Effects of urbanization and environmental heterogeneity on fish assemblages in small streams

Jean C. G. Ortega¹, Iussa Bacani², Tainá F. Dorado-Rodrigues³,  Christine Strüssmann³, Izaias M. Fernandes⁴, Jenny Morales⁵, Lúcia Mateus³, Hugmar Pains da Silva⁶ and Jerry Penha³

PDF: EN    XML: EN | Supplementary:  S1  S2  S3  S4  S5  S6  S7  S8  S9  S10   | Cite this article

Abstract​

Introduction​

The Anthropocene is characterized by an overwhelming global anthropogenic impact that degrades nature and drives increases in species extinction rates (Callisto et al., 2019; Casarim et al., 2020). Urbanization is a process that includes the progressive occupation of the natural landscape by cities, resulting from an increase in human population growth (Seto et al., 2012; Alberti, 2015). This process of occupation includes the removal of native vegetation cover, stream channelization, increases in the amount of impervious surfaces, and the input of untreated sewage, among other habitat disturbances (Booth et al., 2016).

There is a growing body of evidence that the altered physical, chemical, and biological conditions in urban environments affect the integrity of aquatic biota, particularly fish assemblages, by changing the trophic dynamics, diversity, and composition of species assemblages (e.g., Eklöv et al., 1998; Cunico et al., 2006; Ferreira, Casatti, 2006; Felipe, Súarez, 2010; Gebrekiros, 2016; Prado et al., 2020). For example, Meador (2020) observed that fish species loss increased with the proportion of urban land use, due to herbicides and insecticides. Variability in species composition (beta diversity) also tends to be lower in streams under greater urbanization influence, with assemblages dominated by disturbance-tolerant species (Bliss et al., 2017; Bourassa et al., 2017; Meador, 2020). An increase in urbanization often involves the replacement of riparian vegetation by urban structures (Groffman et al., 2003, 2014), which modify physical and chemical factors and reduce habitat availability in aquatic environments, leading to changes in the structure of their biological communities (e.g., Thompson, Parkinson, 2011; Yirigui et al., 2019). This systematic degradation of ecological conditions of streams in urban regions is described as “urban stream syndrome” (Walsh et al., 2005). The effects of urbanization can negatively affect alpha and beta diversity of fish assemblages (Borges et al., 2020). These negative effects may occur because habitat modification and pollution in urban areas filter out more sensitive species from local assemblages, leading to biotic homogenization in the regional pool (McKinney, 2006; Hewitt et al., 2010; Borges et al., 2020).

Urbanization can degrade stream habitats at multiple scales (Engman, Ramírez, 2012). Several studies support the idea that interactions between catchment-scale (physiographic), riparian corridor, and stream-scale environmental variables should be considered to best evaluate the anthropogenic effects on fish assemblages (e.g., Engman, Ramírez, 2012; Marzin et al., 2013; see also Czeglédi et al., 2020, on conflicting results). At small spatial scales, environmental heterogeneity has been demonstrated to be more consistently and strongly correlated with fish assemblage structure (e.g., Engman, Ramírez, 2012 and references therein). There is growing evidence of a positive relationship between environmental heterogeneity and beta diversity, which indicates that as the environmental dissimilarity between sites increases, so does the taxonomic divergence (López-Delgado et al., 2020). Functional and phylogenetic dissimilarities of fish fauna from streams also show this pattern (Roa-Fuentes et al., 2019). This suggests that streams with higher environmental heterogeneity offer a larger variety of suitable environmental conditions for different species (Heino et al., 2014). Positive correlations between environmental heterogeneity and beta diversity have been observed for different aquatic taxa, such as macroinvertebrates, diatoms (Rouquette et al., 2013) and fish (Bourassa et al., 2017), even within an urbanized watershed. Reductions in environmental heterogeneity (environmental homogenization) can increase the dominance of generalist and opportunistic species and enable the invasion of exotic species, leading to the loss of native species (Marchetti et al., 2006; Araújo et al., 2009; Cruz, Pompeu, 2020). Thus, environmental homogenization can reduce biodiversity via impacts on local species richness and beta diversity (Hewitt et al., 2010).

While urbanization and environmental heterogeneity may be related to independent environmental factors, they can be intertwined depending on the spatial scale (Groffman et al., 2014) and the way environmental heterogeneity is measured (Stein, Kreft, 2015). At large spatial scales, urbanization reduces environmental variability, replacing the natural environment with a common urban ecosystem and causing the environment and biota of two disparate regions (even those from different biomes) to reach similar conditions (e.g., McKinney, 2006; Groffman et al., 2014). However, within urban watersheds, there is environmental variation related to the differences in the presence or quantity of urban structures (areas with non-vegetated cover and the predominance of artificial structures, such as streets, roads, highways and buildings; Souza et al., 2020) around streams (e.g., Rouquette et al., 2013; Bourassa et al., 2017). Furthermore, environmental heterogeneity is a broad ecological concept covering several types of variables (for a comprehensive review, see Stein, Kreft, 2015). When urbanization drives local environmental homogenization, chemical variables and some physical variables, particularly those describing habitat structure, are often homogenized. Thus, to disentangle the effects of urbanization and environmental heterogeneity, one would need uncorrelated surrogates of both of these ecological factors.

In this study, we disentangle the effects of urbanization and environmental heterogeneity on stream fish assemblages by using uncorrelated surrogates of variables related to these factors. We tested the hypothesis that increases in urbanization and environmental heterogeneity (EH) would have opposing effects on the structure of fish assemblages. We assessed this hypothesis with the predictions that (i) increases in the proportion of impervious surfaces (a proxy for urbanization) would decrease rarefied fish species richness and local contribution to beta diversity. This may occur because the degradation in stream environmental conditions due to urbanization excludes fish species sensitive to disturbances from local communities (Bliss et al., 2017; Bourassa et al., 2017; Meador, 2020), which would reduce species richness and increase similarity in species composition. Furthermore, (ii) increases in environmental heterogeneity would increase rarefied fish species richness and local contribution to beta diversity. This expectation is justified because habitats with higher environmental heterogeneity provide resources and conditions suitable for a higher number of different species (Tews et al., 2004; Engman, Ramírez, 2012; Heino et al., 2014). We further expected that streams with similar levels of impervious surfaces and EH would have similar fish species compositions.

Material and methods

Study area. The study was carried out in urban streams distributed throughout the 22,851.10 km² area of the city of Cuiabá, situated in central South America (Fig. 1). Cuiabá is one of the oldest Brazilian municipalities, founded in 1719, but only from 1970 to 1990 did it experience exponential population growth, which was not accompanied by urban planning or an increase in wastewater treatment (Brasil, 2019). This city currently has approximately 620,000 inhabitants and a population density of approximately 190 inhabitants per square kilometer. Cuiabá is located in a region with a high density of streams, all tributaries of the Cuiabá River, a tributary of the Paraguay River basin. Vegetation cover is patchy, composed of fragments of Cerrado sensu stricto (a regional savannah vegetation type), tropical dry forests and riparian forests.

Sampling. Data collection was carried out from October 2017 to February 2018 in 45 streams, with one sampling per stream. All the streams are urban and distributed along a gradient of proportions of surrounding impervious surfaces (urban infrastructure) (Hahns, McDonnell, 2006; Meador, 2020; Fig. 1). This urbanization gradient ranged from the downtown area (with the greatest flow of people and a higher proportion of urban structures, e.g., higher proportion of streets and buildings in the landscape) to the peripheral regions of the city (with less urban development and a lower proportion of urban structures).

