The contribution of environmental factors to fish assemblages in the Río Acahuapa , a small drainage in Central America

The ichthyofauna of the Río Acahuapa was analyzed sampling 17 sites that included the basin main channel and its tributaries. Fish were collected using dip-nets, seine-nets and electrofishing. Fish standard length and species abundance were recorded. Species origin and salinity tolerance criteria were used to classify fish species. Water physicochemical variables, habitat structure and sampling sites elevation were recorded. A total of 33 fish species were registered, 12.1% are primary, 45.5% are secondary and 42.4% are of marine derivation. Fish species richness declined with increase of elevation (R2=0.55, p=0.0006). Two assemblages of fishes were identified: the first one associated to sites of low elevations (19-184 masl), composed mainly of secondary and marine-estuarine fish species related with high temperature, water velocity, river width, dissolved oxygen and low sand and silt substrate cover; the second one associated to sites of middle and higher elevations (185-519 masl), composed by primary and secondary freshwater fishes related with high pH, logs and rocks substrate cover. In summary, elevation and environmental variables contributed to the composition and distribution of fish in the Río Acahuapa.


Introduction
The topographic and climatic conditions are important determining geomorphological traits of a basin (Winemiller et al., 2008).Factors such as elevation, slope and rain regimes mainly influence the dynamics of currents in river systems and create habitat variability along rivers longitudinal gradient (Bussing, 2002;Winemiller et al., 2008;Esselman, Allan, 2010).Some authors explain that this habitat variability determines zonation patterns of species (Rahel, Hubert, 1991).For example, fishes adapted to high water velocities habitats, will occur in rapids zones of the rivers, while species that cannot tolerate high water velocities habitats, will occur in zones of backwaters or pools (Bussing, López, 1977).Likewise, many of these geomorphological characteristics of the basin are also influencing water physicochemical factors, such as: dissolved oxygen, pH, temperature, salinity, and conductivity, which can have an effect on fish communities (Bussing, López, 1977;Wootton, 1992).Moreover, landscape factors such as the watershed area or land-use activities can be strongly correlated with the composition, structure, and distribution of species (Schlosser, 1991;Esselman, Allan, 2010;Junqueira et al., 2016).

e170023[2]
Evidence suggests that composition of species may change in a predictable way due to changes in environmental conditions driven primarily by changes along the river continuum (Vannote et al., 1980;Lyons, Schneider, 1990;Lorion et al., 2011).Others suggest that proximity or connectivity with marine or brackish waters may be strong factors that may explain changes in the composition and distribution of species (Esselman et al., 2006;Mercado-Silva et al., 2012).Consequently, these factors are important to explain the dynamics of the fish species, mainly in rivers that have connectivity with brackish waters or altitudinal gradients.However, there are biotic factors that play an important role within the species communities.For example, factors such as migration, competition, predation and mutualism can affect species composition and changes in abundances in a river (Wootton, 1992;Matthews, 1998).The sum of the combined effect of abiotic and biotic factors, may be the main force determining fish communities structure (Schlosser, 1991;Jackson et al., 2001).
The majority of information about the relationships between freshwater fish communities and biotic and abiotic factors is mainly generated in temperate areas (e.g., Wootton, 1992;Matthews, 1998).For the neotropics, information about fish communities is scarce, even more in the Central American region (Lowe-McConnell, 1987;Winemiller et al., 2008;Junqueira et al., 2016).However, in the last decades, some researchers are aiming to improve the understanding of the freshwater fish communities in the area.As a result, there are taxonomic inventories of freshwater fishes for almost all Central American countries (Greenfield, Thomerson, 1997;Kihn-Pineda et al., 2006;Matamoros et al., 2009;Angulo et al., 2013;McMahan et al., 2013).Nonetheless, it is notable a shortage of ecological studies in most of the Central America countries, but Costa Rica and Panama (e.g., Bussing, López, 1977;Angermeier, Karr, 1983;Wootton, Oemke, 1992;Lorion et al., 2011).Particularly in El Salvador a handful of published studies on freshwater fish have been generated (e.g., Hildebrand, 1925;Álvarez et al., 2013;McMahan et al., 2013), focusing on the distribution of species, while the ecological aspects of freshwater fish communities have been neglected.
