Effect of historical contamination in the fish community structure of a recovering temperate coastal lagoon

MPB-07879; No of Pages 10 Marine Pollution Bulletin xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul

Effect of historical contamination in the fish community structure of a recovering temperate coastal lagoon Eva García-Seoane a,⁎,1, João Pedro Coelho b,c, Cláudia Mieiro b,d, Marina Dolbeth a,d, Tiago Ereira b, José Eduardo Rebelo a, Eduarda Pereira b a

Department of Biology, University of Aveiro, Campus de Santiago, 3810-193 Aveiro, Portugal CESAM & Department of Chemistry, University of Aveiro, Campus de Santiago, 3810-193 Aveiro, Portugal Interdisciplinary Centre of Marine and Environmental Research (CIIMAR), University of Porto, Rua das Bragas, 289, 4050-123 Porto, Portugal d Centre for Functional Ecology, Department of Life Sciences, University of Coimbra, Apartado 3046, 3001-401 Coimbra, Portugal b c

a r t i c l e

i n f o

Article history: Received 27 November 2015 Received in revised form 29 June 2016 Accepted 3 July 2016 Available online xxxx Keywords: Coastal lagoon Fish assemblage Ria de Aveiro Trace metals

a b s t r a c t This study aimed to assess the impact of trace element concentrations in fish assemblages of a recovering coastal lagoon. Fish, water and sediment were sampled in winter and summer in the Ria de Aveiro (Portugal). Multivariate analyses were used to examine the relationship between fish assemblages and environmental variables (physical-chemical parameters, contaminants and sediment grain size). In winter, fish density and biomass were mainly affected by the water turbidity, while Li concentration in the water column was found to be significant for fish biomass. During summer, a significant relationship was found between fish density and temperature, Hg, Li and Zn concentration in the sediment. These contaminants were mainly associated with the historically contaminated area, were Liza spp. and Dicentrarchus labrax appeared as dominant species. Environmental variables were not significant for fish biomass. The historical contamination in the Ria de Aveiro still seems to exert some influence on fish community structure. © 2016 Published by Elsevier Ltd.

1. Introduction Transitional areas, such as estuaries and coastal lagoons, are complex systems comprising a natural gradient of physical-chemical characteristics from river to sea without distinct boundaries. These systems undergo permanently variable conditions due to tide hydrodynamics, diverse sediment geomorphology and fluctuating abiotic parameters (e.g. dissolved oxygen and salinity) (Wołowicz et al., 2007). Nevertheless, transitional areas are among the most productive ecosystems, supporting aquaculture, salt-production, fishing activities, port facilities and industries (Dolbeth et al., 2010; Lillebø et al., 2015; Lopes et al., 2008). Regarding fish communities, these areas are important nursery and feeding habitats as well as part of migration routes for several fish species, supporting the offshore stocks of economically valuable species (Dolbeth et al., 2010; Martinho et al., 2013). Owing to transitional areas high value regarding services provided, they are among the most threatened areas, enduring the impacts of multiple anthropogenic stressors, such as eutrophication (Dolbeth et al., 2011), habitat loss (Kennish, 2002) and chemical contamination ⁎ Corresponding author. E-mail address: [email protected] (E. García-Seoane). 1 Present address: Instituto Português do Mar e Atmosfera (IPMA), Av. Brasília, s/n, 1449-006 Lisboa, Portugal.

(Gravato et al., 2010). Among these stressors, contamination has been recognized as contributing significantly as disturbance sources in estuaries and coastal lagoons, since most contaminants will be deposited in sediments, which act as both sink and source of chemicals, increasing their bioavailability to the aquatic biota (Hill et al., 2013). The acknowledgement of the environmental risks through estuarine contamination has resulted in increasing number of research focusing on abiotic quality criteria and establishing biomarkers of responses in order to accurately evaluate the risks to the aquatic biota (Abukila, 2015; Gonçalves et al., 2013; Gravato et al., 2010; Morelli and Gasparon, 2014). The presence, abundance and distribution of fish communities depends on their capacity to respond to a variety of physical and chemical factors to which they are exposed (Dyer et al., 2000). The ichthyofauna composition in estuaries is usually dynamic, reflecting changes in environmental factors and life history patterns of the various species (Whitfield and Elliott, 2002). As such, fish communities have often been used to illustrate changes in the condition of estuarine environments, in particular as they relate to contamination of these systems (Whitfield and Elliott, 2002). The direct and indirect coupling between fish communities and human impacts on estuaries reinforces the choice of this taxonomic group as a biological indicator (Pérez-Domínguez et al., 2012; Whitfield and Elliott, 2002). Many fishes are suitable as early-warning signals of human-induced stress on natural ecosystem dynamics, or conversely, as indicators of ecosystem recovery and

http://dx.doi.org/10.1016/j.marpolbul.2016.07.005 0025-326X/© 2016 Published by Elsevier Ltd.