Each sampling site consisted of a continuous 50 m long reach of a stream (27 first- and 18 second-order streams, following Strahler (1957); Fig. 1). Fish assemblage data were collected over the entire stream reach, while environmental variables were measured at regular intervals (see below).

FIGURE 1 | Location of the 45 streams sampled for fish in urban area within Cuiabá, Mato Grosso state, midwestern Brazil. The gray area indicates impervious surfaces. The color gradient indicates the proportion of impervious surfaces in a 500 m buffer around each stream; the closer to one the value is, the higher the quantity of urban infrastructure around the stream. The continuous lines indicate the hydrography. The dashed lines indicate the political boundary of Cuiabá with the neighboring municipality of Várzea Grande.

Environmental variables. Environmental variables were sampled using the protocol proposed by Mendonça et al. (2005), adapted to urban environments. Stream width and water depth, substrate composition, and canopy cover were used to quantify the local environmental heterogeneity and habitat structure. We measured these local environmental variables because stream fishes often show habitat-specific associations, e.g., higher species richness in sites with a greater proportion of sand, leaves, depth (Kemenes, Forsberg, 2014), and vegetation cover (Cruz, Pompeu, 2020) or species-specific abundance associations with depth and substrate composition (Mendonça et al., 2005).

The width, water depth, and substrate composition were recorded in five transects along each stream (Tab. S1). The water depth, proportion of algae, proportion of woody material (stems and roots), proportion of litter (organic matter, such as leaves and small branches), and substrate composition were estimated from their presence at nine equidistant points along each transect, totaling 45 measurements in each stream reach. The substrate was classified into one of three categories: i) sandy (0.6–2 mm); ii) clay (< 0.6 mm); and iii) gravel and rocks (stones with diameters greater than 2 mm). The contributions of each substrate class, algae, woody material and litter were represented by their relative proportion in each sampling site. The water depth was measured with a graduated ruler.

The canopy cover was measured using five photographs taken in the center of each transect by the same person (IB), approximately 1 m above the water. A cell phone was held in a horizontal position, and the front-facing camera angled directly up was used to take the photos. The photographs were then converted to black and white using GIMP software. From the conversion, a gradation of white, gray, and black tones was obtained, corresponding to pixel values ranging from 0 to 255. The low values (here < 130) represent dark-toned pixels, indicating the presence of canopy coverage. The percentage of canopy cover of each stream section was averaged from the five photographs (for a detailed description, see Arnhold et al., 2019).

We quantified the proportion of impervious surfaces around the stream as a proxy for urbanization (Hahns, McDonnell, 2006; King et al., 2011; Meador, 2020). We used data on land-use and land-cover categories classified by the MapBiomas project (Souza et al., 2020). MapBiomas combines annual Landsat satellite images from 1985 to 2019 (Collection 5), classifying different land-use and land-cover categories in Brazil on a pixel basis (30 × 30 m resolution) with a Random-Forest algorithm (Souza et al., 2020). From the categories available, we used a map with only the urban infrastructure to quantify the extent of urbanization around the Cuiabá streams in 2017. This map had a grid with pixels, with urban infrastructures coded as one (presence) and zero (absence). We built a buffer around the coordinates of each stream, summed the number of pixels inside the buffer that had urban infrastructure, and then divided that number by the buffer area. This process resulted in a simple physical index of urbanization using a relative scale (Hahns, McDonnell, 2006; King et al., 2011), with higher values (closer to one) indicating a greater presence of urban structures around a stream. We calculated the proportions of impervious surfaces with buffers of different radii (100, 200, 300, 400, 500, 800, 1000, 1500, and 2000 m). For inferences related to the estimated proportions of impervious surfaces, we used a buffer of 500 m because the results from all the buffers were highly correlated (Pearson correlation coefficient (r), r > 0.6; Tab. S2), and there was a lower overlap between buffers of this radius than for those with larger sizes (Fig. S3). Additionally, a radius of 500 m resulted in less variation in the estimates of the proportions of impervious surface than smaller buffers (see Fig. S4).

Fish sampling. Fish were sampled using sieves (with 1 mm and 2 mm mesh sizes) and dip nets (1 mm mesh size). Each gear type was used by a different person for 50 minutes along the entire stream section (total sampling effort per section = 100 min). To minimize fish escape and to increase sampling efficiency, we first divided the 50 m section into five subsections of 10 m by using 2.5 mm seine nets and then sampled the subsections sequentially.

The captured fish were euthanized by an anesthetic overdose of clove oil (Fernandes et al., 2017), fixed in 10% formalin solution for five days, and then preserved in 70% ethanol. Fish were identified using regional taxonomic keys (Britski et al., 2007) and subsequently deposited in the fish collection of the Universidade Federal de Mato Grosso, Cuiabá, Mato Grosso, Brazil (CPUFMT 6839 to CPUFMT 6882).

Data analysis. We estimated species diversity as individual-based rarefied species richness (S_rarefied) to the lowest number of individuals sampled (six individuals; no fish were sampled at six sites, and two sites had only one fish; the S_rarefied for these sites was fixed as zero and one, respectively). Rarefaction is a method to control for differences in sampling effort across a set of samples (Gotelli, Colwell, 2011). All 45 streams were used in this analysis.

We estimated beta diversity as the local contribution to beta diversity (LCBD; Legendre, De Cáceres, 2013). LCBD quantifies the relative contribution of each stream to the overall variance in a species composition matrix. In other words, this index expresses the uniqueness of species composition at a given sampling site (Legendre, De Cáceres, 2013). LCBD values vary between 0 and 1, where 0 indicates totally similar assemblages (an assemblage composed of species present in all sampling sites), and 1 indicates totally dissimilar assemblages (a more unique composition; Legendre, De Cáceres, 2013). We estimated LCBD with a Hellinger distance matrix computed with a fish species abundance (columns) by stream (rows) matrix. We excluded from the LCBD computation the six streams where no fish were sampled.

We quantified environmental heterogeneity (EH) as distances to the median in a multivariate space (Anderson et al., 2006; Heino, Grönroos, 2013). We first computed a standardized Euclidean distance matrix among sites using local environmental variables (mean stream width, mean depth, substrate composition, proportion of algae, woody material and litter, and canopy cover). Then, we ordinated the streams with a Principal Coordinate Analysis (PCoA). Finally, we estimated the distance of each stream to the median of the PCoA ordination. Streams with a higher distance to the median (EH henceforth) had a more heterogeneous environment than streams with a smaller distance to the median (the latter were less heterogeneous). It is important to note that EH was computed with stream width, depth, canopy cover and substrate composition; thus, streams with higher EH have a unique combination of these local environmental characteristics compared to the most common combination of these same variables (the multivariate median). Our EH index provides no information regarding the conservation status of the stream (e.g., preserved or degraded) because we did not include any variable describing this status when calculating distances among streams. Furthermore, EH and the proportion of urban structures were not correlated (see below). Thus, degradation had no influence on EH in our study (at least considering a potential measure of degradation around the stream measured by the proportion of urban structures).