For the basin of the Río Acahuapa, there are no publications about ichthyofauna in the area and this study represents the first assessment of the influence of environmental factors on freshwater fish assemblages in El Salvador.Therefore, the main goal of this study was to characterize the composition of species in Río Acahuapa basin.To identify the main environmental factors that affect the composition and distribution of the freshwater fish of the río Acahuapa.Likewise, we evaluated the effect of the elevation with regards to species richness in the basin.It is expected that this information will contribute to the understanding of the factors that locally affect the ichthyofauna and that it can support future studies and management strategies and conservation of biodiversity in the basin of the Río Acahuapa.

Material and Methods
Study area.The Río Acahuapa is located in the Department of San Vicente in the Pacific slope of El Salvador (Fig. 1).The Río Acahuapa belongs to Río Lempa drainage system, the most important and the largest basin of El Salvador.The basin area of the Río Acahuapa is 238.88 km 2 and the length of the river is 35.3 km (Hernández et al., 2010).The basin of the Río Acahuapa has no influence of brackish water, its mouth connects with the Río Lempa 45 km away from the sea.The weather in the area is tropical and the dry season spans from November to April and the rainy season from May to October (McMahan et al., 2013).Most of the Río Acahuapa watershed area is used for livestock and agriculture, which includes crops such as: coffee, sugarcane, several kinds of fruits, vegetables, beans and corn.Also, the Natural Protected Area of La Joya-Sisimico is located within the boundaries of the basin, the most important conservation area of San Vicente Department (Hernández et al., 2010).
Sampling sites.Along the river, 17 sites were sampled; six points were located in the main channel and 11 points in tributaries (Fig. 1).Seven sampling sites were located from 19 to184 m, seven from 185 to 351 m and three locations from 352 to 519 m of elevation.All the sites were sampled in each of the five sampling campaigns that were performed in June, August, andNovember 2011, andFebruary andMay 2012 (Tab. 1).
To determine the sampling area, first we estimated the mean width of the river in each sampling point.Then, the mean width of the river was used to determine the length of the river that was going to be sampled following the criteria: a) if the mean river width was < 2 m = 50 m length of the river were sampled, b) if river width was between 2-5 m = 75 m of the river were sampled, c) if the river width was between 5-10 m = 100 m of the river were sampled, d) if the river width was between 10-20 m = 200 m of the river were sampled, e) if the river width was between 20-30 m = 300 m of the river were sampled.Due to the lack of standardized protocols for the collection of fish in freshwater ecosystems in El Salvador, this methodology was used to standardize the sampling length due to the differences in size among rivers, as some authors argue that increasing the length of the transect based on the river width enhances the probability of catching specimens in the transect and favors species richness comparisons (Lyons, 1992;Hughes et al., 2002;Watkins et al., 2016).
Fishing gear and preservation.Electrofishing (one backpack electrofisher SAMUS 725MP operated by one person and accompanied by two people), a seine-net and dip-nets were used to catch the fishes.For sampling, zigzag sweeps were performed along the transect width in the opposite direction of the current.All the rivers and streams were wadable and it was possible to perform electrofishing.However, due to habitat variability, the sampling effort was standardized until a maximum of one hour of electrofishing per site, except for the smaller sites, where a single F. S. Álvarez, W. A. Matamoros & F. A. Chicas Neotropical Ichthyology, 15(3): e170023, 2017 3 e170023 [3] sweep with electrofishing was carried out in the transect (Angermeier, Smogor, 1995).The captured specimens were identified in the field using the identification keys of Miller et al. (2009) and Bussing (2002).The standard length (SL) was measured before they were released.Each species was classified by salinity tolerance criteria and the species origin following Myers (1949a) criteria.For migratory species (sensu lato McDowall, 2008), usually sub divided in anadromous, catadromous and amphidromous (Myers, 1949b;McDowall, 2008;Acolas, Lambert, 2016), we used only the single category: diadromous; considering direct and/or indirect evidence on these species available up to the date (Cruz, 1987;Lyons, 2005;Ribeiro, Villalobos, 2010;Lorion et al., 2011;Smith, Kwak, 2014a, 2014b).Fish from marine derivation, that can be considered marine-stenohaline or marine-eurihaline (Espinosa-Pérez et al., 2011), were grouped in a single category of marine-estuarine species.This classification was made under a bibliographic revision available for each of the species (Castro-Aguirre et al., 1999;Bussing, 2002;Lyons, 2005;Miller et al., 2009;McMahan et al., 2013;Smith, Kwak, 2014b).Some individuals were kept and deposited in the scientific collection of the Instituto de Ciencias del Mar y Limnología of the Universidad de El Salvador (ICMARES-UES 462-472, 505-538; S1 -Available only as online supplementary file accessed with the online version of the article at http://www.scielo.br/ni).
Environmental variables.In each sampling point three transversal transects were deployed.Within each transect the full set of physicochemical parameters (temperature (ºC), conductivity (uS/cm), salinity (ppt), dissolved oxygen (mg /l), pH), habitat structure (water velocity, depth, river width) and substrate cover (rocks, sand, logs, silt and leaves) were measured in the right-shore, middle and leftshore of the river.The water physicochemical parameters were measured using a YSI Multiparameter probe Model 8510.The width and depth of the river were measured using a measuring tape and the water velocity was recorded with a portable flow meter.The percentages of the substrate coverage were estimated with the use of a quadrat of 100 squares of 4 cm x 4 cm each.We counted the number of squares per substrate and recorded the average percentage of each substrate.The elevation was obtained in each sampling site using a GPS.All the physicochemical parameters and habitat structure variables were measured at each sampling point in each sampling campaign and the averages of each variable for each transects were determined (Tab.1).

e170023[4]
Statistical analysis.To analyze the relationships between environmental variables and sampling sites, a Principal Component Analysis (PCA) was performed.The PCA was carried out using all data of water physicochemical variables (temperature, dissolved oxygen, salinity, pH, and conductivity), habitat structure (water velocity, depth, river width, and substrate types) and elevation (meters above sea level).All the data of the environmental variables were standardized (centering and scaling) to reduce the effect of the measurement units in the PCA analysis.
Subsequently, a Canonical Correspondence Analysis (CCA) was performed to explore the relationship between the fish abundances and the environmental factors (water physicochemical variables, habitat structure and elevation).To describe the environmental factors in the two axes of CCA, we used the PC1 and PC2 scores.In the graphical representation of the CCA, we use the variables that best describe the contribution of the two axes of PCA and a Pearson correlation analysis was performed between the environmental variables and the PC1 and PC2 values.Only the variables with correlation coefficients ≥ 0.60 and statistical significance p> 0.05 were described.
For the CCA, only the data of the most abundant species with presence in two or more sampling points were used, with the purpose of visualizing the contribution of the abundant species in this analysis more clearly.For this study, rare species were considered those species that were captured in a single occasion or in a single location (Cao et al., 2001).Some of the rare species could be the result of natural low abundances influenced by environmental factors or ecological processes that occur in the basin (Magurran, Henderson, 2003).
In order to determine if there were significant differences between fish assemblages detected by the CCA, we applied a one-way analysis of similarity (ANOSIM) (Clarke, 1993).The ANOSIM was implemented using the fish abundances of the identified groups.We used Bray Curtis (Bray, Curtis, 1957) as dissimilarity measure in order to construct our distance matrix, the correction of SidakSS for p-values, 1000 cycles of permutations and non-standardized data.In addition, a linear regression analysis was performed to visualize the effect of elevation on species richness for the Río Acahuapa basin.All the analyses were performed with the statistical package Qeco (Di Rienzo et al., 2010) and InfoStat (Di Rienzo et al., 2011).

Results
Fish diversity and species composition.A total of 6,742 individuals from eight orders, 14 families, 23 genera and 33 species were captured (Tab.2).The families with highest species richness were Cichlidae and Poeciliidae with seven and six species respectively, followed by the families Centropomidae and Eleotridae with three species each.Members of the family Poeciliidae represented 77% of the total abundance, followed by Cichlidae with 6.8%, Characidae with 6.6%, Mugilidae with 3.4% and Gobiidae with 2.7%, that represented 96.6% of the total abundance.According to the classification by origin and tolerance to salinity, four species were primary ones representing the 12.1% of the species richness, followed by 15 species which were secondary and represented the 45.5% of the species richness, and 14 species of marine derivation represented the 42.4% of the species richness.Of the latter, three (9.1%) are diadromous species and 11 (33.3%)marine-estuarine.Changes in species richness were related to the elevation (Fig. 2).It indicates that species richness declines with increasing elevation (R 2 =0.55, p=0.0006).Regarding the species richness registered for the different sampling months, a constant number of primary secondary and diadromous species were observed; whereas an increase in marineestuarine species was observed in February and May between the transition from the dry season to the beginning of the rainy season (Tab.2).Concerning to species distribution, it was observed that many primary and secondary species are present along the basin, while the diadromous and marineestuarine species are present mainly in sites of low elevation.Relationship between samplings sites, physicochemical factors and variables of habitat structure.The Principal Component Analysis (PCA) of the environmental variables explained a 57% of the observed variance (Fig. 3).The PC1 contributed with 35% of the explained variance and it was strongly correlated with water temperature, water velocity, river width, elevation, and dissolved oxygen; whereas, the PC2 contributed with 22% of the explained variance and it was strongly correlated with the following variables: pH, silt, logs, sand, and rocks (Tab.3).The PCA showed a relationship among sampling locations, habitat structure, and physicochemical factors in the basin of the Río Acahuapa.Whereas, variables such as water velocity, river depth, river width, dissolved oxygen, pH, salinity and temperature were associated to areas between 19-184 m above the sea level.On the other hand, variables such as leaves, logs, rocks, silt, and conductivity were associated to sites between 185-519 m above the sea level.Relationship between species, physicochemical factors and variables of habitat structure.According to the fish occurrence data, 15 species were identified as rare, consequently 18 species were used for Canonical Correspondence Analysis (CCA).According to the CCA, the first two axes account for 66% of the variance explained (Fig. 4).The CCA1 contributed with 41% of the explained variance and showed relationship with the variables of the elevation, temperature, water velocity, width, and dissolved oxygen.While the CCA2 contributed with a 25% of the explained variance and showed a relationship with the pH, sand, silt, rocks, and logs.

Discussion
Lack of studies of freshwater fishes in some countries of Central America has created gaps in information that limit the scope of studies in the region (Miller, 1966;Albert, Miller, 1995;Matamoros et al., 2012).But, recent initiatives have tried to address this problem completing and reconstructing the information of freshwater fish in Central America (e.g., Matamoros et al., 2012;2014).In El Salvador, this research represents the first ecological assessment of freshwater fishes and contributes to support future studies at local and regional scales.The Río Acahuapa has an important number of freshwater fish species.In fact, the 33% of the species recorded for El Salvador are reported in this basin.According to the most recent publication listing the freshwater species of El Salvador (McMahan et al., 2013), some of the records of this study become the first record for the Río Lempa basin, such as Cathorops steindachneri, A. monticola, Mugil curema, P. butleri, Bairdiella armata, D. latifrons, Eleotris picta, A. banana, S. multipunctatum and Cyclopsetta panamensis, and in the case of Trinectes fonsecensis this is the first mention in freshwater ecosystems in El Salvador.We also recorded Oreochromis niloticus "tilapia" and the P. managuensis "guapote tigre", two exotic species, whose occurrence can be linked to aquaculture practices in Apastepeque lake, that is located within the basin of the Río Acahuapa.
