Introduction

1Coastal lagoons and estuaries are transitional environments between land and sea permanently or temporarily connected by channels that allow exchange between organisms, nutrients and water. Tidal exchange and rivers discharge produce distinctive spatial and temporal variations that characterize them with complex environmental and ecological processes (Snedden et al., 2013). These systems have a high diversity of invertebrates (mollusks and crustaceans) and vertebrates (birds, mammals and fish) that intensively use these environments for feeding, breeding, and protection (Lara Domínguez et al., 2011; Díaz-Ruiz et al., 2012).

2Because they have a wide range of habitats, high biological diversity and significant primary productivity, however lagoon-estuarine coastal ecosystems of the Gulf of Mexico have characteristics of their own and a complex ecological structure, (Day et al., 2013). Their productivity is due to the subsidies from rivers and tidal exchange, and from vegetation (mangrove forests, marsh grass, others such a freshwater wetlands) that either surrounds them or is submerged (seagrasses, macroalgae), in turn, these communities determine the volume of secondary production (Díaz-Ruiz et al., 2003, Lara-Domínguez et al., 2011).

3Lagoons along the coast of Veracruz cover a 1166 km² area (7.4 % of the national total). Among the most important ones, from North to South, are Pueblo Viejo, Tamiahua, Tampamachoco, Chica-Grande, San Agustín, Salada, Verde, El Llano, El Farallón, La Mancha, Mandinga, Alvarado, Sontecomapan and El Ostión (Lara-Domínguez et al., 2011; Peresbarbosa, 2005).

4Sontecomapan Lagoon, located to the South, is part of the mountain chain of Los Tuxtlas. This system belongs to the Biosphere Reserve of the same name and was decreed on November 23, 1998 (SEMARNAT, 2006). Also, it was declared a Ramsar site in 2004 (Ramsar Site 1342, FIR, 2004) because of the conservation degree of its mangroves and later worldwide acknowledged as a Biosphere Reserve within the “Man and Biosphere” UNESCO program (MAB) in October, 2006 (MAB, 2010).

5Currently, ecological knowledge about fish communities in Sontecomapan Lagoon System is limited since no previous study has quantified their diversity and abundance. Therefore, it is priority to analyze current information on structure and function to better understand their ecological behavior and regional economic value.

Background

6From late seventies to present, research has been conducted on fish that inhabit the lagoon systems in Veracruz, however, the tendency has been to contrast the changes in species composition through different years (Kobelkowsky, 1991; Pérez-Hernández & Torres-Orozco, 2000; Bedia-Sánchez & Franco-López, 2008; Lara-Domínguez et al., 2011; Abarca-Arenas et al., 2012). The work of Lara-Domínguez et al. (2011) summarizes that, for the different coastal environments in Veracruz, a total of 101 families have been recorded, including 478 species. Other contributions to this synthesis, regarding the available literature about Veracruz fish, are from undergraduate or graduate theses, technical reports, conferences and of some scientific articles published. Altogether, these papers comprise a total of 117 references where, only 4 correspond to the Sontecomapan Lagoon; therefore, it is necessary to gather more ecological information that allows a better comparison between systems and fish communities on the coast of Veracruz.

7However, there is some progress on this subject as shown by the various publications for the State coastal systems. Such is the case of works by Chávez-López & Franco-López (1992), Chávez-López et al. (2005) and Franco-López et al. (2012) for the Alvarado Lagoon where different ecological aspects of the system, such as development and the environment regional impact on the fish community are analyzed, summarizing a record of more than 76 species. In Tamiahua Lagoon, Díaz-Ruiz et al. (2003) characterized the correlation of fish associations —including 32 families, 62 genera and 83 species— with environmental variables. Similarly, Castillo-Rivera et al. (2003, 2005, 2011) studied the lagoon system of Pueblo Viejo, analyzing spatial-temporal diversity and abundance patterns of the fish community —30 families, 52 genera and 66 species— in relation to hydro-meteorological variables, considering daytime and nychthemeral scales. For the Chica-Grande lagoon system, Aguirre-León et al. (2014) analyzed the environmental heterogeneity and its relationship with the diversity and fish community abundance, recording 14 families, 20 genera and 22 species, while Mendoza (2015) performed a temporary ecological assessment of the fish community in La Mancha Lagoon, emphasizing environmental processes that determine their behavior; 25 families, 37 genera and 50 species were identified. Also, Juárez (2015) analyzed in Ostión Lagoon, south of Veracruz, the spatial-temporal variations in structure and function of the fish community with relation to habitat dynamics, in this system; 19 families, 23 genera and 34 species were recorded.

8Currently, ichthyological research for Sontecomapan Lagoon System is scarce, most of it is only a list of species. The oldest study is from Reséndez (1983) who related this lagoon’s fish composition and hydrological data, the author recorded 26 families, 43 genera and 52 species; later, Fuentes-Mata et al. (1989) added new records for this specific system. Nava (2005) made a fish catalog for this location, including 18 families, 29 genera and 37 species. Rodríguez-Varela et al. (2010) analyzed the system fish fauna within two climatic seasons, recording 24 families, 42 genera and 52 species. Lara-Domínguez et al. (2011) summarized a total of 91 species for this system. The most recent publication that adds information on coastal fish in Veracruz is from Abarca-Arenas et al. (2012), which makes a synthesis of at least three decades and report for Sontecomapan a total of 33 families, 64 genera and 92 species including marine, estuarine and freshwater. Finally, Castillo-Rivera and Lara-Domínguez (this issue), from a comprehensive literature review on the fish fauna of the lagoon revealed that 115 species have been historically recorded but pointed out possible misidentifications and changes in taxonomic status.

