Effect of local and regional factors on the structure of the fish communities associated with aquatic macrophyte stands in oxbow lakes on the Amazon floodplain

Abstract

The Amazon floodplain is home to an extremely high diversity of fish, with lakes playing an important role in the establishment of this biological richness. These lacustrine environments are subject to constant fluctuations caused by the annual flood pulse, with local factors and other regional patterns also contributing to the variation in fish community structure. The present study verified how local (depth and transparency of the water, the size and species composition of the macrophyte stands) and regional factors (spatial distribution of the stands and the hydrological phase) influence the structure of the fish community of the floodplain lakes of the Môa River, in northern Brazil. Fish species richness was influenced by the depth of the water and the spatial distribution of the macrophyte stands. Fish species composition was influenced by local environmental variables, spatial structure, and the hydrological phase. However, variation partitioning indicated that only the hydrological phase explained the variation in fish composition. These findings indicate that the local environment, the spatial structure, and the hydrological phase drive changes in the structure of the fish communities associated with aquatic macrophytes in the floodplain lakes of the Amazon basin.

Key words
Upper Juruá River; Môa River; Acre; hydrological periods; environmental factors; metacommunity

INTRODUCTION

Floodplains support a high diversity of both plants and animals, which is underpinned by the spatiotemporal heterogeneity of these environments, and is controlled, in turn, by the annual fluctuations in the hydrological cycle (Junk et al. 1989, 2014, Súarez et al. 2001). The alternation between low- and high-water phases is a regional-level factor in floodplain ecology, resulting in the connection of isolated lakes during the high-water phase, which supports the dispersal of aquatic organisms, increases the quantity and diversity of resources, and alters the physicochemical conditions and complexity of the environment (Fernandes et al. 2009, Junk 1989, 2014, Thomaz et al. 2007, van der Sleen & Rams 2023).

Lakes are common features of Neotropical floodplains due to the considerable number of meandering rivers found in this region, and their typical geological formations (Goulding 1980). These environments contain a large part of the biodiversity of the floodplain, by sustaining aquatic habitats during the low-water phase. During this phase, the floodplain lakes become favored sites for the development of many different aquatic taxa due to their relative abundance of feeding resources and refuges from predators, within their complex habitat structure, in comparison with the principal channel of the river (Junk et al. 1989, Fernandes et al. 2009). The spatial arrangement of the lakes on the floodplain also influences the capacity of different fish species to colonize these environments, because lakes further away from the river channel tend to be isolated for longer periods during the low-water phase. These lakes can only be reached by species with an enhanced dispersal capacity, and only when these lakes are connected to the river channel (Penha et al. 2017, Virgilio et al. 2022). Lakes closer to each other may also have a more similar species composition, due to either the spatial similarities of the local environment or the presence of fewer geographic barriers to dispersal (Nekola & White 1999). Given these processes, the spatial structure of these environments may often be an important regional ecological factor determining the fish species composition of floodplain lakes (eg, Virgilio et al. 2022).

Aquatic macrophytes contribute to the complexity of the habitat structure of oxbow lakes (Piedade & Junk 2000, Maltchik et al. 2007, Thomaz et al. 2008, Junk et al. 2012). The dynamics of the colonization of oxbow lakes by aquatic macrophytes is also dependent on the hydrological cycle (Bonetto 1975). These plants are the foundation of food chains (Pott & Pott 2000, Meerhoff et al. 2007), and they also contribute to the formation of commensal interactions among different groups of aquatic organisms, such as the zooplankton (Gazulha et al. 2011, Cabral et al. 2021), macroinvertebrates (Takeda et al. 2003), and fish (Lopes et al. 2015, Freitas et al. 2022, Virgilio et al. 2021). Macrophyte stands provide foraging sites for many fish species, as well as protection from potential predators (Heino 2000, Pelicice et al. 2005, Cunha et al. 2011). The submerged parts of these plants form complex habitats, including stems, roots, and leaves, which facilitate the colonization of algae and associated invertebrates (Junk 1973, Cunha et al. 2011). Aquatic macrophytes also make a significant contribute to the quality of the water by, for example, increasing the amount of oxygen (O₂) dissolved in the water. This facilitates the colonization of these environments by fish species that are sensitive to fluctuations in the availability of O₂ or other ecological parameters (Sánchez-Botero et al. 2001, Soares et al. 2006). Studies of a number of different floodplains (eg, Lopes et al. 2015, Nonato et al. 2021, Virgilio et al. 2021, 2022) have shown that the structural complexity of macrophyte stands and their contribution to the improvement of conditions in the aquatic environment can drive an increase in fish species richness. The macrophyte-fish relationship may be so intimate that the species composition of the macrophyte stands may even correlate with that of the fish species that occupy this vegetation (eg, Suçuarana et al. 2016). Given this, structural attributes of macrophyte stands, such as their biomass, species richness and composition, may be important local ecological factors determining the variation in the composition of the associated fish fauna (Lopes et al. 2015, Suçuarana et al. 2016, Nonato et al. 2021, Virgilio et al. 2021, 2022).

Other local environmental variables, such as the transparency and depth of the water, are also important determinants of the colonization of macrophyte stands by fish, contributing to niche-based species filtering (Miranda & Lucas 2004, Tonn et al. 1990). Predation rates tend to be linked to the transparency of the water (Ortega et al. 2020), given that visually-oriented piscivorous fish species tend to occur in clear waters, whereas small fish may use the cover provided by more turbid water to hide from predators (Tejerina-Garro et al. 1998, Figueiredo et al. 2015, Santos et al. 2017). Studies of floodplain environments have also shown that depth influences both the species richness (Fernandes et al. 2010, Lopes et al. 2015) and composition of fish communities (Virgilio et al. 2022). Depth adds to the space available within a lake, with different depths having varying effects on species composition (Woolnough et al. 2009, Fernandes et al. 2010).

Given these considerations, and the importance of the relationship between fish and aquatic macrophytes, the present study verified how local (depth and transparency of the water, and the characteristics of the macrophytes – species richness and composition, biomass, and stand size) and regional factors (the hydrological cycle and spatial distribution of the stands) influence the structure of the fish community associated with the macrophyte stands found in the lakes of the Môa River floodplain. We evaluated the hypothesis that these local and regional factors have differential effects on the structure of the fish community associated with the macrophyte stands. We expected local environmental variables and the hydrological phase to have a greater influence on the fish composition of the macrophyte stands than their spatial distribution. We expected this because of the relatively small spatial scale of the study, which minimizes the potential influence of geographic barriers on the different fish faunas, and because the flood pulse is the principal factor driving shifts in the composition of different taxonomic groups on the floodplain (Thomaz et al. 2007, Penha et al. 2017, Virgilio et al. 2022). Finally, we examined the relationship between fish species richness and the local and regional factors. We expected fish species richness to be influenced positively by the biomass and species richness of the macrophytes (Lopes et al. 2015, Nonato et al. 2021, Virgilio et al. 2021, 2022). We also expected fish species richness to be correlated positively with the depth of the water, following a species-area relationship (Fernandes et al. 2010), and negatively with its transparency, given the prevalence of predators (Tejerina-Garro et al. 1998, Figueiredo et al. 2015).

MATERIALS AND METHODS

Study area

The Môa River is one of the principal tributaries of the Juruá River in the southwestern Amazon basin of northern Brazil. The floodplain of the Môa River present hundreds of oxbow lakes, which form lentic ecosystems that are permanently or temporarily connected to the channel of the river (Acre 2012; Figure 1). The Môa River has two well-defined hydrological phases: a period of low water from June to October, and a period of high water (the flood phase) from November to May, when the level of the river reaches over 5.5 m (Figure S1). In the present study, the low water phase was sampled in September and October 2015, and the flood phase was sampled between March and May 2016.

Figure 1
Location of the Vermelho, Antas, and Grande lakes on the floodplain of the Môa River in Acre state, northern Brazil.

Sampling

We sampled fish in three oxbow lakes located on the floodplain of the lower Môa River (Figure 1), in the municipality of Cruzeiro do Sul, Acre state, Brazil. In each lake, we sampled five different stands of macrophytes of distinct types (based on the predominance of one or a few plant species) during the day. We collected fish using a floating net measuring 1.5 m × 1.5 m, with a 5-mm mesh, which we deployed nine times during the day and nine times during the night at each stand. The fish captured in this net were anesthetized with Eugenol solution, fixed in 10% formaldehyde, and subsequently conserved in 70% alcohol.

The fish specimens were identified to the lowest possible taxonomic level using the taxonomic keys of Silvano et al. (2001) and Queiroz et al. (2013), as well as comparisons with specimens collected previously from the upper Juruá River and deposited in the collection of the Ichthyology Nucleus of the Upper Juruá Valley (Núcleo de Ictiologia do Vale do Alto Juruá – NIVAJ), in Cruzeiro do Sul, Acre (Brazil). The species list followed the classification of Fricke et al. (2022). The abundance of each fish species was summed across the nine floating net passes for a given macrophyte stand in a given lake to form a sample.

We estimated visually the relative abundance of each macrophyte species using a 0.5 m × 0.5 m square, dropping it haphazardly six times on each macrophyte stand (the abundance data were summed for each site) after the collection of the fish specimens. The macrophytes found within the square were also collected and weighed (g; wet weight of the roots (submerged portion), stems, and leaves (emerged portion)). The species were identified using the taxonomic keys of Pott & Pott (2000) and Guterres et al. (2008). The following environmental variables were measured at each macrophyte stand: the depth (m; using a graduated ruler), the area of the stand (m²; using a surveyor’s tape), and the transparency of the water (m; using a Secchi disk).

Ethics committee

The collection of the specimens was authorized legally by the permit issued by the federal Instituto Chico Mendes de Conservação da Biodiversidade (license number 55808-2, emitted to TRFJ).

Data analysis

We square-root transformed the macrophyte species composition matrix, and then applied a Principal Coordinates Analysis (PCoA; Legendre & Legendre 2012). We also applied the Lingoes correction to minimize the influence of negative eigenvectors on the PCoA (Legendre & Legendre 2012). We then used the envfit routine (Oksanen et al. 2018) to evaluate the contribution of each macrophyte species to the PCoA. We represented the dissimilarities in the macrophyte species composition by using a percentage difference index, and retained the first two PCoA axes for interpretation.

We used a Principal Coordinates of Neighboring Matrices analysis (PCNM; Borcard & Legendre 2002, Borcard et al. 2004) to assess the influence of the spatial distribution of the macrophyte stands on fish community structure. For this, we compiled Euclidean distance matrices with the geographic coordinates of each macrophyte stand. We then computed a PCoA for a truncated distance matrix that connected all the macrophyte stands, that is, the greatest distance in a minimum-spanning tree (Borcard & Legendre 2002). Finally, we selected the PCNM axes for all the analyses by forward variable selection. Only PCNM axes 1 and 3 were selected for analysis.

We developed a Generalized Linear Model (GLM) of fish species richness with a Poisson distribution by testing all the potential combinations of explanatory variables in the glmulti package (Calcagno & de Mazancourt 2010). This model selection tool automatically generates all the possible models (that are within the user-defined constraints) and identifies the best models based on the Akaike Information Criterion (AIC). The best model identified here was the one with the following explanatory variables: (i) depth of the water, (ii) size of the macrophyte stand, and (iii) PCNM axis 3. We checked the residuals of the model visually for the assumptions of normality and homoscedasticity, and used the Variance Inflation Factors (VIFs) between the explanatory variables to assess for potential multicollinearity (Zuur et al. 2010). We considered variables with a VIF of over 3 as collinear (Zuur et al. 2010), although none of the explanatory variables had a VIF value higher than 3. We assessed the potential spatial autocorrelation in the GLM residuals using a Mantel correlogram, although no spatial autocorrelation was detected in this analysis. We used Nagelkerke’s pseudo-R² as a measure of fit.

We employed a Distance-Based Redundancy Analysis (db-RDA; Legendre & Anderson 1999) to assess the effects of the local and regional environmental variables on fish species composition. These local variables were (i) the depth and transparency of the water, (ii) the species richness and composition, and weight of the macrophyte stands, and (iii) the scores of the PCoA 1 and PCoA 2 axes of the macrophyte species composition, while the regional variables were (i) the hydrological phase (a categorical variable with two levels: low water and flood), and (ii) the PCNM axes. We standardized the fish composition (response) matrix by the Hellinger distance to reduce the influence of abundant species, and then applied the Euclidean distances to represent the dissimilarities among the sampling sites (Peres-Neto & Legendre 2010, Legendre & De Cáceres 2013). We used the Lingoes correction to eliminate negative eigenvectors from the db-RDA. We verified the multicollinearity in the set of local environmental variables through their VIF values (Zuur et al. 2010), and we removed the depth variable from the db-RDA because it had a VIF value of over 3. We evaluated the statistical significance of the global and db-RDA axes using permutation tests, with 9999 permutations. We evaluated the relative contributions of the local and regional variables to the db-RDA with the envfit procedure (Oksanen et al. 2018). The significance of the species associations with the environmental variables was determined by 9999 permutations.

Finally, we ran a Partial Redundancy Analysis (pRDA), based on a distance matrix, to assess the exclusive and shared contributions of the set of local and regional variables on the composition of the fish assemblage (Borcard & Legendre 2002). In this analysis, the regional variables were divided between spatial structure – PCNMs – and the hydrological phase. We used the Euclidean distances of the Hellinger-standardized abundance data (response matrix) for the pRDA, and included only the variables selected above to represent the set of local variables and the PCNM. All these analyses were implemented in the vegan package (Oksanen et al. 2018) of the R software (R Core Team 2022). We adopted a significance level of 5% for all the analyses.

RESULTS

Thirteen species of macrophyte were recorded in the present study, belonging to 11 families (Table SI). The species with the highest mean percentage coverage in both the low water and flood phases was Pistia stratiotes L., with 51.88% of the cover in the low water and 30.28% in the flood phase. Ludwigia helminthorrhiza (Mart.) H. Hara was the macrophyte with the second highest mean percentage cover in the low water phase (18.42%), whereas in the flood phase, Ludwigia sedoides (Humb. & Bonpl.) H. Hara had the second-highest percentage cover (20.33%). The negative PCoA 1 scores represent macrophyte stands with a greater relative cover of Paspalum repens P. J. Bergius (correlation calculated by the envfit routine (R_env = -0.997; P = 0.001), while the positive scores represent stands with a greater relative cover of Cyperus blepharoleptos Steud. (R_env = 0.992; P = 0.001) and Pistia stratiotes (R_env = 0.975; P = 0.001). The positive PCoA 2 scores represent macrophyte stands with a greater relative cover of Justicia sp. (R_env = 0.998; P = 0.034) and Ludwigia helminthorrhiza (R_env = 0.975; P = 0.001). Ludwigia sedoides was associated negatively with both PCoA 1 (R_env = -0.816; P = 0.001) and PCoA 2 (R_env = -0.578; P = 0.001), which indicates that macrophyte stands with negative scores on both these axes had a greater cover of this macrophyte species (Table I).

Table I
Macrophytes species average percentage association with Principal Coordinate Analysis axes quantified with ‘envfit’ routine. Significance was assessed with 9,999 permutations. Statistically significant associations (R² with P ≤ 0.05) are highlighted in bold.

Species	PCoA 1	PCoA 2	R2	P
Ludwigia sedoides	-0.81590	-0.57819	0.8719	0.001
Ludwigia helminthorrhiza	-0.22228	0.97498	0.8607	0.001
Pistia stratiotes	0.97513	-0.22161	0.7423	0.001
Cyperus blepharoleptus	0.99167	-0.12877	0.5436	0.001
Paspalum repens	-0.99753	0.07031	0.5213	0.002
Justicia sp.	-0.05463	0.99851	0.2366	0.034
Hydrocotyl ranunculoides	0.99593	-0.09012	0.0774	0.365
Ricciocarpos natans	0.55097	-0.83453	0.0530	0.519
Ludwigia sp.	-0.50639	0.86231	0.0504	0.550
Azolla filiculoides	0.58334	-0.81223	0.0424	0.593
Salvinia minimala	0.62018	-0.78446	0.0392	0.593
Polygonum spectabile	-0.17221	-0.98506	0.0151	0.839

The mean depth of the water was 1.31 m (SD ± 0.74) during the low water phase and 3.20 m (±1.18) during the flood phase. The transparency of the water varied minimally between these periods (low water: 0.88 m ± 0.17 m; flood: 0.91 m ± 0.08 m).

A total of 2,767 fish specimens were collected from the three study lakes, representing four orders (Characiformes, Cichliformes, Gymnotiformes, and Siluriformes), 18 families, and 42 species. The characiforms predominated, with a total of 2,729 individuals, followed by the cichliforms, with 14 individuals (Table SII). Fish were considerably more abundant during the low water phase, with a total of 2,102 individuals, while only 665 individuals were collected during the flood phase. The most abundant species during the low water phase was Bryconops cf. melanurus, with 959 individuals, while Aphyocharax sp. (263 individuals) was the most abundant species in the flood phase.

The multiple regression had a high level of predictive power, explaining 92.2% of the observed variation in fish species richness (Nagelkerke pseudo-R² = 0.922, P < 0.001). The macrophyte stands sampled in deeper water presented lower fish species richness (Figure 2a), while larger stands and those with more positive PCNM 3 scores presented higher fish species richness (Figure 2b, c).

Figure 2
Relationship between fish species richness and the depth of the water (a), the size of the macrophyte stand (b), and PCNM3 (c) in the oxbow lakes surveyed on the Môa River in Acre, Brazil. The thick lines show the curves of the fitted values, while the gray areas indicate the 95% confidence interval.

The distance-based Redundancy Analysis (db-RDA) indicated that the composition of the fish community was influenced by both local and regional variables (F₉, ₁₈ = 1.73; P < 0.001). The first and second db-RDA axes accounted for approximately 25.04% of the total variation in the composition of fish species and were both significant. The db-RDA 1 axis explained 16.49% of the variation in fish species composition (F₁, ₁₈ = 5.54; P < 0.001), and was influenced by the hydrological phase, the size of the macrophyte stands, and their spatial distribution (PCNM 3). The db-RDA 2 axis explained 8.56% of the variation (F₁, ₁₈ = 2.88; P = 0.039), and was influenced by the spatial distribution of the macrophyte stands (PCNM 1) and the stands with a greater relative cover of Justicia sp. and L. helminthorrhiza (positive PCoA 2 scores for macrophyte composition). Hemigrammus hyanuary Durbin 1918, Elachocharax cf. pulcher, and Acestrorhynchus microlepis (Jardine 1841) tended to be more abundant during the low water phase, and in the larger and heavier macrophyte stands. By contrast, Steatogenys elegans (Steindachner 1880), Poptella sp., Aphyocharax sp., Leporinus sp., Moenkhausia melogramma Eigenmann 1908, and Monocirrhus polyacanthus Heckel, 1840 tended to be more abundant in the flood phase, and in stands with higher PCNM 3 values (Figure 3).

Figure 3
Variation in the composition of the fish community in relation to the environmental variables – transparency of the water, macrophyte stand size, wet weight, species richness and composition (PCoA), spatial variables (PCNM), and the hydrological phase – ordered by the distance-based Redundancy Analysis (db-RDA).

The Partial Redundancy Analysis (pRDA) indicated that the local environmental variables (adjusted R² = 0.00; P = 0.599) did not influence the variation in the composition of the fish community among the macrophyte stands. However, the hydrological phase (adjusted R² = 0.13; P = 0.001) and spatial distribution of the macrophyte stands (adjusted R² = 0.04; P = 0.033) did both contribute significantly to the variation in the composition of the fish species in the lakes. Nevertheless, the shared fractions were relatively low, with only 3% and 2% of the variation in fish composition being explained by both spatially-structured local environmental variables, and by the shared fraction between the phases and the local environmental variables. This means that approximately 79% of the variance in the fish species composition of the macrophyte stands remained unexplained.

DISCUSSION

The results of the present study showed that the hydrological phase had a greater effect on the characteristics of the fish community associated with aquatic macrophytes than either the local environmental variables or the spatial distribution of the macrophyte stands. Only the depth of the water, and the size and spatial distribution of the macrophyte stands exerted an influence on fish species richness. By contrast, local variables, such as the species composition and size of the macrophyte stands, and their wet biomass, most influenced the variation in the species composition of the fish communities. Overall, then, the characteristics of the fish communities were influenced primarily by the hydrological phase, while the spatial distribution of the macrophyte stands was only important for the fish species composition in the case of a few of the stands, during the flood phase.

The macrophyte stands surveyed during the present study were relatively homogeneous, with a predominance of only one or a few species, and Pistia stratiotes having the greatest mean cover in both hydrological phases. The predominance of this species is common in the oxbow lakes of the study region, i.e., throughout the state of Acre (Cabral et al. 2021). Secondarily, two Ludwigia species also provided relatively abundant cover, with L. helminthorrhiza prevailing during the low water phase, and L. sedoides during the flood phase. This alternation is likely related to the drought survival strategy of L. helminthorrhiza (Piedade et al. 2010, Bedoya & Madriñán 2015). The macrophytes recorded in this study varied considerably in their morphology, providing a rich diversity of microhabitats for the associated fish fauna (Dibble & Thomaz 2009, Lopes et al. 2015).

The high levels of fish diversity found in the oxbow lakes surveyed in the present study are typical of the pattern observed in most lakes on the Amazon floodplain (eg, Silva et al. 2013, Röpke et al. 2016, Virgilio et al. 2021, 2022). The predominance of characiforms is a common pattern, which has been recorded in numerous studies of the fish associated with macrophyte stands in floodplain lakes (Pelicice et al. 2005, Virgilio et al. 2022). This reflects the many small-bodied species found in this order, such as the “piabas” (eg, Hyphessobrycon sp., Hemigrammus bellottii, Moenkhausia mikia, Ctenobrycon spilurus), which use macrophyte stands for protection, foraging, feeding, and reproduction (Sánchez-Botero et al. 2003, Pelicice et al. 2005). During the low water phase, most oxbow lakes become disconnected from the principal river channel, forming closed ecosystems that are unable to exchange their fish with the river (Penha et al. 2017, Virgilio et al. 2022). As a result, many small species, such as Bryconops cf. melanurus and Carnegiella myersi Fernández-Yépez 1950, use the dense substrate provided by the roots of the macrophytes as a refuge from piscivorous fish such as Acestrorhynchus microleps (Rodríguez & Lewis 1997, Sánchez-Botero et al. 2003, Pelicice et al. 2005).

In the present study, the fish species richness was influenced significantly by local (negatively with depth and positively with macrophyte stand size) and spatial factors (negatively with the PCNM3). The depth of the water acts as an environmental filter for the fish communities in larger aquatic ecosystems, such as rivers and lakes, given that only a few species, such as large predators, are capable of inhabiting deeper waters effectively (Miranda 2011, Arantes et al. 2013, Soares et al. 2021). The fact that the larger macrophyte stands tend to have higher species richness may be accounted for by the typical species-area relationship (MacArthur & Wilson 1967, Drakare et al. 2006), given that larger stands will tend to have more habitats, and greater environmental and structural heterogeneity which will favor a higher number of fish species. The influence of the spatial distribution of the stands may be related to their role as dispersal mechanisms (Schiesari et al. 2003, Virgilio et al. 2022). When these stands become detached during the flood and ebb phases, they can take the associated fish fauna with them (Schiesari et al. 2003, Virgilio et al. 2022).

The relative abundance of fish observed in the low water phase in comparison with the flood phase may be related to an increase in the density of the fish caused by the reduction of the available habitable area. Fish cannot disperse across the floodplain during the low water phase because they are isolated in the oxbow lakes, leading to an increase in the catch per unit of effort (Petry et al. 2003, Penha et al. 2017, Virgilio et al. 2022). During the flood phase, the connection of the lakes to the river channel allows the fish to disperse across the floodplain in search of refuges and feeding resources, which tends to reduce the abundance of fish and their species richness in the lakes (Penha et al. 2017, Virgilio et al. 2022).

While the transparency of the water had no effect on the fish species composition in the present study, the variables related to the structure of the macrophyte population had a clear influence on the fish community. Local environmental variables are considered to be limiting factors for fish, as observed in many previous studies of the lakes of the Amazon floodplain (eg, Röpke et al. 2016, Virgilio et al. 2021, 2022). The structure of the macrophyte stand is crucial to the success of many fish species, given that more heterogeneous stands tend to have a greater diversity of plant species and lifeforms, providing greater habitat complexity in the roots and other submerged structures (Grenouillet et al. 2002, Padial et al. 2009, Virgilio et al. 2021, 2022).

The seasonal dynamics of the rivers of the Amazon floodplain determine the patterns of fish community structure in its lakes, given that the transition from the low water to the flood phases causes shifts in the ecosystem at both local and regional scales (Thomaz et al. 2007, Junk et al. 2014, Virgilio et al. 2021, 2022, van der Sleen & Rams 2023). In the present study, the hydrological phase and spatial distribution of the macrophyte stands were the ecological factors that best explained the variation in the structure of the fish communities, considering the influence of the pure fractions in the partitioning of the variation. During the low water phase, the loss of connectivity constitutes a physical barrier to dispersal, which prevents the fish from exiting the lake, and intensifies many ecological interactions, such as predator–prey relationships (Fernandes et al. 2009, Virgilio et al. 2022). Extended periods of drought may also alter the physicochemical conditions of a lake, for example, shallower waters tends to be warmer, modifying the entire vertical temperature gradient of the lake, which will lead to a reduction in oxygen concentrations, thereby excluding the fish species that are intolerant of low dissolved oxygen concentrations or anoxic conditions (Petry et al. 2003). During the flood phase, the lateral expansion of the river removes the physical barriers, allowing the fish that were previously isolated to disperse across the floodplain (Thomaz et al. 2007, Bozelli et al. 2015). In addition, the supply of feeding resources is increased by the access to new items provided by the flooded forest (Junk et al. 1989, 2014). Finally, the flood pulse homogenizes abiotic conditions across the floodplain, by mixing the physical and chemical conditions of the river and the lakes (Thomaz et al. 2007, Gomes et al. 2012, Bozelli et al. 2015). The potential homogenizing effect on the composition of the fish community may be reflected in the fraction shared between the hydrological phases and the local environmental variables. Furthermore, a spatially-structured environmental variation represented by the shared fraction between space and local environmental variables. Although the significance of the contribution of the shared fractions to the partitioning was not verified, they represented only a very small proportion of the variability in the composition of the fish community in the present study.

The local environment, habitats, and the hydrological phase are factors that explain the variation in the structure of the fish communities of the oxbow lakes of the Amazon floodplain. The shift between the low water and flood phases alters connectivity, affects environmental variables, and the biotic interactions in the oxbow lakes (Thomaz et al. 2007, Fernandes et al. 2009, Petsch 2016). These factors shape the fish communities of the floodplain, favoring fish communities of a specific composition that are characteristics of the different hydrological phases, and are essential for the maintenance of the aquatic biota.

SUPPLEMENTARY MATERIAL

Figure S1.

Tables SI, SII.

ACKNOWLEDGMENTS

We thank the team from the Laboratório de Biologia Animal da Universidade Federal do Acre, and Pablo Silva, Hugo Luis, Roni Zumba, Wirven Lima, Tiago Lucena, Tamires Nascimento, Marilene Santos and André Casas for helping with field sampling, and to Dr. Flávio Lima for confirming fish species. Fundação de Amparo à Pesquisa do Estado do Acre (FAPAC) granted undergraduate research scholarship for RSS; Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) provided a doctoral scholarship for RSS (Process No. 88887.494044/2020-00). FC thanks the Universidade Estadual do Maranhão (Grant No. 18/2023 - PPG/UEMA) for the scholarship grant. We thank Stephen Ferrari for proofreading the English. The dataset is available from the corresponding author upon request.

REFERENCES

• ACRE. 2012. Acre State Water Resources Plan. Rio Branco: SEMA, Rio Branco.
• ANA - AGÊNCIA NACIONAL DE ÁGUAS. 2022. Hidroweb v. 3.2.6. Available at: <https://www.snirh.gov.br/hidroweb/serieshistoricas>. Accessed on April 7th, 2022.
  » https://www.snirh.gov.br/hidroweb/serieshistoricas
• ARANTES CC, CASTELLO L, CETRA M & SCHILLING A. 2013 Environmental factors affecting the distribution of arapaima in floodplains of the Amazon. Environ Biol Fish 96: 1257-1267.
• ARRINGTON DA & WINEMILLER KO. 2003. Diel changeover in sandbank fish assemblages in a neotropical floodplain river. J Fish Biol 63: 442-459.
• BEDOYA AM & MADRIÑÁN S. 2015. Evolution of the aquatic habit in Ludwigia (Onagraceae): Morpho-anatomical adaptive strategies in the Neotropics. Aquat Bot 120: 352-362.
• BONETTO AA. 1975. Hydrologic regime of the Paraná River and its influence on ecosystems. In: HASLER AD (Ed), Coupling of land and water systems. Springer, New York, p. 175-197.
• BORCARD D & LEGENDRE P. 2002. All-scale spatial analysis of ecological data by means of main coordinates of neighbor matrices. Ecol Model 153: 51-68.
• BORCARD D, LEGENDRE P, AVOIS-JACQUET C & TUOMISTO H. 2004. Dissecting the spatial structure of ecological data at multiple scales. Ecology 85: 1826-1832.
• BOZELLI RL, THOMAZ SM, PADIAL AA, LOPES PM & BINI LM. 2015. Floods decrease zooplankton beta diversity and environmental heterogeneity in an Amazonian floodplain system. Hydrobiologia 753: 233-241.
• CABRAL GS, SILVA RS, ARAUJO RMG & GHIDINI AR. 2021. Composition and diversity of phytophilous cladocerans of oxbow lakes of Southwest Amazonian, Acre state, Brazil. Biota Neotrop 21: e20201079.
• CALCAGNO V & DE MAZANCOURT C. 2010. glmulti: An R package for easy automated model selection with (generalized) linear models. J Stat Softw 34: 1-29.
• CUNHA ER, THOMAZ SM, EVANGELISTA HBA, CARNIATO J, SOUZA CF & FUGI R. 2011. Small-sized fish assemblages do not differ between a native and a recently established non-indigenous macrophyte in a neotropical ecosystem. Nat Cons 9: 61-66.
• DIBBLE ED & THOMAZ SM. 2009. Use of fractal dimension to assess habitat complexity and its influence on dominant invertebrates inhabiting tropical and temperate macrophytes. J Freshw Ecol 24: 93-102.
• DRAKARE S, LENNON JJ & HILLEBRAND H. 2006. The imprint of the geographical, evolutionary and ecological context on species-area relationships. Ecol Lett 9: 215-227.
• FERNANDES IM, MACHADO FA & PENHA J. 2010. Spatial pattern of a fish assemblage in a seasonal tropical wetland: effects of habitat, herbaceous plant biomass, water depth, and distance from species sources. Neotrop Ichthyol 8: 289-298.
• FERNANDES R, GOMES LC, PELICICE FM & AGOSTINHO AA. 2009. Temporal organization of fish assemblages in floodplain lagoons: the role of hydrological connectivity. Environ Biol Fish 85: 99-108.
• FIGUEIREDO BRS, MORMUL RP & BENEDITO E. 2015. Structural complexity and turbidity do not interact to influence predation rate and prey selectivity by a small visually feeding fish. Mar Freshw Res 66: 170-176.
• FITZGERALD DB, WINEMILLER KO, SABAJ PÉREZ MH & SOUSA LM. 2017. Seasonal changes in the assembly mechanisms structuring tropical fish communities. Ecology 98: 21-31.
• FREITAS TMDS, PRUDENTE BDS & MONTAG LFDA. 2022. Flood pulse influence on the feeding ecology of two Amazonian auchenipterid catfishes. Neotrop Ichthyol 20: e210103.
• FRICKE R, ESCHMEYER WN & VAN DER LAAN R. 2022. Eschmeyer’s catalog of fishes: genera, species, references. Available at: http://researcharchive.calacademy.org/research/ichthyology/catalog/fishcatmain.asp Accessed 06 April 2022.
  » http://researcharchive.calacademy.org/research/ichthyology/catalog/fishcatmain.asp
• GAZULHA V, MONTÚ M, MARQUES DM & BONECKER CC. 2011. Effects of natural banks of free-floating plants on zooplankton community in a shallow subtropical lake in southern Brazil. Braz Arch Biol Technol 54: 745-754.
• GOMES LC, BULLA CK, AGOSTINHO AA, VASCONCELOS LP & MIRANDA LE. 2012. Fish assemblage dynamics in a Neotropical floodplain relative to aquatic macrophytes and the homogenizing effect of a flood pulse. Hydrobiologia 685: 97-107.
• GOULDING M. 1980. The fish and the forest. University of California Press, Los Angeles, 250 p.
• GRENOUILLET G, PONT D & SEIP KL. 2002. Abundance and species richness as a function of food resources and vegetation structure: juvenile fish assemblages in rivers. Ecography 25: 641-650.
• GUTERRES MGP, MARMONTEL M, ROSAS FCW, PAZIN VFV & VENTICINQUE EM. 2008. Anatomy and morphology of amazonian aquatic plants used as potential food by amazonian manatees. Mamiraua Institute for Sustainable Development, Amazonas, 187 p.
• HEINO J. 2000. Lentic macroinvertebrate assemblage structure along gradients in spatial heterogeneity, habitat size and water chemistry. Hydrobiologia 418: 229-242.
• JUNK WJ. 1973. Investigation of the ecology and production biology of the “Floating meadows” (Paspalo-Echinochloetum) on the Middle Amazon. II. The aquatic fauna in the root zone of floating vegetation. Amazonian, Kiel, p. 9-102.
• JUNK WJ ET AL. 2014. Brazilian wetlands: their definition, delineation, and classification for research, sustainable management, and protection. Aquat Conserv Mar Freshw Ecosyst 24: 5-22.
• JUNK WJ, BAYLEY PB & SPARKS RE. 1989. The food pulse concept in river-foodplain systems. In: DODGE DP (Ed), Proceedings of the international large river symposium. Can Spec Publ Fish Aquat Sci 106: 110-127.
• JUNK WJ, PIEDADE MTF, SCHÖNGART J & WITTMANN F. 2012. A classification of major natural habitats of Amazonian white-water river floodplains. Wetl Ecol Manag 20: 461-475.
• LEGENDRE P & ANDERSON MJ. 1999. Distance-based redundancy analysis: testing multispecies responses in multifactorial ecological experiments. Ecol Monogr 69: 1-24.
• LEGENDRE P & DE CÁCERES M. 2013. Beta diversity as the variance of community data: dissimilarity coefficients and partitioning. Ecol Lett 16: 951-963.
• LEGENDRE P & LEGENDRE L. 2012. Numerical ecology. Elsevier, 1006 p.
• LOPES TM, CUNHA ER, SILVA JCB, BEHREND RD & GOMES LC. 2015. Dense macrophytes influence the horizontal distribution of fish in floodplain lakes. Environ Biol Fish 98: 1741-1755.
• MACARTHUR RH & WILSON EO. 1967. The theory of island biogeography. Princeton University Press, Princeton, 224 p.
• MALTCHIK L, ROLON AS & SCHOTT P. 2007. Effects of hydrological variation on the aquatic plant community in a floodplain palustrine wetland of southern Brazil. Limnology 8: 23-28.
• MEERHOFF M, IGLESIAS C, DE MELLO FT, CLEMENTE JM, JENSEN E, LAURIDSEN TL & JEPPESEN E. 2007. Effects of habitat complexity on community structure and predator avoidance behavior of littoral zooplankton in temperate versus subtropical shallow lakes. Freshw Biol 52: 1009-1021.
• MIRANDA LE. 2011. Depth as an organizer of fish assemblages in floodplain lakes. Aquat Sci 73: 211-221.
• MIRANDA LE & LUCAS GM. 2004. Determinism in fish assemblages of floodplain lakes of the vastly disturbed Mississippi alluvial valley. Trans Am Fish Soc 133: 358-370.
• NEKOLA JC & WHITE PS. 1999. The distance decay of similarity in biogeography and ecology. J Biogeogr 26: 867-878.
• NONATO FAS, MICHELAN TS, FREITAS PV, MAIA C & MONTAG LFDA. 2021. Heterogeneity of macrophyte banks affects the structure of fish communities in flooded habitats of the Amazon Basin. Aquat Ecol 55: 215-226.
• OKSANEN J ET AL. 2018. Vegan: Community Ecology Package. R Package Version 2: 5-2.
• ORTEGA JC, FIGUEIREDO BR, DA GRAÇA WJ, AGOSTINHO AA & BINI LM. 2020. Negative effect of turbidity on prey capture for both visual and non-visual aquatic predators. J Anim Ecol 89: 2427-2439.
• PADIAL AA, THOMAZ SM & AGOSTINHO AA. 2009. Effects of structural heterogeneity provided by the floating macrophyte Eichhornia azurea on the predation efficiency and habitat use of the small Neotropical fish Moenkhausia sanctaefilomenae. Hydrobiologia 624: 161-170.
• PELICICE FM, AGOSTINHO AA & THOMAZ SM. 2005. Fish assemblages associated with Egeria in a tropical reservoir: investigating the effects of plant biomass and diel period. Acta Oecol 27: 9-16.
• PENHA J, LANDEIRO VL, ORTEGA JCG & MATEUS L 2017. Interchange between flooding and drying, and spatial connectivity control the fish metacommunity structure in lakes of the Pantanal wetland. Hydrobiologia 797: 115-126
• PERES-NETO PR & LEGENDRE P. 2010. Estimating and controlling for spatial structure in the study of ecological communities. Glob Ecol Biogeogr 19: 174-184.
• PETRY P, BAYLEY PB & MARKLE DF. 2003. Relationships between fish assemblages, macrophytes and environmental gradients in the Amazon River floodplain. J Fish Biol 63: 547-579.
• PETSCH DK. 2016. Causes and consequences of biotic homogenization in freshwater ecosystems. Int Rev Hydrobiol 3: 113-122.
• PIEDADE MTF & JUNK WJ. 2000. Natural grasslands and herbaceous plants in the Amazon floodplain and their use. In: JUNK WJ, OHLY JJ, PIEDADE MTF & SOARES MGM (Eds), The Central Amazon floodplain: actual use and options for a sustainable management. Leiden, Backhuys Publishers, p. 269-290.
• PIEDADE MTF, JUNK W, D’ÂNGELO SA, WITTMANN F, SCHÖNGART J, BARBOSA KMDN & LOPES A. 2010. Aquatic herbaceous plants of the Amazon floodplains: state of the art and research needed. Acta Limnol Bras 22: 165-178.
• POTT VJ & POTT A. 2000. Plantas aquáticas do Pantanal, Brasília, 404 p.
• QUEIROZ LJ, TORRENTE-VILARA G, OHARA WM, PIRES THS, ZUANON J, DORIA CRC. 2013. Peixes do Rio Madeira. Latin American Documentary Dialect, São Paulo, 402 p.
• R DEVELOMENT CORE TEAM 2022. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna. https://www.r-project.org/
  » https://www.r-project.org/
• RODRÍGUEZ MA & LEWIS JR WM. 1997. Structure of fish assemblages along environmental gradients in floodplain lakes of the Orinoco River. Ecol Monogr 67: 109-128.
• RÖPKE CP, AMADIO SA, WINEMILLER KO & ZUANON J. 2016. Seasonal dynamics of the fish assemblage in a floodplain lake at the confluence of the Negro and Amazon Rivers. J Fish Biol 89: 194-212.
• SÁNCHEZ-BOTERO JI & ARAUJO-LIMA CA. 2001. As macrófitas aquáticas como berçário para a ictiofauna da várzea do rio Amazonas. Acta Amazon 31: 437-437.
• SÁNCHEZ-BOTERO JI, FARIAS MD, PIEDADE MT & GARCEZ DS. 2003. Ictiofauna associada às macrófitas aquáticas Eichhornia azurea (SW.) Kunth. e Eichhornia crassipes (Mart.) Solms. no lago Camaleão, Amazônia Central, Brasil. Acta Sci Biol Sci 25: 369-375.
• SANTOS NCL, SANTANA HS, ORTEGA JCG, DIAS RM, STEGMANN LF, ARAÚJO IMS, SEVERI W, BINI LM, GOMES LC & AGOSTINHO AA. 2017. Environmental filters predict the trait composition of fish communities in reservoir cascades. Hydrobiologia 802: 245-253.
• SAVINO JF & STEIN RA. 1982. Predator-prey interaction between largemouth bass and bluegills as influenced by simulated, submersed vegetation. Trans Am Fish Soc 111: 255-266.
• SCHIESARI L, ZUANON J, AZEVEDO-RAMOS C, GARCIA M, GORDO M, MESSIAS M & VIEIRA EM. 2003. Macrophyte rafts as dispersal vectors for fishes and amphibians in the Lower Solimões River, Central Amazon. J Trop Ecol 19: 333-336.
• SILVA MT, PEREIRA JO, VIEIRA LJS & PETRY AC. 2013. Hydrological seasonality of the river affecting fish community structure of oxbow lakes: A limnological approach on the Amapá Lake, southwestern Amazon. Limnologica 43: 79-90.
• SILVANO RAM. 2001. Fish from the Upper Juruá River (Amazonas, Brazil). EdUSP, São Paulo, 304 p.
• SOARES BE, BENONE NL, BARTHEM RB, MARCENIUK AP & MONTAG LF. 2021. Environmental conditions promote local segregation, but functional distinctiveness allows aggregation of catfishes in the Amazonian estuary. Estuar Coast Shelf Sci 251: 107256.
• SOARES MGM, MENEZES NA & JUNK WJ. 2006. Adaptations of fish species to oxygen depletion in a central Amazonian floodplain lake. Hydrobiologia 568: 353-367.
• SÚAREZ YR, PETRERE JRM & CATELLA AC. 2001. Factors determining the structure of fish communities in Pantanal lagoons (MS, Brazil). Fish Manag Ecol 8: 173-186.
• SUÇUARANA MS, VIRGÍLIO LR & VIEIRA LJS. 2016. Trophic structure of fish assemblages associated with macrophytes in lakes of an abandoned meander on the middle river Purus, Brazilian Amazon. Acta Sci Biol Sci 38: 37-46.
• TAKEDA AM, SOUZA-FRANCO GM, MELO SM & MONKOLSKI A. 2003. Invertebrates associated with aquatic macrophytes of the upper Paraná River floodplain (Brazil). In: THOMAZ SM & BINI LM (Ed), Ecology and Management of Aquatic Macrophytes. Eduem, Maringá, p. 243-260.
• TEJERINA-GARRO FL, FORTIN R & RODRÍGUEZ MA. 1998. Fish community structure in relation to environmental variation in floodplain lakes of the Araguaia River, Amazon Basin. Env Biol Fish 51: 399-410.
• THOMAZ SM, BINI LM & BOZELLI RL. 2007. Floods increase similarity among aquatic habitats in river-floodplain systems. Hydrobiologia 579: 1-13.
• THOMAZ SM, DIBBLE ED, EVANGELISTA LR, HIGUTI J & BINI LM. 2008. Influence of aquatic macrophyte habitat complexity on invertebrate abundance and richness in tropical lagoons. Freshw Biol 53: 358-367.
• TONN WN. 1990. Climate change and fish communities: a conceptual framework. Trans Am Fish Soc 119: 337-352.
• VAN DER SLEEN P & RAMS M. 2023. Flood pulses and fish species coexistence in tropical rivers - a theoretical food web model. Env Biol Fish 106: 1785-1796.
• VIRGILIO LR, RAMALHO WP, SILVA JC, SUÇUARANA MO, GOMES RS & VIEIRA LJ. 2021. Influence of niche and neutral processes on fish communities associated with changes in macrophyte rafts along the hydrological cycle. Biologia 76: 2551-2560.
• VIRGILIO LR, RAMALHO WP, SUÇUARANA MS & VIEIRA LJS. 2022. Effects of hydrological, environmental and spatial factors on fish diversity and community structure in oxbow lakes from the Amazon floodplain. Limnologica 93: 125954.
• WOOLNOUGH DA, DOWNING JA & NEWTON TJ. 2009. Fish movement and habitat use depends on water body size and shape. Ecol Freshw Fish 18: 83-91.
• ZUUR AF, IENO EN & ELPHICK CS. 2010. A protocol for data exploration to avoid common statistical problems. Methods Ecol Evol 1: 3-14.

Publication Dates

• Publication in this collection
  05 Aug 2024
• Date of issue
  2024

History

• Received
  04 May 2023
• Accepted
  10 Feb 2024