The influence of habitat structure on fish assemblages in Amazonian streams of Machado river basin Igor D. da Costa1 & Vanessa M. da Rocha2 1. Laboratório de Ciências Ambientais, Departamento de Engenharia de Pesca, Universidade Federal de Rondônia - UNIR. Rua da Paz, Lino Alves Teixeira, CEP 76196-000, Presidente Médici, RO, Brasil; [email protected] 2. Laboratório de Biologia e Cultivo de Peixes de Água Doce (LAPAD), Departamento de Aquicultura - Centro de Ciências Agrárias (CCA), Universidade Federal de Santa Catarina - UFSC. Rodovia Francisco Thomaz Dos Santos, 3532 - Armação do Pântano Do Sul Florianópolis SC CEP 88066-260; [email protected] Received 27-XI-2015. Corrected 14-VII-2016. Accepted 17-VIII-2016. Abstract: Igarapés are Amazonian streams that are extremely susceptible to environmental changes. Due to the rapidly occurring riparian land use changes, and the several impacts these may have on fish assemblages, it is highly valuable to describe and understand the current relationships between these assemblages and the local environmental conditions, especially in barely know areas. In this research, we studied the taxonomic composi- tion and fish assemblage attributes variation in three streams with different riparian conservation conditions: forest, intermediate and pasture. Samplings were performed every two months from October 2011 to September 2012, in three 1st order streams in the Machado River Basin, Rondônia. Fish were collected using trawls (seine net with a mesh size of 2 mm) and dip nets (2 mm mesh) along the entire stretch; the obtained samples were preserved and identified per site type. A total of 2 141 fish specimens belonging to 59 species, 17 families and five orders were recorded. Unlike the intermediate and pasture streams, the forest stream showed a high richness and low abundance. The forested stream exhibited the highest diversity and evenness value, and had low dominance, unlike the other streams. The variance partitioning and partial Redundancy Analysis (pRDA) indicated that assemblage composition was significantly explained by the environmental variables such as: depth, water velocity, pteridophytes and grasses, but not by spatial predictors. The non-metric multidimensional scaling (NMDS) analysis showed that intermediate and pasture streams separated from the forested stream. We concluded that both, the stream environmental variables and fish assemblage attributes were influenced by the different conservation status and land cover. Given the influence of regional processes, which have a pervasive role in local fish assemblages, land use at the watershed scale is important, especially to explain the higher rich- ness and diversity found in forested streams. Rev. Biol. Trop. 65 (1): 103-115. Epub 2017 March 01. Key words: complexity, conservation, richness, fish assemblages. Tropical rainforest streams are nutrient- (Dembkowski & Miranda, 2012) and influ- poor streams with a dense forest cover that ence aquatic vertebrate fauna (Miserendino prevents light from reaching the water sur- et al., 2011). face (Walker, 1995). Their biological and Physicochemical habitat characteristics are physicochemical features are therefore great- considered to be highly relevant factors affect- ly influenced by the surrounding vegetation ing the distribution of species and the organiza- (Mendonça, Magnusson, & Zuanon, 2005). tion of communities in streams (Hoeinghaus, They are also considered to be extremely Winemiller, & Birnbaum, 2007). Variations in vulnerable to environmental impacts, such species composition, structure and distribution as land-use change by agricultural purposes, in fish assemblages has been attributed to area which may cause environmental degradation (i.e. basin area, lake surface area), regional Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 65 (1): 103-115, March 2017 103 net primary production, habitat heterogeneity The present study aimed to determine the (Tedesco, Oberdorff, Lasso, Zapata, & Hugue- ichthyofauna composition in three streams with ny, 2005), substrate type, microhabitat type and different land cover (forest, mix of forest and water flow (Mendonça et al., 2005). pasture, and pasture) in the central region of the Land use plays a key role in shaping the Machado River basin in the state of Rondônia environment on