Please cite this article as: García-Seoane, E., et al., Effect of historical contamination in the fish community structure of a recovering temperate coastal lagoon, Marine Pollution Bulletin (2016), http://dx.doi.org/10.1016/j.marpolbul.2016.07.005

2

E. García-Seoane et al. / Marine Pollution Bulletin xxx (2016) xxx–xxx

resilience (Holmlund and Hammer, 1999; Mieiro et al., 2010; Mieiro et al., 2014). High levels of metals discharged in aquatic ecosystems may result in selective removal of sensitive life stages of vulnerable fish species, whereas persistent exposure to sub-lethal levels causes reduced growth and condition, among others (Bervoets et al., 2005). Thus, it can be expected that metal contamination will also result in alterations of the fish community (Bervoets et al., 2005). Depending on the tolerances of ichthyofauna, both fish abundance and species diversity can provide managers with a good indication of the ‘stress’ of the system (Whitfield and Elliott, 2002). The Ria de Aveiro is a coastal lagoon located in the north western coast of Portugal. In past years, sewage and effluent discharges from various industries have contributed as point sources of metal contamination to almost all areas of the lagoon (Pacheco et al., 2005). The industrial effluents were the main cause for high levels of metal contamination in sediments and water column, and the most relevant point of entry was the Esteiro de Estarreja, which connects to the Laranjo Basin (Monterroso et al., 2007; Pereira et al., 2009). Nowadays, it is considered historical contamination, restricted to 2-km2 area in Laranjo Basin (Lillebø et al., 2015; Pereira et al., 2009). However, longterm monitoring of the system highlighted persistent risks, despite an overall improvement in local contamination levels, given that metal transport processes were enhanced by hydrological changes and may have increased environmental pressure away from the contamination source (Coelho et al., 2014). Studies in the Ria de Aveiro have focused almost exclusively on mercury (Hg) contamination, reflecting the emission history of a chlor-alkali plant (Pereira et al., 2009). However, Monterroso et al. (2007) found high levels of iron (Fe), manganese (Mn), cadmium (Cd), copper (Cu), lead (Pb) and zinc (Zn) mainly retained in the deep layers of sediments at the Laranjo basin. The same authors also stressed that the surface sediments showed low metal extractability, which indicates low availability to the surrounding biota. Metal mobilization and its transport may occur throughout the lagoon, following sediment disturbance events (e.g. dredging, floods, among others), which may originate sporadic toxicity far from the main polluting source (Coelho et al., 2014). Therefore, it becomes crucial to evaluate the presence of these contaminants along the Ria de Aveiro and their potential impact on biological communities, despite the ongoing recovery of the system. In the Ria de Aveiro, several authors have assessed Hg contamination in fishes at cellular, biochemical and individual level (Abreu et al., 2000; Guilherme et al., 2008; Mieiro et al., 2014; Mieiro et al., 2011b). The effect of Hg was also studied on biological communities, such as macrobenthos (Nunes et al., 2008) and zooplankton (Cardoso et al., 2013). Nevertheless, to our knowledge, the overall metal contamination and potential impacts for the fish community structure have never been evaluated for Ria de Aveiro or other coastal lagoon system. Hence, the objective of this study was to assess the chemical concentrations present in a typical coastal lagoon in the southern Europe and their impact for fish assemblages. This knowledge is valuable to understand the effects of anthropogenic impacts on lagoon ecosystem and thus, relevant to the coastal lagoons management. 2. Materials and methods 2.1. Study area The Ria de Aveiro is a shallow coastal lagoon with a complex morphology. Many branching channels (of which the four main are the Ovar, Murtosa, Ilhavo, and Mira channels) are connected to the ocean by a single tidal outlet, via an intervening tidal lagoon (da Silva et al., 2004; Lillebø et al., 2015). The lagoon has a maximum width and length of 10 and 45 km, respectively, and covers an area of 66–83 km2 during a spring tide. The minimum tidal range is 0.6 m (neap tides), and the maximum tidal range is about 3.2 m (spring tides), corresponding to a

