Introduction

The mechanisms that explain changes in fish assemblage along a river have been well studied in both temperate and tropical rivers (Winemiller and Leslie 1992; Trujillo-Jiménez et al. 2002; Ibañez et al. 2009: Bhatt et al. 2012; Mercado-Silva et al. 2012; Mejía-Mojica et al. 2014; Carvajal-Quintero et al. 2015; Chea et al. 2016; Askeyev et al. 2017). Two hypotheses have been proposed to explain such changes: one that alludes to the concept of “biozonation”, and another that places emphasis on the addition of species. Both share the idea that environmental heterogeneity along the river is the main factor that drives changes in fish assemblage (Petry and Schulz 2006). As a result, the river continuum concept (sensu Vannote et al. 1980) was incorporated to explain the longitudinal changes in fish assemblages on temperate pristine rivers and streams, but have been strongly questioned (Statzner and Higler 1985; Miranda and Raborn 2000). However, this concept inspired theoretical and practical research in streams and rivers around the world (Statzner and Higler 1985; Ibáñez et al. 2011). Previous work on freshwater fish assemblages reported diverse relationships between species diversity and the longitudinal gradient, such as inverse relationships (Jaramillo-Villa et al. 2010), unimodal (optimal in mid-elevations) or nonlinear (without correlation) patterns (Rahel and Hubert 1991; Grenouillet et al. 2004; Li et al. 2009; Askeyev et al. 2017), and even an increase in species diversity in high altitude zones (Carvajal-Quintero et al. 2015). Nonetheless, a case of ecological convergence has recently been found between species richness (assemblage) in four continents with comparable environmental conditions (i.e., assemblage position in the stream’s longitudinal continuum, Ibáñez et al. 2011). It should be obvious that species richness in a river or stream depends on historical events (Ricklefs and Schluter 1993), but also on the constraints that environmental factors exert on them. For example, structural features of the rivers such as the width and depth of the channel, the diversity of the substrate, the order of the currents, etc., as well as biotic relationships such as predation, competition and disease, have been associated with different fish assemblage patterns (Matthews 1998; Tejerina-Garro et al. 2005). In the rivers of northeastern Mexico, such studies are rare, and it is important to know the changes in fish assemblage along the river due to human activity.

The Río Pánuco is located between the Sierra Madre Oriental and the Gulf of Mexico in the Tamaulipean ecoregion of northeastern Mexico (Contreras-Balderas 1969; CONABIO 2000). The main tributary of this river is the Rio Guayalejo-Tamesí system (hereinafter referred to as the river system), which drains an area of 15,257 km2 (INEGI 1983). Darnell (1962) studied the fish community structure of this system; his inspection provided the first extensive inventory (60 species; 23 freshwater and 37 peripheral, of which 12 were exclusive or endemic). However, since this first inspection, no comprehensive studies are available on the fish along the river or their relationship with environmental variables. The basic knowledge of species composition contrasts with the deterioration of habitats, especially in the middle and lower parts of the basin (García-De León et al. 2005). Changes in water flow and the introduction of exotic species pose major threats to native fish assemblages (Hughes et al. 2005; Hermoso et al. 2011; Mendoza et al. 2014). Our study provides updated baseline information that may be useful for future conservation of fish fauna, land use planning and holistic watershed management. We are interested in knowing how the fish assemblage is determined by environmental factors along the channels’ longitudinal gradient and report the new species introduced in the system after Darnell’s (1962) inventory.

Material and methods

Study area

The Río Guayalejo-Río Tamesí system is located in east-central Mexico, in the northern part of the Tampico embayment. In the south, this embayment is drained by the lower section, the Río Pánuco, where these two branches join (22°13′N, 97°51′W) before entering the Gulf of Mexico (INEGI 1983). The northernmost tributary, the Río Guayalejo, originates in the Sierra Madre Oriental at 3400 m. The Sabinas, Frío, Boquilla, and Mante rivers join to form the Río Tamesí, which assumes a meandering course that borders the states of Tamaulipas and Veracruz (Fig. 1). Tributary rivers start as springs, large ponds or wetlands in shallow valleys in the foot-slopes of the Sierra Madre Oriental.

