Highway impact on fish assemblages in Atlantic Forest stream

Thais A. Soinski¹, Mauricio Cetra² and Welber S. Smith^1,3,4

PDF: EN    XML: EN | Supplementary: S1 | Cite this article

Associate Editor: Fernando Carvalho

Section Editor: Fernando Pelicice

Editor-in-chief: José Birindelli

Abstract​

---

Introduction​

---

Highway construction and its environmental impacts are a global concern, drawing significant attention from researchers in various countries, including the United States (Southerland, 1994; Angermeier et al., 2004; Hedrick et al., 2007; Benton et al., 2008; Bouska, Paukert, 2010; Sliger et al., 2024), Canada (Pépino et al., 2012), Chile (Leal et al., 2023), China (Li et al., 2022), and Brazil (Brejão et al., 2020; Azevedo-Santos et al., 2022; De Fries et al., 2023).

It is widely recognized that highway construction and associated interventions generate significant impacts on river ecosystems (Wheeler et al., 2005). These impacts include: (i) geomorphological changes in the river channel upstream and downstream of crossings (Benton et al., 2008); (ii) increased erosion and sedimentation from impervious surface runoff (Hedrick et al., 2007; Ochs et al., 2024); (iii) impairment of fish movement, as roads can act as physical barriers or alter flows (Benton et al., 2008); and (iv) alteration of habitat structure and availability, leading to habitat fragmentation (Bouska, Paukert, 2010).

Trombulak, Frissell (2000) previously highlighted the decline in river health due to highways, reporting that fish abundance decreases in response to higher road densities and more numerous intersections. Further consequences of these new conditions can include reductions in species abundance and diversity, loss of genetic diversity, and even the local extinction of fish species (Bouska, Paukert, 2010).

The removal of riparian vegetation for road construction also leads to significant habitat loss, as the expansion of these cleared areas degrades the surrounding landscape (Ochs et al., 2024). Furthermore, roads can impact previously intact ecosystems by creating new habitat edges along their length, a process that degrades adjacent areas within what is known as the “road-effect zone” (Eigenbrod et al., 2009; Forman, 2000; Forman, Deblinger, 2000). For example, Leal et al. (2023) demonstrated that anthropogenic interventions continually transform the natural characteristics of landscapes along waterways, particularly through the clearing of native vegetation adjacent to roads.

Most research on this topic focuses on the effects of road crossings on rivers and fish movement (Benton et al., 2008; Bouska, Paukert, 2010) or the resulting ecosystem fragmentation that reduces connectivity (Sliger et al., 2024). This trend holds true in Brazil, where studies have concentrated on similar impacts (e.g., Brejão et al., 2020; Azevedo-Santos et al., 2022; De Fries et al., 2023). However, empirical studies employing a before-and-after design to assess the impacts of road construction on streams and their fish fauna are scarce globally, and many existing works are literature reviews (e.g., Angermeier et al., 2004). Such studies are even rarer in the Neotropical region, which underscores the relevance of the present work.

Notably, this study assesses the effects of a highway constructed along a stream, in contrast to most research in Brazil that evaluates the impacts of road crossings. Therefore, our approach considers the “road-effect zone” -the area where ecological impacts extend laterally from the road itself to affect adjacent habitats, in this case, the riparian forest and the stream channel (Forman, Alexander, 1998; Forman, 2000; Forman, Deblinger, 2000).

A clearer understanding of the impacts of highway construction on fish fauna can support environmental agencies in licensing and monitoring these projects, as well as in developing mechanisms to avoid or mitigate their effects. Such projects also present valuable opportunities for the scientific community to learn and provide data that enables governmental agencies and developers to implement future projects with reduced impacts on aquatic ecosystems. Therefore, this study aimed to test the following hypotheses: (1) highway construction negatively affects species richness and alters the composition of the fish assemblage; (2) fine sediment deposition in the stream increases during the construction phase compared to the pre-construction period; (3) key environmental variables altered by the construction, such as sedimentation and water temperature, are significant predictors of changes in the abundance of specific fish species.

Material and methods

---

Study area. This study was conducted in a first-order stream (sensu Strahler, 1957) located in the Serra do Cafezal, an Atlantic Forest mountainous region in the state of São Paulo, Brazil. The study area is situated along the Regis Bittencourt Highway (BR-116), between kilometer markers 336 and 369, a section that traverses the rural municipalities of Juquitiba and Miracatu, São Paulo State (Fig. 1). The stream’s banks are lined with secondary Dense Ombrophilous Forest, which exhibits various stages of natural regeneration (pioneer, initial, medium, and advanced).

FIGURE 1| Geographic location of the studied Atlantic Forest stream along the Regis Bittencourt Highway (BR-116) in the Serra do Cafezal, São Paulo State, Brazil and the four sampled sections (R1–R4).

Our sampling design was specifically structured to evaluate the highway duplication’s effects. We conducted 12 sampling campaigns in this stream from 2011 to 2016. Material collection commenced two years before any site alterations (2011 – July 2012), with these initial four campaigns establishing a pristine baseline. Data acquisition then ceased one year prior to the work’s completion (August 2012 – December 2016), with the subsequent eight campaigns focusing on the direct impact phase of the interventions and acknowledging subsequent, unmonitored changes. To account for seasonal variations, sampling was consistently performed in both July (dry season) and January (rainy season) of each year. Samples were collected from four sections (R1, R2, R3, and R4) within the studied stream. These sections were strategically chosen to align with the highway’s duplication zones, specifically from km 349 to 354 and km 357 to 362 (Tab. 1).