maximum and a minimum water level of 3.5 and 0.3 m, respectively (Dias et al., 2000). Biologically, it is considered rich in nutrients and organic matter and is, therefore, a highly productive environment, providing a habitat for several commercially important fish and invertebrate species (Araújo et al., 2008). 2.2. Sampling strategy Sampling consisted of two campaigns, one conducted in winter (February 2012) and the other in summer (August 2012). The sampling sites (Fig. 1) were located: i) within the first 2–3 km of each main channel entrance (BAR, GAF, SJA), where water residence times are generally lower than 2 days (Dias et al., 2001); ii) at the edges of the main channels (ARE, CAR, VAG), with water residence times of over 14 days (Dias et al., 2001); iii) approximately in the middle of the longest channel (TOR) (approximately 7 days water residence time); iv) in the main freshwater area (RIO); v) in the area that historically showed the highest levels of metal pollution (LAR). Previous studies report that each main channel can be considered as an independent estuary, with almost no particle mixing between them, which excludes secondary pollution from other areas of the system (Dias et al., 2001). Due to bad weather and technical constraints, the sampling was not conducted during winter in CAR. Samples were taken at low tide and during daylight hours. Fish sampling consisted of 3 successive hauls at the same location using a beach seine net, with a final mesh size of 10 mm. The area enclosed by the net was approximately 1550 m2 at all stations except at VAG, were it was 500 m2 due to its narrow topographic configuration. After sampling, fish were taken to the laboratory, where they were frozen (−20 °C) for preservation. At each sampling station, water physico-chemical parameters such as temperature, salinity, dissolved oxygen and pH (WTW Cond. 330i/ set - Tetracon® 325 probe; WTW, model Oxi 330i/SET and WTW pH 330i/set - SenTix® 41 probe) were also measured. Only surface measurements were taken because previous studies point that the Ria the Aveiro should be considered as vertically homogeneous (Dias et al., 1999; Moreira et al., 1993; Rebelo, 1992). At each site, 5 replicate sediment samples were obtained from the top 5 cm layer, and stored in plastic bags and kept cool until arrival at the laboratory. Subsurface water samples were also collected from the adjacent water channel in acid-washed poly(ethylene terephthalate) (PET) bottles and kept on ice during transportation to the laboratory. 2.3. Laboratory methodologies 2.3.1. Sediment analysis At the laboratory, macrodetritus were removed from the sediment and samples were then oven-dried to constant weight at 50 °C, homogenized and sieved through a 2 mm sieve before storage until analysis. Grain size analysis was performed gravimetrically, on a series of sieves, whose fractions were weighted: silt and clay (b 63 μm), very fine sand (63–150 μm), fine sand (150–250 μm), medium sand (250–500 μm), coarse sand (500–1000 μm), very coarse sand and gravel (N1000 μm). These were converted in ϕ scale (−log2 mm) and median grain size determined by GRADISTAT Version 8.0 (Blott and Pye, 2001), using logarithmic Folk and Ward method, with the % of the different sediment fractions. Total mercury in sediments (limit of quantification LOQ – 0.01 μg kg−1) was determined by Atomic Absorption Spectroscopy (AAS) after thermal decomposition and gold amalgamation, using a LECO 254 Advanced Mercury Analyser (AMA) (for more details please see Costley et al. (2000)). The accuracy and precision of the methodology was checked on a daily basis (in the beginning and at the end of the day) through the replicate analysis of Certified Reference Material (CRM), namely Mess-3. The relative standard deviation between replicates was always lower than 9% (n N 3), with a recovery efficiency ranging from 99 to 109% (n = 22).

Please cite this article as: García-Seoane, E., et al., Effect of historical contamination in the fish community structure of a recovering temperate coastal lagoon, Marine Pollution Bulletin (2016), http://dx.doi.org/10.1016/j.marpolbul.2016.07.005

E. García-Seoane et al. / Marine Pollution Bulletin xxx (2016) xxx–xxx

3

Fig. 1. Map of the Ria of Aveiro lagoon. Sampling sites were represented by black points. ARE - Gafanha do Areão, BAR - Costa Nova, CAR - Carregal, GAF - Gafanha de Nazaré, LAR - Largo do Laranjo, RIO - Rio Novo do Príncipe, SJA - São Jacinto, TOR - Torreira and VAG - Vagos. The 2 km2 area of the Laranjo Basin is delimited by the rectangle.