The largest contribution of the abundance and dominance of species recorded in this study corresponds to the families Characidae, Cichlidae and Poeciliidae.The same pattern has been observed in different studies from rivers and streams in the Central American region and some areas of Mexico, which report species corresponding to these families as the most dominant and widely distributed (Bussing, López, 1977;Angermeier, Karr, 1983;Esselman et al., 2006;Espinoza, 2007a;Espinoza, 2007b;Rojas, Rodríguez, 2007;López-López et al., 2009;Trujillo-Jiménez et al., 2010;Matamoros et al., 2012;Pease et al., 2012).Although, it is important to highlight that diadromous and marineestuarine species constitute a significant proportion of the species richness in this study.Despite that a recent research on water quality in the basin suggests that its quality ranges from regular to poor (Hernández et al., 2010), an important number of fish species inhabit the basin.In fact, the current conditions favor the migration and colonization of species of fish of marine origin.An example of this was the collection of a significant number of individuals from A. monticola (222 individuals), a diadromous fish of great value for local subsistence fishery, considered by some authors as an indicator of well-oxygenated waters and low or moderate disturbance (Cruz, 1987;Ribeiro, Villalobos, 2010).Therefore, the basin contributes to the diet of the local people and represents an important habitat for the local ichthyofauna.Consequently, studying distribution patterns in the watershed is relevant for biogeography, management and conservation of these species.
Changes in species richness in the Río Acahuapa indicate higher number of species in the lower areas of the basin, while number of species decreases as the elevation increases.These results resemble other studies in the Central American region where an inverse relationship between the species richness and the increase elevation has been identified (Lyons, Schneider, 1990;Espinoza, 2007a;Lorion et al., 2011).These changes of species composition along the altitudinal gradient of the river may respond to a gradient of environmental factors or changes in the dynamics of the habitat structure along the river.Although our environmental variables may be similar to other studies in small streams or tropical rivers (Pouilly et al., 2006), our data show a difference with respect to river velocity, whose values are high in the intermediate and low areas of the basin, while low areas of rivers usually tend have calm waters.This can be explained because the Río Acahuapa has many tributaries in the intermediate and low areas (Hernández et al., 2010), which allows these unusual characteristics.For example, the areas of middle and low elevation were associated to the areas of riffles with greater water velocity, high concentrations of dissolved oxygen, moderate temperature, greater depth, and river width.This conditions favored some fishes, for example A. monticola and S. multipunctatum, as they can be swimming against current and tolerate high water velocity through physiological adaptations (Bussing, López, 1977;Phillip, 1993;Bussing, 2002;Miller et al., 2009).While in the basin upper reaches, the values of the water velocity and depth of the river are significantly reduced, presenting water with low temperatures, lower dissolved oxygen concentration, and greater substrate of logs and rocks.These conditions favored the presence of fish such as some cichlids, small poeciliids, characids and catfish, conditions where these species are usually very abundant (Bussing, López, 1977).Therefore, our results suggest that the geomorphologic characteristics of the basin driven by changes in elevation and changes in the flow of the rivers condition the habitat structure and the physicochemical factors of the water.Likewise, the change of environmental characteristics between backwaters zones and rapids zones can explain part of the composition and distribution of the fish species in the watershed, since these conditions contributed to the differentiation of two fish associations in the case of the Río Acahuapa.
The basin of the Río Acahuapa has no influence of brackish water, even though, the short distance between the mouth of the Río Acahuapa and the sea (45 km) favors the presence of many fish species of marine origin mainly in the lower areas of the basin.This relation between the distance of river and the sea has been observed in different studies, and some authors argue that the distance from the river to the sea is a factor that can explain the composition and distribution of fish communities in a river, because it allows the addition and flow of species by the colonization of marine origin fishes (Lyons, Schneider, e170023[9] 1990; Lyons, Mercado-Silva, 1999;Esselman et al., 2006;Rojas M., Rodríguez S., 2007;Mercado-Silva et al., 2012).Therefore, the combined effect of distance from river to the sea and the elevation appear to be very important explaining species richness patterns of the Río Acahuapa.Likewise, the elevation also showed a relationship with water physicochemical variables and the habitat structure that allowed to observe changes from the headwater of the river until the mouth of the river.