global, regional and local (Brazil) and to investigate the relationships scales (Krusche et al., 2005). Changes in the between fish species and the environmental type of land cover influence the biodiversity variables. We hypothesized that streams in and ecosystem services (Foley et al., 2007), pasture areas would exhibit altered habitat con- global climate (Hartmann et al., 2013), biogeo- ditions and diversity lower fish assemblages chemical cycles and water cycles (Laudon et compared with sites with preserved and inter- al., 2016). The removal of native forest at the mediate degrees of preservation of land cover, watershed or at the riparian buffer scale, affects and that these would support more balanced stream characteristics at the local scale (Cruz, and diverse assemblages. Miranda, & Cetra, 2013), such as flow, depth, substrate composition, litter amount, stabil- ity of stream banks, and structural complexity MATERIALS AND METHODS (Lorion & Kennedy, 2009; Casatti, Ferreira, & Carvalho, 2009). Study area: This study was conducted Several studies have shown relationships in the Machado River basin, which drains between ichthyofauna and land use in micro- the most populated area of Rondônia, North- basins in temperate regions (Roth, Allan, & ern Brazil, with a total catchment area of Ericson, 1996; Fitzgerald, Kott, Lanno, & 75 400 km2. It rises in the Pimenta Bueno River Dixon, 1998; Rosenvald, Järvekülg, & Lõh- and flows into Madeira River, in the North of mus, 2014). However, in the Neotropics, this the state of Rondônia (Fernandes & Guimarães, issue is not well explored (Levêque, Oberdorff, 2002). The climate is characterized by tempera- Paugy, Stiassny, & Tedesco, 2008) and is a tures ranging from 19 to 33 °C, and the annual critical situation, considering that 40 % of the rainfall is approximately 2 500 mm (Krusche et Neotropical ichthyofauna remains unknown al., 2005). There are two well-defined seasons: (Reis, Kullander, & Ferraris, 2003). In addition the dry season (from late May to September) to a better knowledge of the taxonomy of the and the rainy season (from October to April) ichthyofauna, an understanding of the quantita- (Fernandes & Guimarães, 2002). Sampling tive structure of the ichthyofauna and its rela- was performed every two months from Octo- tionship with the environment is also important ber 2011 to September 2012 in three 1st order to support environmental conservation and streams with different levels of conservation, recovery strategies (Barletta et al., 2010). being a forested stream (FO) (Minuano stream; In the Brazilian Amazon, one of the major 11°01’ S - 61°54’42” W), a stream with inter- causes of deforestation is the logging and burn- mediate (IN) degrees of conservation (Cris ing of tropical forest, to produce pastures for stream; 11°15’ S - 61°51’24” W,) and a stream livestock. This mechanism initially enriches located in pasture (PA) (Vinte e três stream; the soil surface with cations from the burned plant biomass, which is transferred to the soil 11°05’ S - 61°53’29” W), all located in the in the form of ashes (Ferreira, Cyrino, Duarte- Machado River Basin, Rondônia. The study Neto, & Martinelli, 2012). This process is com- sites, located in private areas, showed total monly observed in the Machado River basin absence of domestic/industrial effluents, aqua- where land use resulted in extensive cattle culture, agriculture and dams, and the area farming, which is one of the main economic were formerly used for cattle farming but cur- activities in the Rondônia state (Costa, 2014). rently such activity was absent. 104 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 65 (1): 103-115, March 2017 Environmental data: The average width intertwined of riparian vegetation, grasses, (m) of each stream was based on measure- woody angiosperms and bare bank. ments at six equidistant points along a 50 m stretch. The average depth (m) was calculated Fish sampling: Fish were collected for from random surveys along the same stretch. one hour during the daytime (11:00-12:00) The water velocity (m/s) was determined by in a 50 m stretch. Three collectors used hand measuring the time a floating object traveled gathering techniques with trawls (seine net a known distance (1 m) in the center of the with a mesh size of 2 mm) and dip nets (2 mm channel (Casatti, Langeani, & Castro, 2001). mesh) along the entire stretch, selecting the The average transect area (At) was calculated best technique for each mesohabitat and micro- as the average of the areas of six transects in habitat. Before the collections, the ends of the each section studied using the following for- sampling sections were blocked with fine-mesh nets (5 mm between opposite knots) to prevent mula: At = [(Z + Z ) / 2].L + [(Z + Z ) / 2].L 1 2 2 3 the fish escaping, regardless of the collection +…+ [(Z + Z ) / 2].L, where Z is the depth n n+1 method used. measured in each segment and L is the width of each segment. The mean flow (m3/s) was The collected fish were fixed in 10 % formalin, placed in plastic bags, counted and calculated from the average water velocity and identified and later transferred to 70 % alcohol. the width and depth of the channel using the Taxonomic identification of the species was following formula: Q = At * Vm, where Q = performed using specialized literature (e.g., flow, Vm = mean water velocity, and At = aver- Casatti, Pérez-Mayorga, Carvalho, Brejão, age transect area. & Costa, 2013; Queiroz et al., 2013) and The physical structure of each stream was with the assistance of experts from several evaluated in two dimensions (at the land-water institutions. Voucher specimens were depos- transition zones and on the substrate types), for ited in the fish collection at the Universidade this purpose, each sampled section was divided Federal de Rondônia, Porto Velho, Rondônia into five smaller sections measuring 10 m state, Brazil (Vouchers: UFRO-ICT 023980 to each. To determine the physical structure of the UFRO-ICT 024000). land-water ecotone and the substrate types, the The Shapiro-Wilk normality test and Lev- proportions of the elements making up these ene’s test for homoscedasticity were applied to zones were measured at each section (total the data, to determine whether the assumptions 100 %), a visual-based habitat approach was of the parametric and nonparametric analyses based on previous protocols (Gorman & Karr, for the environmental variables, abundance, 1978; Barbour, Gerritsen, Snyder, & Stribling, species richness, diversity and evenness were 1999). The physical structure of the substrate satisfied. To compare these data within and was determined from the visual estimated pro- between wet and dry periods, the Kruskal- portion of sand (minimum diameter from 0.05 Wallis test was used when the assumptions of mm to 2 mm), clay, trunks (wood larger than normality and homoscedasticity were not met, 10 cm in diameter), coarse litter (composed even after transformation (log-transformed) of leaves and small branches), fine litter (fine of the data. When the data were normally particulate matter), roots (tangled roots of distributed, an analysis of variance (ANOVA) riparian vegetation) and macrophytes (aquatic was performed. Both of the analyses were per- vegetation), which were evaluated by visual formed with Statistica 7.1. estimation. The land-water ecotone (lateral The Shannon-Wiener diversity index distance of 10 m started at the adjacent portion (H), Berger-Parker dominance index (D) and of the stream) structures, on the left and right Pielou’s evenness index (e) (Ludwig & Reyn- banks, were analyzed for the occurrence of five olds, 1988) were calculated for each site and habitat descriptors: pteridophytes, fine roots season, and the species richness was estimated Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 65 (1): 103-115, March 2017 105 using the bootstrap method (1 000 randomiza- that each explained significant additional varia- tions). This is an incidence-based estimator of tion in stream assemblage composition and the species richness of rare species (collected abundance (Ter Braak & Smilauer, 1998). only once or twice in samplings) that is used to Significance of canonical axes and variation estimate the total species richness and pattern explained by environmental variables were deviation (Colwell, 2013). based on 10 000 Monte Carlo permutations. Non-metric multidimensional scaling The RDA decomposes the total explained vari- (NMDS) analysis based on species abundance ation into environmental and spatial compo- was conducted to characterize the composition nents. Then, the unbiased Variation Partitioning of fish assemblages. The