Hydrological Period, Drainage and Local Environmental Conditions Influence Fish Assemblages in Upland Streams in the Eastern Amazon, Brazil

Abstract

Amazon streams are home to a great richness and diversity of fish, having an essential role in maintaining the aquatic ecosystem multifunctionality and global biodiversity. Here, we investigated the structure of the ichthyofauna of upland streams of the Lower Tapajós River and analyzed ecological descriptors of fish assemblages in different drainages in the rainy and dry seasons. A total of 3715 individuals from 110 species were collected. Species richness was higher during the dry season (99 species) than in the rainy season (66 species). Local environmental variables were measured or obtained from publicly accessible databases. Our results showed that ichthyofauna responds to hydrological changes in upland streams in the eastern Amazon. Abundance and richness were greatest during the dry season, with important contributions from representatives of the order Characiformes. Stream structural variables explained most of the variance in assemblage composition (adjusted R² = 0.102, p = 0.004), with channel width, depth, and canopy cover as key factors. The findings underscore the importance of assessing drainage and seasonality effects not only to understand ichthyofaunal biodiversity but also to adequately design research efforts, conservation strategies, and monitoring programs for aquatic environments in the eastern Amazon.

1. Introduction

The Amazon basin has the largest and most diverse freshwater ichthyofauna in the world and is home to a large number of still little-known or undescribed fish species [1,2,3]. Such a megadiversity stems from processes that, although not yet fully understood, are known to be related to the stability of climatic factors, the geomorphological evolution of the region, and its vast, dynamic, and heterogeneous drainage system [3,4,5,6].

The megadiverse Amazonian fish assemblages are annually subjected to large seasonal variations in hydrological regime resulting from wind regimes and, particularly, fluctuations in rainfall that represent the main natural forces that shape the landscape of the basin [7,8]. The intense seasonal variations in water levels may affect the temporal stability of fish communities, making a need to understand their ecological effects as the driving force in the spatial-temporal dynamics of abundance and diversity of species [9].

The current consensus in river ecology emphasizes that the hydrological regime is the principal driver of ecological dynamics in river-floodplain systems, affecting fish assemblages’ taxonomic and functional structure [10,11,12]. The seasonal water level fluctuations are directly associated with some dynamics of ecological processes on fish species, which may include habitat availability and connectivity, reproductive success, and food availability [8,13,14,15,16,17]. Furthermore, these variations promote substantial changes in water physical characteristics and other environmental conditions, such as temperature, turbidity, and oxygenation, which can directly affect the physiology and feeding behavior of aquatic species [8], producing very dynamic and heterogeneous aquatic ecosystems with a wide variety of habitats for fish species [7,18,19,20,21,22].

Small forest streams that make up the complex and vast drainage system of the Amazon are driven by a pattern of variation in hydrological behavior. These watercourses, in general, do not share the annual unimodal flood pulses of floodplain rivers and lakes [8,23,24]. Flood events are commonly short and take place quickly and unpredictably because of intense periods of rainfall, which promote the expansion and contraction of stream networks, modulating the dynamics of aquatic communities [8,19,25].

Some studies in tropical streams demonstrated that seasonal changes do not promote significant alterations in the structure of fish assemblages, due to the high ecological resilience and plasticity of the species that inhabit these environments [19,24]. Studies such as those by [26,27] in Atlantic Forest streams [7,28] in the Amazon argue that, despite seasonal variations in water depth and flow, fish assemblages maintain a stable structure throughout the year, highlighting the importance of evolutionary adaptations of fish species in tropical environments [29,30] have revealed remarkable heterogeneity in species composition among tributaries. Across the broader Amazon Basin, fish diversity follows well-defined biogeographic gradients driven by geomorphological evolution, hydrological connectivity, and climatic stability [2,3,5]. However, despite this growing understanding of large-scale patterns, few studies have explicitly examined how local and regional factors interact to shape stream fish assemblages in upland portions of the eastern Amazon. This study differs from previous Amazonian stream research, e.g., [19,24,30] by integrating spatial (drainage-based) and temporal (hydrological period) factors within a protected upland forest context. Unlike prior works that emphasized central Amazon lowlands, we provide the first assessment of how drainage identity, local habitat structure, and physicochemical conditions jointly influence upland ichthyofauna composition in the eastern Amazon. This multi-scale approach expands the understanding of Amazonian stream biodiversity under different hydrological regimes and contributes to refining conservation and monitoring frameworks for upland aquatic environments, particularly in the eastern portion of the basin, where upland streams represent conservation and monitoring gaps.

In view of the foregoing, our study aimed to investigate the taxonomic structure of the ichthyofauna in upland streams of the lower Tapajós River, analyze possible changes in the composition of fish assemblages between dry and rainy seasons and drainages, and assess associations between local environmental variables and ichthyofaunal composition. Despite extensive research on fish assemblage structure during the last few decades, our understanding of relationships between fish assemblage structure and hydrological dynamics in Amazon upland streams remains limited. Studies conducted on stream systems in central Amazonia, for example, have shown that periods of rain promote intense effects but are insufficient to significantly change the structure of the ichthyofauna, e.g., [24,31]. However, we expect that hydrology and local environmental variations will influence the abundance and diversity of species in upland streams in the eastern Amazon. Upland streams may be more susceptible to hydrological regime than river-floodplain systems since the effects of seasonal variations are specifically pronounced. During the rainfall period, intense and frequent events occur of expansion and contraction of stream networks [25].

2. Methods

2.1. Study Site

The Tapajós River basin, the second largest clear water system in the Amazon basin, is currently impacted by several human activities, mainly derived from advances in the agricultural frontier represented by the significant increase in pasture and agricultural areas, and the significant practice of mining, representing the largest concentration of miners in the entire Brazilian Amazon [32,33].

The streams studied are in a conservation unit, the Tapajós National Forest in the lower portion of the Tapajós River, eastern Amazon, western Pará State, Brazil (3°18′36.36″ S, 55°2′32.28″ W) (Figure 1). It occupies an area of approximately 527,000 ha, encompassing part of the municipalities of Aveiro, Belterra, Placa, and Rurópolis. The predominant vegetation is of the dense ombrophilous forest type. On the banks of the streams that drain the territory, there is an abundance of palm trees of the genus Euterpe (Arecaceae), common components of riparian vegetation in Amazon lowlands [34].