For the other trace elements (arsenic (As, LOQ – 5 mg kg−1), barium (Ba, LOQ – 1 mg kg−1), cobalt (Co, LOQ – 2 mg kg−1), cadmium (Cd, LOQ – 1 mg kg−1), chromium (Cr, LOQ – 2 mg kg−1), copper (Cu, LOQ – 2 mg kg−1), lithium (Li, LOQ – 2 mg kg−1)), manganese (Mn, LOQ – 1 mg kg− 1), nickel (Ni, LOQ – 2 mg kg− 1) and zinc (Zn, LOQ – 1 mg kg− 1) where those quantifiable by ICP-MS (Thermo x Series) quantification, three replicates of both samples and Certified Reference Material (CRM) (Mess-3) were digested in Teflon vessels in a forced air oven, with a mixture of nitric and hydrochloric acid (6 mL HCl and 2 mL HNO3), for 60 min, at 100 °C. The digest was then transferred to 100 mL volumetric flasks and diluted with ultrapure water. Recovery efficiencies ranged between 71% and 98% for the CRM digests.

2.3.2. Water analysis At the laboratory, water samples were filtered through a preweighed, 0.45 μm Millipore cellulose acetate membrane filters. Filtered waters were then acidified with “mercury-free” HNO3 to pH b 2 and stored at 4 °C until analysis. Filters were oven-dried, re-weighed and stored for suspended particulate matter (SPM) analysis. Total dissolved mercury and suspended particulate matter mercury analyses were performed by cold-vapor atomic fluorescence spectrometry (CV-AFS) using a PSA model Merlin 10.023 equipped with a detector PSA model 10.003, with tin chloride as reducing agent (2% in 10% HCl). Total dissolved mercury concentrations were measured after chemical decomposition of sample with potassium persulfate and

Please cite this article as: García-Seoane, E., et al., Effect of historical contamination in the fish community structure of a recovering temperate coastal lagoon, Marine Pollution Bulletin (2016), http://dx.doi.org/10.1016/j.marpolbul.2016.07.005

4

E. García-Seoane et al. / Marine Pollution Bulletin xxx (2016) xxx–xxx

irradiation by a UV lamp. Following irradiation, the excess of oxidant was reduced with a hydroxylamine solution (Pato et al., 2010). Filters were digested with HNO3 4 mol L−1 for determination of the mercury concentration in SPM (for detailed information on the method, please refer to Pato et al. (2010)). Water column Hg concentrations were calculated as the sum of total dissolved + SPM mercury concentrations. Trace elements quantification were performed by analyzing water samples before and after filtration through 0.45 μm pore size Millipore cellulose acetate membrane filters. Both total water column (unfiltered) and dissolved (filtered) trace elements (aluminum (Al, LOQ – 20 μg L−1), barium (Ba, LOQ – 4 μg L−1), iron (Fe, LOQ – 10 μg L−1), lithium (Li, LOQ – 4 μg L−1), manganese (Mn, LOQ – 4 μg L−1) and zinc (Zn, LOQ – 4 μg L−1) where quantifiable in the water samples) were analyzed for trace elements by ICP-MS (Thermo x Series). The relative standard deviation between replicates ranged between 0% and 10%. 2.3.3. Fish community Once defrosted, fish were identified to the lowest possible taxon (usually to species level), counted and weighed. Individuals of the following taxa: Atherina spp., Diplodus spp., Liza spp., Pomatoschistus spp. and Syngnathus spp., were grouped at genus level. 2.4. Data analysis Concentrations of the different trace metals in the sediment were tested with a 2-way ANOSIM for the factors site and season, upon Euclidean distance similarity matrix for each trace metal. For the trace metals in the water column, we tested the differences among seasons with a 1-way ANOSIM. Then, environmental variables were analyzed with a PCA to clarify spatial and seasonal variation patterns. Prior to the analyses, collinear environmental variables were removed, after inspection of the correlation between variables and the variation inflation factor (VIF N 3, Zuur et al. (2010)) and variables were normalised. The relationship between fish and the environmental variables (physical-chemical parameters, trace metals and sediment grain size) was explored with canonical analyses. Initially, we applied a detrended correspondence analysis (DCA) with the fish data (species density and biomass) to evaluate the type of model response for the canonical analyses. Species data were square root transformed, in order to scale down the scores of the highly abundant species. As a linear response was detected with the DCA (length of gradient b 3), a redundancy analysis (RDA) was applied to examine the relationships between biotic and abiotic parameters. The significance of the non-collinear environmental variables was evaluated with the forward selection procedure (Monte Carlo permutation tests). For the fish data without a significant relationship with the environmental data, a non-metric MDS analyses were applied to clarify variation spatial patterns. Correlation and VIF values were determined with R software (R Development Core Team, 2012), PCA and RDA with CANOCO v 4.5 software (Ter Braak and Smilauer, 2002), and ANOSIM and nm-MDS with PRIMER software (Clarke et al., 2014). 3. Results 3.1. Environmental data 3.1.1. Sediments characterization and trace elements concentrations Very fine to fine sands were dominant in the innermost areas of the lagoon, namely in the Ílhavo channel (VAG and GAF) and in the Murtosa channel (LAR, Fig. 2). For RIO, fine sands characterized the grain size during winter, becoming slightly coarser in the summer, i.e., medium sands. For the sampling stations in the western channels (i.e., BAR, ARE, TOR and SJA), coarse and very coarse sands were dominant for both seasons (Fig. 2), with exception of BAR during winter, which median grain size were medium sands. While for BAR, TOR and SJA sites the stronger currents and low water residence times (Dias et al., 2001) may

