Composition of the fish fauna in a tropical estuary: the ecological guild approach

Summary: Ecological guilds have been widely applied for understanding the structure and functioning of aquatic ecosys- tems. This study describes the composition and the spatio-temporal changes in the structure of the fish fauna and the move-ments between the estuary and the coast of a tropical estuary, the Itapissuma/Itamaracá Complex (IIC) in northeastern Brazil. Fish specimens were collected during the dry and rainy seasons in 2013 and 2014. A total of 141 species of 34 families were recorded. Almost half of the species (66 species, 47%) were exclusive to the estuary and 50 species (35%) to the coast; 25 (18%) were common to both environments. Marine species were dominant in both richness and biomass as they explore the environment during part of their life cycle, whereas estuarine species were dominant in abundance. Marine stragglers displayed a higher richness, abundance and biomass in the coastal waters. The estuarine environment was dominated by zoobenthivores in terms of richness, while detritivores prevailed in abundance and biomass. Zoobenthivores had the highest richness and abundance in coastal waters, while piscivores had the highest biomass. The IIC supports a rich fauna with a diverse trophic structure and is an important feeding and development area for migratory species.


INTRODUCTION
The ichthyofauna can be described and classified through the functional attributes of organisms, mainly based on the trophic level, reproductive strategy or use of the environment ). The functional attributes divide the species into guilds, defined as groups of species that exploit the same class of environmental resources in a similar way (Root 1967). The guild approach allows a better understanding of the ecology and role of the biota in the ecosystem ). It may help identify overexploited resources through changes in the composition of the food web (Garrison and Link 2000) and of the energy flows in the system (Harrison and Whitfield 2008). The guild approach also helps to understand the effects of climate changes on the structure and composition of fish fauna (Feyrer et al. 2015).
Trophic and estuarine use guilds have been widely applied to understand the structure and functioning of aquatic ecosystems, the movement pattern between environments and their use as feeding, breeding or development grounds ). Estuarine use guilds reflect migratory patterns and physiological adaptations of species that explore the area throughout their life cycle or part of it . Trophic guilds are useful in the comprehension of the feeding habits of a species ). Its ecological relationships and the energy flows  may reflect the possible strategies for avoiding competition or for optimizing the consumption of available resources (Angel and Ojeda 2001).
Estuaries are important transitional environments for the movement of the ichthyofauna between the continental basins and the ocean (Ray 2005). As an ecotone, estuaries link marine and freshwater ecosystems (Gray and Elliott 2009), and persistent environmental fluctuations place considerable physiological demands on the species inhabiting the area (Elliott and Quintino 2007). Many species are dependent on estuarine environments; several marine species are considered visitors and explore estuarine habitats during their ontogenetic development, evidencing the relationship with coastal environments (Able 2005). Therefore, defining the relationships between species and their functional roles within communities is critical for understanding the dynamics of the ecosystem and fundamental for the implementation of ecosystem-based fisheries management (Buchheister and Latour 2015).
The Brazilian coast hosts large estuarine complexes along the 187 km of the coast of Pernambuco, and several areas are considered of great environmental importance (CPRH 2010). The variety of habitats, along with the complexity of interactions within the fish community and the migratory nature of many species, hampers the assessment of the overall condition of the area .
Using the ecological guilds approach, this study describes the composition and structure of the fish fauna along a tropical estuarine complex in order to identify and explain the main patterns of seasonal and spatial variations in assemblage composition. The study also discusses the importance of the use of the ecological guilds approach to assess the effects of multiple anthropogenic pressures on the structure and functioning of fish communities in tropical estuaries.

Study area
The Itapissuma/Itamaracá Complex (IIC), located in Pernambuco, northeastern Brazil, within the Santa Cruz Environmental Preservation Area (APA Santa Cruz), is considered highly productive (Macêdo et al. 2000), hosting the largest fishery port in the state. Fishery is a very important socio-economical activity in the IIC, generating income and proteins for the local communities (CPRH 2010). Conversely, this ecosystem is exposed to multiple pressures from industrial pollution, domestic sewage discharge, urban expansion, land reclamation and fisheries (Medeiros et al. 2001). In addition, it has a large variety of connecting habitats favouring the development of the ichtyofauna ). The IIC is composed of the estuarine area, the Santa Cruz Channel and the adjacent sea, locally named the "Inner Sea" (Fig. 1). The Santa Cruz channel has a length of 22 km, a width ranging from 0.6 to 1.5 km and a depth ranging from 2 to 5 m in the central part of the channel, reaching 10 m at the northern and southern bars that connect the channel to the sea . The channel bottom consists of quartz sand banks and dark, reductive and dense mud patches. The muddy banks are dominated by Rhyzophora mangle, Laguncularia racemosa, Avicennia sp. and Conocarpus erectus, and by meadows of the marine phanerogam, Halodule wrightii. Surface water temperature varies between 25°C and 31°C and salinity between 18 and 34. The Inner Sea, corresponding to the coastal area hereafter, with a depth of 2 to 5 m, is characterized by a reef barrier parallel to the coast, located 4 km from the beach (Kempf 1970), which functions as a barrier between nearshore and shelf waters. The substrate is formed by terrigenous sediments from the mouth of the Jaguaribe River and the Santa Cruz Channel, and carbonates from the reef barrier (Almeida and Manso 2011), partially covered by large banks of phanerogams (Kempf 1970). The carbonaceous material is the result of the decomposition of rocks and quartz, sand, mollusc shells, foraminifera and calcareous algal fragments. In the Inner Sea, water temperature varies between 27°C and 30.8°C and the average annual salinity is 34.

Data collection
Fish specimens were collected during the dry season (January, February, March, November) and the rainy season (May, July, August) in 2013 and 2014 in the Santa Cruz Channel and the Inner Sea. In order to minimize biases due to gear selectivity, different fishing gears were combined for accessing and sampling different habitats and maximizing the collection of fish individuals (Table S1, Supplementary material). In the estuary, three 25-minute sets with a seine net and one 6-hour set with a block net were carried out quarterly. The seine net was 67.5 m long and had a mesh size of 10 mm. The block net was 348 m long and had a mesh size of 60, 70 and 80 mm. On the coast, samples were obtained quarterly with a gillnet (3 sets of two hours each) and with a fixed tidal trap (6 fishing days). The gill net had mesh sizes of 50, 70 and 80 mm, and was 690 m long, and the fixed tidal trap had a diameter of 27 m and a mesh size of 70 mm.
In the field, the fish fauna was conserved in thermal boxes with ice, and samples were frozen in the laboratory to be identified. Taxonomic classification followed Nelson et al. (2016).

Data analysis
Firstly, we computed a species accumulation curve with the non-parametric Bootstrap method (Smith and van Belle 1984) to assess whether the fish community was exhaustively sampled. This method assumes that all species occur randomly without taking into account species abundance, i.e. the method does not distinguish rare and abundant species (Smith and van Belle 1984). The index and standard deviations of the estimates were obtained through the analytical equation of Colwell et al. (2004) using the EstimateS software v. 9 9.1.0 (Colwell 2013).
The composition of the fish fauna was reported in terms of absolute species richness (S) and, for each species, frequency of occurrence (%FO) and relative abundance in number (%N) and biomass (%B). Species were considered to be abundant according to the Garcia and Vieira (2001) classification when %N was greater than 100/S, where S is the number of species recorded in the area. A species was defined as frequent when its %FO value for a given area was greater than 50%. The combination of these parameters allowed the species to be classified into four categories: abundant and frequent (%N>100/S and %FO≥50%); abundant but infrequent (%N>100/S and %FO<50%); less abundant but frequent (%N<100/S and %FO≥50%) and less abundant and infrequent (%N<100/S and %FO<50%).
Each species was assigned to an estuarine use functional group: marine stragglers, marine migrants and estuarine species, according to the classification proposed by . This classification is based on the type, frequency and period of use of the estuarine environment, and the abundance of the species in the estuary. In addition, each species was assigned to a trophic functional group based on local information about feeding preferences and strategies, according to the categories proposed by . The trophic functional groups were zooplanktivores, detritivores, piscivores, zoobenthivores, herbivores and omnivores. Information on trophic guilds were obtained in studies carried out in the IIC, in the scientific literature or, when not available, based on the WoRMS Editorial Board (2019) and FishBase project (Froese and Pauly 2007) (Table S2, Supplementary material). For each environment (estuary and coast) and season (dry and rainy), the estuarine use guild and the trophic guild were reported in terms of richness (%S), abundance (%N) and biomass (%B).
We computed multivariate analyses to investigate the spatial and temporal variations in the structure of the fish community, considering the absolute richness of estuarine use guild and the richness of trophic guild by environment and by season. To analyse the guild composition, a principal coordinate analysis (PCO) based on Bray-Curtis distances was applied. The differences of the contribution of guilds between environments and seasons was tested by permutational multivariate analysis (PERMANOVA) (Anderson 2001) performed with a Bray-Curtis distance matrix built on square-root-transformed data. Multivariate analyses were performed with the R software (R Core Team 2018).

Fish assemblage
A total of 140 species (135 Actinopterygii and 5 Elasmobranchii) of 34 families were recorded in the IIC (Table 1). For both coastal and estuarine areas, the species accumulation curve did not stabilize towards asymptotic values (Fig. S1, Supplementary material). However, a large portion of the estimated richness was  effectively sampled: 88 species (88% of the estimated richness) were observed in the estuary and 75 species (85% of the estimated richness) on the coast. A total of 25 species (18%) were common to both the estuary and the coast, 65 species (47%) were exclusive to the estuary and 50 species (35%) occurred only on the coast (Table 1). In the estuary, Engraulidae (9 species), Gerreidae (9 species) and Gobiidae (8 species On the coast, Carangidae were dominant in richness (14 species); in abundance, with 288 individuals (49%) and 166 individuals (50%) during the dry and rainy seasons, respectively; and in biomass, with 115 kg (57%) and 203 kg (74%) in the dry and rainy seasons, respectively. In terms of species, Selene brownii (Cuvier, 1816) was dominant with the highest abundance d,uring the dry season (138 individuals; 23%) and Selene vomer (Linnaeus, 1758) during the rainy season (45 individuals, 14%), while Trichiurus lepturus Linnaeus, 1758 (26 kg, 13%) and Caranx hippos (Linnaeus, 1766) (151.59 kg, 55%) dominated in terms of biomass during the dry and the rainy seasons, respectively (Table 1).
Less abundant and infrequent species were dominant in the estuary (85%) and on the coast (77%) (

Estuarine use structure
Richness, abundance and biomass of the estuarine use guilds did not vary by season, but differences were observed between the estuary and the coast. In the estuary, marine stragglers and marine migrants dominated in richness during the dry season (33 species, 43%) and the rainy (23 species, 41%). Estuarine species showed the highest abundance in the dry season (8150 individuals, 66%) and the rainy season (4099 individuals, 64%), but in terms of biomass, marine migrants dominated throughout the year (Fig. 2). On the coast, marine stragglers were dominant in richness (38 species, 70%; 31 species, 65%), in abundance (458 individuals, 78%; 259 individuals, 79%) and in biomass (147 kg, 76%; 238 kg, 90%) in the dry and rainy seasons, respectively (Fig. 2).
The PCO analysis based on the estuarine use guilds revealed that the main effect along the first axis (82.07 %) was spatial as it discriminated the samples from the coast and the estuary. Estuarine samples were very similar between seasons, whereas coastal samples showed a more heterogeneous pattern (Fig. 3). The patterns were tested through PERMANOVA and confirmed the location (estuary and coast) effect (p<0.05). No seasonal effect was observed (Table 2, p=0.01).
According to the PERMANOVA, the environments (estuary and coast) significantly influenced the abundance of the trophic guilds in the IIC (Table 3, p=0.004), confirming the groups formed by PCO (Fig. 5).

DISCUSSION
Overall, species composition of the IIC was similar to that of the fish fauna typically found in other tropical estuaries , Mourão et al. 2014). The observed species richness was close to the estimated richness, indicating that sampling was satisfactory, thanks to the concomitant implementation of active and passive fishing gear. Sampling is known to affect catch composition, especially its diversity (Magurran and McGill 2011), but the use of different gears provides the best estimate of structure (Kwak andPeterson 2007, Mourão et al. 2014) and diversity of fish assemblages (Mérigot et al. 2016). Different gears use different capture processes, mainly based on fish behaviour (Huse et al. 1999). In this study, the use of multiple gears was necessary, considering the differential characteristics of each environment sampled and the fact that fish species explore different habitats of a given environment differently. By exploring multiple gears in different habitats, we improved the estimation of biodiversity, thus providing as wide a variety of guilds as possible.
PCO analysis and PERMANOVA showed spatial differences in the estuarine use functional group between the estuarine and coastal areas, but temporal variations were not evidenced. Spatial segregation processes were observed in other tropical estuaries (Mourão et al. 2014, Loureiro et al. 2016) and may be related to differences in the life cycle and in species tolerance to diverse environmental stresses. Temporal changes in the composition of estuarine fish communities were not observed in the IIC, as reported in other tropical estuaries (Castillo-Rivera et al. 2002, Mendoza et al. 2009).
In the estuary, migrant species predominated in richness and biomass, and estuarine species in abundance. The high richness and biomass of marine species in the estuary can be attributed to the permanent connection between the estuarine area and the Atlantic Ocean throughout the year (Medeiros and Kjerfve 1993), allowing an uninterrupted connectivity with the marine ecosystem (Vasconcelos et al. 2015). Migratory species are of great importance in connected systems, such as estuaries and the adjacent marine area (Harrison and Whitfield 2008). In addition, the IIC is considered a system with high biodiversity and primary and secondary productivity , Mérigot et al. 2016. The positive effect of primary productivity on species richness allows larger populations to persist, thereby reducing extinction risk and supporting a higher diversity of niche specialists (Tittensor et al. 2010). According to Vasconcelos , marine species of the IIC are mostly juveniles, some of which are of commercial value. The high abundance of estuarine species within the estuary of the IIC was mainly due to gobiids. Mérigot et al. (2016) analysed the diversity of fish communities in estuarine complexes in Brazil and revealed differences between assemblages from Itapissuma, especially due to the relatively high abundance of some species of Gobiidae. The high abundance of gobiids in tropical estuaries may be partly due to their prolonged larval duration (Shen and Tzeng 2008), closely linked to the mainly muddy substrate and thus restricting their migrations to the sea .
In the coastal environment of the IIC, the marine stragglers predominated in richness, abundance and biomass in all periods. However, the percentages of resident (estuarine) and dependent (marine migrant) species were also high, thus confirming the dependence between the estuary and coast of the IIC. The connection between continental and marine environments is an essential characteristic, as marine species are important exporters of energy to the adjacent coastal areas ). Also, the coast of the IIC offers favourable conditions for the development of the marine fish fauna as protection and food resource (Medeiros et al. 2001).
In relation to the feeding guild approach, our findings emphasized that the substrate of the IIC is of extreme importance for the high productivity in the system (CPRH 2010), contributing to the high occurrence of species with feeding habits associated with the substrate (i.e. zoobenthivores and detritivores). The high availability of organic rich detritus in mangroves may increase the feeding opportunities for detritivores (Kuo et al. 1999), and can be considered the main trophic contribution factor for the estuarine fish fauna . In north Brazil, Loureiro et al. (2016) observed that fish assemblage was strongly associated with substrates composed of organic matter. The high richness of zoobenthivores in the estuarine area of the IIC can be attributed to the great abundance of available benthic fauna (Silva 2013). Benthos is one of the structuring elements of the food web and plays an important role in the system dynamics (Herman et al. 1999), transferring energy to fishes in estuarine environments (Buchheister and Latour 2015). Detrivores dominated in abundance and biomass mainly due to large supply of organic matter and detritus in the IIC , which support estuarine trophic webs (Hoffman et al. 2008). The estuarine organic material of the IIC originates from various rivers (Eskinazi- Lessa et al. 1999). The river discharge, sediment resuspension, mangrove litter, waste input, terrestrial runoff and atmospheric input are sources of nutrients in the IIC estuary (Medeiros 1991). The highest proportion of detritus usually occurs in environments with great amounts of organic matter. Detritus is consumed, constituting a link between organic production and animal nutrition, and increasing the efficiency of the energy transfer between the trophic levels (Qasim and Sankaranarayanan 1972).
The large supply of zoobenthic fauna (Silva 2013) and the sandy substrate along the coast (Almeida and Manso 2011) favour the high species richness and abundance of zoobenthivores in the IIC coastal area. Benthophagous fish are highly associated with sandy substrates (Loureiro et al. 2016). The dominance of piscivores in biomass is mainly due to large carangids, which benefit from a high supply of food in the coastal area. Carangids are visual, active predators that spend a great part of their time on the reef searching for prey (Cervigón 1972): they feed on fish and also consume benthic prey to complement their diets (Moreno-Sánchez et al. 2016).
Estuaries are dynamic ecosystems subject to notable variability of environmental conditions, and their fish assemblages show within-estuary seasonal and spatial variations, so taking into account this variability should further clarify trait patterns and drivers of estuarine fish (Henriques et al. 2017). The IIC is an important ecosystem for several species that inhabit or visit the area, mainly associated with the substrate. However, coastal areas are exposed to multiple anthropogenic pressures (Blaber and Barletta 2016) that can alter the structure and function of the fish community (Baptista et al. 2015). The anthropogenic stresses and climate changes may facilitate or inhibite the processing of detritus and consequently cause dramatic shifts in species composition, which are often long-lasting and difficult to reverse (Ooi and Chong 2011). The increase in human impacts could significantly affect the topology and functioning of the food web by altering stabilizing elements of the network and decreasing the diversity of trophic flows that ensures the resilience of the trophic structure (Lobry et al. 2008). From the point of view of ecosystem management, it is necessary to identify and understand the biotic and abiotic effects on the distribution of fish fauna as a precursor for the management and monitoring of coastal environments (Pichler et al. 2017).

SUPPLEMENTARY MATERIAL
The following supplementary material is available through the online version of this article and at the following link: http://scimar.icm.csic.es/scimar/supplm/sm04855esm.pdf     Oliveira 1999 Vasconcellos et al. 1998 Chilomycterus spinosus Oliveira 1999 Almeida-Silva et al. 2015