Hydrology and precipitation patterns are fundamental elements for seasonal studies. In the Tapajós basin, the hydrological cycle is characterized by alternating periods of flooding and drought, influenced by seasonal rains. The rainy season, which generally occurs from December to May, promotes changes in the physical structure of streams with lateral flooding, creating and connecting habitats. The dry season occurs from June to November, when the water level drops, concentrating fish in smaller areas and intensifying competition for resources.

The topography is characterized by an undulating relief with low to moderate slopes where waterbodies are common. The streams of the Tapajós National Forest drain into two distinct drainages, namely the Curuá-Una River and Cupari River basins, as well as into direct tributaries of the Tapajós River (Figure 1). The predominant climate is hot and humid, with average annual temperatures ranging from 25 to 27 °C [35].

2.2. Local Environmental Variables

Sampling was carried out in 22 first- to third-order streams from September to December 2015 (dry season) and from February to May 2016 (rainy season). According to data from INMET/INPE, the cumulative rainfall in the study area during the years 2015 and 2016 was 2779.6 mm (Figure 2).

A standardized sampling protocol was used, partially according to the procedures proposed in a previous study [24]. In each stream, a 50 m long transect was demarcated using 5 mm mesh nets. A multiparameter meter (Akso AK 88) was used to measure water pH, electrical conductivity (µS cm⁻¹), dissolved oxygen (mg·L⁻¹), and temperature (°C). Then, four equidistant points were marked along each transect (0, 17, 34, and 50 m) for the measurement of flow velocity (m·s⁻¹), channel width (m), and channel depth (m). Substrate composition was analyzed simultaneously with channel depth and classified into one of six categories, namely sand, clay, tree trunks, leaf litter, tree roots, and macrophytes (

Table 1. Range of variation and standard deviation (in parentheses) of the environmental characteristics of the terra firme streams of the Tapajós National Forest.

Environmental Variable	Drought	Rainy
Litter (%)	8–59 (±14.2)	4.8–66.6 (±18.3)
Liteira Slim (%)	0–37 (±8.7)	2–49.7 (±13.7)
Clay (%)	0–33.9 (±11.3)	0–34 (±11.5)
Pebble (%)	0–25 (±9.7)	0–34 (±10.1)
Sand (%)	0–55.1 (±17.0)	0–69 (±21.1)
Root (%)	0–24.5 (±6.8)	0–12.5 (±4.0)
Macrophyte (%)	0–12.2 (±2.8)	0–16.5 (±7.3)
Trunk (%)	0–9 (±3.0)	0–14.8 (±4.2)
Rock (%)	0–11.3 (±4.0)	0–20 (±5.0)
pH	3.5–6.9 (±0.93)	3.5–7.2 (±1.14)
Conductivity (µS/cm)	9.8–54.8 (±11.0)	7–43.8 (±8.9)
Temperature °C	24. 7–27 (±0.53)	24.4–25.8 (±0.41)
Dissolved oxygen (mg/L)	1.9–8.3 (±1.7)	4.2–8.8 (±1.0)
Chain speed (m/s)	0.02–0.71 (±0.15)	0.03–0.57 (±0.13)
Width (m)	0.88–7.8 (±2.0)	1.1–11.8 (±2.8)
Canopy (%)	6.2–65.9 (±12.3)	5.4–54.9 (±11.3)
Depth (m)	0.1–0.7 (±0.17)	0.16–1.2 (±0.29)
Flow rate	0.01–1.9 (±0.51)	0.03–4.5 (±1.24)

). The proportional substrate composition was estimated as the proportion of each substrate type in relation to all substrates in each transect.

Four photos of the canopy in the direction of cardinal points (16 photos per sampling site) were taken from the riverbank of each transect. The average canopy cover (%) was calculated from the percentage of white pixels using the threshold tool of Adobe Photoshop^®,, version 25.0.

2.3. Fish Collection

We sampled the ichthyofauna in 22 first and second-order streams (50 m stretch per stream) using the standardized active capture method described in [24]. Fish were actively collected by two collectors over a period of approximately 2 h using sieves and seine nets (5 mm mesh size). We sampled the streams in the 2025 dry season (October, September to December) and in the 2016 wet season (March, April and May). 50 m stretches were blocked with fine mesh nets (5 mm stretched mesh size), and fish were sampled with two people/two hours of effort using seine nets and hand nets. The captured fish were anesthetized in eugenol (clove oil) solution, fixed in 10% formalin, and transferred to 70% ethanol. After screening, fish were identified using identification keys [29,30,36,37,38,39,40,41,42,43,44,45].

2.4. Data Analysis

2.4.1. Changes in Community Composition Across Drainages and Periods

The assessment of the ichthyofauna composition in this study was carried out based on the gross abundances of species, recorded in specific sampling periods and segmented by drainage system. This method allows a robust comparative analysis of fish distribution, considering the particularities of each river basin [24].

The changes in ichthyofaunal composition according to period (dry and rainy) and drainage, as well as their interaction (Period × Drainage), were investigated using permutational multivariate analysis of variance (PERMANOVA) [46]. PERMANOVA results were applied to a Bray–Curtis similarity matrix using Type II sums of squares, given that the design was unbalanced [47]. A block design was used for periods (dry and rainy) to limit the 9999 permutations used in the hypothesis test within each period.

Before analysis, the data were tested for homogeneity of multivariate dispersions [48]. Having met the assumptions for each source of variation (period: F = 0.081, p = 0.796; drainage: F = 1.932, p = 0.153; interaction: F = 0.900, p = 0.482), the data were subjected to PERMANOVA. Differences revealed by PERMANOVA were assessed using a paired test for comparisons between pairs of levels for the factor drainage (Anderson et al., 2008) [47].

Ichthyofauna were ordered (abundance) using principal coordinate analysis [49] to visualize the structuring of the community in a multidimensional space. Cailliez’s correction was used to avoid negative eigenvalues. Ichthyofauna were subjected to indicator species analysis (IndVal) for identification of the most important species based on relative abundance and frequency of occurrence for each period–drainage pair, using Bray–Curtis distance matrix. Only species identified as group indicators (p < 0.05) were displayed in the ordination.

