Fish diversity in the cascade of reservoirs along the Paranapanema River , southeast Brazil

The Paranapanema River is a major tributary of the upper Paraná river basin. Eleven hydropower dams regulate its main course, but no study has investigated fish diversity in these impoundments at the basin-scale. The present study investigated spatial patterns of richness, composition, and abundance of native (non-migratory and migratory) and non-native fishes in the cascade of reservoirs along the Paranapanema River. The study is based on data collected from 34 independent studies conducted in nine reservoirs (47 samples). The compilation recorded 161 species, being 111 native (14 migratory) and 50 nonnative. Total richness ranged between 56 and 112 species/reservoir, with a mean of 72 (49.9 non-migratory, 8.1 migratory and 14 non-native). The number of non-migratory species showed no spatial trend along the cascade system, but migratory and non-native richness increased toward downstream reaches. We also observed spatial variation in species composition along the cascade system, but some non-native fishes were widely distributed. Migratory fishes showed low relative abundance (usually < 10%), while non-native species were common and more abundant, especially in reservoirs downstream. Our results revealed a high diversity of fishes in the cascade of impoundments, but indicated that migratory fishes are rare, while non-native species are common or dominant.

The upper Paraná river basin is a special case.This basin is severely regulated by hundreds of large and small dams, which cover almost half of the impounded area in Brazil (Agostinho et al., 2007a).In addition, the main course of all major tributaries (i.e., Paranaíba, Grande, Tietê, Paranapanema) is regulated by cascades of large dams (Agostinho et al., 2016).For example, eleven hydropower dams regulate the main course of the Paranapanema River, creating a cascade of impoundments along the entire river system.Studies that investigated fish diversity in these reservoirs (e.g., Barrella, Petrere Jr., 2003;Hoffmann et al., 2005;Duke Energy, 2008;Orsi, Britton, 2014;Santos et al., 2017;Garcia et al., 2018) indicated that assemblages are composed primarily of smallsized sedentary species, in addition to several non-native fishes.These studies, however, focused on specific impoundments; no study has investigated the whole cascade system, in order to provide a more complete picture about taxonomic diversity and abundance patterns.It is poorly known, for example, whether assemblage structure is homogenized among reservoirs (e.g., Petesse, Petrere Jr., 2012) or contingent to each impoundment.This information is important to guide management and conservation plans, since the Paranapanema basin is severely disturbed by different human activities (e.g., dams, agriculture, urbanization, aquaculture), demanding specific actions to preserve fish diversity and fishing stocks.
In this sense, this study gathered the available information to provide a comprehensive overview of fish diversity in the cascade of reservoirs along the Paranapanema River.
Our main objective was to reveal the structure of fish assemblages in these impoundments, investigating spatial patterns of species richness, composition and abundance of native (non-migratory and migratory) and non-native fishes.In particular, we predict that non-migratory and non-native fishes dominate impoundments in terms of species richness and abundance, as evidenced by previous studies (e.g., Barrella, Petrere Jr., 2003;Petesse, Petrere Jr., 2012;Agostinho et al., 2016).We also expected a trend of increasing species richness, together with changes in composition, along the cascade -a common pattern in large river systems (e.g., Oberdorff et al., 1993;Araújo et al., 2009).Finally, we predict that migratory species are rare (abundance and richness) in impoundments, especially in segments where dams are close to each other (middle/upper reaches) and critical habitats are limited or lacking.

Material and Methods
Study area.The Paranapanema River is a main tributary of the upper Paraná river basin.Its sources are located in São Paulo State, and the river runs west ca.900 km, mostly along the border between São Paulo and Paraná States, to join the Paraná River downstream.Dozens of hydropower dams were built in the basin during the 20 th century.Eleven large dams regulate the main channel (Fig. 1), affecting the upper, middle and lower reaches of the basin.The first large impoundment date back to 1932, but most dams were constructed after 1970 (Tab.1).There are marked variations in the size of reservoirs; the largest are located in upper and lower reaches (Tab.1).Fish passages (ladders) were installed in five dams, namely Canoas I, Canoas II, Ourinhos, Paranapanema and Piraju (Arcifa, Esguícero, 2012;Britto, Carvalho, 2013).Data collection.We gathered data on fish diversity (species richness, composition and abundance) from studies conducted in the basin (Tab.2).We selected only studies conducted in the impounded area (i.e., lentic and semi-lentic habitats) and adjacent environments directly affected by the impoundment (i.e., upstream stretches, tributaries, lateral lakes, and fish passage facilities).We excluded studies or data collected in environments not affect by the impoundment (i.e., free-flowing streams and rivers), in addition to sites downstream from the dam.We searched for studies in different databases (Web of Science, Scopus, SciELO and Google), private libraries, and performed cross-reference searches.We performed exhaustive searches, and used different keywords in databases (e.g., fish, ichthyofauna, diversity, Paranapanema, Paraná).The survey collected published references (scientific journals and book chapters, 78%) and unpublished material (theses, reports and databases, 22%) (Tab.2).