Fig. 1
figure 1

Sites along the Río Guayalejo-Río Tamesí system. See Table 1 for codes

Full size image

Fish collection

Fish were collected in the river system from February 2000 through July 2001. The study area was divided into three elevation zones: mountain and high elevations (200–1500 m), plains and low elevations (50–200 m), and river mouth (0–50 m). The sampling sites are described in Table 1. Ejido Aldama, located ~60 km north of the main course of the river system, was included because it does not belong to any of the major basins of Tamaulipas. This study was conducted at the Technological Institute of Ciudad Victoria in Tamaulipas, where despite the lack of any animal care protocol, the animals were all collected under standard care procedures.

Table 1 Sampling sites for fishes and habitat assessment along the Río Guayalejo-Río Tamesí system, Tamaulipas, Mexico in 2000–2001
Full size table

To collect fish, a combination of fishing gears was used, following standard fish-collecting procedures (Sooley et al. 1998). In the mountain zone, fish were collected with minnow seines, and supported with electrofishing equipment for an average collecting time of 30 min for each sample. In the plains zone, we used electrofishing gear, a minnow seine, and experimental gillnets 180 m long and consisting of eight panels, each 23 m long with 1, 1.5, 2, 2.5, 3, 3.5, 4, and 5-in. meshes. At these sites, the experimental gillnets remained in place for 12 h. In the mouth zone, where the channels are deeper, we used a trammel net (180 m long with a 3-in. internal mesh) and the experimental gillnets described above, both of which were in place for 12 h. The elevation and coordinates of each site were measured with a GPS unit.

The fish were anesthetized and subsequently fixed in 10% formalin neutralized with sodium borate and then preserved in 50% isopropanol solution. The identification of each species was based on Álvarez del Villar (1969), Castro-Aguirre et al. (1999), Page and Burr (1991), McEachran and Fechhelm (1998), and Miller et al. (2005), as well as specific literature for some taxa.

Alkalinity, pH, dissolved carbon dioxide, hardness (amount of dissolved calcium and magnesium in the water, mg/l, Wetzel 2001), transparency, total dissolved solids, and temperature were measured at each site; conductivity (μS/cm) and dissolved oxygen (mg/l) were measured with a conductivity meter and an oxygen meter, respectively. Salinity (ppt) was measured with a refractometer. Water samples collected in each site were fixed immediately in order to determine the quantity of nutrients (ppm) as nitrate, nitrite, and phosphate, all of which were measured with a spectrophotometer (SMART model, LaMotte, Chestertown, MD). To characterize environmental heterogeneity at each site, we considered 6 quantitative and 17 qualitative variables (Moyle and Nichols 1973; Table S1).

Statistical analysis

To describe fish habitats, the presence-absence of each species was recorded (Table S2) and used to quantify the similarity in species composition between sampling sites, using an agglomerative hierarchical two-way cluster analysis. For this analysis, we determined the Euclidian distances between sampling sites and a complete linkage-clustering algorithm, using PRIMER6 software (Clarke 1993; Clarke and Warwick 2001, 2005). The significance of the clusters was tested by similarity profile analysis (SIMPROF, PRIMER6). To detect significant differences in the dissimilarity values between fish habitats, we applied a one-way analysis of similarity of the significance of the habitats defined a priori (ANOSIM). The average similarity and percentage of contribution of each species to the identity of each habitat were determined using the similarity percentage routine (SIMPER) in PRIMER 6 (Clarke and Warwick 2001).

A canonical correspondence analysis (CCA) was performed to define the relationships between each type of habitat and environmental variables. The environmental matrix contained the average of the physicochemical parameters and qualitative variables of the environment that had the highest value in explaining the variables at the 95% confidence level, in an initial CCA (Table 2a). These variables were the maximum depth of the channel, the percentage of water riffles, hardness, the average depth, the percentage of backwater, channel width, and the percentage of floating macrophytes. Diversity (Shannon index), dominance, and evenness of species in each habitat were calculated using PAST 2.07 diversity software (Hammer et al. 2001).

Results

We collected 5918 fish belonging to 23 families, 35 genera, and 48 species, taken at 27 collection sites in the study area (Tables 1 and S2). Of 48 species, 43 were native and five were exotic (Cyprinus carpio, Ctenopharyngodon idella, Hypophthalmichthys molitrix, Micropterus salmoides, and Oreochromis aureus). Exotic species were only present in some places of lower elevation (Micropterus salmoides and Oreochromis aureus) and in virtually all the sites of the mouth region (five species). Exotic species were more numerous in the mouth area, but none was important in terms of abundance with respect to native species (Fig. 2). Globally, four species accounted for 75.9% of abundance (Astyanax mexicanus, Gambusia vittata, Poecilia mexicana, and Herichthys labridens, Table S2). Most river species were restricted in distribution along the longitudinal gradient; 43 species were found at fewer than six sites. Five species were found at many locations along the river: Astyanax mexicanus (24 sites), Poecilia mexicana (19 sites), Herichthys labridens (16 sites), Gambusia vittata (11 sites), and Poecilia formosa (11 sites) (Table S2).