2.4.2. The Role of Local Environmental Variables

Environmental variables were divided into two groups: stream structural variables and water physicochemical variables. The structural variables included litter, fine litter, clay, pebble, root, macrophytes, trunk, rock, channel width, canopy cover, channel depth, and flow velocity. The physicochemical variables were pH, conductivity, temperature, and dissolved oxygen.

To investigate the association between ichthyofauna and environmental variables, we tested environmental variables for collinearity using the variance inflation factor (VIF). The variables sand and flow velocity were excluded from the analysis because they had VIF values greater than 5 [50].

The relationship between ichthyofauna and environmental characteristics of streams was investigated using the variance partitioning method, which evaluates the amount of variance of the response matrix explained by each subset of explanatory variables [51].

Redundancy analysis associated with an analogous analysis of variance, using permutations, was used to assess the significance of each fraction of variance in terms of adjusted R² values [52]. Having identified, through variance partitioning, that the joint fraction between structural and physicochemical variables explained the highest amount of variance in ichthyofaunal composition, we applied redundancy analysis [49] to order ichthyofaunal patterns explained by stream structural variables and water physicochemical variables. In the redundancy analysis graphs, observations and explanatory variables were displayed using ‘scaling = 2′. Samples were coded by colors (for each drainage) and shapes (for each period), and scores resulting from the ordering of explanatory variables (LC scores) were used. Only species with a goodness of fit greater than 0.25 on one of the two canonical axes were presented, using ‘scaling = 1′. Sorting graphs produced with ‘scaling = 1′ show the best display (optimal display) of the distance relationship between objects (sites). On the other hand, graphs with ‘scaling = 2′ result in the best display of covariances between variables [51]. All analyses were performed in the R environment (R Development Core Team, 2020) using the HH package [53,54].

3. Results

Environmental characteristics of the sampled upland streams varied between dry and rainy seasons (Table 1). In the dry season, pH ranged from 3.5 to 6.9 (±0.93), while in the rainy season it ranged from 3.5 to 7.2 (±1.14), with generally more acidic conditions in the dry period. Structural parameters such as channel width (0.88–7.8 m dry; 1.1–11.8 m rainy) and depth (0.1–0.7 m dry; 0.16–1.2 m rainy) were greater in the rainy season, corresponding with increased flow and habitat expansion. Substrate composition included varying proportions of sand, clay, litter, rocks, and woody material, reflecting heterogeneous microhabitats. Canopy cover ranged from approximately 6% to 66%, contributing to light and temperature regimes relevant for fish habitat quality.

A total of 3715 individuals belonging to 110 species were collected The highest richness was recorded in the dry season, with 99 species, 44 of which were exclusive to the period. In the rainy season, 66 species were recorded, 11 of which were exclusive. Ichthyofauna were distributed in 28 families and 6 orders, with 64 (58.2%) species belonging to Characiformes, 19 (17.3%) to Siluriformes, 13 (11.8%) to Cichliformes, 10 (9.1%) to Gymnotiformes, 3 (2.7%) to Cyprinodontiformes, and 1 (0.9%) to Synbranchiformes. The family Acestrorhamphidae, with 27 species (24.5%), had the highest richness, followed by Cichlidae, with 13 species (11.8%), Loricariidae, with 8 species (7.3%), Iguanodectidae, with 7 species (6.4%), and Lebiasinidae and Crenuchidae, with 5 species each (4.5%).

Ichthyofaunal composition differed between hydrological periods and drainages (

Table A1. PERMANOVA results showing significant differences in ichthyofaunal composition between periods and drainages in streams of the Tapajós National Forest, Brazilian Amazon.

Source of Variation	SS	Mean Square	df	F	Pr(>F)
Period	0.3164	0.31638	1	0.8225	0.0001
Drainage	2.6845	1.34224	2	3.4896	0.0001
Interaction	0.6652	0.33262	2	0.8648	0.7685
Residuals	12.6932	0.38464	33
Total	16.3586		38

SS, sum of squares; df, degrees of freedom.

). The pairwise test revealed differences in ichthyofaunal composition between all drainages in the dry period. In the rainy period, ichthyofaunal composition differed significantly only between drainages of the Curuá-Una and Cupari Rivers (Figure 3; Table A1).

In the dry period, the drainage of the Cupari River had six indicator species: Gymnotus carapo, Characidium zebra, Jupiaba acanthogaster, Satanoperca jurupari, Bario oligolepis, and Phenacogaster calverti. In the same period, Saxatilia inpa and Aequidens tetramerus were the indicator species of Curuá-Una, whereas Helogenes marmoratus, Apistogramma agassizii, Copella callolepis, Elachocharax junki, and Crenuchus spilurus were indicators of Tapajós. In the rainy period, Erythrinus erythrinus was the indicator species of Curuá-Una and Curimatopsis macrolepis of Tapajós (Figure 3;

Table A3. Results of indicator species (IndVal) analysis showing indicator species (p < 0.05) for each Period × Drainage group.

Species	Group	IndVal	p-Value	Code
Gymnotus carapo Linnaeus, 1758	Dry_Cupari	34.59	0.043	38
Characidium zebra Eigenmann, 1909	Dry_Cupari	38.33	0.040	42
Jupiaba acanthogaster (Eigenmann, 1911)	Dry_Cupari	36.92	0.039	51
Satanoperca jurupari (Heckel, 1840)	Dry_Cupari	40.00	0.033	53
Moenkhausia oligolepis (Günther, 1864)	Dry_Cupari	40.50	0.023	56
Phenacogaster calverti (Fowler, 1941)	Dry_Cupari	34.29	0.043	57
Crenicichla inpa Ploeg, 1991	Dry_Curuá-Una	48.75	0.006	8
Aequidens tetramerus (Heckel, 1840)	Dry_Curuá-Una	60.73	0.001	97
Helogenes marmoratus Günther, 1863	Dry_Tapajós	39.90	0.027	3
Apistogramma agassizii (Steindachner, 1875)	Dry_Tapajós	43.22	0.021	4
Copella callolepis (Regan, 1912)	Dry_Tapajós	60.80	0.007	11
Elachocharax junki (Géry, 1971)	Dry_Tapajós	47.37	0.012	13
Crenuchus spilurus Günther, 1863	Dry_Tapajós	74.37	0.002	22
Erythrinus erythrinus (Bloch & Schneider, 1801)	Rainy_Curuá-Una	38.83	0.039	2
Curimatopsis macrolepis (Steindachner 1876)	Rainy_Tapajós	45.00	0.010	98

).