We assessed the relationship between S_rarefied or LCBD and the proportion of impervious surfaces and EH using (generalized) linear models. For the model with S_rarefied as the response variable, the proportion of impervious surfaces (a quantitative proxy for urbanization) and EH were the explanatory variables. This model consisted of a multiple regression by ordinary least squares (OLS). To improve the linearity of the relationships, we transformed EH by log_e(x). We assessed the assumption of homogeneity of variance with dispersion plots with residuals and fitted values, the normality of residuals with quantile plots of standardized residuals with fitted values, and the presence of influential observations with a plot with standardized residuals of the leverage function and with Cook’s distance thresholds (Quinn, Keough, 2002; Zuur et al., 2010). We tested for spatial autocorrelation in the residuals with a Mantel correlogram and bubble plot of model residuals (Zuur et al., 2009, 2010; Legendre, Legendre, 2012). The model with S_rarefied as the response variable met linear model assumptions and showed neither influential observations nor spatial autocorrelation (Figs. S5 and S6).

For the model with LCBD as the response variable, the proportion of impervious surfaces and EH were again the explanatory variables. Since LCBD values are bounded between 0 and 1, we used a beta regression (Ferrari, Cribari-Neto, 2004; Douma, Weedon, 2019) to relate LCBD and both explanatory variables. Beta regression is appropriated to model responses for unit intervals, such as rates and proportions, which are typically heteroskedastic (Cribari-Neto, Zeileis, 2010). Also, a link function makes the expected value of the response linear and the expected variances homogeneous (Bolker et al., 2009). For the beta regression, we used a logit link function and reported pseudo-R² as a measure of fit (Ferrari, Cribari-Neto, 2004; Cribari-Neto, Zeileis, 2010). We assessed the assumption of homogeneity of variance with dispersion plots with residuals and fitted values and the presence of influential observations with Cook’s distance (Quinn, Keough, 2002; Zuur et al., 2010). The assumptions of homogeneity of variance and spatial autocorrelation were tested using the same procedures mentioned above. The model with LCBD as the response variable met linear model assumptions and showed neither influential observations nor spatial autocorrelation (Fig. S7 and S8). Multicollinearity was not an issue in either model with S_rarefied or LCBD as response variable because the correlation between the explanatory variables was low and not statistically significant (Pearson correlation coefficient (r) = -0.07, P = 0.627). 

We used distance-based Redundancy Analysis (dbRDA; McArdle, Anderson, 2001) to assess the effects of log-transformed EH and the proportion of impervious surfaces on stream fish species abundances. We used a Hellinger distance matrix (Legendre, Legendre, 2012) computed with fish species abundance (columns) by stream (rows) as the response matrix in the dbRDA. We excluded the six streams with no sampled fish from the dbRDA. We assessed the contribution of each fish species to the dbRDA axes correlating the Hellinger-transformed abundances and stream scores in the dbRDA ordination (“envfit” routine; association significance computed with 9,999 permutations).

We performed all analyses in R software (R Core Development Team, 2020) with the “vegan” (Oksanen et al., 2019), “adespatial” (Dray et al., 2020), ”riverdist” (Tyers, 2020) in the Supplementary material S5. “betareg” (Cribari-Neto, Zeileis, 2010), and “lmtest” (Zeleis, Hothorn, 2002), “riverdist” (Tyers, 2020) packages. We included “riverdist” in the S5. We adopted a 5% significance level for all analyses.

Results​

The proportion of impervious surfaces around the streams varied from 0.002 to 1.00 (mean ± SD = 0.67 ± 0.28). The EH varied from 1.33 to 6.71 (2.73 ± 1.24). Streams with higher EH tended to be deeper and have a higher proportion of clay, litter, or algae. Streams with lower EH tended to be shallow and narrower, to have greater canopy cover, and to have a higher proportion of sand, gravel or rocks (Fig. 2).

FIGURE 2 | Principal Coordinate Analysis (PCoA) of local environmental variables of streams from the urban area of Cuiabá, Mato Grosso, midwestern Brazil. The symbol sizes are proportional to the environmental heterogeneity (EH); PlaMat: plant matter; GraRoc: gravel and rocks; CanCov: canopy cover.

We sampled 6,651 individuals of 31 fish species. Most species were Characiformes (15 spp), with the sixteen remaining species distributed between Siluriformes (10), Cichliformes (three) and Cyprinodontiformes (two). The assemblages were dominated by the exotic invasive poeciliid Poecilia reticulata Peters, 1859(n = 5,091 individuals; 76.54% of total abundance), present at 70.6% of the sites, and by the characids Hemigrammus tridens Eigenmann, 1907 (801 individuals; 12.04%) and Astyanax lacustris (Lütken, 1875) (389 individuals; 5.85%), present at 15.6% and 27.67% of the sites, respectively. Individuals of these three species accounted for 94.44% of the total abundance in the sampled urban streams. We collected only one individual each of the characids Jupiaba acanthogaster (Eigenmann, 1911) and Moenkhausia dichroura (Kner, 1858),the crenuchid Characidium zebra Eigenmann, 1909,and the callichthyid Lepthoplosternum pectorale (Boulenger, 1895).

For local diversity, rarefied species richness (S_rarefied) varied from 0.00 to 4.00 (mean ± SD = 1.68 ± 1.10). The local contribution to beta diversity (LCBD; an index of the uniqueness of species composition) varied from 0.01 to 0.08 (0.03 ± 0.03). There were no strong spatial patterns in the S_rarefied and LCBD distributions because higher and lower values occurred in streams irrespective of their locations (Fig. 3). Streams with higher S_rarefied presented a higher LCBD (Spearman correlation, ρ = 0.73; P < 0.001).

FIGURE 3 | Spatial variation in rarefied species richness (S_rarefied; A) and the local contribution to beta diversity (LCBD; B) in urban streams in Cuiabá (Mato Grosso, Brazil). The circle sizes in the legend indicate the minimum, mean, and maximum values of S_rarefied and the LCBD. The arrow in “A” indicates flow direction.

Multiple regression analysis indicated that variation in the S_rarefied in the urban streams was significantly influenced by the proportions of impervious surfaces (F_{2, 42} = 3.32; P = 0.046) but not affected by EH (P = 0.898). This model had low predictive power, explaining only approximately 10% of the S_rarefied variation (R²_adj = 0.10). The streams surrounded by a higher proportion of impervious surfaces tended to have lower S_rarefied values (Tab. 1; Fig. 4).

FIGURE 4 | Relationship between rarefied species richness (S_rarefied) and the proportion of impervious surfaces in urban streams in Cuiabá. The line indicates fitted values.

TABLE 1 | Relationship between rarefied species richness and environmental heterogeneity (EH) and the proportion of impervious surfaces assessed with an ordinary least squares model. EH consists of distances to a median value estimated from environmental data. The proportion of impervious surfaces was a proxy for urbanization. SE: standard-error.

The beta regression model indicated that the proportion of impervious surfaces explained significantly the variation in the LCBD (Likelihood ratio test, χ² = 13.07; d.f. = 2; P = 0.002). Similar to its influence on S_rarefied, EH did not have an important effect on the LCBD. This model also had low predictive power, explaining approximately 28% of the variation in the LCBD (pseudo-R² = 0.28). The LCBD tended to decline with increases in the proportions of impervious surfaces (Tab. 2; Fig. 5). The results of the two linear models relating the S_rarefied or LCBD to the explanatory variables changed little when we adjusted different buffers applied to estimate the proportion of impervious surfaces (Tabs. S9 and S10).