Fig. 2
figure 2

Species richness and abundance along the Río Guayalejo-Río Tamesí system in northeastern Mexico. The upper figure represents the elevation of each sampling site

Full size image

Fish habitats

The dendrogram of fish species similarity defined four fish habitats (groups of sampling sites) with a separation of 3.6% distance (Fig. 3). These groupings were statistically significant (ANOSIM: R = 0.519, P < 0.05); all pairwise comparisons were also statistically significant (Table S3).

Fig. 3
figure 3

Dendrogram of fish species similarity (presence-absence) based on Euclidian distances between sampling sites in the Río Guayalejo-Río Tamesí system. Red clades indicate significant clusters using the SIMPROF permutation test (5% significance level, 999 permutations) of all specimens

Full size image

Habitat 1 (marine habitat) was mainly represented by marine species and was the most distinct. It included one sampling location (Fig. 3; Tables 1 and 3) of high salinity (Table S4). Habitat 2 (mountain and plain habitat) included eight sites in the mountain zone, six sites in the plains zone, and one site in the mouth zone, and contained mainly freshwater and peripheral species (Fig. 3; Tables 1 and 3). Habitat 3 (close to the mouth) included a few freshwater species and a number of peripheral species at two sites close to the mouth (Fig. 3; Tables 1 and 3), and habitat 4 (plain and mountain habitat) was similar to habitat 2, containing five sites in the plains, three at the mouth, and one in the mountain zone, and was essentially characterized by freshwater fish (Fig. 3; Tables 1 and 3).

Table 2 Relative contribution of each variable by the canonical correspondence analysis (a) and percentage of variance explained by each axis (b)
Full size table
Table 3 Similarity percentages–species contribution (SIMPER) estimated from Euclidean distances and abundance data
Full size table

Since habitat 1 was characterized by high salinity and dominated by marine species, and some of them (euryhaline species) were distributed upstream, it was not included in the CCA. The CCA had three components that accounted for 54.2% of the total variation, considering seven assessed environmental variables (Table 2b). Variation in hardness and percentage of riffles were positively correlated, and channel (width and depth) with percentage of backwater were negative on the first axis. This axis shows no separation between the three habitats (Fig. 4).

Table 4 Level of diversity (H′), dominance (D), and evenness (E) of species per sampling site and fish assemblage along the Río Guayalejo-Río Tamesí. The codes indicate sampling sites and are defined in Table 1. H1, H2, H3, and H4 are the different fish habitats
Full size table
Fig. 4
figure 4

Canonical correspondence analysis showing the effect of the environmental variables on the abundance of fish habitats. Black arrows indicate the environmental variables. Symbols are fish habitats

Full size image

On the second axis, the percentage of floating macrophytes was positively correlated, while maximum depth and average depth were negatively correlated (Fig. 4). This axis shows a better separation between the sites of habitat 4 (correlated with a greater presence of macrophytes) and those of habitat 2, which show greater correlation with maximum depth. Sites in habitat 2 show a higher average maximum depth (2.9 m) in relation to habitat 4 (average 1.27 m). Habitat 3 had sites with broader channels and high percentages of backwaters.

The highest diversity (average Shannon index; H′ = 2.19) and the lowest dominance (average D = 0.17) and evenness (average E = 0.47) were found at sites near the mouth of the river (Table 4); habitat 1 at Puente Moralillo, Table 1). In contrast, the highest dominance (average D = 0. 58) and evenness (average E = 0.58), and the lowest diversity (average H′ = 0.82), were recorded in habitat 2 at higher elevations (average 349 m), where Gambusia vittata and Astyanax mexicanus were the most abundant species (Table 4).

Table 5 Abundance and frequency of occurrence of fish species in the Río Guayalejo-Río Tamesí system in 2000–2001. Absolute abundance (AA) and relative abundance (RA), standard deviation (SD), frequency of occurrence (F). H1, H2, H3, and H4 were determined by cluster analysis (dendrogram)
Full size table

Discussion

Darnell’s (1962) ground-breaking study of fishes from Río Guayalejo and its tributaries was carried out by field collections and complemented with previous records from Jordan and Dickerson (1908). However, Darnell’s study did not assess the component of brackish and marine fishes entering coastal lagoons and near river mouths, or their seasonal variation. Furthermore, Darnell lacked taxonomic revisions for some freshwater fish taxa, mainly in the families Cyprinidae and Poeciliidae. All these reasons account for the differences observed between our study and that of Darnell. Indeed, one notorious change is the occurrence of exotic species in the system (Table S5). The higher abundance and richness of exotic species occurred in sections of the river near the mouth with warmer waters and greater maximum depths and channel width (Table S2). It is important to mention that Darnell (1962) did not record the presence of exotic fish species in the study area due to the high levels of introduction of exotic fishes in Mexico during the 1970s and 1980s (Contreras-Balderas and Escalante 1984; Contreras-Balderas et al. 2004, 2008). We believe that exotic species may have been introduced in the river system in the 1970s, which means that these species have already been interacting with native species for at least 30 years – long enough to disperse throughout the system. However, invasion of the whole river system did not seem to have occurred, at least during sampling. At higher elevations, no exotic species was observed. It is possible that species of tropical origin, such as tilapia, show low yields in northern areas of Mexico with low temperature during winter (Kapetsky 1997), which limits the breeding activity of this African cichlid fish.

Fish habitats and their relationship with environmental factors

Cluster analyses identified four major fish habitats in the river basin. A very different euryhaline marine fish habitat (habitat 1) occurs downstream near the mouth of the river with interchange of flows between the sea and river. Three species (Elops saurus, Bairdiella chrysoura and Brevoortia gunteri) distinguish this type of habitat located in Puente Moralillo, all of which are of euryhaline marine origin and penetrate to coastal lagoons or rivers with sandy bottoms to feed and reproduce (Table S6). Elops saurus and Bairdiella chrysoura feed on small fish and crustaceans, and Brevoortia gunteri feeds on plankton. This habitat also contains two exotic cyprinids in low abundance (Cyprinus carpio and Hypophthalmichthys molitrix). These are related to benthos and are tolerant, over brief periods, to backwaters and low concentrations of oxygen. The former is omnivore and the latter feeds on phytoplankton and zooplankton (Freyhof and Kottelat 2008; Zhao 2011).

Two other habitats (2 and 4) consist essentially of freshwater fish that are distributed along the longitudinal gradient of the channel, but are distinguished by the greater or lesser percentage of floating macrophytes, and the depth of the channel (Fig. 4). Both habitats include generalist species (Astyanax mexicanus, Poecilia mexicana and Herichthys sp.) of neotropical origin as well as species typical to all Atlantic basins, with a high dispersion capacity (Miller and Smith 1986; Miller et al. 2005). This last fish species inhabits practically all the environments along the longitudinal gradient of the river; other species such as Gambusia vitatta show more restricted distribution upstream, in an area characterized by high levels of hardness and presence of riffles or Xiphophorus variatus confined to shallower environments with floating macrophytes (Fig. S1). However, downstream, habitat 4 includes amphidromous species, in which adults are strictly freshwater (Agonostomus monticola and Gobiomorus dormitor, McDowall 1997; Ribeiro and Villalobos 2010), and exotic species with low abundance (Micropterus salmoides and Oreochromis aureus), while habitat 2 has a clear influence of euryhaline marine fish (Dorosoma petenense and Megalops atlanticus) (Table S6). The frequency of occurrence and abundance of some species (Gambusia vittata, Astyanax mexicanus and Xiphophorus variatus) also contributed to the differences between the two habitats (Table 5). Habitat 3 is also composed of generalist freshwater species (Astyanax mexicanus), but shows a large proportion of euryhaline marine species (Dorosoma petenenese, Ariopsis felis, Centropomus undecimalis, and Gobiomorus dormitor) and three exotic cyprinids in sites close to the mouth of the river (Cyprinus carpio, Ctenopharyngodon idella, and Hypophthalmichthys molitrix) with tolerance to the salinities of the estuarine environment (Cudmore and Mandrak 2004; Freyhof and Kottelat 2008; Zhao 2011); however, the habitat also exhibits low species richness and abundance.

Various environmental, geographic and topographic features are often described as determinants of patterns of species richness along longitudinal gradients (Ricklefs and Schluter 1993; Matthews 1998; Tejerina-Garro et al. 2005). How these factors impact the richness of species has been debated (Benke et al. 2011). Our study, the fish species diversity decreases with high elevations, a pattern commonly reported by other authors (Edwards and Contreras-Balderas 1991; Jaramillo-Villa et al. 2010). In order to understand this, we evaluated abiotic factors and found that factors on an intra-basin scale such as salinity, depth, percentage of riffles and backwaters, hardness, channel width and percentage of floating macrophytes could prevent colonization and limit upstream and downstream dispersion, which could act as ecological barriers or filters (Contreras-Balderas 1969; Ruiz-Campos et al. 1985; Contreras-Balderas et al. 2002). For example, Astyanax mexicanus and Herichthys cyanoguttatus are generalist species with dietary habits related to the availability and diversity of food in the particular habitat (Miller 1966; Mitchell et al. 1977). These characteristics allowed them to have an extensive distribution and high abundance, in particular for Astyanax mexicanus, what could be interpreted as species with high tolerance to different environmental conditions, while other species with less environmental tolerance, such as Gambusia affinis and G. regani, are restricted to lenthic biotopes or weedy backwater conditions with virtually no current (Darnell 1962; Miller et al. 2005, Table S6). In addition, many freshwater fish species are bentho-pelagic but have different food requirements; for example, the most representative species of habitat 2, Astyanax mexicanus, feeds on insects, crustaceans and worms; others, such as Gambusia vitatta and Herichthys labridens, are omnivores, and Poecilia mexicana feeds mainly on detritus; this feeding partitioning exhibited by all these species enables them to avoid competition and allows stability in the fish community along the river (Miller and Smith 1986; Tejerina-Garro et al. 2005; Hoeinghaus et al. 2007; Winemiller et al. 2008). The same situation occurs in species of marine origin that enter the river for feeding (Table S6).

In summary, the structure of the fish community of the Guayalejo-Tamesí river system is represented by one upstream fauna composed of mostly freshwater species – some of them generalists that inhabit the entire longitudinal gradient, and others that inhabit more restricted areas – and a downstream assemblage composed of a mixture of the more abundant upstream elements and species of marine origin, both of which are tolerant to brackish water near the mouth. The low abundance of exotic species and the richness of upstream and downstream freshwater species may indicate few changes in this type of rivers in the region (Ruiz-Campos et al. 1985; Edwards and Contreras-Balderas 1991).

The river continuum concept establishes that physical conditions of streams or rivers from upstream to downstream areas create strong constraints on assemblage structure linked to food availability (Vannote et al. 1980). However, the original concept was modified to make it flexible and more applicable (Statzner and Higler 1985). One modification that emerged with studies of modern ecology is that physical connectivity may account for differences in freshwater assemblages between sites (Miranda and Raborn 2000). Recently, Ibañez et al. (2009) provided evidence, from the comparative study of headwater streams in four continents, that the richness of fish assemblage and trophic structure converged along the continuous river to a substantial degree. The Guayalejo-Tamesí river system seems to be consistent with this principle, in particular with regard to species addition and replacement and the occurrence of indicator species along a longitudinal gradient. These findings have been reported by other studies (Grenouillet et al. 2004; Petry and Schulz 2006; Grosmann et al. 2010; Bhatt et al. 2012; Carvajal-Quintero et al. 2015; Chea et al. 2016; Silva et al. 2016; Askeyev et al. 2017).

In terms of conservation, the least disturbed area is the upper part of the river system (habitats 2 and 3); however, without proper regulation and enforcement, tourism and mining could seriously affect fish diversity (García-De León et al. 2005). In the plains and mouth (habitats 1, part of 2, and 3), fish diversity has been heavily affected by several dams, exotic species, pollution with agricultural fertilizers and pesticides, ranching, siltation, and pollution associated with petroleum operations (García-De León et al. 2005; Rodríguez Rodríguez et al. 2012).

The Mexican government continues to deliver programs aimed at introducing exotic fish (Ibáñez et al. 2011). According to our results, exotic species might have a lower impact; however, the effects of exotic species on freshwater ecosystems are widely known (Moyle and Nichols 1973; Contreras-Balderas and Escalante 1984; Mendoza et al. 2014), so it is advisable to continue to monitor the system, since other factors may emerge with long-term ecological data (Grosmann et al. 2010). Unfortunately, at both the federal and state level, the programs that exist in this region for conserving natural resources and maintaining the biodiversity of fish are practically nonexistent (Gobierno de Tamaulipas 2011). Extensive work on the part of the Mexican government is necessary to promote comprehensive strategies in using and managing aquatic resources. This should begin with establishing a baseline of information on the biological diversity and environmental services provided by balanced riverine ecosystems.