TABLE 1 | Geographic location and characterization of the collection points along the stream in the duplication of Regis Bittencourt Highway (BR-116) (stretch from km 344 + 000 to km 363 + 000), including location of each sampling point.

Stretch	Geographical coordinates	Location	Environmental characterization
r1	47°11’07.1”W 24°02’07.6”S	Km 346, south lane	Stretch with presence of dense ciliary forest, narrow bed, presence of rapids and signs of silting.
r2	47°11’52.5”W 24°02’48.4”S	Km 349, south lane	Stretch characterized by high grasso on the banks, narrow bed, presence of boulders and rapids.
r3	47°13’52.8”W 24°03’47.5”S	Km 353 + 100, south lane	Stretch with presence of dense ciliary forest, a larger bed, a section with few rapds, silted and eroded margins.
r4	47°14’33.3”W 24°04’12.4”S	Km 355 + 800, north lane	Stretch with less dense ciliary forest, broader bed, characterized by rapids, boulders and signs of silting.

Environmental characterization of the stream. Prior to fish sampling at each site, an environmental characterization was performed. The following physicochemical water parameters were measured in situ using a YSI 6600V2 multiparameter probe: dissolved oxygen (mg L⁻¹), pH, water temperature (°C), air temperature (°C), turbidity (NTU), total dissolved solids (mg L⁻¹), and electrical conductivity (µS cm⁻¹).

To characterize stream habitats, the following variables were also recorded (summarized in Tab. 2): Canopy Cover: Estimated visually on an ordinal scale to represent the degree of shading over the stream channel (e.g., 0 = no cover, 0.5 = partial cover, 1 = full cover); Channel Width (m) and Average Depth (cm): Measured directly with a tape measure; Water Velocity: Assessed using an ordinal scale: 0 = no perceptible flow; 0.5 = slow/medium flow; 1 = rapid flow; Fine Sediment Deposition: Assessed visually using a five-point ordinal scale: 0 = no fine sediment; 0.5 = low deposition; 1 = moderate deposition; 1.5 = high deposition; 2 = completely silted stream bed.

TABLE 2 | Average of abiotic data, with respective standard deviations, conductivity (µS.cm-1), pH, turbidity (UTN), dissolved oxygen (mg.L-1), total dissolved solids (mg.L-1), shade degrees (%), water temperature (°C), air temperature (°C), width (m), depth (cm), current (m/s) and fine sediment (mm).

Variables	Stretches sample
	R1	R2	R3	R4
Conductivity	19.92 ± 7.31	23.70 ± 9.02	17.63 ± 8.53	21.82 ± 8.21
pH	6.81 ± 0.53	6.94 ± 0.33	6.66 ± 0.49	6.51 ± 0.26
Turbidity	2.75 ± 1.25	9.41 ± 5.23	2.5 ± 1.56	3.33 ± 2.14
Dissolved oxygen	6.15 ± 0.62	5.91 ± 0.69	6.03 ± 0.77	6.18 ± 0.59
Total dissolved solids	10.34 ± 3.65	12.35 ± 3.92	9.11 ± 4.21	9.50 ± 4.68
Shade degrees	0.75 ± 0	0.23 ± 0.07	0.24 ± 0.11	0.70 ± 0.09
Water temperature	19.18 ± 1.79	20.32 ± 2.44	21.02 ± 2.55	19.35 ± 3.16
Air temperature	22.16 ± 2.03	23 ± 2.41	22.92 ± 2.07	22.54 ± 2.44
Width	2.34 ± 0.36	1.9 ± 0.32	4.38 ± 0.66	4.49 ± 1.0
Depth	0.27 ± 0.06	0.22 ± 0.03	0.26 ± 0.04	0.31 ± 0.04
Current	0.91 ± 0.28	1 ± 0	0.62 ± 0.22	1 ± 0
Fine sediment	0.37 ± 0.31	0.25 ± 0.26	1.04 ± 0.86	0.37 ± 0.22

Ichthyofauna sampling. Fish were sampled using two methods. First, a rectangular dip net (100 cm × 70 cm; 1.0 cm mesh size) was used over three consecutive days and nights in each sampling reach. A standardized effort of 12 sweeps during the day and 12 sweeps at night was applied, resulting in a total of 72 sweeps per reach per campaign. Second, electrofishing was performed for approximately 30 min in each reach using a backpack generator (FEG 800) set to direct current (DC) at 750 V and a maximum of 10 A.

Captured specimens were euthanized via an overdose of benzocaine. Subsequently, they were fixed in 10% formalin, preserved in 70% ethanol, and stored in jars labeled with collection date and location information. In the laboratory, specimens were identified to the species level. Voucher specimens were deposited in the fish collection of the Laboratório de Ecologia Funcional e Estrutural de Ecossistemas at the Universidade Paulista, Sorocaba campus (LEEF 60, 61, 62, 64, 65, 66, 67, 68, 69, and 70; Tab. 3).

TABLE 3 | Species of the ichthyofauna sampled during the twelve collection campaigns (C1 to C4, before interventions, and C5 to C12, after the start of interventions). The following classification is based on Reis et al. (2003).

Taxa	Vouchers	Stretch 1 (R1)		Stretch 2 (R2)		Stretch 3 (R3)		Stretch 4 (R4)
		C1 – C4	C5 – C12	C1 – C4	C5 – C12	C1 – C4	C5 – C12	C1 – C4	C5 – C12
CHARACIFORMES
Acestrorhamphidae
Deuterodon janeiroensis (Eigenmann, 1908)	LEEF 61	0	0	6	3	6	11	0	0
Hollandichthys multifasciatus (Eigenmann & Norris, 1900)	LEEF 60	3	11	0	0	2	1	9	11
Erythrinidae
Hoplias malabaricus (Spix & Agassiz, 1829)	LEEF 62	0	0	3	1	0	0	2	1
SILURIFORMES
Callichthyidae
Scleromystax barbatus (Quoy & Gaimard, 1824)	LEEF 70	1	2	5	2	3	1	11	10
Heptapteridae
Rhamdia aff. quelen (Quoy & Gaimard, 1824)	LEEF 64	2	0	6	3	4	5	2	2
Loricariidae
Pseudotothyris obtusa (Miranda Ribeiro, 1911)	LEEF 65	4	1	2	0	7	1	0	0
Isbrueckerichthys duseni (Miranda Ribeiro, 1907)	LEEF 69	16	26	4	3	5	12	5	3
Trichomycteridae
Cambeva zonata (Eigenmann, 1918)	LEEF 66	5	28	3	1	4	57	2	18
CYPRINODONTIFORMES
Poeciliidae
Phalloceros lucenorum Lucinda, 2008	LEEF 67	50	151	12	109	16	65	11	44
CICHLIFORMES
Cichlidae
Geophagus iporangensis Haseman, 1911	LEEF 68	3	0	4	4	3	2	0	0
Total		84	219	45	126	50	155	42	89

Statistical analysis of data. To visualize dissimilarities in species composition between the pre-construction and during-construction periods, we performed a non-metric multidimensional scaling (nMDS). This analysis was based on a Euclidean distance matrix calculated from Hellinger-transformed species abundance data. To formally test for differences in community composition between periods, we used a permutational multivariate analysis of variance (PERMANOVA). Additionally, a permutational analysis of multivariate dispersions (PERMDISP) was used to test for homogeneity of within-group variance. All ordination analyses were conducted in R (R Development Core Team, 2020) using the ‘vegan’ package (Oksanen et al., 2019).

We used linear mixed-effects models (LMMs) to investigate the influence of highway construction (a categorical variable: before/after) and environmental covariables (conductivity, pH, dissolved oxygen, total dissolved solids, water temperature, and width) on response variables. The response variables were species richness (S) and the population abundances of Scleromystax barbatus, Rhamdia aff. quelen, and Cambeva zonata.

All possible sub-models (Garibaldi et al., 2014) were generated using the MuMIn package (Bartón, 2024). The Akaike Information Criterion (AICc) was employed to select the best-fitting model, interpreting AICc as a measure of the distance between each model and an unknown “true” model (Anderson, 2008). The relative importance of independent variables were assessed using Akaike weights for each covariate across all models containing that covariate (Fig. S1). We calculated the ratio of Akaike weights (w/w0) to quantify the evidence that the best-fit model is superior to each reduced model. This ratio represents the relative probability that the best model is the best-fitting model in the candidate set, compared to the corresponding reduced model.

Results​

---

A total of 810 fish were collected, representing 10 species distributed among eight families and four orders (Tab. 3). The orders Cyprinodontiformes and Siluriformes were the most abundant, accounting for 56.54% and 32.82% of the total catch, respectively, followed by Characiformes (8.62%) and Cichliformes (1.97%). Phalloceros lucenorum and Cambeva zonata were the dominant species, with 458 and 118 individuals captured, respectively.

Highway duplication significantly altered the fish community composition between before construction and after construction periods (PERMANOVA, P = 0.01). However, there was no significant difference in the multivariate dispersion of the community between the two periods (PERMDISP, P = 0.97). These results indicate that the primary effect of the construction was a directional shift in community structure (i.e., a change in the group centroid in multivariate space), rather than an increase or decrease in the compositional variability within each period (Fig. 2). Species richness (S) was highest during the pre-construction period, particularly in the dry season sampling campaigns. A clear decline in average species richness was observed after construction began (Fig. 3).

FIGURE 2| NMDS biplot of the fish abundance data. Stress = 0.15.

FIGURE 3| Average species richness captured in four sampled sections before (B) and after (A) the road duplication, categorized by rainy (R) and dry (D) seasons.

The environmental data revealed a clear impact from the construction on in-stream conditions. Following the commencement of the works, both Fine Sediment Deposition and concentrations of Total Dissolved Solids increased across all four sampled reaches (Fig. 4).

FIGURE 4| Average of the Total Dissolved Solids (TDS) variables and Fine Sediments Deposition (FSD) for the four stretches sampled before and after the duplication works. A = After, B = Before.