In total, we gathered 34 independent studies that investigated fish diversity in nine reservoirs, between 1991 and 2012 (Tab.2).Some studies investigated more than one reservoir, so our sample size totaled 47 samples.Sampling effort was distributed heterogeneously: Jurumirim, Salto Grande, Capivara, and Rosana received more effort, while Ourinhos and Taquaruçu were less investigated.We found no study conducted in Piraju and Paranapanema reservoirs.Fish sampling employed different gears, but gill nets and seines prevailed (Tab.2).Most samples (77%) were collected at least 10 years after the construction of the impoundment.It means that fish assemblages have responded to reservoir conditions and community structure have shifted to another regime and stability (Agostinho et al., 2007a).Our study, therefore, focus on reservoir fish assemblages.
We obtained species lists from each sample, which were organized to build a "sample x species" matrix.Each sample was considered an independent unit (Tab.2), and we only considered taxa identified to the species level; those reported as "sp." or "spp." were discarded.Species names and taxonomic classification were carefully checked and, when necessary, updated and corrected following Froese, Pauly (2016), Eschmeyer et al. (2017), Eschmeyer, Fong (2017) and recent literature; updates are shown on Tab.S1 ( Available only as online supplementary file accessed with the online version of the article at http://www.scielo.br/ni).We then assigned species as native (non-migratory or long distance migratory) and non-native.We used pertinent literature to assign species as long distance migratory, i.e., those that migrate more than 100 km (sensu Agostinho et al., 2003).Equally, we followed scientific literature to assign species as non-native, i.e., species introduced to the Paranapanema river basin (sensu Reis et al., 2003;Graça, Pavanelli, 2007;Julio Jr. et al., 2009;Britton, Orsi, 2012;Garavello et al., 2012;Orsi, Britton, 2014;Ortega et al., 2015;Froese, Pauly, 2016;Garcia et al., 2018).Abundance data was recorded whenever available (Tab. 2), and used to calculate species relative abundances (%).
Data analysis.Analyses focused on three groups: (i) native non-migratory species (NT); (ii) native long-distance migratory species (LDM); and (iii) non-native species (NNS).All analyses considered these three groups separately.
To investigate total taxonomic richness in the impoundments, we calculated species accumulation curves based on sampling effort (n = 47).Sample order in the original matrix (sample x species) was randomized 500 times, building expected curves for native (NT + LDM) and non-native species (NNS).We also calculated three nonparametric estimators to estimate total richness: Chao-1, ICE (Incidence-based Coverage Estimator), and Jackknife.These analyses were conducted in the software EstiMateS 5.0 (Colwell, 2006).
Tab. 1. Hydroelectric dams along the Paranapanema River.Information about dams: date of construction, power installed (MW), dam height (m), distance from the mouth (km), presence of fishway.Information about reservoirs: area (km 2 ), volume (x 10 6 m 3 ), residence time (average days), maximum depth (m) and length (km).Source: Agostinho et al. (2007a); Duke Energy (2008); Pelicice et al. (2015a); websites of hydropower companies."-" = unavailable information.Basin subdivision in upper, middle and lower reaches followed the Sistema Integrado de Gerenciamento de Recursos Hídricos do Estado de São Paulo (Governo do Estado de São Paulo, 2018).To investigate spatial variation in species richness and composition, we pooled the 47 samples within each reservoir to build a "reservoir x species" matrix.We then calculated total richness per reservoir for each group (NT, LDM and NNS).To investigate spatial variations in species composition, we used Non-Metric Multidimensional Scaling (NMDS) based on Jaccard distance (presence/absence).We conducted three NMDS, one for each group.These analyses were carried out in the software Past 1.75b (Hammer et al., 2001).