dissimilarity matrix method proposed by Peres-Neto, Legendre, used in the ordination was constructed using Dray and Borcard (2006) was used to obtain the Bray-Curtis index. For the NMDS analy- the variance explained exclusively by envi- sis the data were previously log-transformed ronmental and spatial components, and their (log x+1). The NMDS calculates distortion, respective adjusted coefficients of determina- or stress, between the similarity matrix and tion (R 2). The NMDS and pRDA analyses adj the ordination produced in the graphical rep- were performed only with species greater abun- resentation of the axes. In this study, the stress dance than five individuals. values were below 0.1, which corresponded The VEGAN and PCNM packages were to a good ordering, ensuring good reliability used to run pRDA and PCNM analyses, which in the interpretation of the results (Clarke & were performed with R software (R Develop- Warwick, 1994). ment Core Team, 2009). The NMDS analysis To represent the spatial variation a Princi- was performed with PAST version 2.17 (Ham- pal Coordinates of Neighbor Matrices (PCNM) mer, Harper, & Ryan, 2001), which was also was performed (Dray, Legendre, & Peres-Neto, used to determine the diversity index and rich- 2006). To perform the PCNM analysis, water- ness estimates. The results were considered course distances among sampling points were significant when P ≤ 0.05. measured through Google Earth, using drainage net with 150 ha (Google Earth, 2016), resulting RESULTS in a triangular matrix of distances. Then, a matrix of truncated distance was calculated to The intermediate and pasture streams were retain the distance neighbors sampling points. surrounded by abandoned pastures, while the A Principal Coordinate Analysis (PCoA) of forested stream was surrounded by primary for- the truncated matrix was computed, and its est. There were no macrophytes in the forested positive eigenvectors were retained (Borcard, and intermediate stream, but the pasture stream Gillet, & Legendre, 2011) to be used as spatial showed a large number of emergent rooted predictors (Dray et al., 2006). A Partial Redun- macrophytes. Both banks of the forested stream dancy Analysis (pRDA) with a Variation Par- were composed of shrubby and woody angio- titioning method (Borcard & Legendre, 2002) sperms. The right bank of the intermediate was used to quantify the relative importance of stream was predominantly composed of grasses environmental variables and spatial distances and the left bank was composed of woody (positive PCNM eigenvectors) in explaining angiosperms. The banks of the pasture were the variation in species composition. composed of narrow stretches of woody angio- The pRDA is similar to multiple regres- sperms with stretches of bare bank (Table 1). sions, except that it allows for the analysis of The results of the Kruskal-Wallis test multiple response variables. To remove collin- indicated that the forested stream presented a earity among variables, a forward selection (α = higher percentage of pteridophytes (H = 16.2, 0.05) procedure was applied in pRDA to select P < 0.05), fine root intertwined (H = 16.2, P < and evaluate sets of environmental variables 0.05), woody angiosperms (H = 15.7, P < 0.05), 106 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 65 (1): 103-115, March 2017 TABLE 1 Environmental descriptors of the forested (FO), intermediate (IN) and pasture (PA) streams in the Machado River basin, RO Descriptors FO IN PA Land use in the adjacent matrix Primary forest Pasture Pasture Presence of macrophytes Absent Absent Present Riparian ecotone vegetation Shrubby and woody Grasses Narrow stretches of woody angiosperms on the right bank angiosperms with stretches of bare bank. Riparian ecotone vegetation Shrubby and woody Woody Narrow stretches of woody angiosperms on the left bank angiosperms angiosperms with stretches of bare bank. Physical variables Depth (m) 42.3±21.8 26.0±5.4 24.3±6.4 Width (m) 1.9±0.4 2.0±0.7 2.1±0.6 Water velocity (m/s)* 0.3±0.1a 0.2±0.1a 0.1±0.0b Flow (m3/s) 29.1±21.0 15.4±7.0 11.2±6.8 Ecotone characteristics Pteridophytes (%)** 5.0±0.8 a 0.0 b 0.0 b Fine roots intertwined (%)** 4.8±0.7 a 0.0 b 0.0 b Grasses (%)** 0.0 a 23.4±1.3 b 42.8±2.6 c Woody angiosperms (%)** 65.6±1.7 a 43.0±2.3 b 10.0±1.3 c Bare bank (%)** 0.0 b 0.0 b 42.2±2.0 a Substrate categories Rock (%)** 4.6±0.8 a 0.0 b 19.8±3.5 c Sand (%)** 59.5±8.0 b 89.1±3.3 a 80.