9Considering the ecological information scarce about the fish community structure in Sontecomapan Lagoon, the objectives of this study are: to analyze the variation of physicochemical parameters and define habitats within the system; determine species composition and types of community inhabitants; define trophic levels based on the species feeding habits; determine dominant species on the fish community; analyze diversity, richness, evenness, density, biomass and average weight in the system, and determine which physicochemical factors condition the fish assemblages’ behavior.

Materials and Methods

Study Area

10Sontecomapan Lagoon is in the western of the Gulf of Mexico (018º 32’ N, 095º 02’ W) and is part of the Los Tuxtlas region in southern Veracruz (Fig. 1). The system is roughly 12 km long and 1.5 km wide with a surface area of 8.9 km² and a perimeter of 36.5 km (FIR, 2004). It is shallow, with sediments ranging from sands to silts. It is permanently connected to the sea through an inlet of about 5.5 m depth called “The Bar of Sontecomapan” (Contreras, 1993). The lagoon has a considerable water input from rivers and streams, the most important are: Coxcoapan, Yuhualtijapan, La Palma and El Sábalo. The most significant streams are: Los Pollos, Sontecomapan, El Fraile, La Boya and Basura (Fig. 1). The climate is warm and wet, with an annual rainfall of 4500 mm. The rainy season is from June to September and the dry season from March to May (García, 2004). The driest month is May and the peak of the rainy season is from August to November. The movement of cold and dry air masses from the north affects the area from October to February causing cold fronts that are locally called “Nortes” (Soto-Esparza & Giddings, 2011). There are different aquatic environments in the system, such as the surrounding mangrove vegetation in brackish and low salinity areas mostly with Rhizophora mangle and Avicennia germinans and submerged vegetation represented by Ruppia maritima (Contreras, 1993; Carmona-Díaz et al., 2004).

Field and laboratory activities

Figure 1. Sontecomapan Lagoon System, Veracruz. The location of sampling stations (•), the main physiographic features of the study area and rivers draining into the system are noted. Palma River (LPR), Basura River (BR), Chuniapan River (CHR), Coxcoapa River (CR), Sábalo River (SR), Yuhualtijapan River (YR), Pollos River (PR) Sontecomapan Inlet (SI).

11Field surveys were conducted in the months of October 2014, February and October 2015 and February 2016. A network of 10 sampling stations was set for collecting fish (Fig. 1), they were placed considering the marine inlet and tidal channel, sediment type, submerged and emergent vegetation, and fresh water flow and contribution to the system. Catches were obtained trhough a 5.0 m long shrimp trawl, with a 2.5 m work opening and a 1.8 cm mesh. This was done for two days every month, sampling the first day and taking a replicate the second. Two 10-minute trawls at an average speed of two knots were made at each station —one per working day— for a total of 20 hauls per month. The trawling time corresponds to a sampling area of 1500 m². The fish were stored in labeled plastic bags and fixed with a solution of formaldehyde at 10 % for transport to the laboratory.

12At each sampling site, salinity, temperature and dissolved oxygen in the water column (surface and bottom levels) were measured with a YSI85 multiparametric meter, pH with a Hanna meter, transparency of the water column with a Secchi disc, depth with a digital depth probe, and environment temperature with a pocket weather meter.

13Taxonomic identification of the species was carried out using identification sheets and descriptions based on the work of Reséndez (1981a, 1981b), Castro-Aguirre et al. (1999), Fischer (1995) and Miller (2009), each species was weighed with a digital Ohaus balance (accuracy 0.01 g) and measured using a conventional ichthyometer in cm for the record.

14The community inhabitants types were determined using the Yáñez-Arancibia (1986) criteria, who classifies species as: a) Estuarine-lagoon permanent inhabitant (E), b) Marine euryhaline species (ME), c) Marine stenohaline species (MES), d) Freshwater species (F). The trophic category of species was defined per published information on each species feeding habits, with the following criteria: a) First order consumers, mainly planktivorous fish that consume phytoplankton and zooplankton, b) Second order consumers, omnivorous fish, that include vegetables and detritus in their diet, c) Third-order consumers, fish that are exclusively carnivorous, and their vegetables and detritus consumption is accidental (Yáñez-Arancibia, 1986). The species dominance was determined by using the Importance Value Index (IVI= % N+ % W+ % F) as the sum of percentage in number plus the species weight plus the frequency of their occurrence (De la Cruz-Agüero, 1994).

15Community ecological parameters were estimated: diversity (H’) (Shannon & Weaver, 1963), species richness (D) (Margalef, 1969) and evenness (J)’ (Pielou, 1966), as well as density (ind/m²), biomass (g/m²) (Brower & Zar, 1977) and average fish weight (g/ind) (Aguirre-León et al., 1998).

Statistical analysis

16Spatial (between sampling stations) and temporal (monthly) statistical differences of community’s environmental variables and ecological parameters were estimated with an analysis of variance. Homogeneity of variance between groups was previously evaluated using the Levene test (1960). Logarithmic transformations were made when necessary. When variances were homogenous, an ANOVA F-test was used, otherwise the Kruskal-Wallis H test (K-W) was used (Zar, 1999). A significance level of p <0.05 was considered for all these tests (Zar, 1999; Hair et al., 2001).

17Based on the spatial-temporal values of the physicochemical parameters, habitats were defined using a cluster analysis to determine environmental similarity between stations. The methods used were Euclidean distance coefficient and the clustering algorithm was the weighted average technique (Hair et al., 2001).

18To determine associations of fish and its correlation with environmental factors, a canonical correspondence analysis (CCA) was performed (Ter Braak & Verdonschot, 1995). This was done with an array of abundance of the species per station and month, and an array of average values of the physiochemical parameters. The CCA was done with version 3.22 of the Multivariate Statistical Package (MVSP).

Results

Spatial and temporal variation of environmental parameters

19The analysis of physicochemical data indicated that the system has a very dynamic spatial behavior related to its size and geomorphology. Environmental parameters depend on the seawater through its small inlet and freshwater from the rivers (Fig. 2). The highest average surface salinity (21.4±11.5) was recorded at station 1 (Fig. 2a) and the lowest (2.4±2.3) in station 10, with significant statistical differences between stations, K-W (9,80) H=31.08, P=0.0003. At the bottom, the highest average (29.4±4.5) was recorded at station 3 (Fig. 2a) and the lowest (7.7±5.2) at station 10, with significant differences between stations, K-W (9,80) H =41.20, P<0.0001. There also were differences between surface and bottom (Table 1). The highest average temperatures at the surface (28.2^oC±2.9) and bottom (28.8±2.9) were recorded at station 9 (Fig. 2b), and the lowest (26.4 ±3.5 and 26.5 ±3.6), at station 1. There were no significant differences between stations, neither in surface, bottom or between these two levels (P>0.05) (Table 1). The highest average dissolved oxygen on the surface (5.8mg/L±2.0) and bottom (4.7±1.8) was recorded at station 9 (Fig. 2c), while the lowest values were those from stations 1 (4.7 ±1.5) and 3 (4.8 ±1.3). There were no significant differences between stations, neither on the surface or at the bottom (P>0.05), however, there were differences between these levels (Table 1). The highest average pH values at the surface (7.8±0.3m) and bottom (7.9±0.4) were recorded at station 1 (Fig. 2d), whereas the lowest were at stations 10 (7.3±0.34) and 8 (7.0±0.38). There were significant statistical differences for pH between surface stations, ANOVA (9,70) F=2.46, P=0.0167 and between bottom stations, ANOVA (9,70) F=3.50, P =0.0013; however, there were no differences between levels (Table 1). The largest average depth (3.46m±1.1) was recorded at station 2 (Fig. 2e) and the smallest (1.15 ±0.15) at station 9, with significant differences between stations: K-W (9,80) H=49.74, P<0.0001 (Table 1). The highest average transparency of the water column (1.82m±0.6) was recorded at station 1 (Fig. 2e) and lowest at station 6 (0.78±0.25) and station 8 (0.78±0.15), with significant differences between stations, K-W (9,80) H=36.76, P<0.0001 (Table 1).

20Regarding temporal variation (Fig. 2), physicochemical variables indicated changes between months. Highest average salinity at the surface (16.6ups±13.8) and bottom (21.6±12.9) was recorded in February 2016 (Fig. 2f), while the lowest surface (5.3±6.6) and bottom (15.9±9.3) values were in October 2014, with significant differences only for surface data, K-W (3,80) H =36.17, P=0.00001. Highest average temperatures at the surface (29.9°C±1.5) and bottom (31.2±1.7) were recorded in October (Fig. 2g) and the lowest (24.2 ±1.5 and 23.8±1.6 respectively), in February (2015) thus reflecting monthly temperature variation. There were significant differences for surface, ANOVA (3,76) F =101.39, P<0.0001, and bottom values, ANOVA (3,76) F =132.62, P<0.0001. Average dissolved oxygen concentrations at the surface (6.9mg/L±0.9) and bottom (5.8±0.7) were higher in February 2016 (Fig. 2h), and lower (4.2±1.0 and 2.7±1.0, for surface and bottom respectively), in October 2014, with significant differences for surface ANOVA (3,76) F=21.22, P<0.0001, and bottom, ANOVA (3,76) F=25.76, P<0.0001. Highest average pH values at the surface (7.7±0.4) and bottom (7.8±0.5) were recorded in February 2016 (Fig. 2i), while the lowest values were recorded in October 2015 for the surface (7.4±0.2), and October 2014 for the bottom (7.4±0.4), with significant differences between month for both surface ANOVA (3,76) F=2.85, P=0.043, and bottom, ANOVA (3,76) F=3.80, P=0.013. Average depth of the system ranged from 1.2±2.3m in February 2016 (Fig. 2j) to 1.9±0.7 in October 2014, with no significant differences between months (P>0.05). Highest average value for transparency (0.7±1.4m) was in February 2016 (Fig. 2j) and the lowest (0.9±0.4) in February 2015, with significant differences between months, K-W (3,80) H=19.79, P=0.0002.

Figure 2. Spatial and temporal average values (± 1 SD) of the physicochemical parameters in the Sontecomapan Lagoon System. For each environmental variable, the surface (S) and bottom (B) values are shown.

21Based on the spatial-temporal variation of physicochemical parameters, cluster analysis indicated three groups or habitats (Fig. 3a), the first one went from station 1 to 4, which grouped the marine inlet and the tidal channel of the lagoon where water exchange with the adjacent continental shelf is important. (Fig. 3b). The second group consists of stations 5, 7 and 8 located in the middle section of the lagoon where mixing processes of water from the sea and river drainage happen. The third group includes stations 6, 9 and 10, which corresponds to areas that are mostly influenced by freshwater such as El Sábalo, El Fraile and the mouth of Arroyo Chuniapan (Fig. 3a and 3b). Salinity and pH from surface and bottom, as well as depth and transparency were the defining parameters for these environments since they indicated significant statistical differences at both spatial and temporal scales (Fig. 3, Tables 1 and 2).

Fish composition

22A total of 3,018 individuals with a total weight of 27,500.1 g were captured during the study for an overall of 18 families, 28 genera and 34 species identified (Table 3). The most represented family was Gerreidae with 6 species, Diapterus rhombeus, Diapterus auratus, Eucinostomus argenteus, Eucinostomus gula, Eucinostomus melanopterus and Eugerres plumieri.

23Some of the 34 species were caught in just a month. Such is the case for Bardiella ronchus, Eleotris pisonis and Polydactylus octonemus in October, 2014. During February, 2015, was Mugil curema and Scorpaena plumieri. In contrast, the period with most species was October, 2015 when Caranx crysos, Chloroscombrus chrysurus, Eucinostomus argenteus, Gobiomorus dormitor, Lutjanus analis, Lutjanus synagris, Selene vomer, Sphoeroides splengeri and Synodus foetens were recorded, whereas for February 2016 only Mayaheros urophthalmus was captured. Documented as new records for Sontecomapan Lagoon were Synodus foetens and Sphoeroides splengeri.

Figure 3. (a) Station groups with similar environmental characteristics in the Sontecomapan Lagoon System. (b) Spatial distribution of the three habitats defined in the system.

Table 3. List of the fish species collected at the Sontecomapan Lagoon System, Veracruz

Community inhabitants

24From the species collected in the system, 27 (79.4 %) are Marine euryhaline (ME) with D. rhombeus being the most abundant of this category. The group of Marine stenohaline species (MES) has 3 species (8.8 %), L. synagris, S. plumieri and S. splengeri whereas C. aguadulce, G. dormitor and E. pisonis are estuarine-lagoon permanent inhabitant (E) and only one species, M. urophthalmus, is from freshwater (F) (Table 3).

Trophic categories

25Within the trophic category of first order consumers (planktivorous), 11 species (33 %) were found; 13 species (37 %) are second order consumers (omnivores), and 10 species (30 %) are third order consumers (exclusively carnivores) (Table 3).

Dominant Species

26During this study, two species were established as dominant in the system based on their Importance Value Index (IVI), D. rhombeus (IVI=111.3 %) and C. aguadulce (IVI=57.9 %). These two-species accounted for 80.8 % (2,439 individuals) of numerical abundance and 63.4 % (17,431.0 g) of the total catch of the community (Table 3).

Spatial and temporal variation of diversity

27In general, the spatial analysis of H’ and D showed a tendency with lower values towards the marine inlet and higher inside the system, the value of J’ had an inverse trend (Fig. 4). The average number of species ranged from 2 in station 1, to 8 in stations 9 and 10; the highest average diversity H’ (1.43 ±0.34) was recorded at station 5 (Fig. 4a) and the lowest (0.71 ±0.5) at station 1. Richness D (Fig. 4b) was higher in stations 5 (1.59 ±0.22) and 9 (1.58 ±0.26) and the lowest in station 3 (1.03 ±0.96). The evenness J’ (Fig. 4c) was higher (0.81 ±0.07) at station 4 and the lowest (0.44 ±0.15) in station 8. The behavior of the three indices is related to a low number of species and the dominance of some of them in the sampling stations. There were no significant differences between sites for any of these indices (P> 0.05).

Figure 4. Spatial and temporal average values (± 1 SD) of the diversity indices H’, D and J’ for the fish community in the Sontecomapan Lagoon System.

28The temporal variation of H’, D and J’ is shown in Figure 4. The total number of species changed between sampling periods: 17 in October, 2014; 16 in February, 2015; 25 in October, 2015; and 14 in February, 2016. The highest average H’ (1.28 ±0.28) occurred in February, 2015 (Fig. 4d) and the lowest (0.94 ±0.43) in February, 2016. D index was the highest (1.56 ±0.543) in October, 2015 (Fig. 4e) and the lowest (1.1 ±0.51) in February, 2016. Maximum average J’ (0.74 ±0.17) was in February 2015 (Fig. 4f), while the minimum values occurred in October, 2014 (0.61 ±0.35) and 2015 (0.60 ±0.31). Both H’ and D average monthly values tend to be low. There were no significant differences in the average values of these indices between months (P>0.05).

Spatial and temporal variation of abundance

29Figure 5 shows values of fish density (ind/m²), biomass (g/m²) and average weight (g/ ind). Therefore, the highest average value in density (0.096 ind/m² ±0.12) was recorded at station 10 and the lowest at stations 1 (0.0005 ±0.0006) and 2 (0.001 ±0.001) (Fig. 5a), with significant statistical differences between stations, K-W (9,40) H=29.07, P=0.0006. Species that showed higher density were D. rhombeus, C. aguadulce and D. auratus. The highest average biomass (0.45 g/m² ±0.27) was recorded in station 10 and the lowest (0.014 ±0.02) in station 2 (Fig. 5b), with significant statistical differences between sites, K-W (9,40) H=29.67, P =0.0049. Species with higher biomass were C. aguadulce and D. rhombeus, and, to a lesser amount E. plumieri, Hypanus sabinus, Sciades felis and Centropomus parallelus. Highest average weight (34.2 g/ind ±36.3) was obtained at station 1 and the lowest at station 7 (8.3 ±5.4) (Fig. 5c), showing an inverse relation between this parameter and the other two. Species with the highest average weight were H. sabinus, S. plumieri and Opsanus beta. This parameter showed no significant differences between sites (P> 0.05).