We also investigated trends in abundance.Twenty-two studies (Tab.2) provided abundance data for all species in the assemblage (number, density or capture per effort); we used these data to calculate the relative abundance (%) of each species, generating an abundance-based "sample x species" matrix (32 samples).We used this matrix to investigate spatial variation in the relative abundance of LDM and NNS in assemblages, by plotting samples against reservoirs (scatterplot).We also calculated the range of relative abundance (minimum-maximum) for each species along the cascade system.

Results
We recorded 161 fish species in the cascade of reservoirs along the Paranapanema River, being 111 native and 50 non-native to the basin (Tab.3).Most species belonged to the order Characiformes, followed by Siluriformes and Cichliformes.

Pseudoplatystoma fasciatum 0
Fish diversity in the cascade of reservoirs

Discussion
Our study is the first to compile data on fish diversity in the cascade of reservoirs along the Paranapanema River.This compilation, based on studies conducted in the last decades (1990 to 2012), revealed the presence of 161 species, being 111 native and 50 non-native.The accumulation curve indicated that most native species in the reservoirs were recorded, and we estimated total richness to range between 125 and 132 species.Previous estimates indicated 155 fish species in the Paranapanema river basin (Duke Energy, 2008), and 270 to 310 in the upper Paraná river basin (Agostinho et al., 2007b;Langeani et al., 2007).Impoundments in the Paranapanema River, therefore, hold a high number of fish species and contribute significantly to the basin species pool; a number that may increase if other environments are sampled (i.e., downstream segments, streams, free-flowing stretches).We are aware that misidentifications, re-descriptions and taxonomic revisions may add uncertainty to estimates in impoundments, and we followed recent literature to avoid or minimize nomenclatural inconsistencies.Future studies may refine species lists and distribution ranges, including wholebasin monitoring programs, with standardized protocols and multi-gear sampling strategies.
The number of native species varied among reservoirs in the cascade system (range: 45 to 92), but most impoundments had similar values (ca.50 species).Nonmigratory fishes dominated assemblages, confirming our prediction and previous studies (e.g., Agostinho et al., 2007a).These species prevailed in richness, abundance and occurrence, especially small and medium-sized characids, a common pattern in Neotropical reservoirs (Barrella, Petrere Jr., 2003;Luiz et al., 2005;Agostinho et al., 2016).These values are higher than those reported by Agostinho et al. (2007a), which found between 20 and 40 species in Brazilian reservoirs.However, Agostinho et al. (2007a) estimates were based on gillnets, while studies gathered here used different sampling gears.In addition, some reservoirs were sampled multiple times over a considerable period (e.g., Jurumirim, Salto Grande, Capivara, Rosana).We highlight also that impoundments in the Paranapanema River provide relevant habitats for small-sized species, such as submerged trees, aquatic macrophytes and lateral lagoons (e.g., Casatti et al., 2003;Pelicice et al., 2005;Ferrareze, Nogueira, 2011).In particular, three reservoirs (i.e., Ourinhos, Salto Grande and Capivara) were more species rich, but different mechanisms may be involved.Ourinhos (60 species), for example, was evaluated soon after its filling; in this stage, fishes are disoriented and populations are randomly distributed, with weak environmental selection.Previous studies have reported increases in species richness in the first years after river regulation, with declines in subsequent years (Agostinho et al., 2007a).Salto Grande (67 species), on the other hand, is a small and old impoundment sampled with a variety of gears, what may result in surveys that are more complete.In addition, environmental heterogeneity Tab. 6. (continued).networks (e.g., Vitorino Jr. et al., 2016), probably increasing the similarity among reservoirs (e.g., Petesse, Petrere Jr., 2012).Data collected prior to river regulation is crucial to understand community assembling processes (e.g., Araújo et al., 2013;Lima et al., 2016), but this type of data is rare in the Paranapanema basin.