The linear mixed-effects models (LMMs) revealed the distinct effects of highway construction and environmental variables on the fish fauna. For species richness (S), highway construction had a significant negative effect, while water temperature showed a positive association. The abundance of Scleromystax barbatus was also negatively impacted by the highway but was positively associated with stream width. Similarly, Rhamdia aff. quelen abundance was negatively affected by the highway and positively influenced by water temperature. In contrast, the abundance of Cambeva zonata was positively associated with both highway construction and stream width (Tab. 4; Fig. 5).

TABLE 4 | Best-fitting mixed-effects model results, showing fixed-effect coefficients (standard errors) for road intervention and environmental covariables. The estimated variances for residual and random effects (stream section and sampling period) are also presented for fish richness (S) and selected species abundances.

Variable	Road	Temperature (H2O)	Width	Var Section	Var Year	Var Res	Weigth
Species richness	2.12 (0.77)	0.27 (0.06)		0.10	0.64	1.15	0.21
Scleromystax barbatus	0.36 (0.16)			0.05	0.00	0.27	0.33
Rhamdia aff. quelen	0.36 (0.16)			0.01	0.01	0.16	0.32
Cambeva zonata	-0.47 (0.17)		0.42 (0.11)	0.24	0.00	0.31	0.45

FIGURE 5| Predicted marginal effects from the mixed-effects models (LME), showing the influence of road duplication on fish richness (S) and the abundance of selected species Scleromystax barbatus, Rhamdia aff. quelen, and Cambeva zonata. The figure contrasts estimated values for the before and after intervention periods.

Discussion​

---

The highway duplication analyzed in this study occurred within a highly conserved remnant of the Atlantic Forest (Fundação Florestal, 2011), an ecosystem of critical conservation importance. Road construction in such forested landscapes is a well-documented driver of deforestation and habitat fragmentation. According to the general model of landscape transformation, the greatest ecological impacts from roads occur early in the process (Forman, Alexander, 1998), with increased soil erosion and subsequent stream sedimentation being primary initial consequences.

Our findings are consistent with this framework. Neotropical streams are known to be highly susceptible to disturbances from land-use change, particularly erosion and sedimentation (Santos, Esteves, 2014; Larentis et al., 2022). This study verified that increased sediment transport was indeed a primary impact on the stream following the commencement of construction. The disturbances observed in the studied stream -namely the removal of riparian forest and the subsequent increase in sediment input -are consistent with the well-documented impacts of road construction (Forman, Alexander, 1998; Angermeier et al., 2004). The increased concentrations of fine sediment and Total Dissolved Solids recorded in this study provide direct evidence of these impacts. A key contribution of this study is its focus on the ‘road-effect zone’ (Forman, 2000; Forman, Deblinger, 2000), which considers the lateral extent of ecological impacts from the road into the adjacent landscape. This approach offers a broader perspective than most research, including studies in Brazil, which typically focus on the more localized effects of road crossings. By examining a highway running parallel to a stream, our work captures the influence of processes like surface runoff that extend from the road and degrade the entire riparian corridor and aquatic habitat.

Our results demonstrate a strong link between road-induced disturbances and fish fauna, and these findings are supported by previous research, including the work of Wheeler et al. (2006) and Ottburg, Blank (2015). As in our study, these studies highlight, for example, that changes in environmental variables explain variations in fish abundance. We suggest examining this dynamic within the context of a “road effect zone” created by a parallel highway. We demonstrate how local habitat conditions are directly affected by the removal of riparian vegetation, leading to critical changes such as increased fine sediment input and increased solar radiation reaching the stream. This local-scale approach is particularly salient; for direct and specific impacts, such as those of road construction, local metrics may be more relevant in explaining changes in fish assemblages than broad, watershed-scale analyses, a conclusion supported by the recent work of Welsh et al. (2023).

Highway construction directly influenced the stream’s thermal regime. The removal of riparian vegetation increased solar radiation reaching the water surface, leading to higher average water temperatures. This is a critical alteration, as temperature is a primary factor structuring fish communities in Atlantic Forest streams (Silva et al., 2007). Our models reflected this dynamic, showing that the increased water temperature was positively associated with the abundance of Rhamdia aff. quelen. Conversely, many fish species native to forested neotropical streams are sensitive to thermal increases (Welsh et al., 2023). This sensitivity may have contributed to the negative impacts observed for other species, such as Scleromystax barbatus, which are typically found in cooler, shaded environments. The long-term consequences of such thermal alterations are well-documented and include restricted species distributions, altered migration patterns, and population fragmentation, which can ultimately lead to local extinctions (Park et al., 2008; Beauchene et al., 2014).

The negative impact of the highway was also pronounced for species richness, which was highest during the pre-construction period and subsequently declined. This finding highlights a different mechanism of impact than that reported in studies of highway crossings. While crossings often create physical barriers that impede fish movement and fragment populations, the parallel highway in our study impacted the stream via the ‘road-effect zone’, causing widespread degradation of the physical habitat. Such environmental degradation is known to reduce microhabitat complexity, thereby negatively affecting the availability of food and shelter for fish (e.g., Teresa, Casatti, 2012). This mechanism explains our key finding: while the highway significantly altered overall species richness and composition, it did not change the multivariate dispersion of the communities. The stream is longitudinally connected, allowing fish to move between sampled stretches in search of more favorable conditions. Although small fish species found in Atlantic Forest streams have limited movement, individuals are capable of moving between adjacent stretches. This assertion is reinforced by Mazzoni et al. (2018) and Mazzoni, Barros (2021), who emphasize that stream fish not only move short distances but also undertake long-distance movements. De Fries et al. (2023) found that the probability of movement was higher within open stretches of the studied stream. In streams, fish need to move through different habitats to spawn, grow, feed, and find shelter (Flecker et al., 2010; Mazzoni et al., 2018). However, natural and artificial barriers can limit movement (Mozzaquattro et al., 2020; De Fries et al., 2023) and, consequently, influence the distribution, abundance, and persistence of fish along streams. This ability to move between adjacent reaches was evidently sufficient to prevent the formation of spatially distinct community structures within the impacted zone.

The significant change in fish community composition observed in this study can be understood by considering the specific characteristics of Atlantic Forest headwater streams. These ecosystems, while part of a biome with high overall ichthyofaunal diversity (Oyakawa et al., 2006; Carvalho, Tejerina-Garro, 2022), are typically defined by specialized environmental conditions: clear water, rocky substrates, high current velocity, and cool temperatures. These conditions create a high diversity of microhabitats that are critical in structuring local fish assemblages (Terra et al., 2016; Bonnemann, Silva, 2017). A key feature of these communities is that they generally exhibit low local species richness, with a few species being highly abundant while many others are rare (Matthews, 1998; Silva, 2007; Cetra et al., 2020). This community structure makes the assemblage highly vulnerable to environmental disturbances. By altering the physical habitat – particularly through increased sedimentation that smothers rocky substrates – the highway construction likely disrupted the availability of specialized microhabitats. This disproportionately impacts the system’s many rare species, thus driving the observed shift in community composition.

The multivariate dispersion of the fish communities did not change significantly between before construction and after construction periods. This non-significant result validates the interpretation of the significant PERMANOVA finding as a true shift in community composition (i.e., a change in the multivariate centroid), rather than an artifact of changing within-group variability (Anderson et al., 2006). Two potential factors may explain this stability in community dispersion despite the compositional shift. First, one bank of the stream remained intact throughout the study, which may have buffered the ecosystem from the full impact of the construction and provided refuges, thus enhancing resilience. Second, it is possible that some environmental recovery occurred during the latter part of the monitoring period. For example, Hedrick et al. (2007) found that suspended solids and sediment levels in streams impacted by highway construction returned to near-reference levels within weeks of project completion. However, a critical distinction must be made. The rapid community recovery reported by Hedrick et al. (2007) was for benthic macroinvertebrates, which have much shorter life cycles and faster recolonization dynamics than fish. Therefore, while environmental conditions in our study may have partially improved, the recovery of the fish community would be expected to occur over a much longer timescale.

Abiotic factors play a critical role in the ecological structuring of fish communities (Townsend, Eidels, 2011; Matos et al., 2013). The disturbances observed in this study, where highway construction strongly influenced fine sediment deposition, total dissolved solids, temperature, and stream width, drove the significant changes in species composition. Such impacts must be considered when assessing highway projects. For instance, increased water turbidity, as observed here, can severely affect fish assemblages by impairing the foraging ability of visually oriented species (Matos et al., 2013). Our results reveal a differential response among the ichthyofauna, with some species benefiting from the new conditions while others were negatively impacted, highlighting varying levels of resilience.

We can exemplify Phalloceros lucenorum and Cambeva zonata, which increased in abundance after the construction began. This finding suggests these species are well-adapted to the altered conditions. Cambeva zonata, in particular, appeared to benefit directly from the changes. Its known preference for sandy environments aligns with the increased siltation and deposition of fine substrate observed in the stream bed. This interpretation is supported by field observations of this species frequently burrowing in sandy substrates, a habitat that became more prevalent during the construction phase.

The negative effect of highway construction was particularly evident for species like Rhamdia aff. quelen and Scleromystax barbatus. Although Rhamdia aff. quelen is known to be relatively tolerant of variations in factors like water temperature (Gomes et al., 2000), our findings suggest the overall habitat degradation surpassed its resilience threshold. The suppression or alteration of specific pool habitats, as described by Ferreira, Casatti (2006b), is a plausible explanation for its decline, even though some individual environmental changes, such as reduced shade, might otherwise be tolerated (Esteves et al., 2019). Scleromystax barbatus completely disappeared from the samples at the beginning of the interventions. The sharp increase in total dissolved solids is a likely cause, as this species is known to prefer environments with low levels of total dissolved solids (Giongo et al., 2023). Our own field observations confirm that S. barbatus inhabits clearwater streams with dense riparian vegetation, precisely the conditions that were most degraded by the construction. Notably, this species began to reappear in the samples during the final phase of the study, coinciding with a potential stabilization or reduction in total dissolved solids levels. This pattern reinforces that, after disturbance, some species may begin to recover as environmental conditions improve, and the timeline for this recovery presents a crucial avenue for future research.

Atlantic Forest streams are defined by their habitat heterogeneity, which supports a fish fauna with diverse and specialized adaptations (Oyakawa et al., 2006; Terra et al., 2016). This assemblage is critically dependent on intact riparian vegetation for refuge, feeding, and reproduction (Ferreira, Casatti, 2006a). Therefore, the degradation of riparian and in-stream habitats, as documented here, directly threatens this specialized fauna and explains the observed decline in species richness. Our findings reinforce the urgent need not only for the conservation of these ecosystems but also for continued research to better understand and mitigate the pervasive effects of highways (Mattox et al., 2023).