30Throughout this study, the highest average density (0.055 ind/m² ±0.08) was recorded in October, 2015 and the lowest (0.012 ±0.015) in February, 2016 (Fig. 5d), the rest of the time, it was also low. The highest average biomass and weight (0.22 g/m² ± 0.13 and 22.16 g/ind ±21.8, respectively) were recorded in February, 2015 (Fig. 5f) and the lowest (0.15 g/m² ±0.16 and 12.1 g/ind ±6.6) in February, 2016. There were no significant differences between months for any of these parameters (P>0.05).

Fish-habitat analysis

31The ordination diagram (Fig. 6) shows the relationship between species, stations and environmental variables as well as the relative importance of each environmental variable that determines the composition and distribution of fish in the community. The CCA showed that 71.9 % of the total variance was explained by the relationship between the abundance of species and environmental variables, with eigenvalues of 0.348 (axis 1) and 0.105 (axis 2). CCA explained 55.2 % of the variation for species-environment axis 1, indicating that the most important factors in the distribution of species are transparency (r=0.80), salinity at the bottom (r=0.65), depth (r=0.57) and pH at the bottom (r=0.45). Species that showed a high correlation with axis 1 were, Citharychthys spilopterus, E. plumieri and E. melanopterus from the euryhaline species, L. synagris, S. plumieri and S. splengeri from the stenohaline, and E. pisonis among estuarine fish. Low and high abundances of these species were associated to areas with higher and intermediate depths (stations 2, 3, 4, 5, 6, 7) with higher salinity and transparency.

Figure 5. Spatial and temporal average values (± 1 SD) in density, biomass and average weight for the fish community in the Sontecomapan Lagoon System.

Figure 6. Canonical Correspondence Analysis Plot of species abundance, environmental variables and sampling stations in the Sontecomapan Lagoon System. Arrow length and direction indicate the relative importance of environmental variables. Species abbreviations (AB) are listed in Table 3.

32Axis 2 explained 16.68 % of species-environment variation with bottom dissolved oxygen (r=0.48) and bottom temperature (r=-0.14) as the most important factors that explain the distribution of species along this axis. Euryhaline species that showed high correlation with axis 2 were D. rhombeus, D. auratus and S. felis, the estuarine species were C. aguadulce and G. dormitor, and M. urophtalmus from freshwater. The abundance of these species was intermediate and high, strongly related to temperature and also associated to shallow areas (stations 8, 9 and 10) with lower salinity and high oxygen levels.

Discussion

33The variability of physicochemical parameters in Sontecomapan Lagoon System is conditioned by its shape and size, by sea water input through the marine inlet and tidal channel from the Gulf of Mexico, and by freshwater from rivers and streams that flow into the lagoon. Salinity had a spatial gradient with higher values close to the sea channel and lower near the system header. This was previously described by Martínez (1987), Morán (1994), Castellanos (2002) and Rodríguez-Varela et al. (2010) who indicated high values in the inlet and the tidal channel, and lower into the system. The highest average salinities were recorded in February, 2015 (rainfall 58.5 mm; CONAGUA, 2015) and 2016 (rainfall 55.0 mm; CONAGUA, 2016) due tu reduced precipitation and increased tide effect into the system, while the lowest were found in October, 2014 (rainfall 660.5 mm; CONAGUA, 2014) and 2015 (rainfall 872.3 mm; CONAGUA, 2015). The variation in salinity was thus closely related to the climatic changes of this coastal region (Morán, 1994; Castellanos, 2002; and Rodriguez-Varela et al., 2010). In this study, the data showed a vertical stratification of the whole lagoon within the analyzed months.

34Water temperature showed little variation throughout the system, with differences of 1.1°C to 1.3°C between recording stations and between surface and bottom. Martínez (1987), Castellanos (2002) and Rodríguez-Varela et al. (2010) also recorded an average fluctuation of ±1°C due to solar radiation heating the water column of this shallow system. In relation to the thermal change, the highest values were in October, 2014 and 2015 and the lowest in February 2015, and 2016 when the reduction of air temperature extends to the lagoon. Aguirre-León et al. (1998) in a coastal system in Campeche and Castellanos (2002) in Sontecomapan Lagoon, both described similar conditions.

35Dissolved oxygen showed little spatial variation in the system, however, higher values were recorded at the surface and lower at the bottom. This stratified distribution was related to the salt wedge described above and associated with the tide and local winds effect that exchange oxygen in the water column (Morán, 1994). Castellanos (2002) reported high dissolved oxygen values into the system, due to photosynthetic activity of submerged vegetation in shallow areas. Monthly changes were observed, with the lowest values recorded in October, 2014 and 2015 and the highest in February, 2015 and 2016. This has a relation with the thermal variations described above and the frequency and intensity of the “nortes” wind that increase oxygen exchange in the water column (De la Lanza, 1994).

36Depth and transparency showed a similar spatial trend, with higher values in the marine inlet and tidal channel (stations 1-4) and lower towards the head of the system (stations 8-10). Currently the Sontecomapan Lagoon System has an excess of sediment, due to the channel narrowness in its marine inlet and to the accumulation of deposits that are stream-carried from its eroded banks as a result of animal breeding and agriculture (SEMARNAT, 2006). Monthly depth has not significant differences, however, the lowest was registered in October, 2014 and the highest in February, 2016. Transparency values were inverse to depth in October, 2014, February and October, 2015. Both parameters are influenced by the intensity of precipitation and the “Nortes” phenomena; Zamora (2002), Castellanos (2002) and Rodríguez-Varela et al. (2010) also observed this behavior.

37The cluster analysis of the system’s environmental data showed that salinity, pH, depth and transparency defined three environments or habitats with statistically significant differences between them. The spatial behavior of these three environments remained unchanged in the analyzed months. Morán (1994), García-Cubas & Reguero (1995) and Castellanos (2002) described a similar environmental pattern according to the relationship between the distribution of habitats and the physicochemical variables, water movement in the lagoon, nutrient concentration or input from rivers and seasonality.

38The fish community of the Sontecomapan lagoon has been under study since 1979 (Zamora, 2002), with most of the reports being only species listings without any additional ecological analysis, and no attempt to analyze why species composition has spatial-temporal changes. Therefore, there is no retrospective analysis of the change in the number of species at the community or any certainty over the past three decades. Although previous studies are available, they have been conducted with different sampling techniques, which results in different fish composition, so it is reasonable to assume that there is still no complete record of all species in the Sontecomapan system. The latest integrated species publication for this lagoon reported 92 (Abarca et al., 2012) however, Fuentes-Mata et al. (1989) had reported 97; recently, Espinoza (2017) indicates that the historical record of fish in the system counts more than 100 species. Based on available literature on ichthyology and data from this research, S. foetens and S. spengleri species are new records for the Sontecomapan Lagoon System.

39Yáñez-Arancibia & Nugent (1977) used the concept of the ecological role that community inhabitants play in a system and this concept was later expanded by criteria from Castro-Aguirre et al. (1999). In tropical coastal ecosystems, it is common to find the four types of community components found in Sontecomapan using these criteria. The ecological behavior of these species is due to physiological adaptations that allow them to use saline environments with large temperature changes and to move into the continental and marine environment as eury-thermohaline species (Able, 2005), such is the case of D. rhombeus that was the dominant species (69 % abundance, IVI=111.3 %) in the four months analyzed. Conversely, from an evolutionary perspective, there are stenohaline species that are only adapted to enter the ecosystem when salinity conditions only show slight changes and favor them. Generally, these are rare species (Cowan et al., 2013) such as the stone fish S. plumieri (IVI=1.5 %) and the puffer S. spengleri (IVI=0.6 %), which were captured in the tidal channel near the marine inlet. Other species can remain throughout their life cycle and move across all the physical and chemical system gradients all year long, these so-called estuarine species occur in low proportion, as is the case of C. aguadulce (IVI=57.9 %). Additionally, the species M. urophthalmus (IVI=0.5 %) (sensu lato Cichlasoma urophthalmus, Rícan et al., 2016) is a typical freshwater component captured in the habitat only under fluvial influence. However, in other coastal systems of the southern Gulf of Mexico and the Caribbean Sea, this species has been collected in brackish and marine environments, demonstrating its great adaptability to habitats other than their original (Miller, 2009). Studies that analyzed the proportion of community inhabitants for Sontecomapan Lagoon reported that most of the species were from the euryhaline component (Martínez, 1987; Zamora, 2002). For other systems, such as the Alvarado and Chica-Grande lagoons, there were also more euryhaline inhabitants and less stenohaline and estuarine ones (Sánchez, 2003; Aguirre-León et al., 2014).

40Based on the criteria proposed by Yáñez-Arancibia (1986), and the available information on feeding habits, we found that the ichthyofauna of Sontecomapan was represented by the three categories of defined consumers, where the number of species belonging to the second-order consumers (omnivores; n=13) was higher than the number of first-order (planktivorous; n=11) or third-order (carnivorous; n=10) consumers. However, the second-order has the lowest total abundance (total individual number= 143 and total weight= 2,682.0 g), compared to trophic level 1 (total individual number=2,398, total weight=12,525.0 g) and third-order (total individual number= 475, total weight= 12,294.0 g). Based on stomach contents, Alarcón (2007) determined, 6 first-order, 19 second-order and 10 third-order consumers and stated that the plant-derived organic matter, detritus and zooplankton are the main food sources that support the trophic structure of this system. In other coastal systems in the Gulf of Mexico, as Términos Lagoon, the most abundant orders were first and second (Álvarez-Guillén et al., 1985). The same pattern of trophic categories was found for Raya Lagoon in Venezuela (Ramírez Villaroel, 1994). Likewise, Alvarez-Rubio et al. (1986), have pointed out that the spatial-temporal variation of the trophic levels is related to the species’ ecological behavior, the environment they use and the selection of fishing nets. The same authors indicated that second-order species have a broad trophic spectrum, allowing them to be dominant (as is the case of C. aguadulce and D. rhombeus in Sontecomapan) since, because of their adaptive capabilities, they have different feeding options in different situations. Identifying what type of consumers use a coastal system could tell several aspects of the energy flow (food webs) such as producer-consumer and prey-predator relations, food availability and adaptations for consumption (Zárate-Hernández et al., 2007; Hagy & Kemp, 2013). Another aspect of the trophic levels structure in a system is the size of the species, as it indicates what stage of their life cycle develops in the ecosystem. In Sontecomapan Lagoon, small sized of D. rhombeus (5.0 to 6.0 cm) to medium sized of C. aguadulce (16 to 18 cm) predominate, suggesting that habitats are used as nursery areas, as well as for feeding and growth. This ecological behavior has been well documented in tropical coastal systems (Yáñez-Arancibia et al., 1985; Able, 2005; Cowan et al., 2013; Hagy & Kemp, 2013).

41The dominance analysis is a fundamental aspect in fish communities, since it allows to define the ecological importance of species through their interactions and habitats at a spatial-temporal scale (Sheridan & Minello, 2003; Yáñez-Arancibia et al., 2007). In Sontecomapan Lagoon, previous fish research did not clearly analyze this aspect, so the system’s natural variation of dominance reamained unknown, especially for species that are fishery resources or indicators of adverse environmental conditions (Aguirre-León et al., 2010). The Importance Value Index (IVI) indicated that D. rhombeus and C. aguadulce were dominant in the system, the latter has been reported as Cathorops melanopus in most previous ichthyological studies in the Gulf of Mexico, and now it is recognized for Mexican coastal waters as a synonym of C. aguadulce (Marceniuk & Bentacour, 2008; Miller, 2009).

42It is necessary to unify the numeric criteria upon which dominance is calculated throughout its latitudinal distribution and to report the specific fishing gear used to estimate it, however, each author has used different indicators and nets for this purpose. Martínez (1987) reported D. rhombeus as one of the most abundant species in the Sontecomapan system, while Castillo-Rivera et al. (2005), Pérez-Ponce (2007), Aguirre León et al. (2014) and Juárez Bautista (2015) indicate the dominance of C. aguadulce (sensu lato Cathorops melanopus) and D. rhombeus in several lagoon systems of the Gulf of Mexico. The same authors also reported O. beta, C. spilopterus, C. parallelus., A. mitchilli, D. auratus, E. melanopterus, M. curema and A. lineatus as dominant.

43In other coastal States of the Gulf of Mexico, C. aguadulce has been reported as an abundant species in lagoons of Yucatan (Vega-Cendejas, 2004). Aguirre-León & Díaz-Ruiz (2006) found that D. rhombeus is also abundant in coastal systems from Campeche. Similarly, in other latitudes, D. rhombeus, is dominant in the Raya Lagoon in Venezuela (Ramírez, 1994). The wide distribution of these two species in coastal lagoons is due to the adaptive success that allows them to move, feed and reproduce within them (Cowan et al., 2013).

44Values for the H’ and D indexes, showed similar spatial trends in the system and were generally higher on the connecting channel and middle part of the system where the tide influence and salinity is more pronounced, and temperature is warmer. The lagoon system has a geomorphology with a narrow marine inlet and a long tidal channel, this determines the exchange of water with the sea and species point of entry. In Chica-Grande lagoon in the north of Veracruz, Aguirre-León et al. (2014) observed similar behavior since it also has a sinuous tidal channel, and a tidal effect that allows the access of euryhaline and stenohaline species to the system. Other variables that influence diversity in coastal systems are depth, freshwater discharge and transparency associated with temporal precipitation. Also, in Pueblo Viejo Lagoon, Veracruz, Castillo-Rivera et al. (2005) established that the changes of fish diversity are determined by organic production and rainfall, which increases the availability of food and controls="true" the immigration of species. In other tropical latitudes, it has also been established that the diversity and richness of the fish community is influenced by low transparency and changes in salinity that have a significant effect on their spatial-temporal distribution (Blaber, 1985; Cyrus & Blaber, 1987; Blaber et al., 1995; Blaber, 2002). Spatially, evenness J’ was opposite to H’ and D. This trend was related to the low diversity near the marine inlet, however, the lowest values towards the system header were due to the abundance of species such as C. aguadulce and D. rhombeus, which behaved as dominant. Velázquez-Velázquez et al. (2008), Díaz-Ruiz et al. (2012) and Cowan et al. (2013) mention that evenness is inversely related to dominance, so that high values of dominance reveal low evenness values, as recorded for Sontecomapan.

45There were no marked changes in the indexes between months, however, small H’ changes were related to the evenness and dominance of the species in each month, as in February, 2015, where only D. rhombeus dominated and the average value of evenness was higher, in addition it was the month with the lowest average water temperature. Also, local changes in salinity and transparency influenced the monthly richness D of the system. Research from other lagoon systems of the Gulf of Mexico indicate that temporal values with highest diversity are also related to the input of nutrients from the rivers and an increase of primary production (Díaz-Ruiz et al., 2003; Aguirre-León et al., 2010; Arceo-Carranza et al., 2010; Castillo-Rivera et al., 2011).

46The spatial variations of density and biomass were similar, with low values towards the marine inlet and higher values towards the head of the system. This trend was associated with the low diversity in the areas closest to the inlet, as noted in the previous section. In contrast, average weight showed an opposite trend, except in October, 2014 when values close to the marine inlet were higher due to the presence of rare big species such as S. felis (sensu lato Ariopsis felis, Marceniuk & Menezes, 2007), S. plumieri and H. sabinus (sensu lato Dasyatis sabina, Last et al., 2016). Other features of the system that determine variation in abundance are, the spatial distribution of sediments (sandy in the sea mouth and silty in the header), the submerged vegetation (that functions as protection areas for growth and feeding), and spatial changes on physicochemical parameters (Martínez, 1987).

47The highest average values of density, biomass and weight occurred in February and October, 2015 when D. rhombeus and C. aguadulce had the highest abundance due to their life cycles and the system’s environmental characteristics. For Pueblo Viejo Lagoon, Castillo-Rivera et al. (2005) reported that lifecycles can be favored by organic production of the ecosystem when river discharge and runoff into the lagoon increase, providing allochthonous organic matter and nutrients to the system thus increasing the availability of food for consumer species. Several authors (Azevedo et al., 2007; Aguirre-León et al., 2010; Arceo-Carranza et al., 2010; Lara-Domínguez et al., 2011; Díaz-Ruiz et al., 2006, 2012) have reported this seasonal pattern of abundance in the community related to environmental conditions and the seasonal migration of species linked to their lifecycles. In this sense, the species that contributed the most to the temporal changes in abundance at Sontecomapan Lagoon were D. rhombeus, C. aguadulce, D. auratus, C. spilopterus, S. felis and E. plumieri. These are mostly estuarine or euryhaline species which develop much of their life cycle in the lagoon system since both juveniles and pre-adult sizes were collected during this research. Particularly, the stripped mojarra E. plumieri is a marine species with commercial value in Sontecomapan and is often sold and consumed locally in the nearby town of Catemaco (personal observation from authors). However, biological and ecological knowledge of this and other fishery resources such as snooks of the genus Centropomus is scarce. Therefore, it is necessary to gather more information about the distribution, abundance and seasonality of fish community as soon as possible.

48CCA results allowed to identify two groups of species related to the environmental system gradients. The first group included species of three community inhabitants (ME, E and MES), some of them were scarce and distributed in the first 4 sampling sites (stations 1 to 4). These sites correspond to the connection inlet and tidal channel of the system, where salinity, dissolved oxygen, transparency and depth were greater, thus the species S. plumieri, S. spengleri and S. foetens were collected in this habitat. The rest of this group’s species had greater abundance and were associated mainly to sites 5, 6 and 7 with smaller values of the physicochemical variables previously indicated. A second group of species was closely related to higher salinity changes during the months of study, which tolerate conditions of low salinity, higher temperature, lower depth and transparency in areas of fluvial influence. There are abundant euryhaline and estuarine species such as S. felis, C. spilopterus, C. aguadulce, D. rhombeus, D. auratus and E. plumieri; freshwater species like M. urophtalmus were present as well. Both groups species use the system for protection, growth, feeding or maturation purposes in the “Nortes” season. Castillo et al. (2005), Mendoza et al. (2009), Arceo-Carranza et al. (2010), Díaz-Ruiz et al. (2012), Cowan et al. (2013) and Aguirre-León et al. (2014) reported a similar behavior of fish groups as observed in Sontecomapan for different costal lagoons of Mexico and the United States. It is known that in coastal lagoons fish use multiple habitats as a series of complex responses to environmental factors thus causing variations in species abundance and distribution. Seasonal migrations within and outside the system, as well as resident populations determined by environmental variables also cause these variations (Able, 2005; Cowan et al., 2013; Velázquez-Velázquez et al., 2008; Aguirre- León et al., 2014).

49Our analysis showed that the major changes in the distribution and composition of the fish community, are highly associated with changes in the hydrological behavior of the system that determines the variation of physicochemical variables (transparency, bottom salinity, pH, temperature and dissolved oxygen). In turn, this explains the distribution and migration patterns of fish during their lifecycles. Similar studies have emphasized that the variation of environmental factors such as salinity, temperature and depth determined by the rainy season and the entry of seawater have a significant effect on the distribution and abundance of fish in coastal ecosystems (Cyrus & Blaber, 1992; Barletta et al., 2005; Pombo et al., 2005; Cowan et al., 2008; Velázquez-Velázquez et al., 2008). Biological factors also play an important role in the distribution and abundance of species, in terms of availability changes in food and space during the year. The lifecycle of most fish in the coastal systems is typically estuarine, with seasonal migrations that concur with periods of high productivity and food availability (Díaz-Ruiz et al., 2012; Aguirre-León et al., 2014).

50In Sontecomapan Lagoon the fish have adapted migration strategies and routes for the use of lagoon-estuarine habitats that offer protection and food for juveniles, facilitate spawning, and provide feeding of adult individuals both of marine and estuarine species linked to spatial-temporal changes of the environmental conditions. This latter has also been analyzed by Able (2005), Guevara et al. (2007), Kopp et al. (2007), Díaz-Ruiz et al. (2004, 2012), Zárate-Hernández et al. (2012) and Cowan et al. (2013). At the same time, biological information on species feeding, reproduction and spawning is required to know its relationship with the fish distribution and abundance in the lagoon (Aguirre-León et al., 2010). Therefore, knowledge of the ecological structure of fish community and their interaction with environmental factors in the Sontecomapan lagoon is relevant, since there is no previous record of medium or long term studies in this system. Consequently, the information offered in this study is a starting point for further evaluations and a tool for monitoring the environmental status and ecological behavior with the purpose of preserving its biodiversity and protection status as an important area for breeding, feeding, protection and/or reproduction of fishery resources in this Biosphere Reserve that is also a Ramsar site.

Acknowledgement

51This study was supported by the Universidad Autónoma Metropolitana-Xochimilco during the field surveys and laboratory work. We thank the local fishermen for their support during the fish captures.