Fig. 2. Sediment median grain size (φ) in Ria de Aveiro, for winter (February) and summer conditions (August), with indication of the sediment classification. ARE - Gafanha do Areão, BAR - Costa Nova, CAR - Carregal, GAF - Gafanha de Nazaré, LAR - Largo do Laranjo, RIO - Rio Novo do Príncipe, SJA -São Jacinto, TOR - Torreira and VAG - Vagos.

justify the coarser sediments, in ARE this fact is justified by the sea intrusion during winter storms in 2011, transporting coarser beach sediments into the system (personal observation). Details on median grain size analyses, including information on the sorting, skewness and kurtosis are available in Table 2S (Supplementary material). In most samples, grain size was well to moderately sorted, but some samples were poorly sorted (e.g. LAR, SJA, Table 2S). In sediments, trace element concentrations were lower than 190 mg kg−1 for Mn, Zn and Li, lower than 50 mg kg−1 for Cu, Ba, Ni, Pb and As and lower than 8 mg kg−1 for Hg, Cd and Co (Fig. 3). The highest values for Hg, As and Zn were registered in LAR sampling point. The trace elements concentration in the sediments varied among sites and seasons, particularly for some sampling stations (Fig. 3). In fact, we found statistically significant differences for all elements for the differences between site groups (global R N 0.3, p = 0.001) and between seasons (global R N 0.4, p = 0.001). For Ba, Pb, Ni, Mn, Li and Zn we found significant differences within almost all sites (global tests and pairwise tests for each trace element are available in Table 3S from Supplementary material). However, for Zn, the global R for the differences between season groups were low (0.4), indicating a low discrimination between seasons. For As and Cd, the global R for the differences between sites was 0.3 and 0.4 respectively, also indicating a low discrimination between sites, resulting in non-significant differences among several sites (Table 3S). 3.1.2. Trace elements in the water column In the water column, trace elements concentration was always lower than 500 μg L−1 for Al, Li and Fe and lower than 100 μg L−1 for Hg, Mn, Zn, and Ba (Fig. 4). Seasonal patterns of trace element contamination in water are statistically significant only for Zn, which was higher in winter (ANOSIM, global R = 0.4, p = 0.001). Station LAR showed the highest values for Hg contamination. For the other elements, more than one sampling station presented high values, for instance, Ba was highest in ARE and VAG (Fig. 4). The overall environmental data was analyzed, after removing collinear variables, to depict on spatial variation patterns (PCA, Fig. 5). The non collinear variables for winter conditions were: Li, Al, Zn, Hg, Mn in the water column; Mn, Pb, Hg in the sediment; grain size above 1000 μm, between 250 and 150 μm and below 0.63 μm; pH, salinity, temperature and turbidity (Fig. 5a). For the summer, the variables tested were Fe, Mn, Zn in the water column; Hg, Cd and Li in the sediment, grain size above 1000 μm, between 500–250 μm and 250–125 μm; pH, salinity, temperature and turbidity (Fig. 5b). For winter the PCA ordination plot explained 60% of variation from the two first axes (Fig. 5a) and for summer 57.3%, also from the two first axes (Fig. 5b, Table 4S from Supplementary material). LAR was always associated with higher contamination by Hg in both seasons, and lower sediment fractions (mainly silts, Fig. 5). Other contaminants in sediment also associate to LAR, which was similar to VAG with regard to the environmental descriptors.

Please cite this article as: García-Seoane, E., et al., Effect of historical contamination in the fish community structure of a recovering temperate coastal lagoon, Marine Pollution Bulletin (2016), http://dx.doi.org/10.1016/j.marpolbul.2016.07.005

E. García-Seoane et al. / Marine Pollution Bulletin xxx (2016) xxx–xxx

5

Fig. 3. Concentration of trace elements (mg kg−1, dry weight) found in sediments along the Ria de Aveiro sampling sites: ARE - Gafanha do Areão, BAR - Costa Nova, CAR - Carregal, GAF Gafanha de Nazaré, LAR - Largo do Laranjo, RIO - Rio Novo do Príncipe, SJA - São Jacinto, TOR - Torreira and VAG - Vagos.

During winter, turbidity, salinity and concentrations of Li, Zn and Al in the water column were generally higher and similar for the stations closer to the lagoon's entrance, i.e., BAR, TOR and SJA (Fig. 5a). However, this pattern changed in the summer, as higher turbidity was found for LAR and VAG, together with several contaminants in the sediments and a tendency for higher salinity (Fig. 5b). The other stations were discriminated more by the dominance of a particular grain size, rather than by contaminants concentrations (e.g. ARE characterized by coarse sands and GAF by very fine sands). Still, CAR and TOR were similar with regard to the concentrations of Mn and Zn in the water column (Fig. 5b). 3.2. Relationship between environmental and biological variables Initially, 30 environmental variables were explored for the RDA analyses: dissolved oxygen, water temperature, salinity, pH, turbidity,

different percentage of grain sizes and all trace elements concentrations in the water column and in the sediment. These variables were first checked for co-linearity and if appropriate removed from the forward selection procedure described in the Methods section. The noncollinear variables tested were the ones mentioned in the PCA analyses. A second analysis was performed only with the significant variables from the non-collinear ones, defined with the Monte Carlo permutations tests (Fig. 6). In the winter, fish density and biomass were mainly affected by the water turbidity (Fig. 6a, b). For the fish density, the RDA explained 46% of total variability (Table 5S from Supplementary material for more details). Both SJA and TOR were characterized by higher turbidity (Fig. 5a), but their fish composition taking into account density levels were different: Pomatoschistus spp., European pilchard Sardina pilchardus (Walbaum, 1792), Tub gurnard Chelidonichthys lucerna (Linnaeus,

Please cite this article as: García-Seoane, E., et al., Effect of historical contamination in the fish community structure of a recovering temperate coastal lagoon, Marine Pollution Bulletin (2016), http://dx.doi.org/10.1016/j.marpolbul.2016.07.005

6

E. García-Seoane et al. / Marine Pollution Bulletin xxx (2016) xxx–xxx

Fig. 4. Mean concentration of trace elements (μg L−1) found in the water column along the Ria de Aveiro sampling sites: ARE - Gafanha do Areão, BAR - Costa Nova, CAR - Carregal, GAF Gafanha de Nazaré, LAR - Largo do Laranjo, RIO - Rio Novo do Príncipe, SJA - São Jacinto, TOR - Torreira and VAG - Vagos.

1758) and Black goby Gobius niger Linnaeus, 1758 were associated to SJA, while Great sandeel Hyperoplus lanceolatus (Le Sauvage, 1824) and Atherina spp. had higher abundance in TOR (Fig. 6a). The stations LAR, BAR, GAF and VAG had similar fish composition, while RIO and ARE were distinct from all other samples and associated to lower turbidity (Fig. 6a). For the fish biomass, Li concentration in the water column also becomes significant, and together with turbidity explained 70% of total variation (Fig. 6b, Table 5S). Again, SJA and TOR were associated to higher turbidity and were similar regarding the fish biomass distribution of their communities. Li concentration was higher in BAR, which had several associated species (Fig. 6b). These species were the same as for the density plot, but with higher biomass expression for BAR sampling station alone. During summer, a significant relationship was found between fish density and temperature, Hg, Li and Zn concentration in the sediment,

which together with the biological data explained 61% of total variability (Fig. 6c, Table 5S). Higher concentrations of those metals were associated to LAR, a pattern also revealed before (Fig. 5b), with Liza spp. as dominant species in that area and those environmental conditions (Fig. 6c). Several species were associated to CAR, which registered high temperature during summer (but always lower than 27 °C). TOR, VAG and GAF were also associated to high temperature and to a similar fish composition dominated by Atherina spp. (Fig. 6c). The remaining stations were associated with lower temperature and lower Hg, Li and Zn concentration, with a different fish composition, such as Diplodus spp., Senegalese sole Solea senegalensis Kaup, 1858 and Corkwing wrasse Symphodus melops (Linnaeus, 1758) (Fig. 6c). For fish biomass patterns, environmental variables were not significant. Regarding fish biomass composition, CAR clearly distinguished from the remaining sampling stations, with several species occurring in that area (Fig. 6d).

Please cite this article as: García-Seoane, E., et al., Effect of historical contamination in the fish community structure of a recovering temperate coastal lagoon, Marine Pollution Bulletin (2016), http://dx.doi.org/10.1016/j.marpolbul.2016.07.005

E. García-Seoane et al. / Marine Pollution Bulletin xxx (2016) xxx–xxx

Fig. 5. PCA ordination of samples for winter and summer conditions with regard to the non-collinear environmental variables (grey vectors).

The remaining stations had a different fish biomass composition among each other, except for RIO, ARE and LAR, which presented more similarity (Fig. 6d). 4. Discussion In this work, we evaluated the impact of historical metal contamination in the fish community structure, taking into account a wide range of trace metals, which have not been studied before in the system. This sort of studies is of extreme importance in the frame of European policies aiming for the ecological protection of transitional and coastal areas, such as the Water Framework Directive (WFD). In addition, we highlighted which species or taxonomic groups were associated to the areas with higher contamination. Overall, sediments of the Ria de Aveiro presented low levels of trace elements contamination, compared to the Portuguese legislation for dredged material (Portaria no. 1450/2007). The levels of As, Cd, Cr, Cu,

7

Hg, Pb, Ni and Zn are considered clean material for the entire lagoon, except for Hg at LAR and Cd at VAG, ranging from vestigial (summer) to moderately contaminated (winter). Nevertheless, having in mind the Canadian Interim Quality Criteria for Marine sediments (ISQC) for the protection of the aquatic life, all sites exceeded this threshold (0.7 mg kg−1, dry weight) for Cd, LAR and VAG exceed the limit for Cu (18.7 mg kg−1, dry weight), LAR exceeded the ISQC for Zn and Hg (124 and 0.13 mg kg−1 dry weight, respectively) and VAG exceed the limit for Pb (30.2 mg kg−1, dry weight). Regarding the probability of risk to biota (PEL), only the levels of Hg found at LAR exceeded this threshold (0.7 mg kg−1, dry weight). There are no comparative concentrations for Li and Mn, while Ba and Co can only be compared with the Canadian limits allowed for soil, being lower than the levels for all kind of soil utilization. Water column trace elements concentrations were also within the Environmental Quality Standards (EQS) established in the WFD, except for mercury in Laranjo Basin (LAR), which exceed the permitted value (0.07 μg L−1, 2008/105/EC). These results support the idea of an ongoing recovery process in the lagoon, after industrial point source contamination in the Ria de Aveiro was controlled (Coelho et al., 2014). Currently, metal contamination can be considered historical and restricted to a small area (2 km2 in Laranjo Basin). Chemical contamination may impact fish communities' structure and function, conducting, for e.g., to decreased diversity levels and of particular fish ecological guilds (Fonseca et al., 2013) and fish biochemical responses (Guilherme et al., 2008; Mieiro et al., 2014; Souid et al., 2013). In the Ria de Aveiro, the levels for the trace metals analyzed were within the permitted thresholds in most of the system, and previous studies have showed that the ecological quality based on the fish communities was not affected (e.g. Ria de Aveiro's fish communities classified in a good ecological status (Cabral et al., 2012; Fonseca et al., 2013). However, there are still risks of impact for biota, particularly due to accidental dredging or high precipitation events inducing strong currents, which might increase during winter periods. In this study, metal contamination during winter did not show a significant influence for the fish density distribution. Turbidity was the only factor studied that explained variations in the fish community structure. The effect of turbidity on the fish community structure might be related to an increase of food supply and a reduction of predation pressure (Blaber and Blaber, 1980; Costa et al., 2002). Higher turbidity was found in the stations nearest to the lagoon's mouth of the North Channel, SJA and TOR. The species associated to those stations were mainly typical estuarine residents, such as the Common goby Pomatoschistus microps (Krøyer, 1838), the Sand goby Pomatoschistus minutus (Pallas, 1770), G. niger and the Big-scale sand smelt Atherina boyeri Risso, 1810, but also some juveniles of marine migrants such as the Sand smelt Atherina presbyter Cuvier, 1829, C. lucerna and S. pilchardus. In the Ria de Aveiro, the diet of S. pilchardus and P. microps is strongly dominated by detritus (Pombo et al., 2005), which explains, in part, the association of those species with higher turbidity areas, particularly during the winter periods when food sources may decrease. Overall, turbidity was lower in winter than in summer, when this factor is no longer significant, in agreement to Marshall and Elliott (1998), who reported that the effects of turbidity in fish distribution are negligible at high turbidity levels. The same winter analyses with fish biomass revealed that Li concentration in the water column, together with turbidity, could be influencing the fish assemblages of the Ria de Aveiro. While not being regulated and therefore usually not monitored, industrial and consumer use of Li in drugs, batteries, and alloys has increased dramatically over the past decade, which can potentially cause environmental effects, given its psychoactive characteristics (Tkatcheva et al., 2015). Li has been shown to cause a similar toxic effect to that of copper in Daphnia magna, such as energy production and ionoregulation impairment (Nagato et al., 2013). Additionally, in a short-term environmentally

Please cite this article as: García-Seoane, E., et al., Effect of historical contamination in the fish community structure of a recovering temperate coastal lagoon, Marine Pollution Bulletin (2016), http://dx.doi.org/10.1016/j.marpolbul.2016.07.005

8

E. García-Seoane et al. / Marine Pollution Bulletin xxx (2016) xxx–xxx

Fig. 6. Ordination plots of fish density and fish biomass, with the significant environmental variables when a significant relationship was found, in winter and summer conditions. a, b, c) RDA ordination triplot relating species density or biomass (grey vector lines) and significant environmental variables (after Monte Carlo permutation tests, black vector lines, whose length is proportional to their relative significance); d) nm-MDS ordination plot of the fish community biomass, with indication of the species associated to samples (Spearman correlation N 0.5).

relevant waterborne exposure, Li was found to be highly bioavailable to fish and to efficiently cross the blood–brain barrier, exerting effects on ion regulation (Tkatcheva et al., 2015). Therefore, the high water column concentrations of this metal in winter, mainly in the more urban sampling stations (BAR, SJA and TOR), seem to be significantly affecting the resident fish communities. Contrary to what was observed in winter, and despite slightly lower concentrations, in summer sediment contamination (Hg, Li and Zn) significantly influenced the fish community structure in terms of density patterns. These contaminants were mainly associated with the restricted historically contaminated area by industrial effluents, the Laranjo Basin. Differences between seasons are probably related with several environmental parameters, rather than sediment contamination levels per se. Temperature plays a significant role in contaminant bioavailability and toxicity to biota, and therefore may indirectly affect the fish assemblages. Mercury toxicity is known to increase with temperature (Boening, 2000), as a result of increased bacterial mediated methylation rates in the sediments (Mauro et al., 1999), and hence can be excluding the least resistant species from the impacted area. Additionally, the significance of sediment contamination in the summer fish community structure may be a result of increased turbidity. Despite not being a significant parameter for fish distribution patterns in summer, higher turbidity reflects increased sediment re-suspension rates, which will affect the release and bioavailability of contaminants from sediments (Eggleton and Thomas, 2004) and may hence justify the significant effect of sediment associated metals in the summer campaign. Sediment contamination has been associated with effects on both benthic and pelagic species (Chapman and Wang, 2001). In the lagoon,

the species that associated the most with this contaminated area were the pelagic mullets (Liza spp.). Mullets are often found in impacted areas and have been previously recorded in the most contaminated site, suggesting some degree of tolerance to metal contamination, particularly Hg (Mieiro et al., 2012; Mieiro et al., 2011a). The seabass Dicentrarchus labrax (Linnaeus, 1758) was also found in the areas with Hg, Li and Zn contamination, but considering the low overall contamination of the system, the absence of other fish species should not be attributed solely to metal concentrations, as it may occur through a combination of both environmental parameters and geographical/topographical constrains. A higher diversity of species was associated to CAR, the uppermost area of the North Channel. CAR was characterized by fine sands and higher concentrations of Mn and Zn in the water column. However, these concentrations did not seem to affect the fish diversity. None of the environmental factors studied explained the variations in the fish biomass community structure during summer. As such, there was no evidence for a contaminant effect on the fish biomass levels, despite its effect on density levels. Still, the fish distribution patterns was somewhat similar to the one found with biomass, with several species associated to CAR. Other unmeasured physicochemical factors and/or biological interactions may be responsible for the observed fish abundance and biomass distribution. For example, depth possibly affects the fish community (Pombo et al., 2005). Environmental factors play the initial role in structuring the fish community, but biological interactions are then superimposed on that structure (Marshall and Elliott, 1998). Abiotic parameters appear to have a greater effect on fish distribution than

Please cite this article as: García-Seoane, E., et al., Effect of historical contamination in the fish community structure of a recovering temperate coastal lagoon, Marine Pollution Bulletin (2016), http://dx.doi.org/10.1016/j.marpolbul.2016.07.005

E. García-Seoane et al. / