resource to stream habitat complexity), these marginal grasses can promote physical habitat heterogeneity and increase the availability of microcrustaceans (CORTELEZZI et al., 2013) and aquatic insects (CENEVIVA-BASTOS; CASATTI, 2014), food resources consumed by small opportunistic species, such as Hyphessobrycon eques (PELICICE; AGOSTINHO, 2006) and Serrapinnus notomelas (ZENI; CASATTI, 2014), respectively.The presence of marginal grasses can also favor medium and large-sized species, since juveniles of Acestrorhynchus lacustris (Lütken, 1875), Erythrinus erythrinus, Galeocharax gulo (Cope, 1870), Hoplias cf.malabaricus (Block, 1794), Myleus tiete (Eigenmann & Norris, 1900), and Pterygoplichthys ambrosettii (Holmberg, 1893) were found along the banks during the period of fish migration and reproduction ('piracema').According to Rozas and Odum (1988), submerged aquatic vegetation can provide refuges from predators, reducing predation pressure and increasing survival of large-bodied species juveniles.One of them, Myleus tiete, just as well Aphyocheirodon hemigrammus, are listed as threatened for Brazil and São Paulo State (AGOSTINHO et al., 2003;OYAKAWA et al., 2009;ICMBio, 2014).In this context, marginal areas occupied by grasses in large rivers can play a nursery role, besides increasing spatial heterogeneity, important factors for the maintenance of different fish species, which can play different functional roles in the community.
Significant differences in community structure between the sections may result from the interaction between structural characteristics (physical heterogeneity and microhabitat availability), food resource availability and from their consequences on the species fitness.Species are usually more abundant at sites where energy gain is maximized, i.e., where the ratio of the energy obtained vs the energy spent on maintenance is high (HALL et al., 1992).In this way, habitat structural characteristics, such as water velocity, substrate and depth, can be 'species-selective' through their influence on this energetic relationship, increasing or decreasing the maintenance costs of the organism.Thus, it is possible to observe a set of morphological features shared among the most frequent and abundant species in each site.In R1, Loricariidae (i.e., Hypostomus spp.) with a low body depth, ventral mouth and pectoral fins (used to deflect water flow) can attach to rocky substrates, decreasing the energy required for establishment and maintenance at sites with high water velocity.A shallow depth and a rocky substrate provide suitable surfaces for periphyton growth, a food item generally used by these species (GARAVELLO; GARAVELLO, 2004).Furthermore, drift feeders such as Astyanax fasciatus, Bryconamericus stramineus, and Planaltina britskii obtain the maximum input of food in sites with running water flow.For these species, energy efficiency, namely, the energy acquired from food resources minus the energy spent in uninterrupted swimming, is high in riffles, increasing the abundance of these species in R1.The differences in these species abundance between R1 and R2 were the major factor accounting for the dissimilarity on the fish community structure.
In this context, spatial heterogeneity of the mesohabitat represents a crucial conservation measure, because it allows the coexistence of different species in the same environment.However, a general trend towards the homogenization of aquatic ecosystems has been observed worldwide (RAHEL, 2002) and dam-building is one of the most significant impacts related with all levels of homogenization: physical, biological, and functional.The loss of the natural river mesohabitat mosaic caused by an artificial reservoir, usually lead to changes on species composition by replacing high trophic level specialist into low trophic level generalist species (HOEINGHAUS et al., 2009).The negative consequences of dam-building on fish community structure can be observed in the upper Preto River, which has a complex of three dams used to supply water to São José do Rio Preto city, São Paulo State.In a study performed in these artificial reservoirs, Andrade (2003) found higher frequency of non-native species, especially Tilapia rendalli (Boulenger, 1897) and Oreochromis niloticus (Linnaeus, 1758).At the present moment, the construction of Small Hydroelectric Plants on the lower Preto River (R1) has not been approved by the "Companhia Ambiental do Estado de São Paulo" (CETESB) because the alterations caused by these projects are environmentally impracticable (CETESB, 2012).According to our findings, the presence of different mesohabitats in the lower Preto River can support higher fish richness with occurrence of rare species or threatened with extinction, such as Aphyocheirodon hemigrammus and Myleus tiete.Thus, the fish community that will be affected by this impoundment has to be taken in consideration, especially because this change may affect the local and regional pool of species.In this way, the fish fauna surveys in these areas are important not only in terms of the diverse and unknown fauna of the region, but can also guide and/or support management and conservation measures of aquatic systems.As a matter of fact, this paper can bring one important contribution that can support government decisions regarding fish conservation.

Conclusion
The results demonstrate that the mosaic of mesohabitats along a river can contribute to the maintenance of a diverse fish assemblage, consequently promoting higher local and regional fish richness.It is possible that abiotic characteristics, such as water velocity, substrate composition, and marginal area interact with biotic aspects (competition and predation pressure) and influence species occurrence.In this context, the physical habitat homogenization due to the suppression of these mesohabitat patches by impoundments will threaten not only fish richness and diversity in the lower Preto River, but also the regional pool of species of the upper Paraná River basin.Thus, studies conducted before damming can represent a valuable tool for governmental decisions toward fish conservation.

Figure 1 .
Figure 1.Location of the two stretches (R1 -Cachoeira de São Roberto and R2 -near the mouth of the Botelho stream) in the lower Preto River, Northwestern São Paulo State, Brazil.

Figure 2 .
Figure 2.Stretches sampled in the lower Preto River.R1 -Cachoeira de São Roberto, a riffle mesohabitat with high water velocity with some slope, substrate primarily composed of slab and basaltic boulders; R2 -near the mouth of the Botelho stream, a run mesohabitat with low water velocity, sandy substrate and marginal grasses.