As expected, migratory fishes showed low species richness and abundance in the cascade system.Although we recorded the presence of 14 species, especially in lower reaches, the abundance of migratory fishes was very low in all impoundments.The decline/collapse of migratory species is recurrent in regulated rivers (Petrere Jr., 1996;Okada et al., 2005;Godinho, Kynard, 2008;Pelicice, Agostinho, 2008), since these fishes demand different fluvial habitats for feeding and reproducing, migrating over variable distances to complete their life cycle (Pompeu et al., 2012).The cascade of dams along the Paranapanema River prevents long-distance migration and creates environments adverse to migratory fishes (i.e., lentic areas, flow regulation).Pimelodus maculatus was the most important, but it also summed low abundance in most samples.This catfish is usually common in reservoirs of the upper Paraná basin, probably because it is small-sized, migrates over shorter distances and displays opportunistic feeding behavior (Lima-Junior, Gointein, 2004;Paschoalini et al., 2013).Large-sized species, in particular, were recorded only downstream Salto Grande dam.This pattern may be related to biogeographic factors, since waterfalls were present in the area before dam construction, constraining fish movements.However, the position of reservoirs in the cascade must also be considered to explain the distribution of migratory fishes.The upper/middle reaches (between Jurumirim and Canoas I) are the most affected by river regulation, because 8 dams are distributed over 200 km.This sequence of impoundments creates sections where lotic segments are short or absent, with no critical habitat for migratory and rheophilic fishes (e.g., lateral lagoons, backwaters, free-flowing stretches, tributaries).The area with greatest potential to support migratory fish populations is the upper section of the Capivara reservoir, which preserves large tributaries (Hoffmann et al., 2005;Vianna, Nogueira, 2008;Bennemann et al., 2013), especially the Tibagi river (Bennemann et al., 2000;Shibatta et al., 2007).This impoundment showed the highest species richness, including virtually all migratory species of the basin.Early life forms of large-size species have been recorded in the area, indicating the presence of spawning areas and nursery grounds (Orsi et al., 2016).However, the construction of fish ladders at Canoas I and II dams affected the distribution of migratory fishes.Populations that recruited in the area of Capivara passed upstream and remained confined in short sections between Canoas I and Salto Grande dam (Lopes et al., 2007;Pelicice, Agostinho, 2008;Britto, Carvalho, 2013).The greater abundance of migratory species in Canoas I and II indeed suggest that fishes migrated upstream and evaded Capivara reservoir soon after the installation of ladders, as indicated previously may have played a role, considering that Salto Grande has relevant habitat diversity (Brandão et al., 2009).This mechanism seems to be even more important in Capirava (80 species), which preserves remnant lotic areas and large tributaries (Hoffmann et al., 2005;Pelicice, Agostinho, 2008;Bennemann et al., 2013).The remaining reservoirs showed similar values of species richness (ca.50 species), although we expected increasing values along the cascade system (e.g., Oberdorff et al., 1993;Araújo et al., 2009;Foubert et al., 2018).The convergence of species richness among impoundments (see also Agostinho et al., 2007a) is probably linked to the loss of fluvial heterogeneity along the main channel, together with the creation of similar and homogeneous habitats (i.e., impoundments).Neotropical fishes are primarily rheophilic, demanding specific habitats and flow conditions to survive and reproduce; few species evolved traits to cope with lentic habitats (Gomes, Miranda, 2001).Declines in species richness is a common pattern in rivers impacted by dams (e.g., Barrella, Petrere Jr., 2003;Agostinho et al., 2007a;Hoeinghaus et al., 2009;Petesse, Petrere Jr., 2012;Araújo et al., 2013).Long-term studies, in particular, have shown that fish assemblages in impounded areas are substantially different from the structure found in pristine conditions (Mol et al., 2007;Orsi, Britton, 2014;Lima et al., 2016).Future studies, therefore, must improve our understanding about factors that determine species richness in Neotropical reservoirs, as a means of improving predictions about diversity losses in new impoundments.
While most reservoirs were similar in respect to species richness, the composition of native non-migratory fishes changed considerably along the cascade.Some species were widely distributed, but each reservoir tended to have specific composition.Variation in fish composition along longitudinal gradients is a common pattern in large river systems (e.g., Araújo et al., 2009;Foubert et al., 2018), usually in response to spatial and environmental components.Species are deleted/added to local assemblages, creating turnover and/or nestedness patterns along the river corridor.Impoundments, however, add discontinuity to the river continuum (Ward, Stanford, 1983) and provide common habitats irrespective to their position in the basin, usually associated with lentic conditions, shallow shores, vast water column and deep sites.These new environmental conditions select common traits and homogenize fish diversity in functional terms (Santos et al., 2017).In this sense, reservoirs promote similar pressures upon the fish fauna and favor opportunistic or lentic-adapted species from the local species pool, forming simplified versions of the original riverine community.However, species are assembled from different species pools along the cascade system, resulting in divergent species composition among reservoirs.Therefore, even though species richness and functional traits are convergent among impoundments, taxonomical composition is divergent (i.e., beta diversity), explaining the high species richness observed in the cascade system.However, the sequence of impoundments must have changed beta diversity patterns that characterize riverine 14 e170150[14] (Pelicice, Agostinho, 2008;Britto, Carvalho, 2013).Another region with potential to maintain migratory fishes is the upper section of Rosana reservoir, which has a short lotic stretch with fluvial lagoons.However, Rosana dam is located ca. 25 km from the confluence with the Paraná river; this dam blocked upward fish movements and isolated the Paranapanema basin (Antonio et al., 2007).In addition, Ferrareze, Nogueira (2011) captured no young migratory fish in lateral lagoons, possibly due to the lack of spawning grounds upstream.Similarly, Kipper et al. (2011) reported very low abundance of larvae across the reservoir, indicating that recruitment is negligible.Finally, large scale monitoring did not record eggs and larvae of important large-size species (e.g., Piaractus mesopotamicus, Prochilodus lineatus) in this reservoir (Orsi et al., 2016).Therefore, migratory fish populations in reservoirs along the Paranapanema River are significantly disturbed, probably with low or no recruitment in many sections and impoundments, particularly in middle and upper reaches.The maintenance of migratory fish populations, much appreciated in riverine and floodplain fisheries, will depend on the preservation of lotic remnants and fluvial habitats (e.g., Vianna, Nogueira, 2008;Suzuki et al., 2013;Barzotto et al., 2015), together with the revision of current management practices (i.e., Pompeu et al., 2012).In this sense, the area affected by Capivara reservoir should receive conservation efforts, particularly to maintain ecosystem functionality in tributaries (e.g., Tibagi, Cinzas).
Another important result is the high number of non-native species in the impoundments (50 species), which accounted for 30.5 % of total richness.This is an impressive high number, confirmed by another recent study that reported 47 non-native species in the basin (Garcia et al., 2018).This scenario is not an exception, considering that the upper Paraná river basin is highly invaded by multiple non-native species (e.g., Britton, Orsi, 2012;Vitule et al., 2012;Orsi, Britton, 2014;Ortega et al., 2015).Our data showed that some species have wide distribution (e.g., P. squamosissimus, H. eques, O. niloticus, S. marginatus, Schizodon borelli, Metynnis lippincottianus, C. kelberi), explaining why most reservoirs were similar in composition.However, many species were restricted to one or a few reservoirs, probably because of sporadic introductions that occurred in specific impoundments.This trend created a compositional turnover along the cascade, increasing the richness of non-native species at the basin level.In general, introductions occur due to different reasons (Ortega et al., 2015;Azevedo-Santos et al., 2015), but aquaculture, fish stocking, sport fishing, aquarium dumping, and the removal of natural barriers probably caused massive fish introductions in the Paranapanema River (Júlio Jr. et al., 2009;Agostinho et al., 2010;Britton, Orsi, 2012;Garcia et al., 2018).These actions caused the release of species with high invasive risk, which are currently widespread (e.g., O. niloticus, Coptodon rendalli, C. kelberi, S. marginatus, P. squamosissimus).It is not possible to confirm that all 50 species are established in these impoundments (i.e., recruiting), but the presence of early life forms collected in different localities (Almeida et al., 2018) indicate that some are reproducing.Furthermore, non-native species were abundant in most samples, summing relevant fraction or dominating local assemblages.Reservoirs located downstream were more invaded (i.e., species richness and abundance), supporting previous studies that reported high invasion rates (Garcia et al., 2018) or strong impacts (Orsi, Britton, 2014;Pelicice, Agostinho, 2009).Jurumirim reservoir was the less invaded and, consequently, has been considered a model for sustainable artisanal fisheries (Novaes, Carvalho, 2013), because fishing activities are based on native species, with higher prices, and may be sustained in the long term.The current scenario in the cascade of impoundments, therefore, demands immediate management actions to control populations (e.g., selective fishing) and to prevent new introductions (e.g., education, inspection), mainly because impoundments are vulnerable to new introductions (e.g., Espinola et al., 2010;Franco et al., 2018).Non-native species impact native biodiversity in multiple ways, causing population declines, extirpation or homogenization, as recorded by studies conducted in the Paraná river basin (e.g., Pelicice, Agostinho, 2009;Vitule et al., 2012;Orsi, Britton, 2014;Daga et al., 2015) and elsewhere (Cucherousset, Olden, 2011;Azevedo-Santos et al., 2017).Equally important is the implementation of long-term monitoring in the cascade system, to perform robust studies about impacts, invasion dynamics (e.g., establishment, invasion) and broad surveys.The difference in total richness between our study and Garcia et al. (2018), for example, suggests that the real number of non-native species is uncertain.Updated surveys are, therefore, crucial to understand biological invasions and minimize negative impacts.
In general, our results confirm trends reported in other regulated rivers (Petrere Jr. 1996;Barrella, Petrere Jr. 2003;Agostinho et al., 2007b;Daga, Gubiani, 2012;Petesse, Petrere Jr., 2012;Agostinho et al., 2016), where the diversity of migratory fishes is low and non-native species are common.The present study cannot infer about demographic trends and long-term dynamics, since our data come from multiple studies that applied different sampling methodologies.However, fish assemblage structure is probably stable in the studied reservoirs, especially because most are old, constructed decades ago.Specific environmental filters created by the impoundment constrain fish assemblages to persist in another state (Agostinho et al., 2016;Santos et al., 2017).In this sense, the current flow regime in the Paranapanema river, ruled by the cascade of dams, is the main force behind ecosystem state and dynamics, determining patterns of diversity and productivity.Conservation plans must consider this scenario to set appropriate management actions, especially because a number of studies have questioned the efficacy of current practices, i.e., fish passages and stocking (Pelicice, Agostinho, 2008;Agostinho et al., 2010;Pelicice et al., 2015a;Lira et al., 2017).Fish passages in Canoas dams, in particular, caused significant impacts e170150 [15] on fish populations (Pelicice, Agostinho, 2008;Casimiro et al., 2017), declining the conservation status of migratory fishes.In addition, stocking practices, conducted for decades (Agostinho et al., 2010), contributed to introduce several non-native species (Britton, Orsi, 2012;Ortega et al., 2015;Garcia et al., 2018), but was unable to restore populations of migratory fishes in impoundments.This situation should concern authorities about the conservation of Neotropical fishes in light of severe human disturbances (Pelicice et al., 2017), because rivers are extensively regulated, while the watershed is subjected to extensive agriculture and urban development, with activities that promote land cover changes and use non-native species (e.g., aquaculture).Conservation will depend on measures that preserve fluvial remnants (e.g., tributaries, such as the Tibagi and Cinza rivers), regulate human activities in surrounding areas, curb the introduction of non-native species, and control/ eradicate populations already introduced (see Bennemann et al., 2000;Britton, Orsi, 2012;Orsi et al., 2016;Pelicice et al., 2017).We emphasize that native fishes in these impoundments must receive conservation efforts, since they represent a relevant fraction of regional diversity, in addition to sustaining ecosystem services that are valued by artisanal and recreational fisheries.

Fig. 4 .
Fig. 4. Non-Metric Multidimensional Scaling (NMDS) showing variation in assemblage composition in the cascade of reservoirs along the Paranapanema River, considering native nonmigratory species (a) and native long-distance migratory species (b).The line connecting reservoirs indicates the sequence in the cascade system.Reservoir names (codes) are shown in Fig. 3.

Fig. 5 .
Fig. 5. Relative abundance (%) of long-distance migratory fishes (a) and non-native species (b) in the cascade of reservoirs along the Paranapanema River.Each dot is a sample (n = 32).Reservoir names (codes) are shown in Fig. 3.

F
. M.Pelicice, V. M. Azevedo-Santos, A. L. H. Esguícero, A. A. Agostinho & M.  Distribution of long-distance migratory species in the cascade of reservoirs along the Paranapanema River (gray = presence), based on all studies (47 samples).Numbers are the relative abundance range (percentage, min-max), based on studies that provided abundance data (32 samples); the most abundant species in the cascade are highlighted in bold.Reservoir names (codes) are shown in Fig.Distributionof non-native species in the cascade of reservoirs along the Paranapanema River (gray = presence), based on all studies (47 samples).Numbers are the relative abundance range (percentage, min-max), based on studies that provided abundance data (32 samples); the most abundant species in the cascade are highlighted in bold.Reservoir names (codes) are shown in Fig.