Acknowledgments​

---

The authors thank the UNIP/Laboratório de Ecologia Estrutural e Functional de Ecossistemas and INSITU AMBIENTAL for support and infrastructure for developing this study. We would also like to thank Luís G. N. Carvalho for creating the map.

References​

---

DAnderson DR. Model-based inference in the life sciences: a primer on evidence. New York, NY: Springer; 2008.

Anderson MJ, Ellingsen KE, McArdle BH. Distance-based tests for homogeneity of dispersions across groups. Ecol Lett. 2006; 9(6):683–93. https://doi.org/10.1111/j.1461-0248.2006.00926.x

Angermeier PL, Wheeler AP, Rosenberger AE. A Conceptual framework for assessing impacts of roads on aquatic biota. Rev Fish Sci. 2004; 29(12):19–29. https://doi.org/10.1080/10641260590964449

Azevedo-Santos VM, Daga VS, Fearnside PM. Roads to pollution: Brazil’s aquatic biodiversity affected by truck leaks. Oecol Austr. 2022; 26(3):483–93. https://doi.org/10.4257/oeco.2022.2603.07

Bartón K. MuMIn: Multi-Model Inference. R package version 1.48.4. 2024. Available from: https://CRAN.R-project.org/package=MuMIn

Beauchene M, Becker M, Bellucci CJ, Hagstrom N, Kanno Y. Summer thermal thresholds of fish community transitions in Connecticut streams. North Am J Fish Manag. 2014; 34(1):119–31. https://doi.org/10.1080/02755947.2013.855280

Benton PD, Ensign WE, Freeman BJ. The effect of road crossing on fish movements in small Etowah sasin streams. Southeast Nat. 2008; 7(2):301–10. https://doi.org/10.1656/1528-7092(2008)7[301:TEORCO]2.0.CO;2

Bonnemann ST, Silva JFM. Ecologia de riachos de montanha de Mata Atlântica. Londrina: Eduel; 2017.

Brejão GL, Teresa FT, Gerhard P. When roads cross streams: fish assemblage responses to fluvial fragmentation in lowland Amazonian streams. Neotrop Ichthyol. 2020; 18(3):e200031. https://doi.org/10.1590/1982-02224-2020-0031

Bouska WW, Paukert CP. Road crossing designs and their impact on fish assemblages of great plains streams. Trans Am Fish Soc. 2010; 139:214–22. https://doi.org/10.1577/T09-040.1

Carvalho RA, Tejerina-Garro FL. Headwater stream fish assemblages: what its diversity and the environment tell us about their structure and organization? Front J Soc Technol Environ Sci. 2022; 11(4):134–52. http://dx.doi.org/10.21664/2238-8869.2022v11i4.p134-152

Cetra M, Mattox G, Romero PB, Escobar SH, Guimarães EA, Turin RAF. Ichthyofauna from “serranias costeiras” of the Ribeira de Iguape River basin, Southeast Brazil. Biota Neotrop. 2020; 20(4):e20200094. https://doi.org/10.1590/1676-0611-BN-2020-0994

De Fries L, Camana M, Guimarães M, Becker FG. Road crossings hinder the movement of a small non-migratory stream fish. Environ Biol Fish. 2023; 106:1295–311. https://doi.org/10.1007/s10641-023-01416-y

Esteves KE, Silva MHC, Nanini-Costa MH, Petesse ML. Organization of fish assemblages in blackwater Atlantic Forest streams. Neotrop Ichthyol. 2019; 17(1):e180120. https://doi.org/10.1590/1982-0224-20180120

Ferreira CP, Casatti L. Integridade biótica de um córrego na bacia do Alto Rio Paraná avaliada por meio da comunidade de peixes. Biota Neotrop. 2006a; 6(3):1–25. http://dx.doi.org/10.1590/S1676-06032006000300002

Ferreira CDP, Casatti L. Influência da estrutura do habitat sobre a ictiofauna de um riacho em uma micro bacia de pastagem, São Paulo, Brasil. Rev Bras Zool. 2006b; 23(3):642–51. http://dx.doi.org/10.1590/S0101-81752006000300006

Flecker AS, McIntyre PB, Moore JW, Anderson JT, Taylor BW, Hall RO Jr. Migratory fishes as material and process subsidies in riverine ecosystems. Am Fish Soc Symp. 2010; 73:559–92.