FIGURE 5 | Relationship between the local contribution to beta diversity (LCBD) and the proportion of impervious surfaces in urban streams in Cuiabá. The line indicates fitted values.

TABLE 2 | Relationship between the local contribution to beta diversity and environmental heterogeneity (EH) and the proportion of impervious surfaces estimated with a beta regression model. EH consists of distances to a median value estimated from environmental data. The proportion of impervious surfaces was a proxy for urbanization. SE: standard-error.

EH and the proportion of impervious surfaces explained the variation in the fish species composition of the urban streams significantly (dbRDA significance test assessed with 9,999 permutations: F_{2, 36} = 3.21; P = 0.011; R²_adj = 0.104). The first and second dbRDA axes explained approximately 15.15% of the distances in species composition between the streams. The proportion of impervious surfaces had a strong and negative association with dbRDA 1, and EH presented a strong and negative association with dbRDA 2. None of the species showed a strong preference for streams with lower or higher EH; however, streams located in regions with higher proportions of impervious surfaces tended to have high abundances of Poecilia reticulata, while Serrapinnus microdon (Eigenmann, 1915), S. calliurus (Boulenger, 1900), Hemigrammus tridens and Astyanax lacustris were more abundant in streams located in areas with lower proportions of impervious surfaces (Fig. 6). On the other hand, Corydoras aeneus (Gill, 1858), Phenacogaster jancupa Malabarba & Lucena, 1995, Astyanax abramis (Jenyns, 1842), Hoplias malabaricus (Bloch, 1794) and Hypostomus khimaera Tencatt, Zawadzki & Froehlich, 2014 tended to be more abundant in streams with lower EH.

FIGURE 6 | Ordination of fish species composition by distance-based redundancy analysis (dbRDA) in relation to environmental heterogeneity (EH; estimated from environmental data and transformed by log_e(x) prior to dbRDA) and the proportion of impervious surfaces (ImpSurf). Species codes, 1: Poecilia reticulata; 2: Corydoras aeneus; 3: Phenacogaster jancupa;4: Astyanax abramis;5: Hoplias malabaricus;6: Hypostomus khimaera;7: Serrapinnus calliurus;8: S. microdon;9: Hemigrammus tridens; 10: A. lacustris.

Discussion​

We set out to investigate whether environmental heterogeneity and urbanization affect fish assemblages in streams in a medium-sized Brazilian city. Our assessment indicates that urbanization had a negative effect and was more important than EH for explaining the spatial variation in S_rarefied and LCBD, at least with the surrogate variables we used. Furthermore, increases in urbanization were related with increases in the abundance of Poecilia reticulata, an introduced fish species dominant in our samples, and decreases in native species occurrences. These results are consistent with growing evidence of the negative effect of urbanization on aquatic fauna worldwide (e.g., Groffman et al., 2014; Borges et al., 2020; Cruz, Pompeu, 2020; Meador, 2020).

The negative effects of urbanization on all the attributes of fish assemblage structure may occur for several reasons. First, impervious surfaces decrease water infiltration and increase surface runoff, reducing flood event intervals in urban streams (White, Greer, 2006). Constant and intense flood events can affect aquatic biodiversity, reducing system productivity, food resources, trophic structure, species distribution and fish assemblage composition (Hakamada, Penha, 2014; Fraley et al., 2018). Second, riparian vegetation in urban regions is normally limited and composed of invasive plant species, thus modifying natural streamflow dynamics due to reduced rates of infiltration and high runoff (Groffman et al., 2003; White, Greer, 2006). Changes in plant species composition and the canopy openness of riparian vegetation in urban areas, compared to those in conserved areas, may also change food resource availability and increase water temperature (Oliveira, Bennemann, 2005; Godinho, 2008). Altered conditions and changes in resource availability can limit colonization by fish species. Irrespective of the specific mechanism, the stressful conditions imposed by urbanization are likely to exclude more sensitive fish species from streams with a greater presence of urban structures via local extinction or by precluding colonization by such species (Hewitt et al., 2010; Bourassa et al., 2017). This form of species exclusion results in a lower number of species. Additionally, the local contribution to beta diversity is reduced when streams with higher levels of urbanization are all occupied by the same set of disturbance-tolerant species (Hewitt et al., 2010; Petsch, 2016; Bourassa et al., 2017).

Water depth, stream width, substrate composition, and canopy cover may constitute a greater variety or diversity of habitats (Bojsen, Barriga, 2002; Peláez, Pavanelli, 2019) and high EH, which provides protection against predation and adverse environmental conditions and may support a larger area for colonization (MacArthur, MacArthur, 1961; Tews et al., 2004; Ortega et al., 2018; Ben-Hur, Kadmon, 2020). Consequently, assemblages at locations with higher EH may be richer because those locations can accommodate the niche requirements of a greater number of species (MacArthur, MacArthur, 1961; Tews et al., 2004; Ortega et al., 2018). EH is also a factor that often explains spatial variation in fish community beta diversity in the natural environment (Peláez, Pavanelli, 2019; Roa-Fuentes et al., 2019); however, EH did not influence the S_rarefied or LCBD and had only a small effect on species composition in our study. It is noteworthy that our proxy for EH included only distances considering the physical and biotic characteristics of the streams, such as substrate composition, width, depth, and vegetation cover. Thus, if an effect of EH on diversity does exist in these streams and was not detected by our study, it is likely that diversity may correlate with the EH measured with other variables, such as chemical (e.g., differences in pH, dissolved oxygen and nutrients) or biotic (e.g., macrophyte cover or richness) variables. For example, Stein, Kreft (2015) observed that 165 different variables were used as a proxy for EH in the ecological literature, and they represented measures of different types, such as biotic, chemical or physical. Another possibility is that diversity may correlate with the EH measured at larger spatial scales than we used. For example, Stein et al. (2014) observed that EH had a pervasive positive effect on species richness at large spatial scales. A confounding factor to consider when assessing the effect of EH on diversity at large spatial scales is the indirect effect of urbanization. At large spatial scales, urbanization tends to homogenize both the environment and the biota (McKinney, 2006; Groffman et al., 2014).

Unfortunately, by expressing EH only with physical environmental characteristics (and not water chemical conditions) we might have underestimated EH’s influence in urban streams. Normally, urban streams are polluted by domestic and industrial sewage, and surface runoff of chemicals is released directly into the ground and water bodies. Untreated domestic and industrial wastewater are rich in matter and organic compounds (Lee, Rasmussen, 2006), the degradation of which by microorganisms consumes much of the dissolved oxygen in the water column (Seitzinger, 1994; Daniel et al., 2002). The resulting decrease in oxygen availability often results in fish die-offs (e.g., Starling et al., 2002; Wepener et al., 2011; Ram et al., 2014). Furthermore, our study encompassed first- and second-order streams (Strahler, 1957); thus, it is likely that the environmental conditions and the biota from this system were naturally homogeneous. In other words, it is likely that we sampled a short gradient of EH that would be important for fish diversity. 

Streams with higher S_rarefied tended to have higher LCBD values. This result has management implications because conserving streams with higher species richness would help to conserve locations with higher contributions to beta diversity. Other studies have found a negative relationship between species richness and LCBD (Legendre, De Cáceres, 2013; Heino et al., 2017). In these cases, management actions to conserve both of these characteristics of diversity would need to maintain a balance between sites with high species richness and those with a high contribution to beta diversity.

The abundance of some fish species was correlated with EH and urbanization. Interestingly, the two most dominant species tended to be correlated with different variables. Astyanax lacustris tended to occur in streams with high EH, those with greater depths and substrates with higher proportions of clay, litter or algae. Astyanax species often have small body sizes and large spatial distributions in many Neotropical basins, occurring in small or large rivers and marginal lagoons (Lima et al., 2003; Mehanna, Penha, 2011; Costa-Pereira et al., 2017). Astyanax lacustris has an omnivorous diet, feeding on algae, seeds, fruits, other plant parts and even invertebrates (Costa-Pereira et al., 2017). Other studies reported species of Astyanax to be indicators of well-conserved streams or of those impacted by pastures and to be absent from urban streams (Casatti et al., 2010); the reproductive activity a close relative species Psalidodon fasciatus (former Astyanax fasciatus) has been shown to be influenced by the degree of pollution in streams (Schulz, Martins-Júnior, 2001). In contrast, Poecilia reticulata, the most abundant species in our sample (present at 70.6% of sites), presented higher abundances in streams located in highly urbanized regions. This species is considered an indicator of degraded aquatic environments (de Carvalho et al., 2017) and is highly invasive, replacing native species in ecosystems with various degrees of contamination from industrial and domestic sewage (Gomes-Silva et al., 2020). Poecilia reticulata can consume insects and debris, food resources commonly available in aquatic environments (Oliveira, Bennemann, 2005; de Carvalho et al., 2019). Other potential traits that may favor the persistence of high abundances of P. reticulata include internal fertilization, livebearing (Magurran, 2005; El-Sabaawi et al., 2016) and broad tolerance to both abiotic conditions (Chervinski, 1984; Araújo et al., 2009) and predation pressure (Magurran, 2005).

In summary, we observed a negative effect of urbanization on the S_rarefied and LCBD of fish in urban streams. Furthermore, EH was associated with the abundance distribution of only a few native fish species. Our results highlight the negative effect of urbanization on fish assemblage structure and show that maintaining high environmental heterogeneity can help native fish species to persist in urban ecosystems. Specifically, we showed that an increase in impervious surfaces around streams reduced fish species richness and community uniqueness. Thus, to maintain native fish assemblages in urban areas, it is important to avoid expanding impermeable surfaces around streams. Additionally, reducing impermeable surfaces around streams located in more urbanized areas seems to be a good strategy to restore fish communities. Such results are likely to become increasingly significant in the near future, given the increasing presence of urban ecosystems in the landscape (Seto et al., 2012).

Acknowledgments​

We thank “Projeto Água para o Futuro” — a multidisciplinary conservation initiative focusing on the assessment, recovery, and protection of streams throughout the Cuiabá urban area — and the Public Ministry of the state of Mato Grosso for supporting field activities. We thank all reviewers and the editor for their constructive review. CS and JP thank the Conselho Nacional de Desenvolvimento Científico e Tecnológico for research fellowships (CNPq #3123038/2018–1 and #307002/2019–5, respectively).

References​

Alberti M. Eco-evolutionary dynamics in an urbanizing planet. Trends Ecol Evol. 2015; 30(2):114–26. https://doi.org/10.1016/j.tree.2014.11.007

Anderson MJ, Ellingsen KE, McArdle BH. Multivariate dispersion as a measure of beta diversity. Ecol Lett. 2006; 9(6):683–93. https://doi.org/10.1111/j.1461-0248.2006.00926.x

Araújo FG, Peixoto MG, Pinto BCT, Teixeira TP. Distribution of guppies Poecilia reticulata (Peters, 1860) and Phalloceros caudimaculatus (Hensel, 1868) along a polluted stretch of the Paraíba do Sul River, Brazil. Braz J Biol. 2009; 69(1):41–48. https://doi.org/10.1590/S1519-69842009000100005

Arnhold TR, Penha J, Peoples BK, Mateus LAF. Positive co-occurrence between feeding-associative savannah fishes depends on species and habitat. Freshw Biol. 2019; 64(5):1029–39. https://doi.org/10.1111/fwb.13283

Ben-Hur E, Kadmon R. Heterogeneity–diversity relationships in sessile organisms: A unified framework. Ecol Lett. 2020; 23(1):193–207. https://doi.org/10.1111/ele.13418

Bliss SM, Lennox RJ, Midwood JD, Cooke SJ. Temporally stable and distinct fish assemblages between stream and earthen stormwater drain reaches in an urban watershed. Urban Ecosyst. 2017; 20(5):1045–55. https://doi.org/10.1007/s11252-017-0663-4

Bojsen BH, Barriga R. Effects of deforestation on fish community structure in Ecuadorian Amazon streams. Freshw Biol. 2002; 47(11):2246–60. https://doi.org/10.1046/j.1365-2427.2002.00956.x

Bolker BM, Brooks ME, Clark CJ, Geange SW, Poulsen JR, Stevens MHH, White J-SS. Generalized linear mixed models: A practical guide for ecology and evolution. Trends Ecol Evol. 2009; 24(3):127–35. https://doi.org/10.1016/j.tree.2008.10.008

Booth DB, Roy AH, Smith B, Capps KA. Global perspectives on the urban stream syndrome. Freshw Sci. 2016; 35(1):412–20. https://doi.org/10.1086/684940

Borges PP, Dias MS, Carvalho FR, Casatti L, Pompeu PS, Cetra M et al. Stream fish metacommunity organisation across a Neotropical ecoregion: The role of environment, anthropogenic impact and dispersal-based processes. PLoS ONE. 2020; 15(5):e0233733. https://doi.org/10.1371/journal.pone.0233733

Bourassa AL, Fraser L, Beisner BB. Benthic macroinvertebrate and fish metacommunity structure in temperate urban streams. J Urban Ecol. 2017; 3(1):jux012. https://doi.org/10.1093/jue/jux012

Brasil. Ministério do Desenvolvimento Regional. Secretaria Nacional de Saneamento – SNS. Sistema Nacional de Informações sobre Saneamento: 24º Diagnóstico dos Serviços de Água e Esgotos – 2018. Brasília: SNS/MDR; 2019. Available from: http://www.snis.gov.br/downloads/diagnosticos/ae/2018/Diagnostico_AE2018.pdf

Britski HA, Silimon KZS, Lopes BS. Peixes do Pantanal: Manual de identificação. 2nd ed. Brasília: Embrapa Informação Tecnológica; 2007.

Callisto M, Solar R, Silveira FAO, Saito VS, Hughes RM, Fernandes GW et al. A Humboldtian approach to mountain conservation and freshwater ecosystem services. Front Environ Sci. 2019; 7:195. https://doi.org/10.3389/fenvs.2019.00195

de Carvalho DR, Leal CG, Junqueira NT, de Castro MA, Fagundes DC, Alves CBM et al. A fish-based multimetric index for Brazilian savanna streams. Ecol Indic. 2017; 77:386–96. https://doi.org/10.1016/j.ecolind.2017.02.032

de Carvalho DR, Flecker AS, Alves CBM, Sparks JP, Pompeu PS. Trophic responses to aquatic pollution of native and exotic livebearer fishes. Sci Total Environ. 2019; 681:503–15. https://doi.org/10.1016/j.scitotenv.2019.05.092

Casarim R, Caldeira YM, Pompeu PS. Representativeness of national parks in protecting freshwater biodiversity: A case of Brazilian savanna. Ecol Freshw Fish. 2020; 29(4):705–21. https://doi.org/10.1111/eff.12547

Casatti L, Romero RM, Teresa FB, Sabino J, Langeani F. Fish community structure along a conservation gradient in Bodoquena Plateau streams, central West of Brazil. Acta Limnol Bras. 2010; 22(1):50–59. https://doi.org/10.4322/actalb.02201007

Chervinski J. Salinity tolerance of the guppy, Poecilia reticulata Peters. J Fish Biol. 1984; 24(4):449–52. http://doi.org/10.1111/j.1095-8649.1984.tb04815.x

Costa-Pereira R, Tavares LER, de Camargo PB, Araújo MS. Seasonal population and individual niche dynamics in a tetra fish in the Pantanal wetlands. Biotropica. 2017; 49:531–38. https://doi.org/10.1111/btp.12434

Cribari-Neto F, Zeileis A. Beta regression in R. J Stat Softw. 2010; 34(2):1–24. https://doi.org/10.18637/jss.v034.i02

Cruz LC, Pompeu PS. Drivers of fish assemblage structures in a Neotropical urban watershed. Urban Ecosyst. 2020; 23(4):819–29. https://doi.org/10.1007/s11252-020-00968-6

Cunico AM, Agostinho AA, Latini JD. Influência da urbanização sobre as assembléias de peixes em três córregos de Maringá, Paraná. Rev Bras Zool. 2006; 23(4):1101–10. http://doi.org/10.1590/S0101-81752006000400018

Czeglédi I, Kern B, Tóth R, Seress G, Erõs T. Impacts of urbanization on stream fish assemblages: The role of the species pool and the local environment. Front Ecol Evol. 2020; 8:137. https://doi.org/10.3389/fevo.2020.00137

Daniel MHB, Montebelo AA, Bernardes MC, Ometto JPHB, de Camargo PB, Krusche AV et al. Effects of urban sewage on dissolved oxygen, dissolved inorganic and organic carbon, and electrical conductivity of small streams along a gradient of urbanization in the Piracicaba River basin. Water Air Soil Pollut. 2002; 136(1):189–206. https://doi.org/10.1023/A:1015287708170

Douma JC, Weedon JT. Analysing continuous proportions in ecology and evolution: A practical introduction to beta and Dirichlet regression. Methods Ecol Evol. 2019; 10(9):1412–30. https://doi.org/10.1111/2041-210X.13234

Dray S, Bauman D, Blanchet G, Borcard D, Clappe S, Guenard G et al. Adespatial: Multivariate Multiscale Spatial Analysis. Version 0.3-8 [Internet]. 2020. Available from: https://CRAN.R-project.org/package=adespatial

Eklöv AG, Greenberg LA, Brönmark C, Larsson P, Berglund O. Response of stream fish to improved water quality: A comparison between the 1960s and 1990s. Freshw Biol. 1998; 40(4):771–82. https://doi.org/10.1046/j.1365-2427.1998.00370.x

El-Sabaawi RW, Frauendorf TC, Marques PS, Mackenzie RA, Manna LR, Mazzoni R et al. Biodiversity and ecosystem risks arising from using guppies to control mosquitoes. Biol Lett. 2016; 12:20160590. https://doi.org/10.1098/rsbl.2016.0590

Engman AC, Ramírez A. Fish assemblage structure in urban streams of Puerto Rico: The importance of reach- and catchment-scale abiotic factors. Hydrobiologia. 2012; 693(1):141–55. https://doi.org/10.1007/s10750-012-1100-6

Felipe TRA, Súarez YR. Caracterização e influência dos fatores ambientais nas assembléias de peixes de riachos em duas microbacias urbanas, alto rio Paraná. Biota Neotrop. 2010; 10(2):143–51. https://doi.org/10.1590/S1676-06032010000200018

Fernandes IM, Bastos YF, Barreto DS, Lourenço LS, Penha JM. The efficacy of clove oil as an anaesthetic and in euthanasia procedure for small-sized tropical fishes. Braz J Biol. 2017; 77(3):444–50. https://doi.org/10.1590/1519-6984.15015

Ferrari S, Cribari-Neto F. Beta regression for modelling rates and proportions. J Appl Stat. 2004; 31(7):799–815. https://doi.org/10.1080/0266476042000214501

Ferreira CP, Casatti L. Integridade biótica de um córrego na bacia do alto rio Paraná avaliada por meio da comunidade de peixes. Biota Neotrop. 2006; 6(3):bn00306032006. https://doi.org/10.1590/S1676-06032006000300002

Fraley KM, Warburton HJ, Jellyman PG, Kelly D, McIntosh AR. Responsiveness of fish mass–abundance relationships and trophic metrics to flood disturbance, stream size, land cover and predator taxa presence in headwater streams. Ecol Freshw Fish. 2018; 27(4):999–1014. https://doi.org/10.1111/eff.12410

Gebrekiros ST. Factors affecting stream fish community composition and habitat suitability. J Aquac Mar Biol. 2016; 4(2):00076. https://doi.org/10.15406/jamb.2016.04.00076

Godinho FN. The influence of riparian vegetation on freshwater fish. In: Arizpe D, Mendes A, Rabaça JE, editors. Sustainable riparian zones: A management guide. Generalitat Valenciana; 2008. p.96–100.

Gomes-Silva G, Pereira BB, Liu K, Chen B, Santos VSV, de Menezes GHT et al. Using native and invasive livebearing fishes (Poeciliidae, Teleostei) for the integrated biological assessment of pollution in urban streams. Sci Total Environ. 2020; 698:134336. https://doi.org/10.1016/j.scitotenv.2019.134336

Gotelli NJ, Colwell RK. Estimating species richness. In: Magurran AE, McGill BJ, editors. Biological diversity: Frontiers in measurement and assessment. Oxford: Oxford University Press; 2011. p. 39–54.

Groffman PM, Bain DJ, Band LE, Belt KT, Brush GS, Grove JM et al. Down by the riverside: Urban riparian ecology. Front Ecol Environ. 2003; 1(6):315–21. https://doi.org/10.2307/3868092

Groffman PM, Cavender-Bares J, Bettez ND, Grove JM, Hall SJ, Heffernan JB et al. Ecological homogenization of urban USA. Front Ecol Environ. 2014; 12(1):74–81. https://doi.org/10.1890/120374

Hahns AK, McDonnell MJ. Selecting independent measures to quantify Melbourne’s urban–rural gradient. Landsc Urban Plan. 2006; 78(4):435–48. https://doi.org/10.1016/j.landurbplan.2005.12.005

Hakamada KYP, Penha J. Occupancy dynamics in small-stream habitats: Niches define the responses to floods by two neotropical fishes. Popul Ecol. 2014; 56(1):139–50. https://doi.org/10.1007/s10144-013-0395-0

Heino J, Bini LM, Andersson J, Bergsten J, Bjelke U, Johansson F. Unravelling the correlates of species richness and ecological uniqueness in a metacommunity of urban pond insects. Ecol Indic. 2017; 73:422–31. https://doi.org/10.1016/j.ecolind.2016.10.006

Heino J, Grönroos M. Does environmental heterogeneity affect species co-occurrence in ecological guilds across stream macroinvertebrate metacommunities? Ecography. 2013; 36(8):926–36. https://doi.org/10.1111/j.1600-0587.2012.00057.x

Heino J, Melo AS, Bini LM. Reconceptualising the beta diversity-environmental heterogeneity relationship in running water systems. Freshw Biol. 2014; 60(2):223–35. https://doi.org/10.1111/fwb.12502

Hewitt J, Thrush S, Lohrer A, Townsend M. A latent threat to biodiversity: Consequences of small-scale heterogeneity loss. Biodivers Conserv. 2010; 19(5):1315–23. https://doi.org/10.1007/s10531-009-9763-7

Kemenes A, Forsberg BR. Factors influencing the structure and spatial distribution of fishes in the headwater streams of the Jaú River in the Brazilian Amazon. Braz J Biol. 2014; 74(3 Suppl. 1):S23–32. https://doi.org/10.1590/1519-6984.06812

King RS, Baker ME, Kazyak PF, Weller DE. How novel is too novel? Stream community thresholds at exceptionally low levels of catchment urbanization. Ecol Appl. 2011; 21(5):1659–78. https://doi.org/10.1890/10-1357.1

Lee CJ, Rasmussen TJ. Occurrence of organic wastewater compounds in effluent-dominated streams in Northeastern Kansas. Sci Total Environ. 2006; 371(1–3):258–69. https://doi.org/10.1016/j.scitotenv.2006.07.023

Legendre P, De Cáceres M. Beta diversity as the variance of community data: Dissimilarity coefficients and partitioning. Ecol Lett. 2013; 16(8):951–63. https://doi.org/10.1111/ele.12141

Legendre P, Legendre L. Numerical ecology. developments in environmental modelling, Vol. 24. 3rd ed. Amsterdam: Elsevier; 2012.

Lima FCT, Malabarba LR, Buckup PA, da Silva JFP, Vari RP, Harold A et al. Genera incertae sedis in Characidae. In: Reis RE, Kullander SO, Ferraris CJ, Jr., organizers. Check list of the freshwater fishes of South and Central America. Porto Alegre: Edipucrs; 2003. p.106–69.

López-Delgado EO, Winemiller KO, Villa-Navarro FA. Local environmental factors influence beta-diversity patterns of tropical fish assemblages more than spatial factors. Ecology. 2020; 101(2):e02940. https://doi.org/10.1002/ecy.2940

MacArthur RH, MacArthur JW. On bird species diversity. Ecology. 1961; 42(3):594–98. https://doi.org/10.2307/1932254

Magurran AE. Evolutionary Ecology: The Trinidadian guppy. Oxford: Oxford University Press; 2005.

Marchetti MP, Lockwood JL, Light T. Effects of urbanization on California’s fish diversity: Differentiation, homogenization and the influence of spatial scale. Biol Conserv. 2006; 127(3):310–18. https://doi.org/10.1016/j.biocon.2005.04.025

Marzin A, Verdonschot PFM, Pont D. The relative influence of catchment, riparian corridor, and reach-scale anthropogenic pressures on fish and macroinvertebrate assemblages in French rivers. Hydrobiologia. 2013; 704(1):375–88. https://doi.org/10.1007/s10750-012-1254-2

McArdle BH, Anderson MJ. Fitting multivariate models to community data: A comment on distance-based redundancy analysis. Ecology. 2001; 82(1):290–97. https://doi.org/10.1890/0012-9658(2001)082[0290:FMMTCD]2.0.CO;2

McKinney ML. Urbanization as a major cause of biotic homogenization. Biol Conserv. 2006; 127(3):247–60. https://doi.org/10.1016/j.biocon.2005.09.005

Meador MR. Historical changes in fish communities in urban streams of the south-eastern United States and the relative importance of water-quality stressors. Ecol Freshw Fish. 2020; 29(1):156–69. https://doi.org/10.1111/eff.12503

Mehanna MN, Penha J. Fatores abióticos afetam a distribuição do gênero Astyanax Baird & Girard, 1854 em riachos de cabeceiras de Chapada dos Guimarães, bacia do Rio Cuiabá, Mato Grosso. Bioscience Journal. 2011; 27(1):125–37. 

Mendonça FP, Magnusson WE, Zuanon J. Relationships between habitat characteristics and fish assemblages in small streams of Central Amazonia. Copeia. 2005; 2005(4):751–64. https://doi.org/10.1643/0045-8511(2005)005[0751:RBHCAF]2.0.CO;2

Oksanen J, Blanchet FG, Friendly M, Kindt R, Legendre P, McGlinn D et al. vegan: Community ecology package. Version 2.5-6 [Internet]. 2019. Available from: https://cran.r-project.org/web/packages/vegan/index.html

Oliveira DC, Bennemann ST. Ictiofauna, recursos alimentares e relações com as interferências antrópicas em um riacho urbano no sul do Brasil. Biota Neotrop. 2005; 5(1):95–107. https://doi.org/10.1590/S1676-06032005000100011

Ortega JCG, Thomaz SM, Bini LM. Experiments reveal that environmental heterogeneity increases species richness, but they are rarely designed to detect the underlying mechanisms. Oecologia. 2018; 188(1):11–22. https://doi.org/10.1007/s00442-018-4150-2

Peláez O, Pavanelli CS. Environmental heterogeneity and dispersal limitation explain different aspects of β-diversity in Neotropical fish assemblages. Freshw Biol. 2019; 64(3):497–505. https://doi.org/10.1111/fwb.13237

Petsch DK. Causes and consequences of biotic homogenization in freshwater ecosystems. Int Rev Hydrobiol. 2016; 101(3–4):113–22. https://doi.org/10.1002/iroh.201601850

Prado MR, de Carvalho DR, Alves CBM, Moreira MZ, Pompeu PS. Convergent responses of fish belonging to different feeding guilds to sewage pollution. Neotrop Ichthyol. 2020; 18(1):e190045. https://doi.org/10.1590/1982-0224-2019-0045

Quinn GP, Keough MJ. Experimental design and data analysis for biologists. Cambridge: Cambridge University Press; 2002.

R Core Development Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2020. Available from: https://www.R-project.org

Ram A, Jaiswar JRM, Rokade MA, Bharti S, Vishwasrao C, Majithiya D. Nutrients, hypoxia and mass fishkill events in Tapi Estuary, India. Estuar Coast Shelf Sci. 2014; 148:48–58. https://doi.org/10.1016/j.ecss.2014.06.013

Roa-Fuentes CA, Heino J, Cianciaruso MV, Ferraz S, Zeni JO, Casatti L. Taxonomic, functional, and phylogenetic â-diversity patterns of stream fish assemblages in tropical agroecosystems. Freshw Biol. 2019; 64(3):447–60. https://doi.org/10.1111/fwb.13233

Rouquette JR, Dallimer M, Armsworth PR, Gaston KJ, Maltby L, Warren PH. Species turnover and geographic distance in an urban river network. Div Distrib. 2013; 19(11):1429–39. https://doi.org/10.1111/ddi.12120

Schulz UH, Martins-Júnior H. Astyanax fasciatus as bioindicator of water pollution of rio dos Sinos, RS, Brazil. Braz J Biol. 2001; 61(4):615–22. https://doi.org/10.1590/S1519-69842001000400010

Seitzinger SP. Linkages between organic-matter mineralization and denitrification in eight riparian wetlands. Biochemistry. 1994; 25(1):19–39. https://doi.org/10.1007/BF00000510

Seto KC, Güneralp B, Hutyra LR. Global forecasts of urban expansion to 2030 and direct impacts on biodiversity and carbon pools. Proc Natl Acad Sci USA. 2012; 109(40):16083–88. https://doi.org/10.1073/pnas.1211658109

Souza CM, Jr., Shimbo JZ, Rosa MR, Parente LL, Alencar AA, Rudorff BFT et al. Reconstructing three decades of land use and land cover changes in Brazilian biomes with Landsat archive and Earth Engine. Remote Sens. 2020; 12(17):2735. https://doi.org/10.3390/rs12172735

Starling F, Lazzaro X, Cavalcanti C, Moreira R. Contribution of omnivorous tilapia to eutrophication of a shallow tropical reservoir: Evidence from a fish fill. Freshw Biol. 2002; 47(12):2443–52. https://doi.org/10.1046/j.1365-2427.2002.01013.x

Stein A, Gerstner K, Kreft H. Environmental heterogeneity as a universal driver of species richness across taxa, biomes and spatial scales. Ecol Lett. 2014; 17(7):866–80. https://doi.org/10.1111/ele.12277

Stein A, Kreft H. Terminology and quantification of environmental heterogeneity in species-richness research. Biol Rev Camb Philos Soc. 2015; 90(3):815–36. https://doi.org/10.1111/brv.12135

Strahler AN. Quantitative analysis of watershed geomorphology. Trans Am Geophys Union. 1957; 38(6):913–20. https://doi.org/10.1029/tr038i006p00913

Tews J, Brose U, Grimm V, Tielbörger K, Wichmann MC, Schwager M, Jeltsch F. Animal species diversity driven by habitat heterogeneity/diversity: The importance of keystone structures. J Biogeogr. 2004; 31(1):79–92. https://doi.org/10.1046/j.0305-0270.2003.00994.x

Thompson R, Parkinson S. Assessing the local effects of riparian restoration on urban streams. N Z J Mar Freshwater Res. 2011; 45(4):625–36. https://doi.org/10.1080/00288330.2011.569988

Tyers M. Riverdist: River Network Distance Computation and Applications. R package version 0.15.1. Available from: https://CRAN.R-project.org/package=riverdist; 2020.

Walsh CJ, Roy AH, Feminella JW, Cottingham PD, Groffman PM, Morgan RP, II. The urban stream syndrome: Current knowledge and the search for a cure. J North Am Benthol Soc. 2005; 24(3):706–23. https://doi.org/10.1899/04-028.1

Wepener V, van Dyk C, Bervoets L, O’Brien G, Covaci A, Cloete Y. An assessment of the influence of multiple stressors on the Vaal River, South Africa. Phys Chem Earth. 2011; 36(14–15):949–62. https://doi.org/10.1016/j.pce.2011.07.075

White MD, Greer KA. The effects of watershed urbanization on the stream hydrology and riparian vegetation of Los Peñasquitos Creek, California. Landsc Urban Plan. 2006; 74(2):125–38. https://doi.org/10.1016/j.landurbplan.2004.11.015

Yirigui Y, Lee S-W, Nejadhashemi AP, Herman MR, Lee J-W. Relationships between riparian forest fragmentation and biological indicators of streams. Sustainability. 2019; 11(10):2870. https://doi.org/10.3390/su11102870

Zeleis A, Hothorn T. Diagnostic checking in regression relationships. R J. 2002; 2(3):7–10. https://CRAN.R-project.org/doc/Rnews

Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM. Mixed effects models and extensions in Ecology with R. New York: Springer; 2009. https://doi.org/10.1007/978-0-387-87458-6

Zuur AF, Ieno EN, Elphick CS. A protocol for data exploration to avoid common statistical problems. Methods Ecol Evol. 2010; 1(1):3–14. https://doi.org/10.1111/j.2041-210X.2009.00001.x

Authors

Jean C. G. Ortega¹ , Iussa Bacani², Tainá F. Dorado-Rodrigues³,  Christine Strüssmann³, Izaias M. Fernandes⁴, Jenny Morales⁵ , Lúcia Mateus³, Hugmar Pains da Silva⁶ and Jerry Penha³ .

^[1]    Programa de Pós-Graduação em Ecologia e Manejo de Recursos Naturais, Universidade Federal do Acre, BR-364, Km 4, Campus Universitário, 69915-900 Rio Branco, AC, Brazil. ortegajean@gmail.com (corresponding author).

^[2]    Faculdade de Medicina Veterinária, Universidade Federal de Mato Grosso, Av. Fernando Correia da Costa 2367, 78060-900 Cuiabá, MT, Brazil. iussa.bacani2014@gmail.com.

^[3]    Centro de Biodiversidade, Instituto de Biociências, Universidade Federal de Mato Grosso, Av. Fernando Correia da Costa 2367, 78060-900 Cuiabá, MT, Brazil. (TFDR) tainadorado@gmail.com, (CS) christrussmann@gmail.com, (LM) lafmateus@gmail.com, (JP) jpenha.bio@gmail.com.

^[4]    Laboratório de Biodiversidade e Conservação, Universidade Federal de Rondônia, Av. Norte Sul, 7300, 76940-000 Rolim de Moura, RO, Brazil. biomedice@gmail.com.

^[5]    Programa de Pós-Graduação em Ecologia e Conservação da Biodiversidade, Instituto de Biociências, Universidade Federal de Mato Grosso, Av. Fernando Correia da Costa 2367, 78060-900 Cuiabá, MT, Brazil. jennymoralesp@gmail.com.

^[6]    Laboratório de Citogenética e Genética Animal, Instituto de Biociências, Universidade Federal de Mato Grosso, Av. Fernando Correia da Costa 2367, 78060-900 Cuiabá, MT, Brazil. painsbio@gmail.com.

Authors Contribution

Jean C. G. Ortega: Conceptualization, Formal analysis, Writing-original draft, Writing-review and editing.

Iussa Bacani: Data curation, Investigation, Writing-original draft, Writing-review and editing.

Tainá F. Dorado-Rodrigues: Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing-original draft, Writing-review and editing.

Christine Strüssmann: Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Writing-original draft, Writing-review and editing.

Izaias M. Fernandes: Conceptualization, Formal analysis, Methodology, Writing-original draft, Writing-review and editing.

Jenny Morales: Data curation, Writing-original draft, Writing-review and editing.

Lúcia Mateus: Conceptualization, Methodology, Writing-original draft, Writing-review and editing.

Hugmar Pains da Silva: Data curation, Investigation, Writing-original draft.

Jerry Penha: Conceptualization, Formal analysis, Investigation, Methodology, Supervision, Writing-original draft, Writing-review and editing.

Ethical Statement​

Specimens were collected under the legal authority of permits issued by the Chico Mendes Institute for Biodiversity Conservation (License number 11729–2 to JP).

Competing Interests

The authors declare no competing interests. 

How to cite this article

Ortega JCG, Bacani I, Dorado-Rodrigues TF, Strüssmann C, Fernandes IM, Morales J, Mateus L, Silva HP, Penha J. Effects of urbanization and environmental heterogeneity on fish assemblages in small streams. Neotrop Ichthyol. 2021; 19(3):e210050. https://doi.org/10.1590/1982-0224-2021-0050

Copyright​

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Distributed under

Creative Commons CC-BY 4.0

© 2021 The Authors.

Diversity and Distributions Published by SBI