Stream structural variables changed significantly across periods (Figure 4). Variance partitioning identified stream structural variables as the pure fraction explaining the variance in ichthyofaunal composition between periods and drainages (

Table 2. Variance partitioning results, adjusted R², and significance for each testable fraction df, degrees of freedom. a + b + c = matrizes de dados.

Variable	Fraction	R2	Adj. R2	df	F	Pr(>F)
Physicochemical (PC)	[a + b]	0.148	0.019	5	1.149	0.111
Structural (S)	[b + c]	0.346	0.080	11	1.303	0.002
PC + S	[a + b + c]	0.480	0.102	16	1.269	0.004
PC (pure effect)	[a]		0.021	5	1.127	0.169
Joint effect	[b]		−0.001	0		Not testable
S (pure effect)	[b]		0.082	11	1.275	0.017
Residuals	[d]		0.898			Not testable

). The effect of water physicochemical variables on ichthyofauna was significant only when added to the effect of structural variables (adj. R² = 0.102, F = 1.269, Pr (>F) = 0.004).

Figure 3. Principal coordinate analysis (PCoA). (A) plot of ichthyofauna collected in different periods and drainages and (B) identification of indicator species according to Table 3. Polygons (dashed lines) represent the space occupied by samples from each drainage, regardless of the period.

Figure 3. Principal coordinate analysis (PCoA). (A) plot of ichthyofauna collected in different periods and drainages and (B) identification of indicator species according to Table 3. Polygons (dashed lines) represent the space occupied by samples from each drainage, regardless of the period.

The portion explained by environmental variables at redundancy analysis showed that canonical axis 1 explained 18.5%, associating the ichthyofauna with the high channel depth, channel width, pebbles, trunks and canopy cover. Streams that drain into the Curuá-Una River, which have high flow velocity and dissolved oxygen content, were plotted on the negative side of axis 1. Canonical axis 2, which explained 16.5% of the variance in the dataset, represented the environments of the Cupari River, characterized by high conductivity and pH, clay substrate, rocks, and litter (Figure 5; Table 2).

Figure 5. Relationship between ichthyofauna and environmental variables, resulting from the Redundancy Analysis carried out for the dry and rainy periods. Biplot of the explanatory variables and the observations by drainage with the respective numbers of the sampling points of the igarapés (A). Explanatory variables and species with the highest cumulative proportion of variance in each canonical axis (B) as shown in Table 3.

Figure 5. Relationship between ichthyofauna and environmental variables, resulting from the Redundancy Analysis carried out for the dry and rainy periods. Biplot of the explanatory variables and the observations by drainage with the respective numbers of the sampling points of the igarapés (A). Explanatory variables and species with the highest cumulative proportion of variance in each canonical axis (B) as shown in Table 3.

Table 3. List of species occurring by drainage and their respective codes in the redundancy analysis.

CHARACIFORMES	DRAINAGES
Acestrorhamphidae	Tapajós	Cupari	Curua-una	Code
Hemigrammus ulreyi (Boulenger, 1895)			x	5
Moenkhausia comma Eigenmann, 1908	x		x	9
Hemigrammus vorderwinkleri Géry, 1963	x		x	21
Moenkhausia conspicua Soares & Bührnheim, 2016			x	34
Hyphessobrycon sp.		x	x	35
Hyphessobrycon heterorhabdus (Ulrey, 1894)		x		44
Jupiaba acanthogaster (Eigenmann, 1911)		x		51
Moenkhausia oligolepis (Günther, 1864)	x	x		56
Brachychalcinus sp.		x		58
Hemigrammus marginatus Ellis, 1911		x		70
Jupiaba potaroensis (Eigenmann, 1909)		x		72
Hyphessobrycon agulha Fowler, 1913		x		74
Moenkhausia gr. lepidura (Kner, 1858)		x		75
Jupiaba zonata (Eigenmann, 1908)		x		76
Hyphessobrycon copelandi Durbin, 1908		x		77
Moenkhausia ceros Eigenmann, 1908		x		79
Hemigrammus bellottii (Steindachner, 1882)	x	x		81
Hemigrammus analis Durbin, 1909	x			86
Hemigrammus stictus (Durbin, 1909)	x			87
Moenkhausia collettii (Steindachner, 1882)	x		x	92
Hyphessobrycon wosiackii Moreira & Lima, 2017	x			95
Hemigrammus ocellifer (Steindachner, 1882)	x	x	x	96
Jupiaba sp.		x		99
Moenkhausia sp.		x		106
Moenkhausia jamesi Eigenmann, 1908		x		107
Hemigrammus sp.	x		x	108
Hemigrammus geisleri Zarske & Géry, 2007	x			109
Acestrorhynchidae
Gnathocharax steindachneri Fowler, 1913	x			64
Heterocharax virgulatus Toledo-Piza, 2000	x			88
Acestrorhynchus falcatus (Bloch, 1794)		x		39
Anostomidae
Anostomus ternetzi Fernández-Yépez, 1949		x		104
Characidae
Phenacogaster calverti (Fowler, 1941)		x		57
Chilodontidae
Chilodus punctatus Müller & Troschel, 1844		x		43
Crenuchidae
Characidium sp. 1	x			17
Melanocharacidium sp.		x	x	33
Characidium zebra Eigenmann, 1909		x		42
Crenuchus spilurus Günther, 1863	x			22
Curimatidae
Cyphocharax gangamon Vari, 1992	x			85
Curimatopsis macrolepis (Steindachner, 1876)	x			98
Erythrinidae
Erythrinus erythrinus (Bloch & Schneider, 1801)	x	x	x	2
Hoplias malabaricus (Bloch, 1794)	x	x	x	10
Hoplias curupira Oyakawa & Mattox, 2009	x	x		48
Gasteropelecidae
Carnegiella strigata (Günther, 1864)	x			62
Iguanodectidae
Bryconops aff. colaroja Chernoff & Machado-Allison, 1999	x			18
Bryconops allisoni Silva-Oliveira, Canto, & Ribeiro, 2019	x			19
Iguanodectes variatus Géry, 1993	x			20
Bryconops imitator Chernoff & Machado-Allison, 2002		x		41
Bryconops munduruku Silva-Oliveira, Canto, & Ribeiro, 2015	x			69
Bryconops melanurus (Bloch, 1794)	x	x		94
Bryconops sp.			x	105
Lebiasinidae
Copella callolepis (Regan, 1912)	x			11
Nannostomus marginatus Eigenmann, 1909			x	26
Nannostomus trifasciatus Steindachner, 1876			x	31
Nannostomus eques Steindachner, 1876	x			63
Pyrrhulina capim Vieira & Netto-Ferreira, 2019	x	x	x	66
Serrasalmidae
Myloplus rubripinnis (Müller & Troschel, 1844)		x	x	25
Serrasalmus rhombeus (Linnaeus, 1766)	x			90
Stevardiidae
Knodus sp. 2		x	x	23
Creagrutus sp.			x	27
Knodus sp. 1		x	x	30
Hemibrycon surinamensis Géry, 1962		x		52
Creagrutus petilus Vari& Harold, 2001		x		55
Knodus sp. 4		x		71
CICHLIFORMES
Cichlidae
Apistogramma agassizii (Steindachner, 1875)	x		x	4
Saxatilia inpa Ploeg, 1991	x	x	x	8
Crenicichla johanna Heckel, 1840	x			12
Satanoperca jurupari (Heckel, 1840)		x		53
Taeniacara candidi Myers, 1935	x			61
Crenicichla pellegrini Ploeg, 1991	x			68
Geophagus sp.		x		73
Hypselecara coryphaenoides (Heckel, 1840)	x			83
Acaronia nassa (Heckel, 1840)	x			84
Aequidens tetramerus (Heckel, 1840)	x	x	x	97
Apistogramma regani Kullander, 1980	x		x	100
Caquetaia spectabilis (Steindachner, 1875)		x		103
Acarichthys heckelii (Müller &Troschel, 1849)	x			110
Crenuchidae
Elachocharax junki (Géry, 1971)	x			13
CYPRINODONTIFORMES
Rivulidae
Anablepsoides urophthalmus (Günther, 1866)	x	x	x	37
Anablepsoides sp.			x	89
Anablepsoides ornatus (Garman, 1895)	x	x	x	93
GYMNOTIFORMES
Gymnotidae
Gymnotus coropinae Hoedeman, 1962	x		x	7
Gymnotus carapo Linnaeus, 1758		x	x	38
Gymnotus coatesi LaMonte, 1935	x	x		80
Hypopomidae
Steatogenys duidae (LaMonte, 1929)		x	x	14
Hypopygus sp.	x			15
Hypopygus lepturus Hoedeman, 1962	x		x	16
Microsternarchus bilineatus Fernández-Yépez, 1968	x			65
Rhamphichthyidae
Gymnorhamphichthys petiti Géry & Vu, 1964	x		x	32
Sternopygidae
Eigenmannia macrops (Boulenger, 1897)	x	x		45
Sternopygus macrurus (Bloch & Schneider, 1801)	x	x	x	91
SILURIFORMES
Auchenipteridae
Tatia intermedia (Steindachner, 1877)			x	24
Cetopsidae
Denticetopsis seducta Vari, Ferraris, & de Pinna, 2005			x	28
Heptapteridae
Brachyglanis microphthalmus Bizerril, 1991			x	29
Callichthyidae
Callichthys callichthys (Linnaeus, 1758)		x		102
Cetopsidae
Helogenes marmoratus Günther, 1863	x		x	3
Doradidae
Acanthodoras sp.	x			67
Heptapteridae
Rhamdia quelen (Quoy & Gaimard, 1824)	x	x	x	36
Phenacorhamdia sp.		x		47
Pimelodella cristata (Müller & Troschel, 1849)			x	82
Loricariidae
Hypostomus sp.	x	x	x	40
Farlowella smithi Fowler, 1913		x		46
Ancistrus sp. 1		x		49
Ancistrus sp.		x		50
Harttia dissidens Rapp Py-Daniel & Oliveira, 2001		x		54
Rineloricaria lanceolata (Günther, 1868)		x		59
Curculionichthys sp.		x		60
Farlowella reticulata Boeseman, 1971		x		78
Pseudopimelodidae
Batrochoglanis raninus (Valenciennes, 1840)			x	101
Trichomycteridae
Ituglanis amazonicus (Steindachner, 1882)			x	1
SYNBRANCHIFORMES
Synbranchidae
Synbranchus marmoratus Bloch, 1795	x	x	x	6

x represents the presence/occurrence of the species in the drainage area.

Table 3. List of species occurring by drainage and their respective codes in the redundancy analysis.

CHARACIFORMES	DRAINAGES
Acestrorhamphidae	Tapajós	Cupari	Curua-una	Code
Hemigrammus ulreyi (Boulenger, 1895)			x	5
Moenkhausia comma Eigenmann, 1908	x		x	9
Hemigrammus vorderwinkleri Géry, 1963	x		x	21
Moenkhausia conspicua Soares & Bührnheim, 2016			x	34
Hyphessobrycon sp.		x	x	35
Hyphessobrycon heterorhabdus (Ulrey, 1894)		x		44
Jupiaba acanthogaster (Eigenmann, 1911)		x		51
Moenkhausia oligolepis (Günther, 1864)	x	x		56
Brachychalcinus sp.		x		58
Hemigrammus marginatus Ellis, 1911		x		70
Jupiaba potaroensis (Eigenmann, 1909)		x		72
Hyphessobrycon agulha Fowler, 1913		x		74
Moenkhausia gr. lepidura (Kner, 1858)		x		75
Jupiaba zonata (Eigenmann, 1908)		x		76
Hyphessobrycon copelandi Durbin, 1908		x		77
Moenkhausia ceros Eigenmann, 1908		x		79
Hemigrammus bellottii (Steindachner, 1882)	x	x		81
Hemigrammus analis Durbin, 1909	x			86
Hemigrammus stictus (Durbin, 1909)	x			87
Moenkhausia collettii (Steindachner, 1882)	x		x	92
Hyphessobrycon wosiackii Moreira & Lima, 2017	x			95
Hemigrammus ocellifer (Steindachner, 1882)	x	x	x	96
Jupiaba sp.		x		99
Moenkhausia sp.		x		106
Moenkhausia jamesi Eigenmann, 1908		x		107
Hemigrammus sp.	x		x	108
Hemigrammus geisleri Zarske & Géry, 2007	x			109
Acestrorhynchidae
Gnathocharax steindachneri Fowler, 1913	x			64
Heterocharax virgulatus Toledo-Piza, 2000	x			88
Acestrorhynchus falcatus (Bloch, 1794)		x		39
Anostomidae
Anostomus ternetzi Fernández-Yépez, 1949		x		104
Characidae
Phenacogaster calverti (Fowler, 1941)		x		57
Chilodontidae
Chilodus punctatus Müller & Troschel, 1844		x		43
Crenuchidae
Characidium sp. 1	x			17
Melanocharacidium sp.		x	x	33
Characidium zebra Eigenmann, 1909		x		42
Crenuchus spilurus Günther, 1863	x			22
Curimatidae
Cyphocharax gangamon Vari, 1992	x			85
Curimatopsis macrolepis (Steindachner, 1876)	x			98
Erythrinidae
Erythrinus erythrinus (Bloch & Schneider, 1801)	x	x	x	2
Hoplias malabaricus (Bloch, 1794)	x	x	x	10
Hoplias curupira Oyakawa & Mattox, 2009	x	x		48
Gasteropelecidae
Carnegiella strigata (Günther, 1864)	x			62
Iguanodectidae
Bryconops aff. colaroja Chernoff & Machado-Allison, 1999	x			18
Bryconops allisoni Silva-Oliveira, Canto, & Ribeiro, 2019	x			19
Iguanodectes variatus Géry, 1993	x			20
Bryconops imitator Chernoff & Machado-Allison, 2002		x		41
Bryconops munduruku Silva-Oliveira, Canto, & Ribeiro, 2015	x			69
Bryconops melanurus (Bloch, 1794)	x	x		94
Bryconops sp.			x	105
Lebiasinidae
Copella callolepis (Regan, 1912)	x			11
Nannostomus marginatus Eigenmann, 1909			x	26
Nannostomus trifasciatus Steindachner, 1876			x	31
Nannostomus eques Steindachner, 1876	x			63
Pyrrhulina capim Vieira & Netto-Ferreira, 2019	x	x	x	66
Serrasalmidae
Myloplus rubripinnis (Müller & Troschel, 1844)		x	x	25
Serrasalmus rhombeus (Linnaeus, 1766)	x			90
Stevardiidae
Knodus sp. 2		x	x	23
Creagrutus sp.			x	27
Knodus sp. 1		x	x	30
Hemibrycon surinamensis Géry, 1962		x		52
Creagrutus petilus Vari& Harold, 2001		x		55
Knodus sp. 4		x		71
CICHLIFORMES
Cichlidae
Apistogramma agassizii (Steindachner, 1875)	x		x	4
Saxatilia inpa Ploeg, 1991	x	x	x	8
Crenicichla johanna Heckel, 1840	x			12
Satanoperca jurupari (Heckel, 1840)		x		53
Taeniacara candidi Myers, 1935	x			61
Crenicichla pellegrini Ploeg, 1991	x			68
Geophagus sp.		x		73
Hypselecara coryphaenoides (Heckel, 1840)	x			83
Acaronia nassa (Heckel, 1840)	x			84
Aequidens tetramerus (Heckel, 1840)	x	x	x	97
Apistogramma regani Kullander, 1980	x		x	100
Caquetaia spectabilis (Steindachner, 1875)		x		103
Acarichthys heckelii (Müller &Troschel, 1849)	x			110
Crenuchidae
Elachocharax junki (Géry, 1971)	x			13
CYPRINODONTIFORMES
Rivulidae
Anablepsoides urophthalmus (Günther, 1866)	x	x	x	37
Anablepsoides sp.			x	89
Anablepsoides ornatus (Garman, 1895)	x	x	x	93
GYMNOTIFORMES
Gymnotidae
Gymnotus coropinae Hoedeman, 1962	x		x	7
Gymnotus carapo Linnaeus, 1758		x	x	38
Gymnotus coatesi LaMonte, 1935	x	x		80
Hypopomidae
Steatogenys duidae (LaMonte, 1929)		x	x	14
Hypopygus sp.	x			15
Hypopygus lepturus Hoedeman, 1962	x		x	16
Microsternarchus bilineatus Fernández-Yépez, 1968	x			65
Rhamphichthyidae
Gymnorhamphichthys petiti Géry & Vu, 1964	x		x	32
Sternopygidae
Eigenmannia macrops (Boulenger, 1897)	x	x		45
Sternopygus macrurus (Bloch & Schneider, 1801)	x	x	x	91
SILURIFORMES
Auchenipteridae
Tatia intermedia (Steindachner, 1877)			x	24
Cetopsidae
Denticetopsis seducta Vari, Ferraris, & de Pinna, 2005			x	28
Heptapteridae
Brachyglanis microphthalmus Bizerril, 1991			x	29
Callichthyidae
Callichthys callichthys (Linnaeus, 1758)		x		102
Cetopsidae
Helogenes marmoratus Günther, 1863	x		x	3
Doradidae
Acanthodoras sp.	x			67
Heptapteridae
Rhamdia quelen (Quoy & Gaimard, 1824)	x	x	x	36
Phenacorhamdia sp.		x		47
Pimelodella cristata (Müller & Troschel, 1849)			x	82
Loricariidae
Hypostomus sp.	x	x	x	40
Farlowella smithi Fowler, 1913		x		46
Ancistrus sp. 1		x		49
Ancistrus sp.		x		50
Harttia dissidens Rapp Py-Daniel & Oliveira, 2001		x		54
Rineloricaria lanceolata (Günther, 1868)		x		59
Curculionichthys sp.		x		60
Farlowella reticulata Boeseman, 1971		x		78
Pseudopimelodidae
Batrochoglanis raninus (Valenciennes, 1840)			x	101
Trichomycteridae
Ituglanis amazonicus (Steindachner, 1882)			x	1
SYNBRANCHIFORMES
Synbranchidae
Synbranchus marmoratus Bloch, 1795	x	x	x	6

x represents the presence/occurrence of the species in the drainage area.

4. Discussion

Ichthyofauna responds to hydrological changes in upland streams in the eastern Amazon. The dissimilarity in the composition of the ichthyofauna between in upland streams of the Tapajós National Forest reinforces the hypothesis that the composition of the fish assemblages in the Amazon basin responds significantly to the hydrological regime [19,55]. Our results showed that abundance and richness were greatest during the dry season, with important contributions from representatives of the order Characiformes, and that stream structural variables were the most influential factors affecting ichthyofauna.

Our study differs from studies conducted in Amazonian streams that support the hypothesis that seasonal effects in the basin are unable to cause significant changes in the ichthyofauna structure in these environments [7,19,24,28]. Although our study was conducted in minimally preserved streams located within a conservation unit, we observed that structural changes in the streams, such as width, depth, substrates, and canopy openness (Figure 4) were the most influential factors affecting the ichthyofauna between the dry and rainy seasons.

Stream morphological changes, such as variations in channel width, depth, substrate heterogeneity, and canopy cover, influenced the distribution and abundance of specific fish species. For example, benthic insectivores like Characidium zebra prefer shallow, fast-flowing riffles with stable coarse substrates, common in structurally complex reaches [56,57]. Piscivorous and invertivorous cichlids such as Satanoperca jurupari and Aequidens tetramerus select habitats with abundant cover and nesting sites during low water periods, consistent with findings from [19,58]. These species exemplify contrasting habitat-use strategies along structural gradients. Such habitat preferences underscore the importance of habitat complexity in maintaining species diversity and spatial segregation within upland Amazonian streams. Our results demonstrate that associations between fish taxonomic composition and hydrological changes are an important component for understanding relationships between fish lateral and longitudinal movements and hydrological fluctuations in upland streams in the Amazon [58]. The Amazonian seasonal hydrological cycle promotes significant changes in the structure of streams, stimulating the displacement of species to more suitable environments [59]. Our results show that abundant fish species changed locations during seasonal events, such as Elachocharax junki, Crenuchus spilurus, Aequidens tetramerus, Satanoperca jurupari, Gymnotus carapo, and Characidae zebra move to flooded marginal habitats during the rainy season to exploit increased food availability and shelter, returning to main channels during drought to maintain territories [58,60,61,62]. Reproductive cycles synchronized with the onset of rainy season, evidenced in most cichlids of the genera Aequidens and Satanoperca and for some Siluriformes such as Helogenes marmoratus, correspond with habitat shifts that optimize offspring survival under fluctuating environmental conditions [63,64]. These life history traits highlight adaptive strategies that facilitate resilience to seasonal habitat variability in Eastern Amazon fish assemblages.

The records show that, during the rainy season, two indicator species, Erythrinus erythrinus and Curimatopsis macrolepis, were found to be more abundant. This suggests that, similar to other members of the Curimatidae family, C. macrolepis may follow a similar pattern, moving from rivers and lakes into flooded forest areas or upland streams. during high-water periods to exploit abundant resources and favorable breeding conditions. As water levels recede, these movements are likely followed by a return to more stable environments. However, specific studies focusing on the movement ecology of C. macrolepis are required to confirm these behaviors. Meanwhile, Erythrinus erythrinus takes advantage of the rainy season by moving into prey-rich flooded areas, where it benefits from the increased availability of resources [65]. This dynamic highlights the behavioral and ecological adaptability of species to seasonal fluctuations in Amazonian upland streams.

The indicator species of this study (Figure 3; Table A3) share ecological and biological characteristics that enable them to move between dry and rainy seasons in Amazonian streams. Furthermore, some indicator species have their reproductive cycle beginning at the end of the drought (December) and extending until the rainy season, for example, Aequidens tetramerus, Helogenes marmoratus, and Gymnotus cf. carapo [63,64,66]. For example, Characidium zebra, a benthic insectivore adapted to clear, fast-flowing environments with coarse substrates [7], was characteristic of the Cupari drainage during the dry period, when flow velocity increases locally and water transparency is highest. This reflects its ecological preference for shallow riffle habitats and its territorial behavior restricted to stable microhabitats. Similarly, Gymnotus carapo and Satanoperca jurupari indicate structurally complex reaches where shelter and nesting sites are abundant during low water levels, reinforcing the role of habitat structure as a driver of assemblage differentiation.

On the other hand, species such as Gymnotus carapo and Helogenes marmoratus move to flooded environments during the rainy season, where they find abundant shelter and food. During the dry season, these species concentrate in the streambed, using marginal areas as refuges. Piscivorous and invertivorous species, such as Saxatilia inpa, Apistogramma agassizii, Satanoperca jurupari and Aequidens tetramerus, actively move to flooded environments during the rainy season to take advantage of the greater availability of prey [65], in addition to spawning during this period, taking advantage of the abundance of refuges and resources available to protect eggs and larvae [67]. During the dry season, these species tend to return to the stream-bed, where the concentration of prey is more abundant during this period. Apistogramma agassizii and Aequidens tetramerus, in addition to their predatory behaviors, exhibit intense parental care, ensuring greater survival of the offspring even in highly dynamic environmental conditions [68,69].

The dietary plasticity observed in species such as Bario oligolepis, Phenacogaster calverti, and Jupiaba acanthogaster, which feed on a wide variety of resources (Goulding, 1980), is further enhanced by reproductive strategies tightly synchronized with hydrological cycles [67,70]. This combination of traits promotes mobility and enables these species to thrive in unstable environments, allowing them to efficiently exploit seasonal resources. During floods, they capitalize on the abundant opportunities provided by the inundation, while during the dry season, they adapt to adverse conditions by seeking refuge and sustaining their survival strategies, as described by Junk [8].

This combination of feeding, reproductive and behavioral strategies reflects the remarkable adaptability of these species to the pronounced effects of seasonal variations in Amazonian highland streams, which promote environmental changes in physical–chemical variables and, mainly, structural characteristics of streams, affecting the gradients of resource availability, such as shelter, food, or abiotic conditions, such as habitat stability and connectivity, causing these migrations [58,71,72,73].

Considering that local environmental characteristics are true filters of biodiversity [74,75], we demonstrated that structural characteristics associated mainly with deeper and wider streams, with pebbles, logs and greater canopy cover were important for the variability of ichthyofauna between seasonal periods. In addition to structural characteristics, water physicochemical variables influence the presence of species indicators of spatio-temporal variability in upland streams in the eastern Amazon (Table 2; Figure 3; Table A3).

The influence of water’s physical and chemical attributes is the result of changes in environmental patterns promoted by seasonality in Amazon River systems [76]. We identified significant differences in physicochemical characteristics between drainages of the Cupari, Curuá-Una and Tapajós rivers during both periods. Electrical conductivity and pH were the variables that most contributed to differentiating the drainage of the Cupari River from drainages of the Curuá-Una and Tapajós Rivers with high values in the Cupari drainage during the dry and rainy seasons. Furthermore, the drainage of the Cupari River had wider and deeper channels, some classified as third order [77], capable of supporting a rich and abundant ichthyofaunistic community, as observed in stream 9, where the highest values of abundance and richness were recorded during the study. Environments with greater support capacity are more able to sustain a rich community of local fish throughout the seasons, providing diverse microhabitats that serve as shelter and foraging for many fish species [78,79,80].

Our results demonstrate that spatial factors influence the change in ichthyofaunal composition in streams that drain the Tapajós National Forest, complementing previous findings, e.g., [43], that show hydrological events and the resulting changes in local environmental variables and habitat characteristics are not the sole determinants of the structure of fish assemblages in Amazon streams and upland environments [19,81]. Other factors, such as the geographical isolation promoted by drainage systems, can determine the composition of fish assemblages through physical barriers that, from an ecological perspective, can be said to play the role of reproductive barriers, directly influencing the reproductive chain of species. In addition, other environmental variables not considered in this study also vary between drainages, making them significant. Drainages with distinct characteristics offer different ecological conditions for each fish group [3,82]. This effect is particularly pronounced in upland streams, where the environment is more susceptible to seasonal variations.

The evidence supporting the hypothesis that local environmental variables determine patterns of changes in ichthyofaunal composition throughout the year is associated with results showing that spatial heterogeneity on a regional scale modulates environmental characteristics of the Amazon basin [5]. This set of environmental factors (e.g., geographic, climatic, habitat, vegetation, and soil) that explain the historically unequal distribution of ichthyofauna at the basin level indicates that the composition of fish assemblages in upland streams also undergoes significant changes throughout the year in response to environmental events imposed by the seasonal cycle.

The interplay between seasonal hydrological regimes, spatial heterogeneity among drainages, and the isolation imposed by upland interfluves drives the spatiotemporal dynamics of fish assemblages in the eastern Amazon. Variability in stream morphology and physicochemical conditions promotes ecological specialization and allopatric speciation, reflecting basin-wide evolutionary patterns described for Amazonian freshwater biota [2,5]. Understanding these processes is essential for designing conservation strategies that maintain both local habitat integrity and regional connectivity. Conservation and monitoring programs in the eastern Amazon should prioritize the protection of headwaters, the maintenance of riparian forests, and the preservation of hydrological regimes that sustain habitat diversity, biodiversity, and evolutionary potential. Given the high sensitivity of upland streams to hydrological variation, monitoring frameworks should adopt multi-year approaches to encompass interannual variability and inform adaptive management under climate change scenarios. Linking species’ life-history traits and ecological roles to hydrological cycles will further enhance predictive capacity for community responses to environmental alteration.

In view of the above, we can infer that the variability in ichthyofaunal structure in the Tapajós National Forest is explained by the joint effect of stream structural and water physicochemical variables, demonstrating that seasonal events (drought and rain) are capable of promoting significant changes in the fish assemblage of Amazonian streams. The high abundance of species during periods of low rainfall may be interpreted as a strong indication that most of the fish species occurring in drainage systems depend on stream integrity. Studies focusing on the conservation of fish species in streams of the lower Tapajós basin should consider the specificities of drainages, observing the response of these environments to the hydrological cycle.

Our results provide important contributions to understanding the complex and dynamic relationship involving the Amazonian seasonal cycle (rainy and dry periods) and its influences on the abiotic parameters that lead to changes in the dynamics of abundance and diversity of species in upland streams in the eastern Amazon. However, a limitation of our study is that sampling encompassed only one hydrological cycle (one dry and one rainy season). Given the well-known interannual variability of precipitation and discharge in the eastern Amazon [9], interannual differences in rainfall intensity or duration could alter stream connectivity and habitat stability, influencing fish assemblage responses. Therefore, while our findings capture clear seasonal and drainage-related patterns, they should be interpreted as a representative snapshot rather than a long-term average. Continuous multi-year monitoring is required to confirm the temporal consistency of the observed patterns.

5. Conclusions

This study demonstrates that the ichthyofaunal composition of upland streams in the Tapajós National Forest is strongly influenced by seasonal hydrological events, with both stream structural attributes and water physicochemical variables playing a central role in shaping fish assemblages. The pronounced differences observed between the dry and rainy seasons, particularly the greater abundance and richness during the dry period, highlight the dynamic responses of fish communities to fluctuations in habitat availability, connectivity, and resource distribution.

Our findings emphasize spatial heterogeneity among drainages, coupled with differences in channel morphology, substrate composition, canopy cover, and physicochemical parameters, drives distinct patterns of species distribution and seasonal movement. The presence of indicator species with diverse feeding, reproductive, and behavioral strategies underscores the ecological adaptability of Amazonian stream fishes to the pronounced variability of their environments.

These results contribute to a broader understanding of how the interplay between hydrological cycles, habitat structure, and regional-scale spatial factors determines ichthyofaunal dynamics in the eastern Amazon. They also reinforce the need for conservation approaches that consider not only the temporal variability imposed by seasonal cycles but also the specific ecological characteristics of individual drainages. Maintaining the structural integrity and environmental quality of these streams is essential for sustaining the rich and diverse fish communities that characterize the lower Tapajós basin.