Forman RTT. Estimate of the area affected ecologically by the road system in the United States. Conserv Biol. 2000; 14(1):31–35. http://dx.doi.org/10.1046/j.1523-1739.2000.99299.x

Forman RTT, Alexander LE. Roads and their major ecological effects. Ann Rev Ecol Evol Syst. 1998; 29(1):207–31. https://doi.org/10.1146/ANNUREV.ECOLSYS.29.1.207

Forman RTT, Deblinger RD. The ecological road-effect zone of a Massachusetts (U.S.A.) suburban highway. Conserv Biol. 2000; 14(1):36–46. https://doi.org/10.1046/j.1523-1739.2000.99088.x

Fundação Florestal. Fundação Florestal examina impactos do projeto de duplicação da Régis Bittencourt, na Serra do Cafezal. Secretaria de Infraestrutura e Meio Ambiente. 2011, [updated in 2019]. Available from: https://www.infraestruturameioambiente.sp.gov.br/fundacaoflorestal/2011/08/fundacao-florestal-vai-examinar-impactos-do-projeto-de-duplicacao-da-regis-bittencourt-na-serra-do-cafezal/

Garibaldi L, Casas C, Biganzoli F. Datos jerárquicos en ciencias ambientales: ejemplos prácticos y análisis de modelos jerárquicos en lenguaje R. 2014. Available from: https://rid.unrn.edu.ar/handle/20.500.12049/4232

Giongo ML, Petesse ML, Esteves KE. Fish responses to multiple scales in coastal blackwater Atlantic Forest streams in Southeast Brazil. Neotrop Ichthyol. 2023; 21(2):e230030. https://doi.org/10.1590/1982-0224-2023-0030

Gomes LC, Golombieski JI, Gomes ARC, Baldisserotto B. Biologia do jundiá Rhamdia quelen (Teleostei, Pimelodidae). Ciênc Rural. 2000; 30(1):179–85. https://doi.org/10.1590/S0103-84782000000100029

Hedrick LB, Welsh AS, Anderson JT. Effects of highway construction on sediment and benthic macroinvertebrates in two tributaries of the Lost River, West Virginia. J Freshw Ecol. 2007; 22(4):561–69. https://doi.org/10.1080/02705060.2007.9664817

Larentis C, Kotz-Kliemann BC, Neves MP, Delariva RL. Effects of human disturbance on habitat and fish diversity in neotropical streams. PLoS ONE. 2022; 17(9):e0274191. https://doi.org/10.1371/journal.pone.0274191

Leal CGO. Assessment of landscape changes caused by highway construction: case study of ruta del Canal Pargua Highway in Chile. Landsc Online. 2023; 98:1113. https://doi.org/10.3097/LO.2023.1113

Li C, Zhang J, Philbin SP, Yang X, Dong Z, Hong J, Ballesteros‑Pérez P. Evaluating the impact of highway construction projects on landscape ecological risks in high altitude plateaus. Sci Rep. 2022; 12:5170. https://doi.org/10.1038/s41598-022-08788-8

Matthews WJ. Padrões em ecologia de peixes de água doce. 1ª ed. Springer US; 1998.

Matos PR, Carmo CM, Melo CE. Relação entre variáveis ambientais e a estrutura da comunidade de peixes em córregos das bacias do rio das Mortes e do rio Xingu – MT, Brazil. Biotemas. 2013; 26(3):139–51. https://doi.org/10.5007/2175-7925.2013v26n3p139

Mazzoni R, Barros TF. Ecologia do movimento em peixes de riacho. Oecol Austr. 2021; 25(2):381–97. https://doi.org/10.4257/oeco.2021.2502.10

Mazzoni R, Pinto MP, Iglesias-Rios R, Costa R. Fish moviment in na Atlantic Forest stream. Neotrop Ichthyol. 2018; 16(1):e170065. https://doi.org/10.1590/1982-0224-20170065

Mozzaquattro LB, Dala-Corte RB, Becker FG, Melo AS. Effects of spatial distance, physical barriers, and habitat on a stream fish metacommunity. Hydrobiologia. 2020; 847:3039–54. https://doi.org/10.1007/s10750-020-04309-8

Ochs AE, Swihart, RK, Saunders MR. A comprehensive review of the effects of roads on salamanders. Landsc Ecol. 2024; 39:77. https://doi.org/10.1007/s10980-024-01867-3

Oksanen J, Blanchet FG, Friendly M, Kindt R, Legendre P, Mcglinn D et al. Vegan: community ecology package. R package version 2019. 2019. Available from: https://cran.r-project.org/web/packages/vegan/index.html

Ottburg F, Blank M. Solutions to the impacts of roads and other barriers on fish and fish habitat. In: Ree RVD, Smith DJ, Grilo C, editors. Handbook of road ecology. Oxford: Wiley Blackwell; 2015. p.364–72. https://doi.org/10.1002/9781118568170.ch45

Oyakawa OT, Akama A, Mautari KC, Nolasco JC. Peixes de riachos da Mata Atlântica nas Unidades de Conservação do Vale do Rio Ribeira de Iguape no Estado de São Paulo. São Paulo: Editora Neotrópica; 2006.

Park D, Michael S, Bayne E, Scrimgeour G. Landscape level stream fragmentation caused by hanging culverts along roads in Alberta’s boreal forest. Can J For Res. 2008; 38(3):566–75. https://doi.org/10.1139/X07-179

Pépino M, Rodríguez MA, Magnan P. Impacts of highway crossings on density of brook charr in streams. J Appl Ecol. 2012; 49(2):395–403. https://doi.org/10.1111/j.1365-2664.2012.02108.x

R Development Core Team. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. 2015. Available from:
http://www.R-project.org/

Reis RE, Kullander SO, Ferraris Jr. CJ. Check list of the freshwaterfishes of South and Central America.PortoAlegre: Edipucrs; 2003.

