Life on the road: fish communities composition in roadside ditches of the Atlantic Forest

João Henrique Alliprandini da Costa^1,2 , Amanda Selinger², Francisco Langeani³, Ursulla Pereira Souza² and Rafael Mendonça Duarte^1,2

PDF: EN    XML: EN | Supplementary: S1 S2 | Cite this article

Associate Editor: Franco Teixeira de Mello

Section Editor: Fernando Pelicice

Editor-in-chief: José Birindelli

Abstract​

---

Introduction​

---

Temporary aquatic habitats are dynamic ecosystems that often harbor high levels of endemism and a specialized ichthyofauna adapted to extreme environmental conditions (Williams, 2007). Survival in such ephemeral environments requires physiological and behavioral adaptations to cope with stressful environmental conditions, with several fish species exhibiting exceptional physiological tolerance to sharp fluctuations in water temperature and hypoxic conditions (Polačik, Podrabsky, 2015), while others have evolved amphibious lifestyle to exploit these unstable habitats (Turko et al., 2021). In the Neotropical region, studies on fish from temporary aquatic habitats have primarily focused on temporary pools formed by rainfall and river flooding. However, most ichthyological research emphasizes taxonomic descriptions of new species (Costa, 2006; Drawert, 2022; Ramos et al., 2023), rather than understanding the structure and composition of fish communities, also an essential step for assessing ecological processes, evolutionary strategies, and conservation needs. Therefore, filling these knowledge shortfalls, particularly regarding species distribution and habitat associations, is critical (Freitas et al., 2021; Reis et al., 2024).

Despite their ecological importance, studies on fish community composition in temporary aquatic habitats remain scarce and are largely restricted to the Amazon region (Pazin et al., 2006; Espírito-Santo et al., 2009). In contrast, in the Atlantic Forest, research has predominantly focused on taxonomy, genetics, and population-level dynamics (Costa, 2002, 2006, 2009; Abilhoa et al., 2010; Contente, Stefanoni, 2010; Berbel-Filho et al., 2021; Guedes et al., 2023), with a notable absence of studies addressing community-level patterns in temporary environments. This lack of ecological knowledge is particularly concerning in a biome that is both a global biodiversity hotspot and home to approximately 36% of Brazil’s threatened small freshwater fish species (Castro, Polaz, 2020).

Among these temporary habitats, roadside ditches represent an overlooked but potentially important artificial temporary habitat, particularly in coastal restinga forests of the Atlantic Forest, where such temporary waters may host threatened or narrowly endemic species (Giongo et al., 2023; Costa et al., 2024a). Some fish species have even been documented exclusively in artificial habitats, such as highway gutters (Costa, 2002), reinforcing the urgent need to investigate these systems. Although artificial temporary environments are rarely addressed in fish community ecology, evidence from macroinvertebrate studies shows that even human-made ephemeral habitats, such as tire tracks, can harbor unique and rare species (Armitage et al., 2012). Similarly, a study conducted in Florida, USA, on fish assemblages in artificial ditch systems highlighted their potential for ecological research and conservation applications (Hohausová et al., 2010). Fish assemblages from these artificial temporary habitats requires specialized traits or even amphibious lifestyles such as: overland displacement capacity for fish species that are good colonizers (Baber et al., 2002; Gibb et al., 2011), adaptations to cope with hypoxia such as air-breathing strategies that involve breathing through gas bladders, gastrointestinal tract or even the skin (Graham, 1997; Turko, Wright, 2015, Turko et al., 2021), or desiccation tolerance during the embryonic phase, as is the case for seasonal killifishes (Furness et al., 2015). This makes it even more necessary for these environments to be further investigated, since these roadside ditches, excavated in what was once a floodable forest area and is now a dirt road, are artificial systems that may alter the basin’s natural hydrological function. This is because they can serve as an entry point for artificial water input and introduced species, which may influence natural systems, such as temporary pools, after enhanced connectivity. This can threaten these peculiar adaptations and their hydrological dynamics, with the possibility of even affecting the stream network. Despite this, no study has yet examined fish community composition in artificial temporary habitats within the Neotropical region, leaving a significant gap in our understanding of fish living in these environments.

Given the rapid fragmentation of natural landscapes (Vancine et al., 2024), artificial ephemeral habitats, though often dismissed, can serve as natural laboratories for understanding the processes that structure species composition and how community assembly occurs in new habitats in the absence of natural ones. So, documenting fish communities in such habitats is critical not only for biodiversity assessments, but also for informing conservation actions in a changing and increasingly fragmented world, since highly threatened species can use these artificial environments (Costa et al., 2024a), and also to investigate possible pathways of exotic species introductions (Brisson et al., 2010), once roadside ditches are human made and highly connected to human activity, favoring high tolerant invasive species that can cope with these extreme environments.

A key environmental trait in such systems is the oscillation in water volume, which varies from zero to almost full hydrological connectivity depending on the rainfall season, and this must also be considered when accounting for community assembly in new habitats. Dispersal limitation can be a relevant force during low-water periods (Declerck et al., 2011), while fish communities in ephemeral habitats become homogenized during periods of high water (Bokhutlo et al., 2021), with similar patterns of community structuring through hydrological connectivity being found in wetlands of the Neotropical region (Fernandes et al., 2009; Macedo-Soares et al., 2010).

In this study, we investigate the species composition of fish communities from roadside ditches in the Preto River microbasin, a part of the Itanhaém River basin located in Itanhaém, São Paulo State, Brazil. This area is characterized by blackwater streams flowing through restinga forests of the coastal Atlantic Forest. The fish composition and distribution in the streams of the Itanhaém River basin have been documented by Ferreira, Petrere (2009), and Ferreira et al. (2014), who identified distinct fish zones within this important segment of the Atlantic Forest’s coastal basins, marked by high endemism. This endemism stems from the region’s natural isolation, with mountainous terrain acting as a physical barrier upstream and downstream by the Atlantic Ocean (Menezes et al., 2007; Giongo et al., 2023), which makes it suitable as a natural laboratory to investigate hydrological connectivity and the use of these artificial environments as ecological traps or refuges in the face of the loss of natural habitats. The Preto River microbasin blackwater streams ichthyofauna is mainly composed by few species (approximately 19), dominated by Characiformes typical of restinga streams, such as Mimagoniates lateralis (Nichols, 1913) and Hollandichthys multifasciatus (Eigenmann & Norris, 1900), followed by Loricariidae and Callichthyidae, respectively exemplified by Pseudotothyris obtusa (Miranda Ribeiro, 1911) and Scleromystax macropterus (Regan, 1913), in addition to other recurring groups such as Gymnotidae, Cichlidae, Crenuchidae (Ferreira et al., 2014). However, no information is available regarding the fish fauna inhabiting temporary habitats in this region, particularly artificial temporary environments such as roadside ditches.

Thus, we aimed to: (i) describe the fish species inhabiting these ephemeral artificial environments; (ii) assess differences in species richness and composition between Wet and Drier periods, given the seasonal influence of rainfall on hydrological connectivity; and (iii) examine how environmental conditions and spatial proximity between ditches influence community structure. We hypothesize that species richness and composition differ between rainfall periods, with higher richness during the Wet Period due to increased connectivity with nearby streams, which may facilitate colonization by stream-dwelling species. In contrast, the Drier Period is expected to harbor mainly resident species adapted to extreme conditions, such as non-seasonal killifishes and air-breathing catfishes, since during periods of low connectivity, species probably need to deal with high temperatures, partial or total drought, limited dispersal, and amphibious lifestyle may be a way to circumvent such conditions and even thrive in them.

Material and methods

---

Study area and sampling. The study was conducted in the Preto River microbasin, part of the Itanhaém River basin, located in the coastal Atlantic Forest of São Paulo State, Brazil (Fig. 1A). The Itanhaém River basin, the second largest coastal basin in the state, is situated in the central-southern region of São Paulo’s coastline. The Preto River microbasin lies predominantly within the coastal plain, characterized by sandy-clay soils and remnant restinga forests. Its waters are dark, with high levels of dissolved organic carbon (DOC) resulting from the microbial decomposition of lignin-rich materials. These blackwaters are also highly acidic and exhibit very low oxygen concentrations (Camargo, Cancian, 2016).

FIGURE 1| A. Sampled roadside ditches at the Preto River microbasin, part of the Itanhaém River basin, Itanhaém, São Paulo, orange represent densely urbanized areas and the red line represent the highway. B. Roadside ditch filled with water after stream flooding and rainfall. C. Another roadside ditch dried after the water recedes, as an example of the ephemeral nature of these environments, even though they were not sampled due to the absence of fish when dry.

A total of 36 roadside ditches were sampled across the Preto River microbasin from January to December 2024, with three different ditches sampled each month. These ditches are excavated alongside unpaved roads and have been maintained for several years (Fig. 1B). They are temporarily flooded by rainfall and overflow from nearby streams, followed by a dry phase when water recedes (Fig. 1C). The same ditch (or another within the same temporarily flooded area) was only resampled in a later month if it had undergone complete drying followed by full reflooding, ensuring the site could be recolonized naturally. This criterion was adopted to avoid carryover effects from previous sampling on the structure of newly colonized communities in these artificial habitats. All sampling locations were georeferenced using GPS to allow continuous monitoring. Although 36 samples were obtained over the year, the monthly samples were deliberately limited to minimize disturbance in this ecosystem, which contains highly vulnerable species. This precaution was reinforced after a Critically Endangered killifish species was rediscovered at one roadside ditch during the first three months of sampling (Costa et al., 2024a), with the joint occurrence of other species with a high level of threat (Costa et al., 2024b).

To standardize sampling effort across sites, each ditch was sampled for 15 minutes by two researchers using hand nets with an oval mouth (50 x 40 cm) with a 1 mm mesh size, appropriate for capturing small-bodied fish. The sampling method was specifically designed for this study, as traditional seines were impractical due to the abundance of submerged branches and human litter in these habitats. The chosen method is also well-suited for small aquatic environments, facilitating future comparisons with other temporary habitats, such as temporary pools. This is particularly relevant given the scarcity of previous studies investigating fish community composition in similar artificial environments in the Neotropical region.

Following sampling, specimens were anesthetized using 1 to 1.5 mL of eugenol per liter of water, fixed in 10% formalin for 48 hours, and subsequently preserved in 70% ethanol. Collected specimens were identified and deposited in the Fish Collection from Departamento de Ciências Biológicas, Universidade Estadual Paulista “Júlio de Mesquita Filho”, Campus São José do Rio Preto (DZSJRP) of UNESP in São José do Rio Preto.

In the field, ditches environmental variables were recorded at each site, including maximum length (m), maximum width (m), maximum depth (m), and distance to the nearest stream (m), with an open reel fiberglass measuring tape. Distance to the nearest stream was measured directly using a reel tape only when the stream was nearby, or obtained from Google Maps satellite imagery for distances greater than 100 meters. Physicochemical parameters were also recorded using a Hanna portable multiprobe (HI98194), including pH, temperature (°C), dissolved oxygen (mg.L^-1), conductivity (µS.cm⁻¹), and total dissolved solids (mg.L^-1). The water volume of each ditch (m³) was estimated assuming a half-ellipsoid shape, using the formula: V = 2/3 x x a x b x c, where V is the volume, a is the maximum length (m), b is the maximum width (m), and c is the maximum depth (m).

Rainfall patterns, which play a crucial role in the dynamics of temporary environments, were used to classify months into two distinct periods: the Wet Period, encompassing November to March, and the Drier Period, from April to October. Although the dry season in Brazilian coastal streams is not strongly pronounced due to consistent rainfall throughout the year (Tonhasca, 2005), regional precipitation data indicate monthly variations in rainfall intensity, justifying this classification to better understand patterns in temporary environments, which remain largely unexplored. Therefore, we decided to use the term Drier Period instead of Dry Period because even in the period of less rainfall, there is still considerable rain in the region. Rainfall data were obtained from the Centro Nacional de Monitoramento e Alertas de Desastres Naturais (CEMADEN), specifically from the Balneário Gaivota station (24°14’27.6”S 46°53’34.8”W) in Itanhaém-SP, which is located near the sampled sites. The classification of each month into two rainfall periods was based on the accumulated precipitation per month (average between years) (mm) (Fig. 2), using monthly precipitation data from 2016 to 2024, with months with an average greater than 150 mm being grouped in the Wet Period. To ensure data reliability, raw 10-minute interval rainfall records were aggregated into hourly sums. A day was excluded from analysis if it contained fewer than 18 hours of available functional rainfall data. Likewise, a month was excluded from the final monthly precipitation sum if fewer than 20 days met this data availability criterion.

FIGURE 2| Mean annual rainfall per month (mm) in the Preto River microbasin region, Itanhaém, São Paulo, based on monthly precipitation data from 2016 to 2024. Bars represent the mean monthly rainfall, with error bars indicating standard deviations. Months were classified as part of the Drier Period (DP) or Wet Period (WP). Data were obtained from the Centro Nacional de Monitoramento e Alertas de Desastres Naturais (CEMADEN), using records from the Balneário Gaivota station, 24°14’27.6”S 46°53’34.8”W.

Statistical analysis. To explore the relationships among environmental variables across the Drier and Wet periods a Principal Component Analysis (PCA) was conducted, including the following environmental variables: ditch volume (m³), temperature (°C), pH, dissolved oxygen (mg.L^-1), stream distance (m), total solids (mg.L^-1) and conductivity (µS.cm^-1). The variables maximum length (m), maximum width (m), and maximum depth (m) were excluded from the analysis because they were used to calculate the volume of the water bodies, avoiding redundancy. Prior to the analysis, all environmental variables were standardized (mean = 0, standard deviation = 1) to ensure comparability among variables measured on different scales.

Also, to assess and compare the species richness between the periods, rarefaction and extrapolation curves were generated using the iNEXT package in R (Hsieh et al., 2016). This analysis was based on incidence data, which accounts for species presence/absence across sampling units, and was performed using Hill numbers (q = 0), which corresponds to species richness. Rarefaction and extrapolation were estimated up to 40 sampling units to project the species accumulation. The variability in the estimates was assessed using 999 bootstrap replications, providing 95% confidence intervals.

We also performed a Permutational Multivariate Analysis of Variance (PERMANOVA) using the Bray-Curtis dissimilarity index to assess differences in fish community composition between the two periods, and the analysis was conducted with 999 permutations to evaluate the statistical significance. To verify the assumption of homogeneity of multivariate dispersions we applied a Permutational Multivariate Analysis of Dispersion (PERMDISP) using the Bray-Curtis dissimilarity matrix. This step ensured that any detected difference was not due to differences in group dispersion. To visualize patterns in fish community composition between the periods a Non-metric Multidimensional Scaling (NMDS) ordination was performed based on the Bray-Curtis dissimilarity index. The NMDS was computed using two dimensions (k = 2) to effectively represent the multivariate space, with a maximum of 100 random starts to ensure convergence to a global solution.

To explore how environmental variables influence fish community composition a Redundancy Analysis (RDA) was performed using the Bray-Curtis dissimilarity index. Prior to the analysis, the community matrix was transformed using the Hellinger transformation to minimize the impact of rare species and zero inflation, which is appropriate for abundance data. Environmental variables were selected based on the Variance Inflation Factor (VIF), with a threshold of VIF < 3 to avoid multicollinearity. The final model included the following variables: stream distance (m), volume (m³), dissolved oxygen (mg.L^-1), pH and temperature (°C).

We also conducted a Mantel test to investigate the influence of spatial structure on fish community composition by assessing the correlation between geographic distances and community dissimilarity. Geographic coordinates from each sampling site were used to generate a pairwise geographic distance matrix (in kilometers). After that, the Hellinger transformed community matrix was used to calculate the Bray-Curtis dissimilarity to represent pairwise differences in community composition. The Mantel test was performed with 999 permutations to assess the significance of the correlation between geographic distance and community dissimilarity. PERMANOVA, PERMDISP, NMDS, RDA and Mantel test analyses were performed using the vegan R package (Oksanen et al., 2022). In addition, all raw community data, environmental variables, and coordinates of each sampling point, as well as code with the step-by-step data analysis, are available in a Git Hub repository for better reproducibility (https://github.com/JH-All/fish_community_roadside_ditches).

Results​

---

The studied roadside ditches exhibited considerable environmental variability between periods, as evidenced by extremely low dissolved oxygen levels, ranging on average from 1.74 mg/L during the Wet Period (WP) to 0.34 mg/L in the Drier Period (DP). This marked variability in oxygen levels was also seen within periods, varying from 0 to 5.20 mg/L at WP, and from 0 to 2.62 mg/L at DP (Tab. 1). Additionally, extreme temperature oscillation was recorded throughout the year, reaching a maximum of 32.95°C (WP) and a minimum of 16.14°C (DP), along with substantial variability in water volume (on average 27.36 and 56.88 m³ in DP and WP, respectively). The pH ranged from very acidic (2.78) to circum-neutral (6.49) with average of 4.86 in WP and 5.20 in DP (Tab. 1). There was partial overlap in environmental properties between the periods; however, it was possible to observe higher values of dissolved oxygen, temperature, and volume during the Wet Period samples when compared to the Drier Period (Fig. S1).

TABLE 1 | Summary of environmental variables measured in roadside ditches during the Wet Period and Drier Period in the Atlantic Forest. Values represent minimum (Min), maximum (Max), mean and standard deviation (SD) for each variable (N = 36).

	Wet Period			Drier Period
Variables	Min	Max	Mean (SD)	Min	Max	Mean (SD)
Distance to nearest stream (m)	0.30	851	206.14 (292.70)	0.30	703.08	367.61 (313.07)
Maximum length (m)	2.20	57.65	23.93 (15.61)	1.12	49.10	20.25 (14.87)
Maximum width (m)	0.81	9.0	2.75 (1.82)	1.10	4.70	2.23 (0.80)
Maximum depth (m)	0.11	0.80	0.28 (0.15)	0.13	0.80	0.32 (0.17)
Volume (m3)	1.49	293.40	56.88 (78.49)	0.63	165.37	27.36 (34.94)
Temperature (°C)	20.89	32.95	26.78 (3.11)	16.14	27.06	21.47 (3)
Dissolved oxygen (mg.L-1)	0	5.20	1.74 (1.75)	0	2.62	0.34 (0.78)
pH	2.78	6.30	4.86 (1.09)	3.33	6.49	5.20 (1.03)
Conductivity (µS.cm-1)	47	162	85.13 (32.26)	48.6	154	100.74 (40.17)
Total dissolved solids (mg.L-1)	30	105	54.20 (21.87)	26	77	50.52 (17.91)

A total of 787 individuals were collected, representing 17 species from 10 families and five orders (Tab. 2). The most abundant species was Atlantirivulus santensis (Köhler, 1906) (N = 212), followed by Hyphessobrycon boulengeri (Eigenmann, 1907) (N = 131), Phalloceros reisi Lucinda, 2008 (N = 104), Poecilia reticulata Peters, 1859 (N = 77), Hyphessobrycon bifasciatus Ellis, 1911 (N = 72), and Mimagoniates lateralis (N = 62). Several species were exclusively recorded during the Wet Period, including Hollandichthys multifasciatus, Geophagus iporangensis Haseman, 1911, Hoplosternum littorale (Hancock, 1828), Scleromystax macropterus and Synbranchus marmoratus Bloch, 1795. In contrast, only Pseudotothyris obtusa was exclusive to the Drier Period (Tab. 2).

TABLE 2 | Fish species recorded in 36 roadside ditch samples in the Atlantic Forest during the Wet Period (WP) and Drier Period (DP). The table presents the number of individuals sampled in each period, the total abundance, and the corresponding voucher numbers. Species marked with an asterisk are classified as exotic. Threatened species are marked with their category in superscript: VU = Vulnerable, CR = Critically Endangered. Conservation status of each species was based on the Biodiversity Extinction Risk Assessment System – SALVE (ICMBio, 2025) and in state decrees (Estado de São Paulo, 2014).

Order/Family	Species	WP	DP	Total	Voucher DZSJRP
CHARACIFORMES
Stevardiidae
	Mimagoniates lateralis (Nichols, 1913) VU	41	21	62	25059
Acestrorhamphidae
	Hyphessobrycon boulengeri (Eigenmann, 1907)	45	86	131	25058
	Hyphessobrycon bifasciatus Ellis, 1911	18	54	72	25070
	Rachoviscus crassiceps Myers, 1926 VU	1	8	9	23340
	Hollandichthys multifasciatus (Eigenmann & Norris, 1900)	1	–	1	25062
Erythrinidae
	Hoplias malabaricus (Bloch, 1794)	5	3	8	25065
Lebiasinidae
	Nannostomus beckfordi Günther, 1872*	29	6	35	25069
CICHLIFORMES
Cichlidae
	Geophagus iporangensis Haseman, 1911	4	–	4	25066
CYPRINODONTIFORMES
Poeciliidae
	Phalloceros reisi Lucinda, 2008	42	62	104	25064
	Poecilia reticulata Peters, 1859*	4	73	77	25068
Rivulidae
	Atlantirivulus santensis (Köhler, 1906)	86	126	212	25056
	Leptopanchax itanhaensis (Costa, 2008) CR	8	5	13	24845; 24846
SILURIFORMES
Callichthyidae
	Callichthys callichthys (Linnaeus, 1758)	45	4	49	25057
	Hoplosternum littorale (Hancock, 1828)	1	–	1	25173
	Scleromystax macropterus (Regan, 1913)	1	–	1	25060
Loricariidae
	Pseudotothyris obtusa (Miranda Ribeiro, 1911)	–	6	6	25073
SYNBRANCHIFORMES
Synbranchidae
	Synbranchus marmoratus Bloch, 1795	2	–	2	25072

Two exotic species appeared with notable abundance: Nannostomus beckfordi Günther, 1872 (N = 35), recorded for the first time in temporary habitats of the region, and Poecilia reticulata (N = 77), marking its first record in the basin. Although the majority of the native species are classified as Least Concern (LC), two are categorized as Vulnerable (VU) – M. lateralis and Rachoviscus crassiceps Myers, 1926, and one as Critically Endangered (CR) – Leptopanchax itanhaensis (Costa, 2008) (Tab. 2). Photographs of all species are provided (Fig. 3).

FIGURE 3| Species sampled in the roadside ditches from the Preto River microbasin, Itanhaém, São Paulo. A. Mimagoniates lateralis. B. Atlantirivulus santensis. C. Leptopanchax itanhaensis. D. Phalloceros reisi. E. Hyphessobrycon boulengeri. F. Hoplias malabaricus. G. Poecilia reticulata. H. Hollandichthys multifasciatus. I. Hoplosternum littorale. J. Nannostomus beckfordi. K. Rachoviscus crassiceps. L. Callichthys callichthys. M. Geophagus iporangensis. N. Hyphessobrycon bifasciatus. O. Scleromystax macropterus. P. Pseudotothyris obtusa. Q. Synbranchus marmoratus. Photographs of live specimens taken by Amanda Selinger during fieldwork.

Species richness increased more rapidly during the Wet Period compared to the Drier Period as sampling effort expanded. While species accumulation in the Drier Period showed signs of stabilization close to 10 interpolated samples, the Wet Period continued to show potential for additional species even after 15 samples (Fig. 4). Despite these patterns, the community composition did not differ significantly between periods (PERMANOVA: F = 1.08, R² = 0.03, p = 0.34; PERMDISP: F = 0.06 p = 0.80). This result is further supported by the NMDS, which revealed a partial overlap in species composition between the periods (Fig. 5A). However, some individual sampling points appeared to differ, with a slight association of C. callichthys, L. itanhaensis, and R. crassiceps with the Wet Period (Fig. 5B).

FIGURE 4| Rarefaction and extrapolation curves of fish species richness in roadside ditches during the Wet Period (WP) and Drier Period (DP) in the Atlantic Forest. Solid lines represent the rarefaction curves based on the observed data, while dashed lines indicate the extrapolated species richness beyond the observed sampling effort. Shaded areas correspond to the 95% confidence intervals.

FIGURE 5| Non-metric Multidimensional Scaling (NMDS) ordination of fish assemblages in roadside ditches during the Wet Period and Drier Period in the Atlantic Forest. A. Distribution of sampling sites for both periods, where overlapping points show greater similarity. Ellipses represent 95% confidence intervals for each period. B. Positioning of fish species within the ordination space.

The RDA revealed that environmental variables collectively explained only a modest portion of the variation in fish community composition, with RDA1 accounting for 14.44% and RDA2 for 5.31% of the total variance. Despite the relatively low explained variance, the observed patterns hold important ecological significance. For instance, increases in distance from the nearest stream were positively associated with the occurrence of P. reticulata, but showed a negative association with species such as M. lateralis and H. boulengeri. Atlantirivulus santensis exhibited a negative relationship with increasing water volume and a positive association with higher pH levels. Additionally, N. beckfordi and M. lateralis exhibited a small – but positive – association with higher levels of dissolved oxygen (Fig. 6). However, a significant positive correlation was found between spatial distance among samples and community dissimilarity (Mantel r = 0.31, p = 0.001), suggesting that geographically closer sites tend to have more similar fish communities (Fig. S2), where limited dispersion might exert an important role in structuring the community of these artificial temporary environments.

FIGURE 6| Redundancy Analysis (RDA) biplot showing the relationship between fish species composition and environmental variables in roadside ditches of the Atlantic Forest. Temp = Temperature (°C). DO = Dissolved oxygen (mg.L^-1). Volume = Volume in m³. Stream_dist = Distance to nearest stream (m).

Discussion​

---

In this study we provide the first description of the fish communities inhabiting roadside ditches in the Atlantic Forest, artificial and temporary habitats that, despite their instability, support a distinct and surprisingly diverse ichthyofauna. Even under extreme oscillations in physicochemical conditions, as indicated by dissolved oxygen and pH levels, and elevated temperatures during certain periods of the year, we recorded a total of 17 fish species, including threatened taxa and species likely introduced through human activities. Our results support the hypothesis that species richness increases during the Wet Period, likely due to enhanced hydrological connectivity with nearby streams. However, contrary to our expectations, species composition did not differ significantly between rainfall periods, regardless of the total richness observed. Finally, we also take initial steps toward understanding how environmental variables and spatial proximity between ditches shape fish communities in these overlooked ecologically unique environments.

The roadside ditches demonstrated harsh environmental conditions typical of extreme temporary habitats, particularly in terms of dissolved oxygen, which remained critically low throughout the year, averaging 0.32 mg/L during the Drier Period and 1.74 mg/L during the Wet Period. These values are substantially lower than those reported for blackwater streams in the region, which average 5.55 mg/L (Giongo et al., 2023), indicating a significant physiological challenge for fish dispersing from streams into ditches. Temperature oscillations were also marked, ranging from 16.14°C (DP) to 32.95°C (WP), far exceeding the thermal variability observed in nearby streams, ranging from 17°C to 22°C (Ferreira et al., 2014; Giongo et al., 2023). Notably, the maximum temperature recorded in ditches was only ~5°C below thresholds linked to mass fish mortality events in the Amazon (Braz-Mota, Val, 2024), suggesting that extreme summer heatwaves, as seen during 2024, might push many fish species close to their thermal limits. Although environmental conditions showed an overlap between periods (as indicated by the PCA), the Wet Period consistently exhibited, on average, higher oxygen levels, temperatures, and water volumes, likely reflecting the combined effects of increased rainfall and seasonal warming.

We recorded a surprisingly rich ichthyofauna for such a spatially restricted, ephemeral, and environmentally harsh habitat, with 17 fish species found in the roadside ditches. Of the 19 species previously documented in blackwater streams of the Preto River microbasin (Ferreira et al., 2014), 11 were also found in ditches, and more two other species that were reported in a stream fish zone that overlaps with the sampled area (H. littorale, N. beckfordi) (Ferreira, Petrere, 2009). Four additional species were recorded in ditches but had not been previously reported in community composition published in literature for this microbasin, detailed for the first time in this context (L. itanhaensis, R. crassiceps, P. reticulata, H. bifasciatus). These findings highlight the potential of artificial temporary habitats to contribute to regional species inventories, revealing both floodplain-associated taxa, highly threatened species and even introduced ones.

The most abundant species was A. santensis, a non-seasonal killifish broadly distributed along the São Paulo coast (Contente, Stefanoni, 2010). Although rarely abundant in stream communities (Ferreira, Petrere, 2009; Ferreira et al., 2014; Giongo et al., 2023), its high abundance in ditches suggests a preference for small flooded areas. The second and third most abundant species, H. boulengeri and P. reisi, respectively, are common in local streams and likely reach ditches during flooding, becoming trapped when water levels recede. Other stream-dwelling species, H. multifasciatus, G. iporangensis, H. littorale, S. macropterus and S. marmoratus, were observed only during the Wet Period, indicating that increased rainfall and flooding facilitate their entry into temporary habitats. This same aspect likely reflects the difference in accumulated species richness between periods, with the Wet Period exhibiting greater richness and delayed stabilization compared to the Drier Period, which supports our hypotheses.

However, remains unclear if this pattern reflects accidental drift or intentional use of flooded areas for reproduction, although previous studies suggest that some species may exploit these environments for reproduction (Espírito-Santo et al., 2013). Interestingly, P. obtusa was found only during the Drier Period. As a stream-associated species, its presence may reflect delayed movement or failure to detect environmental cues to exit temporary habitats, as observed in Amazonian species that avoid entrapment (Espírito-Santo et al., 2017). Such responses to hydroperiod and environmental variability can influence local species richness (Pazin et al., 2006) and likely play a role in shaping these ditch communities as well.

Although species richness varied between rainfall periods, our results did not support the hypothesis that species composition in roadside ditches differs significantly between the Wet and Drier periods, a pattern also reported in coastal streams (Tonhasca, 2005). This may be due to several factors. The first is that the Drier period is insufficient to generate compositional changes in the ichthyofauna. Although some ditches dry out, others remain even with reduced water volume due to year-round rainfall (Tonhasca, 2005), with the addition of new species being restricted to post-flood dynamics. Another factor that may be acting in combination is artificial water inputs from the urban area (such as well water derived from residents’ use, since much of the sampled region lacks a sewage system), which could be maintaining some volumes in the ditches near homes, masking possible changes in assemblage composition. However, the NMDS revealed non-random associations between certain species and the Wet Period, which, although not driving overall community turnover, suggest ecologically meaningful patterns. For instance, L. itanhaensis, an annual killifish recently rediscovered in this region (Costa et al., 2024a), was mostly found during the Wet Period. Although five individuals were recorded during the Drier Period, this may reflect the species life cycle, as even within the dry season, occasional rainfall events may trigger hatching.

A similar association was observed for R. crassiceps, a small tetra with a highly fragmented distribution and poorly understood ecology (Costa et al., 2024b), despite this, its total abundance was greater in the Drier Period, with the association found in the NMDS having to be interpreted with caution, since the species occurs in few sampling points, due to the very nature of its rarity, since it has not been reported for a long time in the region (Costa et al., 2024b). Despite its absence in regional stream surveys (Ferreira et al., 2014), it appears to favor flooded environments such as roadside ditches and temporary pools (Costa et al., 2024b), highlighting the need for further research on its habitat use, life history and persistence in ephemeral systems. Callichthys callichthys, another species associated with the Wet Period, possesses physiological traits such as aerial respiration – using intestine as air breathing organ – and terrestrial locomotion ability (Le Bail et al., 2000), which likely facilitate its dispersal across flooded landscapes during heavy rains.

The presence of threatened and endangered species in these artificial temporary habitats underscores the conservation relevance of considering these roadside ditches in species inventories. Species such as L. itanhaensis (Critically Endangered – CR), R. crassiceps and M. lateralis (Vulnerable – VU) were recorded, suggesting that these environments may act as refugees or transient habitats. This reinforces the need for inventorying and monitoring artificial water bodies, especially before road paving, drainage works or routine maintenance, as such activities may inadvertently eliminate important species. At the same time, the occurrence of these species in artificial habitats raises concerns. It may reflect the loss or degradation of natural temporary environments, such as temporary pools in restinga forests, forcing species to occupy suboptimal or exposed habitats. These environments have extreme physicochemical conditions and exotic species, as seen in our results, adding further stress to already vulnerable taxa. Future research should focus on assessing the impacts of biological invasions and the adaptive physiological or behavioral mechanisms that enable these species to thrive in such challenging habitats. Additionally, due to the nature of ditches, which are often located near dirty roads, they are directly exposed to surface runoff. This means they can accumulate diffuse contaminants present on the road, such as pesticides (Zhang et al., 2024), hydrocarbons (Liu et al., 2019), heavy metals (Pereira et al., 2007), and microplastics (Lin et al., 2024). This further raises the possibility that these environments may act as ecological traps. Despite supporting all these fish fauna, they also pose specific risks. The effect of chronic exposure to these pollutants on native and endangered species requires further investigation in future studies. Future regional inventories should also attempt to integrate the environmental quality of these ditches, as understanding this will be essential to investigate the balance between their use as refuges and ecological traps (with pollution risk), especially the effect of hydrological connectivity in such patterns.

The detection of the exotics N. beckfordi and P. reticulata, with the last one recorded for the first time in the entire Itanhaém River basin, raises additional conservation concerns. Notably, P. reticulata emerged as the fourth most abundant species in the community, highlighting its apparent tolerance to habitat degradation and pollution (Kennard et al., 2005; Silva et al., 2023). The RDA results further revealed a positive association between P. reticulata and greater distances from the nearest stream, suggesting that its presence is more frequent in more isolated roadside ditches, which are closer to areas with higher urbanization, where the occurrence of native species is likely lower. This spatial pattern may reflect intentional introductions, either by local residents or public authorities, as a mosquito control strategy (Miraldo, Pecora, 2017). Such introductions are particularly worrisome, as subsequent flooding events could facilitate the dispersal of P. reticulata into nearby temporary pools or streams, potentially disrupting native fish communities through competition or disease transmission.

Although the RDA explained a limited portion of the variation in community composition, some ecologically meaningful patterns emerged. Atlantirivulus santensis showed a negative association with ditch volume and a positive association with pH, a pattern that may reflect the characteristics of smaller ditches filled by recent small rainfall events, which are less influenced by acidic floodwaters from nearby blackwater streams. The use of shallow, low-volume habitats by A. santensis aligns with its non-annual life cycle and potential for terrestrial jumps to escape predators and competition, as observed in other rivulid species (Espírito-Santo et al., 2019). Its amphibious lifestyle may also confer an advantage in hypoxic, ephemeral environments prone to desiccation (Turko et al., 2021), allowing it to move terrestrially through jumps to environments that still retain water when its habitat begins to dry out (Espírito-Santo et al., 2019). However, these behavioral and physiological traits still require further investigation in the context of this species and region.

Other species, such as M. lateralis and H. boulengeri, exhibited a negative relationship with distance from the nearest stream and a positive association with higher dissolved oxygen levels, consistent with their ecology as typical stream dwellers in the region (Ferreira et al., 2014), and with their physiology as pelagic species with high swimming activity level that requires a higher metabolic rate to sustain its lifestyle (Campos et al., 2018). Their presence in roadside ditches likely reflects dispersal during flooding, especially near permanent water bodies. Although some of these stream dwellers species are acidophilic (e.g., M. lateralis and S. macropterus), and thus potentially sensitive to the higher pH observed in some ditches, our data suggest that this factor does not act as a short-term environmental filter. Nonetheless, the effects of such novel conditions on their fitness remain an important question for future research.

Although environmental filtering often plays a key role in structuring fish assemblages in coastal streams (Silva et al., 2023), our results suggest a different dynamic in roadside ditches. The low explanatory power of environmental variables in the redundancy analysis, combined with a significant positive correlation between geographic distance and community dissimilarity, indicates that dispersal limitation may be a dominant force shaping fish communities in these ephemeral habitats. For species lacking amphibious adaptations, colonization likely depends on complete flooding events, making geographically closer ditches more similar in species composition. In other cases, dispersal can be aided by eggs stuck to the feet and feathers of birds, or even through ingestion by birds, in the case of resistant diapause eggs of seasonal killifishes (Lovas-Kiss et al., 2020). This spatial structuring pattern resembles those observed in odonate communities in artificial ponds (McCauley, 2006) and in freshwater fish assemblages in lakes and reservoirs, where dispersal limitation has a greater influence on species richness than environmental heterogeneity (Drakou et al., 2009). While our findings represent an initial step toward understanding fish metacommunity dynamics in artificial temporary habitats, they also demonstrate the need to consider anthropogenic waterbodies as components of larger metacommunity networks when developing ecosystem management strategies, since human-modified landscapes are changing mechanisms of community assembly (Pecl et al., 2017), and these new systems need to be taken into account to result in effective ecological conservation. However, further research on the topic is urgently needed; studies focusing on both taxonomic and functional diversity, along environmental and spatial gradients, such as distance from streams, hydroperiod and temperature variation, are essential to clarify the mechanisms driving community assembly in these increasingly common and ecologically relevant systems. Another point that needs to be considered is the natural isolation of these coastal basins, which can be disrupted by artificial environments such as ditches, often associated with road networks, allowing for connectivity to neighboring hydrological basins during periods of high connectivity. This can result in a homogenization of regional ichthyofauna, either through the introduction of exotic species (Toussaint et al., 2016) or even the dispersion of native species endemic to each basin (Wang et al., 2025). This concern requires monitoring these ditches on a regional and temporal scale.

Altogether, our findings offer a foundational understanding of fish community composition in artificial temporary habitats within a blackwater basin from the Atlantic Forest, revealing that even human-made environments with extreme physicochemical conditions can harbor high species richness, including taxa of conservation concern. This study represents the first step toward understanding how fish utilize such habitats in the absence of natural environments, such as temporary pools. As these artificial systems may be acting as refugees or ecological traps, their role in regional biodiversity dynamics deserves urgent attention. In a world facing accelerating habitat fragmentation, climate-driven hydrological shifts and widespread loss of natural wetlands, exploring how fish communities persist and adapt in ephemeral human-modified landscapes is not only timely but essential for guiding effective conservation strategies.

Acknowledgments​

---

We thank the Núcleo de Pesquisas Hidrodinâmicas at Universidade Santa Cecília team for the support with rainfall data, and Thomas A. Vidal and Esli Mosna for their support during the fieldwork. We thank the associate editor and reviewers for their comments that helped improve the quality of the manuscript.

References​

---

Abilhoa V, Vitule JRS, Bornatowski H, Lara FB, Kohler GU, Festti L et al. Effects of body size on the diet of Rivulus haraldsiolii (Aplocheiloidei: Rivulidae) in a coastal Atlantic Rainforest island stream, southern Brazil. Biotemas. 2010; 23(4):59–64. https://doi.org/10.5007/2175-7925.2010v23n4p59

Armitage PD, Hawczak A, Blackburn JH. Tyre track pools and puddles – Anthropogenic contributors to aquatic biodiversity. Limnologica. 2012; 42(4):254–63. https://doi.org/10.1016/j.limno.2012.07.002

Baber MJ, Childers DL, Babbitt KJ, Anderson DH. Controls on fish distribution and abundance in temporary wetlands. Can J Fish Aquat Sci. 2002; 59(9):1441–50. https://doi.org/10.1139/f02-11

Berbel-Filho WM, Tatarenkov A, Pacheco G, Espírito-Santo HMV, Lira MG, Leaniz CG et al. Against the odds: hybrid zones between mangrove killifish species with different mating systems. Genes. 2021; 12(10):1486. https://doi.org/10.3390/genes12101486

Bokhutlo T, Cunha ER, Winemiller KO. Ephemeral habitat supports high fish α-diversity and β-diversity during drought in a subtropical semiarid wetland. Biotropica. 2021; 53(3):778–85. https://doi.org/10.1111/btp.12937

Braz-Mota S, Val AL. Fish mortality in the Amazonian drought of 2023: the role of experimental biology in our response to climate change. J Exp Biol. 2024; 227(17):jeb247255. https://doi.org/10.1242/jeb.247255

Brisson J, Blois S, Lavoie C. Roadside as invasion pathway for common reed (Phragmites australis). Invasive Plant Sci Manage. 2010; 3(4):506–14. https://doi.org/10.1614/ipsm-09-050.1

Camargo AFM, Cancian LF. Ecologia da bacia do rio Itanhaém: características limnológicas e uso do solo. In: Moraes MEB, Lorandi R, editors. Métodos e técnicas de pesquisa em bacias hidrográficas. Ilhéus: Editus; 2016. p.197–218.

Campos DF, Val AL, Almeida-Val VMF. The influence of lifestyle and swimming behavior on metabolic rate and thermal tolerance of twelve Amazon forest stream fish species. J Therm Biol. 2018; 72:148–54. https://doi.org/10.1016/j.jtherbio.2018.02.002

Castro RMC, Polaz CNM. Small-sized fish: The largest and most threatened portion of the megadiverse neotropical freshwater fish fauna. Biota Neotrop. 2020; 20(1):e20180683. https://doi.org/10.1590/1676-0611-BN-2018-0683

Contente R, Stefanoni M. Diet of the Atlantic rainforest killifish Rivulus santensis (Rivulidae, Cyprinodontiformes) in southeastern Brazil. J Appl Ichthyol. 2010; 26(6):930–32. https://doi.org/10.1111/j.1439-0426.2010.01465.x

Costa JHA, Duarte RM, Vidal TA, Souza UP, Langeani F. Rachoviscus crassiceps Myers, 1926. Bol Soc Bras Ictiol. 2024b; 144:51–55.

Costa JHA, Souza UP, Selinger A, Vidal TA, Langeani F, Duarte RM. Rediscovery of the critically endangered annual killifish Leptopanchax itanhaensis (Costa, 2008) in a temporary pool and roadside ditch from the Atlantic Forest of Southeast Brazil. Biota Neotrop. 2024a; 24(3):e20241637. https://doi.org/10.1590/1676-0611-BN-2024-1637

Costa WJEM. Peixes anuais Brasileiros: diversidade e conservação. Curitiba: Editora da UFPR; 2002.

Costa WJEM. Taxonomy and phylogenetic relationships among species of the seasonal, internally inseminating, South American killifish genus Campellolebias (Teleostei: Cyprinodontiformes: Rivulidae), with the description of a new species. Zootaxa. 2006; 1227:31–55. https://doi.org/10.11646/zootaxa.1227.1.2

Costa WJEM. Peixes aploqueilóides da Mata Atlântica Brasileira: história, diversidade e conservação. Rio de Janeiro: Série Livros 34, Museu Nacional, UFRJ; 2009.

Declerck SAJ, Coronel JS, Legendre P, Brendonck L. Scale dependency of processes structuring metacommunities of cladocerans in temporary pools of High-Andes wetlands. Ecography. 2011; 34(2):296–305. https://doi.org/10.1111/j.1600-0587.2010.06462.x

Drakou EG, Bobori DC, Kallimanis AS, Mazaris AD, Sgardelis SP, Pantis JD. Freshwater fish community structured more by dispersal limitation than by environmental heterogeneity. Ecol Freshw Fish. 2009; 18(3):369–79. https://doi.org/10.1111/j.1600-0633.2009.00354.x

Drawert HA. A new species of the seasonal killifish genus Moema (Cyprinodontiformes: Rivulidae) from the Piraí watershed in the Southwest Amazon basin. Neotrop Ichthyol. 2022; 20(4):e220067. https://doi.org/10.1590/1982-0224-2022-0067

Espírito-Santo HMV, Magnusson WE, Zuanon J, Mendonça FP, Landeiro VL. Seasonal variation in the composition of fish assemblages in small Amazonian forest streams: Evidence for predictable changes. Freshw Biol. 2009; 54(3):536–48. https://doi.org/10.1111/j.1365-2427.2008.02129.x

Espírito-Santo HMV, Rodríguez MA, Zuanon J. Reproductive strategies of Amazonian stream fishes and their fine-scale use of habitat are ordered along a hydrological gradient. Freshw Biol. 2013; 58(12):2494–504. https://doi.org/10.1111/fwb.12225

Espírito-Santo HMV, Rodríguez MA, Zuanon J. Strategies to avoid the trap: stream fish use fine-scale hydrological cues to move between the stream channel and temporary pools. Hydrobiologia. 2017; 792(1):183–94. https://doi.org/10.1007/s10750-016-3054-6

Espírito-Santo HMV, Sodré JG, Zuanon J. He leaps, she beats: The role of social interactions on the overland movements of an Amazonian amphibious killifish. Ecol Freshw Fish. 2019; 28(3):356–64. https://doi.org/10.1111/eff.12458

Estado de São Paulo. Decreto N^o 60.133, de 7 de fevereiro de 2014. Declara as espécies da fauna silvestre ameaçadas de extinção, as quase ameaçadas e as deficientes de dados para avaliação no Estado de São Paulo e dá providências correlatas [Internet]. 2014.

Fernandes R, Gomes LC, Pelicice FM, Agostinho AA. Temporal organization of fish assemblages in floodplain lagoons: The role of hydrological connectivity. Env Biol Fish. 2009; 85(2):99–108. https://doi.org/10.1007/s10641-009-9466-7

Ferreira FC, Petrere M. The fish zonation of the Itanhaém River basin in the Atlantic Forest of southeast Brazil. Hydrobiologia. 2009; 636:11–34. https://doi.org/10.1007/s10750-009-9932-4

Ferreira FC, Silva AT, Gonçalves CS, Petrere M. Disentangling the influences of habitat structure and limnological predictors on stream fish communities of a coastal basin, southeastern Brazil. Neotrop Ichthyol. 2014; 12(1):177–86. https://doi.org/10.1590/S1679-62252014000100019

Freitas TMS, Stropp J, Calegari BB, Calatayud J, Marco P, Montag LFA et al. Quantifying shortfalls in the knowledge on Neotropical Auchenipteridae fishes. Fish Fish. 2021; 22(1):87–104. https://doi.org/10.1111/faf.12507

Furness AI, Reznick DN, Springer MS, Meredith RW. Convergent evolution of alternative developmental trajectories associated with diapause in African and South American killifish. Proc R Soc B Biol Sci. 2015; 282(1802):20142189. https://doi.org/10.1098/rspb.2014.2189

Gibb AC, Ashley-Ross MA, Pace CM, Long JH. Fish out of water: terrestrial jumping by fully aquatic fishes. J Exp Zool A Ecol Genet Physiol. 2011; 315A(10):649–53. https://doi.org/10.1002/jez.711

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

Graham JB. Air-breathing fishes: evolution, diversity, and adaptation. San Diego: Academic Press; 1997.

Guedes GHS, Luz CHP, Mazzoni R, Lira FO, Araújo FG. New occurrences of the endangered Notholebias minimus (Cyprinodontiformes: Rivulidae) in coastal plains of the State of Rio de Janeiro, Brazil: populations features and conservation. Neotrop Ichthyol. 2023; 21(3):e230013. https://doi.org/10.1590/1982-0224-2023-0013

Hohausová E, Lavoy RJ, Allen MS. Fish dispersal in a seasonal wetland: Influence of anthropogenic structures. Mar Freshw Res. 2010; 61(6):682–94. https://doi.org/10.1071/MF09140

Hsieh TC, Ma KH, Chao A. iNEXT: an R package for rarefaction and extrapolation of species diversity (Hill numbers). Methods Ecol Evol. 2016; 7(12):1451–56. https://doi.org/10.1111/2041-210X.12613

Instituto Chico Mendes de Conservação da Biodiversidade (ICMBio). Sistema de avaliação do risco de extinção da biodiversidade – SALVE. 2025. Available from: https://salve.icmbio.gov.br/

Kennard MJ, Arthington AH, Pusey BJ, Harch BD. Are alien fish a reliable indicator of river health? Freshw Biol. 2005; 50(1):174–93. https://doi.org/10.1111/j.1365-2427.2004.01293.x

Le Bail PY, Keith P, Planquette P. Atlas des poissons d’eau douce de Guyane. Tome 2, Fascicule II: Siluriformes. Paris: Publications scientifiques du Muséum national d’Histoire naturelle; 2000.

Lin Y, Wang Y, Ho Y, Fang JK, Li Y. Characterization and ecological risks of microplastics in urban road runoff. Sci Total Environ. 2024; 954:176590. https://doi.org/10.1016/j.scitotenv.2024.176590

Liu A, Hong N, Zhu P, Guan Y. Characterizing petroleum hydrocarbons deposited on road surfaces in urban environments. Sci Total Environ. 2019; 653:589–96. https://doi.org/10.1016/j.scitotenv.2018.10.428

Lovas-Kiss A, Vincze O, Löki V, Paller-Kapusi F, Halasi-Kovács B, Kovács G et al. Experimental evidence of dispersal of invasive cyprinid eggs inside migratory waterfowl. PNAS. 2020; 117(27):15397–399. https://doi.org/10.1073/pnas.2004805117

Macedo-Soares PHM, Petry AC, Farjalla VF, Caramaschi EP. Hydrological connectivity in coastal inland systems: lessons from a Neotropical fish metacommunity. Ecol Freshw Fish. 2010; 19(1):7–18. https://doi.org/10.1111/j.1600-0633.2009.00384.x

McCauley SJ. The effects of dispersal and recruitment limitation on community structure of odonates in artificial ponds. Ecography. 2006; 29(4):585–95. https://doi.org/10.1111/j.0906-7590.2006.04787.x

Menezes NA, Weitzman SH, Oyakawa OT, Lima FCT, Castro RMC, Weitzman MJ. Peixes de água doce da Mata Atlântica. São Paulo: Museu de Zoologia da Universidade de São Paulo; 2007.

Miraldo MC, Pecora IL. Efficiency of Brazilian native ornamental fishes as mosquito larvae predators. Bol Inst Pesca. 2017; 43:93–98. https://doi.org/10.20950/1678-2305.2017.93.98

Oksanen J, Blanchet FG, Kindt R, Legendre P, Minchin P, O’Hara B et al. Vegan: an R package for community ecologists. R Package Version 2.6-4. 2022. Available from: https://cran.r-project.org/web/packages/vegan/index.html

Pazin VFV, Magnusson WE, Zuanon J, Mendonça FP. Fish assemblages in temporary ponds adjacent to “terra-firme” streams in Central Amazonia. Freshw Biol. 2006; 51(6):1025–37. https://doi.org/10.1111/j.1365-2427.2006.01552.x

Pecl GT, Araújo MB, Bell JD, Blanchard J, Bonebrake TC, Chen IC et al. Biodiversity redistribution under climate change: impacts on ecosystems and human well-being. Science. 2017; 355(6332):1–09. https://doi.org/10.1126/science.aai9214

Pereira E, Baptista-Neto JA, Smith BJ, McAllister JJ. The contribution of heavy metal pollution derived from highway runoff to Guanabara Bay sediments – Rio de Janeiro / Brazil. An Acad Bras Ciênc. 2007; 79(4):739–50. https://doi.org/10.1590/S0001-37652007000400013

Polačik M, Podrabsky JE. Temporary Environments. In: Riesch R, Tobler M, Plath M, editors. Extremophile fishes. Switzerland: Springer International Publishing; 2015. p.193–215. https://doi.org/10.1007/978-3-319-13362-1_9

Ramos TPA, Nielsen DTB, Abrantes YG, Lira FO, Lustosa-Costa SY. A new species of cloud fish of the genus Hypsolebias from Northeast Brazil (Cyprinodontiformes: Rivulidae). Neotrop Ichthyol. 2023; 21(3):e230068. https://doi.org/10.1590/1982-0224-2023-0068

Reis GS, Tejerina-Garro FL, Dagosta FCP, Teresa FB, Carvalho RA. Seeking for gaps in taxonomic descriptions of endemic fishes: a pathway to challenge the Linnean shortfall in a Neotropical basin. Neotrop Ichthyol. 2024; 22(2):e230128. https://doi.org/10.1590/1982-0224-2023-0128

Silva RGA, Pérez-Mayorga MA, Romero RM. Coastal freshwater stream fish fauna from a threatened estuarine lagoon complex in northeastern Brazil. Environ Biol Fishes. 2023; 106(12):2141–56. https://doi.org/10.1007/s10641-023-01495-x

Tonhasca A. Ecologia e história natural da Mata Atlântica. Rio de Janeiro: Editora Interciência; 2005.

Toussaint A, Beauchard O, Oberdorff T, Brosse S, Villéger S. Worldwide freshwater fish homogenization is driven by a few widespread non-native species. Biol Invasions. 2016; 18(5):1295–304. https://doi.org/10.1007/s10530-016-1067-8

Turko AJ, Rossi GS, Wright PA. More than breathing air: evolutionary drivers and physiological implications of an amphibious lifestyle in fishes. Physiology. 2021; 36(5):307–14. https://doi.org/10.1152/physiol.00012.2021

Turko AJ, Wright PA. Evolution, ecology and physiology of amphibious killifishes (Cyprinodontiformes). J Fish Biol. 2015; 87(4):815–35. https://doi.org/10.1111/jfb.12758

Vancine MH, Muylaert RL, Niebuhr BB, Oshima JEF, Tonetti V, Bernardo R et al. The Atlantic Forest of South America: spatiotemporal dynamics of the vegetation and implications for conservation. Biol Conserv. 2024; 291:110499. https://doi.org/10.1016/j.biocon.2024.110499

Wang J, Tao J, Brosse S, Liu C, Chen Q, Cheng G et al. Translocated species exceed alien species in homogenizing freshwater fish assemblages. Glob Chang Biol. 2025; 31(9):e70506. https://doi.org/10.1111/gcb.70506

Williams DD. The biology of temporary waters. London: Oxford: Oxford University Press; 2007.

Zhang Y, Li J, Wang J, Li Y, Kallenborn R, Xiao H et al. High-throughput screening of 222 pesticides in road environments in a megacity of northern China: a new approach to urban population exposure. Environ Res. 2024; 257:119379. https://doi.org/10.1016/j.envres.2024.119379

Authors

---

João Henrique Alliprandini da Costa^1,2 , Amanda Selinger², Francisco Langeani³, Ursulla Pereira Souza² and Rafael Mendonça Duarte^1,2

^[1]    Programa de Pós-Graduação em Biodiversidade de Ambientes Costeiros, Universidade Estadual Paulista “Júlio de Mesquita Filho” – UNESP, Instituto de Biociências, Praça Infante Dom Henrique, s/n, Parque Bitaru, 11330-900 São Vicente, SP, Brazil. (JHAC) jh.costa@unesp.br (corresponding author), (RMD) r.duarte@unesp.br.

^[2]    Laboratório de Biologia de Organismos Marinhos e Costeiros, Universidade Santa Cecília, Rua Oswaldo Cruz, 277, Boqueirão, 11045-907 Santos, SP, Brazil. (AS) amandaselinger@gmail.com, (UPS) upsouza@gmail.com.

^[3]    Laboratório de Ictiologia, Universidade Estadual Paulista “Júlio de Mesquita Filho” – UNESP, Instituto de Biociências, Letras e Ciências Exatas, Rua Cristóvão Colombo, 2265, Jardim Nazareth, 15054-000 São José do Rio Preto, SP, Brazil. (FL) francisco.langeani@unesp.br.

^[4]    Laboratório de Ecofisiologia e Evolução Molecular, Instituto Nacional de Pesquisas da Amazônia – INPA, Av. André Araújo, 2936, Petrópolis, 69011-830 Manaus, AM, Brazil.

Authors’ Contribution

---

João Henrique Alliprandini da Costa: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Resources, Writing-original draft, Writing-review and editing.

Amanda Selinger: Investigation, Methodology, Resources, Writing-original draft.

Francisco Langeani: Investigation, Methodology, Validation, Writing-review and editing.

Ursulla Pereira Souza: Conceptualization, Investigation, Supervision, Writing-review and editing.

Rafael Mendonça Duarte: Conceptualization, Funding acquisition, Investigation, Methodology, Resources, Supervision, Writing-original draft.

Ethical Statement​

---

All procedures were approved by the Ethics Committee for the Use of Animals at Universidade Estadual de São Paulo (UNESP), Bioscience Institute (CEUA – IB/CLP #15/2023), and sampling was authorized under SISBIO license #90241–1.

Competing Interests

---

The author declares no competing interests.

Data availability statement

---

The datasets generated during the current study are available in the https://github.com/JH-All/fish_community_roadside_ditches repository.

AI statement

---

The authors used generative AI (GPT-5) to improve english grammar. After using this service, the authors reviewed and edited the content as needed.

Funding

---

This study was financed, in part, by the Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP #2023/14344–5) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES Finance Code 001). We appreciate the aid from INCT-ADAPTA II project that is supported by CAPES (Finance Code 001), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq #465540/2014–7) and Fundação de Amparo à Pesquisa do Estado do Amazonas (FAPEAM #06201187/2017).

Supplementary Material

---

Supplementary material S1
Supplementary material S2

How to cite this article

---

Costa JHA, Selinger A, Langeani F, Souza UP, Duarte RM. Life on the road: fish communities composition in roadside ditches of the Atlantic Forest. Neotrop Ichthyol. 2025; 23(4):e250117. https://doi.org/10.1590/1982-0224-2025-0117

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 14, 2025

Submitted July 3, 2025

Epub February 02, 2026