This research shows a great sampling effort during almost a year of surveys, nonetheless, low abundances were recorded for some species in the different sampling events, mainly for marine-estuarine species.Due to lack of studies in freshwater ecosystems in El Salvador, it is difficult to compare our results at local scales.However, our results suggest that for many species of marine derivation, the low basin is an important site that favors its occurrence mainly at the beginning of the rains.This pattern has been observed in other studies in the region, where the rainy season between May and June increases the presence of marine-estuarine fish (Rojas, Rodríguez, 2007).Probably the presence of marine-estuarine species is favored by an increment in river flow during the rainy season and related to feeding excursions or shelter (Espinosa-Pérez et al., 2011).However, these hypotheses should be tested in future studies in the basin.Diadromous species, were present all year round, mainly in areas of low elevation (19-352 m).According to studies in the Central American region, diadromous species can be an important component within species richness in freshwater ecosystems (Bussing, López, 1977;Esselman et al., 2006;Rojas, Rodríguez, 2007;Lorion et al., 2011) and tend to be more abundant and diverse in high areas of the rivers (Lorion et al., 2011).Possibly the habitat structure and physicochemical factors of water in the río Acahuapa, basin did not favor their presence and development in high areas.
The records of temporal and seasonal distribution for this area are relevant and can be useful in future research that determines the temporal and seasonal patterns of these species with greater scientific evidence.Also, we recognize the importance of incorporating other factors that can explain the movement of species in rivers, mainly of these fish groups that depend on connected habitats between marine systems and freshwater ones.For example, some biotic factors may explain the flow and spatial distribution of some species along the basin, such as migrations along the fluvial system for breeding and mating or search for shelter (Schlosser, 1991;McDowall, 1992;Matthews, 1998;Lucas, Baras, 2001;McDowall, 2007).Also, trophic aspects may also explain associations of fish and can be related to environmental aspects in rivers (Pouilly et al., 2006).Therefore, it is important to consider these and other additional factors to complete information on the ecological dynamics in the Río Acahuapa basin.
Our findings describe some of the components that can influence the composition and distribution of the freshwater fish of the Río Acahuapa.Mainly, we evidenced a strong relationship between patterns of species richness and elevation.Also, we identified two different assemblages along the altitudinal gradient of the basin.Likewise, we evidenced that much of species richness come from primary and secondary species; however, marine-estuarine species contribute greatly to species richness in this basin.We hope that our result will be the starting point for future research.Also, we hope this information will be of paramount interest to decision makers on the management and conservation of biodiversity.

Tab. 3 .
Results of Principal Components Analysis (PCA), showing the loadings of environmental variables on the first two principal components and the percent of total variance explained by these components.Shaded values in bold represent values with correlation coefficients >0.60 and statistical significance (p> 0.05).

Fig. 3 .
Fig. 3. Principal Component Analysis (PCA) of four physicochemical variables of the water, eight variables of habitat structure and 17 sampling sites.Localities separated by the two groups identified by CCA in the Río Acahuapa basin, San Vicente, El Salvador, 2011-2012.
Tab 1. Habitat characteristics of the 17 study sites in the Río Acahuapa, San Vicente, El Salvador, 2011-2012.Average values of each of the environmental variables for five sampling campaigns.List of species classified according with their tolerance to salinity, abundance, standard length (maximum, minimum, average and standard deviation) and occurrence by elevation; High (352-519),meters above the sea level in the main rivers of the Río Acahuapa basin, San Vicente, El Salvador, 2011-2012.Shaded data represents the month of occurrence, the black circles represent occurrence by elevation.