Ecological Impacts of the African Catfish Clarias Gariepinus at Environmental Protection Area in Southeastern Brazil


 The African catfish (Clarias gariepinus) is considered one of the most important species of catfish for aquaculture. It has a great capacity to withstand several stress factors, such as harsh abiotic conditions, in addition to wide feeding flexibility. However, the species was detected in the Guapimirim Environmental Protection Area in southeastern Brazil, threatening native fish diversity and ecosystem functioning of this ecosystem. In 2018, during the dry and wet seasons, samples of the fish community were collected at thirty-two sites of the Guapi-Macacu River, in addition to abiotic variables (salinity, pH, temperature, turbidity, dissolved oxygen, and transparency) to diagnose which factors influence the distribution of the alien species along the river. Multivariate analyses indicated that African catfish dominate the region in the buffer zone to the Environmental protection area, benefiting from higher levels of dissolved oxygen and temperature. However, C. gariepinus does not dominate yet the most protected area of Guapimirim, where the highest percentage of native fish species inhabit. Climate change associated with changes in abiotic factors might significantly contribute to the dominance of the invasive alien species in this protected area, which might colonize the entire river.


Introduction
Invasive alien species (IAS) are considered one of the main threats to biodiversity and an important component of global environmental change (Clavero & García-Berthou, 2005). Once present and acclimated, IAS becomes di cult to control and result in marked changes to the native fauna, thus compromising the balance and stability of the ecosystem through reductions in their stocks, extinctions, competition, spreads of pathogens, and hybridization (Agostinho & Julio, 1996).
The African cat sh, Clarias gariepinus (Burchell, 1822), is native to much of the African continent and parts of southwestern Asia such as Israel, Syria, and southern Turkey (Graaf & Janssen, 1996). It is considered one of the most important species of cat sh for aquaculture (Osibona et al., 2006) and the third most important commercial species in countries such as Egypt (Khallaf & Gaber, 1991).
Furthermore, its ability to move to the terrestrial environment and breathe atmospheric air through pseudo lungs confers exceptional capacity to tolerate adverse conditions of extreme temperatures and low oxygen concentrations (Graaf & Janssen, 1996).
Clarias gariepinus was introduced to Brazil in the 1980s for aquaculture and pay shing (Agostinho et al., 2007). Currently, this species is produced commercially, mainly in Santa Catarina state (58.9% of the total sh production in 2007), followed by Espirito Santo (39.7%) and Rio de Janeiro states (1.3%) (IBAMA, 2009). However, because C. gariepinus is considered a carnivorous, vigorous, and highly resistant species, its escape from production systems into the natural environment has threatened several native sh species (Umbria, 2008), becoming a potent IAS for that ecosystems.
In Brazil, there are still few studies about invasive populations of C. gariepinus (Weyl et al., 2016). The riverside population knows the presence of African cat sh of the Guapi-Macacu River and in the We collected The sh in February (wet season) and August 2018 (dry season) in the Guapi-Macacu River, using nets with different meshes, traps, holes, hooks, and lines. Fishing gear was installed and used at night because the species under study has a nocturnal habit. At the same time, using a multiparameter probe Hanna model HI 9828, we measured, in situ, the physicochemical variables of water: temperature (ºC); pH; dissolved oxygen (mg/L); transparency (cm); turbidity (FNU); and salinity (PSU). The sampling did distribute in thirty-two points: ten points between the mouth of the river and the inner limit of Guanabara ESEC (downstream area); 12 points between the outer boundary of Guanabara ESEC and the inner perimeter of the Guapimirim EPA (intermediate zone); and ten points distributed between the outer limit of the Guapirim EPA to the dam (upstream area), within the Buffer Zone of the Guanabara Ecological Station (Guanabara ZA), totaling 64 samples points in both seasons ( Figure 1). We removed the waiting nets 24 hours after being installed. The sh were grouped in CPUE (Capture per Unit of Effort) at each sampling point and placed in plastic bags, labeled and refrigerated on ice, and then transferred to the Laboratory of Applied Ecology, UFF.
Laboratory activities, data processing, and statistical analysis The sh collected were identi ed according to Reis et al. (2003); Figueiredo and Menezes (1978; and Figueiredo (1980, 1985), weighed and measured. To characterize the sh communities were analyzed: richness, abundance, biomass, diversity, dominance, and evenness indices. We calculated the richness through the total number of species collected at each sampled point and the abundance from the total number of individuals collected. For diversity analysis, the Shannon-Wiener index, the Simpson dominance index, and Pielou's evenness were used (Magurran, 1988).
We the Multivariate Permutation Analyses of Variance (PERMANOVA) in Program R (R Core Team, 2020), available in the VEGAN package (Oksanen et al., 2020), to test whether spatial and seasonal differences in physicochemical variables of water and ecological descriptors of ichthyofauna were signi cant (p<0.01). PERMANOVA was also applied to test the spatial and temporal variation of species richness, abundance, and biomass of sh and the diversity, equitability, and dominance indices. PERMANOVA is similar in strength to traditional ANOVA and does not require the assumptions of normality and homoscedasticity (Anderson, 2001;McArdle & Anderson, 2001). Therefore, we use the Bray-Curtis distance in all PERMANOVA tests, with data exchanged 4999 times per analysis, as Manly (2018) recommended for tests with p<0.01.
The Multinomial Species Classi cation Method (CLAM) was used, through the "CLAMTEST" function available in the VEGAN package (Oksanen et al., 2020), to classify sh species into generalists or specialists in the rainy and dry seasons without excluding rare species (Chazdon et al., 2011). This method uses a multinomial model to estimate the relative abundance of species in two groups (A, B), minimizing adverse effects due to sampling differences or insu ciency within each habitat (Solymos, 2020). We stipulate a limit of 50% of specialization in each season, with a signi cance level of 95% for individual tests.
To analyze the spatial proximity and formation of groups, according to the abundance found, taking into account the river as a whole, Cluster Analysis was used, through the "HCLUST" function of the DENDEXTEND package, using the "ward. D2" (Galili, 2015), and the Bray-Curtis distance in the dissimilarity matrix of the VEGAN package (Oksanen et al., 2020). The sum of squares criterion was utilized for this agglomerative method, producing groups that minimize the dispersion within the group in each binary fusion (Murtagh & Legendre, 2014). In addition, the speci city and delity (IndVal) of each species we calculated using a permutation test (with a signi cance of 0.05), using the "MULTIPATT" function, available in the INDICSPECIES package (De Cáceres et al., 2020). Finally, we used the delity and speci city test to determine which indicator species are simultaneously between seasons (wet and dry) and river areas and only between river segments, regardless of the season.
Canonical Correspondence Analysis (CCA), considered one of the best methods for direct gradient analysis in community ecology (Ter Braak 1986, Rodriguez & Lewis, 1997), was used in the R program version 4.0.2 (R Core Team, 2020), available in the VEGAN package (Oksanen et al., 2020), and applied to the matrix of environmental and biological data (excluding the rare species previously selected in the CLAM model), to identify which physical-chemical variables of water contributed most to the characterization of the points of the river in the wet and dry seasons, as well as to verify the distribution of the abundances of the sh species found, correlating them with the environmental variables (complete model). The multicollinearity of physicochemical variables was diagnosed to select redundant variables and obtain an adequate model. For this, the "ORDISTEP" function was used, which performs an automatic selection, which compares the model in which no environmental variable explains the variation in species composition and abundance (null model) with the complete model (where we used all environment variables), based on permutation test using the P-value. The variable selection procedure based on Pvalue seeks to nd the ideal model, in which only the most signi cant environmental variables explain the model (Blanchet et al., 2008). Subsequently, the signi cance of both models (complete and reduced model) did test, as well as which axes and terms were signi cant, by analysis of variance (ANOVA), using the "ANOVA.CCA" function. This function allows determining whether the environmental variables have an essential relationship and differences in the composition and abundance of species and how many axes are needed to present this relationship. Thus, an ideal and reduced model for the Guapi-Macacu River was obtained, which exposes which environmental variables can predict changes in the composition and abundance of the ichthyofauna and show which species will be affected in this relationship.

Results
PERMANOVA identi ed signi cant differences for all environmental variables among areas and seasons studied, except for temperature that did not show clear differences between the river segments (Table 1).
Water temperature was higher in the rainy season. Salinity was much higher downstream of the river in the dry season but much lower in the rest of the groups (site and season combinations). Transparency was higher in the dry season and in the upstream sites. Dissolved oxygen (DO) was higher in the dry season than in the rainy season. Higher DO values occurred in the upstream segment during the dry season, followed by the intermediate and downstream areas. In the rainy season, DO was higher in the upstream region, followed by the middle area and downstream. The pH was also higher in the dry season than in the rainy season. Higher values were recorded in the upstream segment during the dry season, followed by the downstream and intermediate zones. During the wet season, the pH remained the same for the three parts of the river. The highest turbidity values occurred in the rainy season downstream, followed by the intermediate and low segments. During the wet season, turbidity did not vary signi cantly between the three areas of the river ( Figure 2). Thirty-one species were collected regarding the ichthyofauna, referring to 427 specimens, distributed in nine orders and 21 families ( Table 2). The distribution of the relative abundance of the sh community along the Guapi-Macacu River between the areas and sampled seasons is exhibited in Figure 3.
PERMANOVA detected signi cant differences for sh abundance in relation to the area (F= 2.6214; p = 0.0454) and the season (F= 6.9902; p = 0.0020). The highest abundance of medians occurred in the dry season downstream. PERMANOVA did not see signi cant differences between the other zones and seasons, as well as for diversity and evenness (p>0.05, for all) ( Figure 4).
The CLAM classi cation portrays that 22.6% of the species collected were classi ed as generalists in the dry and wet seasons, frequently occurring in the two seasons studied, while 9.7% were classi ed as specialists in the rainy season and 9.7% in the dry season. Rare species in the samples corresponded to 58.1% ( Figure 5).
As for the cluster analysis results, this coincided with the previous demarcations shown in Figure 1, portraying the physiography of the Guapimirim EPA in the spatial separation of the ichthyofauna ( Figure  6). The main species that make up the Guapimirim EPA were more on the left of the dendrogram. The species in the right part of the dendrogram are distributed in the upstream area and cover the entire river, and the rare species that make up the last two groups formed. In this classi cation, the target species C. gariepinus, appears close to groups with wide distribution in the river, such as L. castaneus, T. striatulus, and H. auroguttatus. Furthermore, C. gariepinus occurs similarly to R. quelen and T. striatulus, sharing the same habitats.
The speci city and delity analysis (IndVal) showed as indicators the most important species that make up the two main groups formed in the cluster analysis (continuous line and dotted line the river segments and seasons rainy and dry ( Figure 7). The complete model ( Figure 7A)  p=0.017), being related to the distribution and abundance of sh species, as well as the river segments. ANOVA did not show signi cant differences in CCA for turbidity (F = 1.9987 and p = 0.061), and salinity (F = 1.0788; p = 0.372).
On the other hand, the signi cance of the axes is changed in the reduced model, obtained through the "ORDSTEP" function ( Figure 7B)

Abiotic variables in the Guapi-Macacu River
In the Guapi-Macacu River, salinity was higher downstream in both seasons, while the values of DO and pH were higher in the dry season upstream. Partially similar results were reported by Macêdo et al. (2000) in the Rio Formoso estuary (Pernambuco, Brazil), which detected the highest pH values in the lower estuarine area, beyond has higher DO and salinity. According to Macêdo et al. (2000), salinity and oxygenation levels are in uenced by tidal cycles and photosynthesis and respiration rates. The existing neutralization capacity in the aquatic ecosystem due to the buffer effect prevents wide pH variations. Therefore, maximum values did obtain in areas with more signi cant saline in uence. In the Guapi-Macacu River, the highest salinity at the mouth of the Guanabara Bay is expected because it is a coastal segment, and its intensity may vary with the rainy season, which favors its dilution. The higher DO and pH rates re ect a season without rain, with greater water transparency, favoring photosynthesis that removes CO 2 raising the pH of the water due to the consumption of H + ions.
When analyzed individually, it observed that temperature and pH were the environmental attributes that were not related to the different segments of the river but seasons. The highest temperature in the rainy season corresponds to the hottest season of the year in South America. However, there is collinearity of environmental attributes. When they are presented in the reduced model and correlated with species abundances, pH reveals as a variable with high signi cance for the downstream area, probably due to its higher value in the dry season. The association of the other variables analyzed, such as the predominance of increased transparency in the intermediate region of the river, contributed to a better distinction between the river segments in the dry and rainy seasons, corroborating the environmental characteristics of the Guapi-Macacu River in the Guapimirim EPA, and serving as predictors for related sh species in this habitat. According to Blaber (2000), sh from tropical estuaries are subject to a range of interactions of physicochemical and biological factors that determine their patterns of occurrence, distribution, and movement. According to this author, in the Rio Formoso estuary (Pernambuco, Brazil), the temperature, salinity, pH, and dissolved oxygen were higher in the lower estuarine zone and dry season. In estuarine regions of Pernambuco, seasonal variations in water temperature and salinity are well evidenced, with the highest values being recorded in the dry season (Macêdo et al., 2004).
However, de cient levels of DO can be seen in the area downstream of the river, especially in the rainy season, coinciding with the lowest richness and abundance recorded. For Edokpayi et al. (2017), levels of DO concentrations below 5.0 mg/L adversely affect aquatic life. Furthermore, during the rainy season, Guanabara Bay drags organic materials and concentrates contamination during the dry season to the area downstream of the Guapi-Macacu River. In addition to the resuspension of sediments, explain less oxygenation during the wet season in this area.

The sh assemblage of the Guapi-Macacu River
The sh assemblage of the Guapi-Macacu River, within the Guapimirim EPA and its Buffer Zone, has freshwater species with marine species, many of which are euryhaline. Marine species were concentrated in the lower part of the river (downstream), at the mouth of the Guanabara Bay, with some species migrating to inland areas, such as bass, mullet, carapebas, and croakers. This segment was constituted by resident species, marine and freshwater migrants, which use the estuaries as feeding areas, for rearing larvae and juveniles, or for reproduction (Blaber, 2000). These habitats favor the presence of various sh populations on their margins (Vidy, 2000), consisting mainly of juveniles of marine species (Rozas & Zimmerman, 2000). The greater abundance of sh downstream of the Guapi-Macacu River is probably due to the food availability from primary production, structural complexity of mangrove vegetation, which provides refuge, especially for young sh, and high water turbidity. According to Loebman & Vieira (2005), the structure of the tropical estuarine sh fauna varies with the type of estuary and with differences in the spatial and temporal pattern of the community.
The lower abundance of sh in the rainy season may be related to the season of greater water volume in the river, providing greater availability of more sheltered areas, facilitating dispersal, and hindering sh collection. In estuarine environments, mangroves provide a natural refuge for young individuals due to the protection provided by the root structure of their trees. Most of the sh caught in tropical coastal areas enjoy this protection during their young phase and at the time of laying and thus depend closely on the integrity of this ecosystem (Lacerda, 1984). In general, the ichthyofauna remains with the most abundant species still well preserved, especially in the area further downstream from the river. For Teixeira et al. (2005), the determination of biodiversity, especially of the sh community and its spatial and temporal variation patterns, is of great relevance to assessing environmental quality.
The Siluriformes was the order with the most incredible abundance of sh in the Rio Guapi-Macacu. The dominance of Siluriformes over the others constitutes a characteristic pattern of the eastern region of Brazil, being particularly accentuated in areas of high river courses, where the condition of high hydrodynamics favors the occupation by demersal species (Bizerril & Primo, 2001). The site with the most increased occurrence of native species is also located downstream of the river, within Guanabara ESEC, the area of most preservation within the Guapimirim EPA and with better ecological indices, such as abundance and richness of the species found, especially in the dry season. Genidens genidens, representative of the second most abundant family, Ariidae, occurs in coastal areas and is generally more signi cant in shallow coastal waters, on a muddy or sandy bottom (Araujo, 1988; Andreata et al., 1989).
The presence of G. genidens downstream of the river may be related to the spawning season. The species seek the mouth of the rivers performing oral incubation with males and rarely females, carrying eggs and initial forms of offspring until they complete embryonic development (Yanez-Arancibia & Sanchez-Gil 1988; Reis, 1986b), which explains the presence of specimens downstream in the Guapi-Macacu River. In addition, G. genidens had high levels of delity (59.09%) and speci city (100%), considered a perfect indicator species in the IndVal analysis. The IndVal method has some advantages compared to other bioindication methods, is calculated for each species independently, where the categorization of habitats occurs without restrictions and can be grouped subjectively or quantitatively (McGeoch & Chown, 1998).
Thus, the IndVal analysis establishes that for this speci c area, regardless of the season analyzed, three species were considered indicators of this habitat G. genidens, B. pectinata, and M. furnieri ( Figure 8A). In addition to these species, E. saurus was selected only for the dry season in the downstream area. All species chosen for the downstream area were considered asymmetric indicators, as they contribute more to habitat speci city than delity (Dufrêne & Legendre, 1997). This segment was also particularly evidenced in the cluster analysis, which showed that the main species selected in IndVal share this river segment.
The area within the Guapimirim EPA, disregarding the Guanabara ESEC (i.e., the intermediate segment of the river), revealed ecological indices similar to the other parts of the river. However, greater abundance and richness can be observed ( Figure 8A). Thus, the downstream and intermediate segments (i.e., comprising Guanabara ESEC and EPA de Guapimirim) presented only C. undecimalis indicative of the EPA regardless of the time of the year approached ( Figure 8A; 8B). Centropomus undecimalis belongs to the order Perciformes, the second most abundant in the Guapi-Macacu River. According to Peterson & Gilmore (1991), sea bass does not undergo large migratory cycles, being a relatively fast-growing sh that spawn a large number of eggs in brackish waters during late spring and early summer. Sea bass juveniles show a greater a nity for fresh water and survive in waters with lower oxygen levels than adults, being found upstream of rivers at all times of the year (Ager et al., 1976). Its primary or nursery habitat has been described as warm shallow streams or drainage channels, with low current and unvegetated bottoms or bordering the mangrove (Mcmichael et al., 1989). As they develop, they move from shallow water habitats to estuaries, mangroves, and deeper water (Tucker & Campbell, 1988).
According to the cluster analysis, the river segment shared with C. undecimalis and other species provides this species with a habitat with abundant food resources and protection for its development.
The river's intermediate region was characterized by a very winding zone with greater diversity in physiography, constantly ooded with deeper portions, and a salinity gradient that decreases from the river's mouth towards the interior. This segment presents vegetation composed of mangrove forests, riverside types. This area is directly related to larger transparency samples, evidencing E. brasilianus and O. niloticus with intermediate values of this attribute ( Figure 8B). Eugerres brasilianus is a species of marine origin that tolerates signi cant variations in salinity (Ramos et al., 2006), is anadromous, migrating from the sea to rivers, living in coastal waters of warm seas, penetrating coastal lagoons and estuaries to complete its life cycle (Yañez-Arancibia, 1986). In addition, E. brasilianus is a species with nocturnal habits, generalist and opportunistic; it is epibenthic and demersal. That is, it exhibits patterns strictly linked to the substrate, being did consider an excellent biological resource, mainly because it did regard as an abundant shery resource (Cyrus & Blaber, 1983; Tapia-García & Ayala Pérez, 1996; 1997; Araújo & Santos, 1999; Barletta & Blaber, 2007;Barletta & Costa, 2009).
On the other hand, the cluster analysis did not evidence the intermediate region with a speci c community for this segment of the river, showing itself as a transition area and occupied mainly by species that travel throughout the river, such as H. littorale, H. auroguttatus, and L. castaneus ( Figure 8C). Hypostomus auroguttatus and L. castaneus belong to the most abundant family, Loricaridae, common in areas with muddy river bottoms and may even occur in lentic environments. That two species recorded in the three segments of the Guapi-Macacu River during the dry and wet seasons, considered generalists in the present study.
The species A. lacustris ( Figure 5), considered rare in the analysis shown in CLAM TEST, was selected in IndVal for the upstream segment of the river, as well as in the rainy season, with high speci city and low delity, for Dufrêne & Legendre (1997), rare species may receive the same IndVal value as indicator species and are called asymmetric indicators. On the other hand, in this same river segment, without considering the attributes of the seasons (dry or wet), IndVal selected T. striatulus and C. gariepinus ( Figure 8D), both with high speci city. Meaning that these sh can also be considered indicators asymmetric, which contribute to the speci city of the habitat, but they do not serve to predict groups (Dufrêne & Legendre, 1997).
From the CCA, we can see that the segment upstream of the river presents higher values of DO and abundance of C. gariepinus. The upstream part is the river's headwaters, which have a humid tropical climate, high and variable slope, determining the dynamic character of the uvial system, with the presence of rapids, characteristic of mountain and plateau regions. Barella (2000) discusses the vital role of riparian forests in providing resources for feeding aquatic fauna and attracting dispersers, making the riparian environment a fundamental element in the sustainability of rivers and lakes and in the connection between the different systems that make up the rural landscape. The CCA also highlights the most protected area concerning the Guapimirim EPA, downstream of the river, with the most diversity and richness, exhibiting greater salinity, pH, and turbidity attributes. with the dominance of G. genidens, as it presents greater abundance in this group. The CCA also highlights the similar spatial distribution of T. striatulus with C. gariepinus (with more representation in abundance), corroborating the cluster analysis.

African cat sh in the Guapi-Macacu River
The African cat sh, C.gariepinus was the third most abundant in the river, however in the most preserved area of the Guapimirim EPA, it still does not show signi cant abundance. It also notices that the species that present ecological equivalence to the African cat sh, R. quelen, and T. striatulus, despite showing a similar distribution, give a lower abundance, which denotes the habitat overlap of the IAS over the native ones ( Figure 9A). In addition, the population of C. gariepinus exhibited greater abundance contributions in the dry season, especially upstream, compared to the other species, demonstrating a high potential to colonize the entire river ( Figure 9B).
Even though the CCA analysis does not show similar correlations of the abiotic factors R. quelen with Clarias gariepinus, the group exposed in the cluster analysis, it is clear that these species share the same segments of the river. Rhamdia quelen prefers lakes and river bottoms, preferring calmer water environments with a sand and mud bottom along the banks and vegetation (Gomes et al., 2000). They are omnivores with a clear preference for sh, crustaceans, insects, plant remains, and organic debris (Guedes, 1980;Meurer & Zaniboni Filho, 1997). Therefore, they are considered a generalist about food choice (Guedes, 1980). Among the many biological similarities that R. quelen has with C. gariepinus, highlighting the food preference and the use of habitat, however R. quelen has a disadvantage about its development, since females can reach up to 66.5 cm and males up to 52.0 cm (Gomes et al., 2000).
According to our analysis, dissolved oxygen was the attribute directly correlated with the distribution of C. gariepinus ( Figure 7B). Thus, dissolved oxygen (DO) plays a fundamental role in regulating the body's metabolic functions, including the aquaculture community, in addition to being an environmental indicator of water quality (Anyachor & Sikoki, in press). On the other hand, African cat sh can withstand low concentrations of dissolved oxygen (Adewolu et al., 2008) due to an air-breathing accessory organ, which can absorb oxygen from the atmospheric air (Moussa 1956), allowing sh to survive for many hours or weeks out of water or in muddy swamps (Idahor et al., 2014). Thus, the most signi cant contribution of DO may have acted as a facilitator for the development of the IAS, serving as an attraction for other species that could migrate to this segment of the river, in search of better environmental conditions, and probably due to the greater dominance of the IAS, have failed to establish themselves in this area.
Another signi cant environmental attribute in our analysis was temperature, and this seems to be an essential factor in the distribution of the species in the Rio Guapi-Macacu. The highest temperature values in the upstream area corroborate the results found in the CCA, where the greatest abundance of the species did correlate with this segment of the river. For Hecht (2013), African cat sh larvae exhibit an optimal development around 28ºC, whose value was recorded upstream in the rainy season. And this could be seen in the dry season, which showed a high increase in the IAS population, favored due to the temperature rise previously recorded. On the other hand, our studies indicate that, despite the average temperature values measured in the Guapi-Macacu River not being considered ideal for the development of cat sh, the species' adaptability has allowed it to develop well in this ecosystem and colonize other segments of the river. However, climate changes resulting from global warming can contribute to the gradual increase in temperature, signi cantly altering this scenario, which may favor the rapid development of this alien species and, possibly, the decline of native populations.
Salinity is also an attribute that limits the occurrence of alien species in the mouth of Guanabara Bay.
That is because African cat sh are stenohaline, with a limited capacity to withstand increased salinity in the environment, in more advanced stages (Hoogendoorn, 1981). Borode  Even so, we can observe that despite the signi cant abundance of this species, the analysis of speci city and delity does not list C. gariepinus as an indicator species, taking into account both studied seasons and collection areas. However, this pattern diverges when considering only the upstream zone, when the species speci city index rises to 81% and the delity to 36% for this area, demonstrating C. gariepinus with the probability of 53% being an asymmetric indicator species of this area. For Dufrêne & Legendre (1997), a species can be an asymmetric indicator without high delity.

Conclusion
The analyses of the biological attributes of the African cat sh populations, together with the environmental and biological descriptors of the segments of the Guapi-Macacu River, reveal that this species does not colonize Guanabara Bay. That is can also be explained by potential predators and competitors that occupy the downstream part of the river. The shallow areas downstream of the Guapi-Macacu River also harbor a larger contingent of typically estuarine juvenile sh, portraying this area as natural breeding and shelter for several sh species of ecological and economic importance. Imminent climate change factors can irreversibly alter this scenario, with more prolonged droughts in the hottest season and, consequently, increasing the temperature in other river segments to favor the habitat and, hence, the successful development of African cat sh. As it is a generalist species found in the intermediate and upstream zones of the Guapi-Macacu River, has food versatility and robustness, it has probably been competing for food and habitats with native species, which have ecological equivalence, such as R. quelen, and T. striatulus and thus, impacting the local ichthyofauna, with the visible decline in the abundance of native species where African cat sh establish themselves. Global climate change continues to impact sh habitat quality and biodiversity, especially about the dynamics of IAS. This study suggests that African cat sh can bene t from the high temperatures predicted by climate change, facilitating their further dispersal and subsequent establishment in these environments.

DATA AVAILABILITY
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.     Table 2.

Figure 4
Boxplots with the distribution of abundance, richness, Shannon-Wiener diversity, Pielou Equitativity, Dominance, and total Biomass (g) between segments and seasons in the Guapi-Macacu River. DD:      Table 2.