±3.5 a Fine litter (%)** 11.3±1.9 a 6.3±1.0 b 0.0 c Coarse litter (%)** 16.7±2.2 a 0.0 b 0.0 b Large branches and trunks (%)** 10.8±1.3 a 4.5±2.9 b 0.0 c * = mean ± standard deviation and P < 0.05, and ** = median ± standard deviation and P < 0.05. Different letters indicate the significant differences between pairs of means in the ANOVA and Kruskal-Wallis tests. rock-type substrate (H = 16.0, P < 0.05), fine forested (n = 476) (Appendice 1). The forested litter (H = 15.9, P < 0.05), coarse litter (H = stream had higher richness (S = 41) (ANOVA, 16.1, P < 0.05) and large branches and trunks F = 15.0, p < 0.05) than the intermediate and (H = 15.9, p < 0.05) than the other analyzed pasture streams, which had 27 and 25 species, streams. The pasture stream had the highest respectively. The estimated richness generated percentage of grasses (H = 15.7, P < 0.05) and by the bootstrap method was 47 species in the bare bank (H = 16.1, P < 0.05) on the land-water forested and 31 in both the intermediate and ecotone. The forested and intermediate streams pasture stream. Forested stream exhibited the had higher water velocities (ANOVA, F = 3.85, highest diversity (H’ = 1.10) and evenness (e p < 0.05) than the pasture stream, whereas the = 0.21) value, and had a low dominance (D = intermediate and pasture stream had higher per- 0.31), unlike with intermediate (H’ = 0.81; D = centages of sand (Kruskal-Wallis, H = 14.6, P < 0.54; e = 0.17) and pasture stream (H’ = 0.52; 0.05) than the forested stream (Table 1). D = 0.35; e = 0.14). A total of 2 141 specimens belonging The temporal analysis of abundance (F = to 59 species, 17 families and 5 orders were 1.62; P > 0.05) and evenness (F = 0.99; P > recorded. The intermediate (n = 544) and pas- 0.05) did not show any differences among the ture (n = 1 119) streams had higher abundances streams for either period (Fig. 1A, 1D). The (Kruskal-Wallis, H = 11, 5; P < 0.05) than the species richness of the forested stream in the Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 65 (1): 103-115, March 2017 107 2.8 1.5 A B 2.6 1.4 +1) 2.4 1)1.3 og x 2.2 g x +1.2 L o Abundance ( 211...086 Richness (L110...109 1.4 0.8 1.2 0.7 INW IND PAW PAD FOW FOD INW IND PAW PAD FOW FOD Samples Samples 3.0 0.9 2.8 C D 2.6 0.8 2.4 0.7 2.2 2.0 0.6 H’1.8 e0.5 1.6 1.4 0.4 1.2 1.0 0.3 0.8 0.2 0.6 0.4 0.1 INW IND PAW PAD FOW FOD INW IND PAW PAD FOW FOD Samples Samples Fig. 1. Species abundance (A), richness (B), diversity (C) and evenness (D) in streams in the Machado River Basin, RO, during the dry and wet seasons. H’ = Shannon-Wiener index. e = Pielou’s evenness index. INW = intermediate stream/wet, IND = intermediate stream/dry, PAW = pasture stream/wet, PAD = pasture stream/dry, FOW = Forested stream/wet and FOD = Forested stream/dry. wet season (F = 4.01, P < 0.05) and the dry The comparison between fish species com- season (F = 4.74, P < 0.05) was higher than the position among stream groups, environmen- species richness of the pasture stream during tal variables and spatial distances explained the dry season (Fig. 1B). The diversity of the together 28 % of the variation (R 2 = 0.27; F adj forested stream in the wet season (F = 3.05, P = 2.27; P = 0.001). Assemblage composition < 0.05) and dry season (F = 3.75, P < 0.05) was was significantly influenced by environmental higher than the diversity of the pasture stream variables accounted for 15 % (R 2 = 0.15; adj during the dry season (Fig. 1C). The forested F = 1.83; P = 0.001), but not spatial predic- and pasture streams did not show any sig- tors (R 2= 0; F = 0.73; P = 0.75). The pRDA adj nificant differences in the parameters analyzed indicated that ten variables were redundant, related to the intermediate stream. and thus these variables were excluded from The NMDS analysis of the samples based the environmental data set. The forward selec- on species abundance showed that intermedi- tion procedure showed depth, water velocity, ate and pasture streams were separated from pteridophytes and grasses were the environ- forested stream from the test mean of the scores mental variables that accounted for significant of axis 1 (Kruskal-Wallis, H = 10, 1; P < 0.05), (P < 0.05) portions of the total variance in which was retained for interpretation (Stress = fish species composition. The pRDA with 0.09) (Fig. 2). these four environmental variables produced 108 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 65 (1): 103-115, March 2017 0.16 0.08 -0.4 -0.3 -0.2 -0.1 0.1 2 S D M N -0.08 -0.16 -0.24 NMDS 1 Fig. 2. Projection of the first two axes of non-metric multidimensional scaling based on the abundance of species in each stream. □ = forested stream, ○ = intermediate stream and ∆ = pasture stream. an ordination in which all canonical axes were Ferreira, & Esteves, 2015; Casatti et al., 2015), significant (Monte Carlo test; P < 0.05). confirming the initial hypothesis of this study The first axis of pRDA separated the forest- in other regions and ecosystems. ed and intermediate stream with predominance The greatest richness, diversity and even- of Knodus heteresthes, Crenicichla lepidota ness values were observed in the forested and Rineloricaria phoxocephala from pasture stream. According to Mendonça (2002) and stream with predominance of Creagrutus anary Naiman, Décamps and Mcclain (2005), the presence of riparian forests along all of the and Imparfinis stictonotus (Fig. 3A). The most streams were responsible for the internal het- important abiotic variables for species compo- erogeneity of the aquatic habitats, interception sition were depth (F = 2.83; P = 0.009), water of sediments, fertilizers and pesticides that velocity (F = 2.25; P = 0.024), pteridophytes would otherwise enter the streams through (F = 2.02; P = 0.040), associated with K. het- surface or subterranean drainage. The riparian eresthes, C. lepidota and R. phoxocephala, and forests also exchanged organic matter between grasses (F = 3.50; P = 0.002), associated with the terrestrial and aquatic ecosystems (Casatti C. anary and I. stictonotus (Fig. 3B). et al., 2012), as a result of which the latter sup- ported a greater diversity of species. A study DISCUSSION with streams in the Adolfo Ducke Reserve in Manaus, found that the composition of the Our variance tests, NMDS and pRDA fish community was influenced by different seems to support the importance of forest to types of substrate, especially trunks and litter improve environmental complexity and spe- banks (Mendonça, 2002), corroborating the cies diversity/composition corroborating with results found for forested stream. The presence previous studies (i.e. Galuch, 2007; Teresa of large amount of submerged litter derived & Casatii, 2010; Cruz et al., 2013; Santos, from leaves that fell from the riparian forest Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 65 (1): 103-115, March 2017 109 Fig. 3. Ordination of the Partial Redundancy Analysis (pRDA) on fish species composition (see abbreviations on Appendice 1) with sites (■ = forested stream, ● = intermediate stream and ▲ = pasture stream) (A) and abiotic variables relationships (arrows) (B). De = Depth, Wv = Water velocity, Pt = Pteridophytes and Gr = Grasses. resulted in greater number of microhabitats, In contrast, the removal of riparian for- which ensured greater fish species richness and ests changes the trophic relationships between diversity (Ferreira & Marinelli, 2001). Even so, organisms (Pusey & Arthington, 2003) and studies on the effects of riparian vegetation on affects the structure and composition of aquatic aquatic communities have stressed the impor- communities (Teresa & Casatti, 2010). The tance of local influence, because riparian buf- amount of grass on the stream banks is the fers help the preservation of mesohabitats and main predictor of functional redundancy, which found for forested stream in our study (Lorion indicates that despite being more numerous & Kennedy 2009; Casatti et al. 2012; Casatti et and diverse, the groups of species living al. 2015; Santos et al. 2015). in this condition perform similar functions 110 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 65 (1): 103-115, March 2017 in the ecosystem (Casatti et al., 2015). This than the pasture stream during the dry season, vegetation comes from adjacent pastures in and no differences were observed for this areas of limited shade from the riparian zone, parameter between the forested and pasture and it is a good indicator of low habitat integ- streams in relation to the intermediate stream. rity (Casatti et al., 2009) and habitat homog- Comparisons made by Santos et al. (2015), enization (Zeni & Casatti, 2014; Casatti et in streams located in area of native forest al., 2015). The occurrence of grass, as pasture (comprised dense ombrophilous forests, low- stream, at the border and in the instream habi- land vegetation, dense woody vegetation on tat, can provide additional microhabitats that both banks, with buffer widths of 12-18 m), favor the colonization, reproduction, and feed- secondary forest (comprised semi-deciduous ing of aquatic insects, playing a similar role to and secondary dense ombrophilous vegeta- that of aquatic macrophytes, which increase tion, were located within sugarcane proper- macroinvertebrate abundance in streams (Casa- ties showing clear signs of deforestation with tti et al., 2015). High macroinvertebrate avail- a relatively open understory and 6-12 m of ability can in turn favor the persistence of riparian buffer zones) and sugarcane (riparian fish populations of the opportunistic feeding. vegetation consisted primarily of Brachiaria Among the submerged roots of grass, large type grasses, severe erosion with the presence amounts of sediment can be trapped with detri- of gullies), suggested that this component had tus that (Casatti et al., 2015), together with high significant effects on some aspects of the fish respiration rates, may reduce dissolved oxygen assemblages, such as species composition and (Bunn, Davies, & Kellaway, 1997). In this type trophic structure. Although, richness, diversity, of environment, midge larvae are abundant and biomass and total density were not influenced represent, along with detritus, the most con- by riparian condition in Santos et al. (2005) sumed resources by local fish (Zeni & Casatti, study, these parameters of the assemblages 2014). Among the species able to live in such were influenced mainly by the dry and rainy conditions, we can cite Creagrutus anary and periods (seasonality). Imparfinis stictonotus. The species C. anary Seasonal flooding in small streams of (Characidae), which has the highest species the Amazon forest are less predictable, with abundance in the pasture stream, swims in hydrological changes more related to the inci- mid-waters, and most of them are small-sized dence of local rainfall (flash system) than the (Buckup, 1999) with high feeding plasticity regular and predictable monomodal pulse of and a wide variety of dietary strategies (Casa- the great rivers (smooth system; Welcomme & tti et al., 2001). Many species of Characidae Halls, 2004). In small streams, local rainfall are opportunistic as the use of food resources events entail an almost immediate response in in environments heavily modified by human increased water flow (Tomasella et al., 2008). activity (Melo, Machado, & Pinto-Silva, 2004; Ceneviva-Bastos & Casatti, 2007). The inci- Thus, the connection to the lateral region of dence of I. stictonotus can be related to feeding the stream can act as an important refuge for tactics (autochthonous invertivore) and the fish during the period of high water, provid- psammophilous characteristics (sand-dwelling) ing shelter regarding the sudden increase in found for some species of the genus (Carvalho, discharge of streams during heavy rain events Zuanon, & Sazima, 2008). Sandy bottoms and disturbances (e.g. reduced availability of are often found in Amazonian streams and invertebrates and increased energy expenditure harbor specialized sand-dwelling fish assem- for maintenance position) (Piccolo, Hughes, & blages composed mostly of catfishes (Carv- Bryant, 2008). In addition, the marginal region alho et al., 2008). flooded potentially serves as an important The forested stream in the wet and dry sea- habitat for feeding and reproduction of many sons had higher species richness and diversity species (Espírito-Santo, 2015). Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 65 (1): 103-115, March 2017 111 Our results from the pRDA were similar Willan Ohara and João Filho for their help with to those of Barros (2008), in studies in streams identification of the taxonomic material. of the Madeira-Purus basins, who observed that the community structure was significantly RESUMEN affected by depth, flow, woody angiosperms and the presence of bare bank/grasses. An Influencia de la estructura del hábitat en el increase in stream dimensions provided greater ensamble de comunidades de peces en ríos de la cuen- environmental heterogeneity and availability ca amazónica Machado. Igarapés son ríos amazónicos extremadamente susceptibles a los impactos ambientales. of habitats, thus increasing the species abun- Debido al impacto de los cambios en el uso de las zonas dance and richness (Vannote, Minshall, Cum- riparias en los ensambles de peces, es importante describir mins, Sedell, & Cushing, 1980). The presence y entender las relaciones entre los ensambles de peces y los of forests in the forested and intermediate factores ambientales, especialmente en áreas poco conoci- stream probably provided a higher availability das. Se estudió la composición taxonómica y los atributos of microhabitats for species associated with del ensamble de peces en ríos bajo diferentes condiciones de conservación de la zona riparia: bosque, zona intermedia trunks (Rineloricaria phoxocephala), backwa- y pastos. Los muestreos se realizaron cada dos meses entre ter areas (Crenicichla lepidota) (Mendonça, Octubre 2011 y Septiembre 2012, en tres ríos de primer 2002) and for species that feed on allochtho- orden en la cuenca del río Machado, Rondonia. Los peces nous resources (Knodus heteresthes) (Nogueira fueron recolectados con redes de cerco y de profundidad & Costa, 2014). In the pasture stream, the (2 mm) a lo largo de todo el tramo. Se registró un total de 2 141 ejemplares pertenecientes a 59 especies, 17 familias presence of grasses generates environmental y 5 órdenes. A diferencia de la zona intermedia y pasto, conditions that favored the occurrence of toler- el bosque mostró una alta riqueza y baja abundancia. El ant species (e.g., S. notomelas, C. anary and B. área boscosa exhibió el valor más alto de diversidad y caudomaculatus). This indicated that changes uniformidad, y baja dominancia, a diferencia de los otros in the fish community followed the sensitive sitios. La varianza particionada y el análisis de redundancia modifications occurring in the vegetation cover parcial (pRDA) indicaron que la composición del ensamble fue significativamente explicado por las variables ambien- of the streams, the presence or absence of trees, tales como: profundidad, velocidad del agua, pteridofitas in each margin, in the same stream. y pastos, pero no por predictores espaciales. El análisis Our results indicated that the stream fish de escalamiento multidimensional no métrico (EMNM) fauna responded to the modification of the mostró que los ríos de sitios intermedios y pastos estaban riparian vegetation through a change in its separados de los ríos del bosque. Llegamos a la conclusión de que las variables ambientales de los ríos y los atributos species composition and ecological parameters del ensamble de peces se ven influenciados por los diferen- of assemblages. The streams with preserved tes niveles de conservación y cobertura del suelo. En vista riparian vegetation represented refuges for fish de la influencia de los procesos regionales en el ensamble fauna from the taxonomic and functional point de peces locales, el uso del suelo a escala de cuenca es of view. With the loss of sensitive and spe- importante, especialmente para explicar la alta riqueza y diversidad que se encuentra en el los ríos del área boscosa. cialized species richness in degraded riparian vegetation and in stream habitat, it was pre- Palabras clave: complejidad, conservación, riqueza, sumed that biotic homogenization occurred, comunidad de peces. which may reduce the species diversity and ecosystem services. REFERENCES ACKNOWLEDGMENTS Barbour, M. T., Gerritsen, J., Snyder, B. D., & Stribling, J. B. (1999). Rapid bioassessment protocols for use in streams and wadeable rivers. Washington: USEPA. We are grateful to Instituto Chico Mendes de Conservação da Biodiversidade – ICMBio Barletta, M., Jaureguizar, A. J., Baigun, C., Fontoura, for Collection permits (no. 311560-1/2011) and N. F., Agostinho, A. A., Almeida-Val, V. M. F., Val, A. L., Torres, R. A., Jimenes-Segura, L. F., Laboratório de Ictiologia e Pesca da Univer- Giarrizzo, T., Fabre´, N. N., Batista, V. S., Lasso, sidade Federal de Rondônia and in particular C., Taphorn, D. C., Costa, M. F., Chaves, P. T., 112 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 65 (1): 103-115, March 2017