Santos JAP, Esteves KE. The fish fauna of an Atlantic Forest conservation area in the largest urban center of South America (São Paulo, SP, Brazil) and its relationship to some environmental factors. Stud Neotrop Fauna Environ. 2014; 49(2):135–50. http://dx.doi.org/10.1080/01650521.2014.941739

Silva EF, Melo CE, Venere PC. Fatores que influenciam a comunidade de peixes em dois ambientes no baixo Rio da Mortes, Planície do Bananal, Mato Grosso, Brazil. Rev Bras Zool. 2007; 24(2):482–92. http://dx.doi.org/10.1590/S0101-81752007000200029

Sliger R, Graham J, Hoenke K, Kimball ME, Sterling KA, Peoples BK. Integrating fish swimming abilities into rapid road crossing barrier assessment: case studies in the southeastern United States. PLoS ONE. 2024; 19(2):e0298911. https://doi.org/10.1371/journal.pone.0298911

Southerland M. Evaluation of ecological impacts from highway development. U. S. Environmental Protection Agency. 1994.

Strahler AN. Quantitative analysis of watershed geomorphology. Trans Am Geophys Union. 1957; 38:913–20.

Teresa FB, Casatti L. Influence of forest cover and mesohabitat types on functional and taxonomic diversity of fish communities in Neotropical lowland streams. Ecol Freshw Fish. 2012; 21(3):433–42. https://doi.org/10.1111/j.1600-0633.2012.00562.x

Terra BF, Hughes RM, Araújo FG. Fish assemblages in Atlantic Forest streams: the relative influence of local and catchment environments on taxonomic and functional species. Ecol Freshw Fish. 2016; 25(4):527–44. http://dx.doi.org/10.1111/eff.12231

Townsend JT, Eidels A. Workload capacity spaces: a unified methodology for response time measures of efficiency as workload is varied. Psychon Bull Rev. 2011; 18:659–81. https://doi.org/10.3758/s13423-011-0106-9

Trombulak SC, Frissell CA. Review of ecological effects of roads on terrestrial and aquatic communities.Conserv Biol. 2000; 14(1):18–30.

Welsh DP, Ludlam JP, Downs EL, Gordon ES, Clark EV, Levy B et al. Stream fish community structure across an urban gradient in a northeastern US watershed. Environ Biol Fish. 2023; 106:1579–95. https://doi.org/10.1007/s10641-023-01438-6

Wheeler AP, Angermeier PL, Rosenberger AE. Impacts of new highways and subsequent landscape urbanization on stream habitat and biota. Rev Fish Sci. 2005; 13(3):141–64. https://doi.org/10.1080/10641260590964449

Authors

---

Thais A. Soinski¹, Mauricio Cetra² and Welber S. Smith^1,3,4

^[1]    Laboratório de Ecologia Estrutural e Funcional de Ecossistemas, Universidade Paulista – UNIP, Câmpus Sorocaba, Av. Independência, 752, Iporanga, 18103-000, Sorocaba, SP, Brazil. (TAS) thaissoinski@outlook.com, (WSS) welber_smith@uol.com.br (corresponding author).

^[2]    Departamento de Ciências Ambientais, Universidade Federal de São Carlos, Câmpus Sorocaba, Rodovia João Leme dos Santos, km 110, Bairro do Itinga, 18052-780, Sorocaba, SP, Brazil. (MC) mcetra@ufscar.br

^[3]    Programa de Pós-Graduação em Patologia Ambiental e Experimental, Universidade Paulista – UNIP, Rua Doutor Bacelar, 1212, 04026-002, São Paulo, SP, Brazil.

^[4]    Programa de Pós-Graduação em Aquicultura e Pesca, Secretaria de Agricultura e Abastecimento, SP – Agência Paulista de Tecnologia dos Agronegócios – Instituto de Pesca, Av. Conselheiro Rodrigues Alves, 1252, Moema, 04014-002, São Paulo, SP, Brazil.

Authors’ Contribution

---

Thais A. Soinski: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing-original draft.

Mauricio Cetra: Conceptualization, Data curation, Methodology, Writing-original draft.

Welber S. Smith: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing-original draft, Writing-review and editing.

Ethical Statement​

---

All collection activities were authorized by SISBIO (license number 6017122) and approved by the institutional Ethics Committee. The procedures complied with all relevant national regulations, including Law N° 11.794/2008, Decree N° 6.899/2009, and the norms established by the National Council for Animal Experimentation Control (CONCEA).

Competing Interests

---

The author declares no competing interests.

Data availability statement

---

The authors confirm that the data supporting the findings of this study are available within the article

AI statement

---

The authors did not use artificial intelligence in this article.

Funding

---

The authors gratefully acknowledge the Technological Initiation Scholarship Program (Santander) for financial support to the first author and the Vice-Rectory of Graduate and Research for financial support to the third author.

Supplementary Material

---

Supplementary material S1

How to cite this article

---

Soinski TA, Cetra M, Smith WS. Highway impact on fish assemblages in Atlantic Forest stream. Neotrop Ichthyol. 2025; 23(4):e250055. https://doi.org/10.1590/1982-0224-2025-0055

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

© 2025 The Authors.

Diversity and Distributions Published by SBI

Accepted October 24, 2025

Submitted March 30, 2025

Epub February 2, 2026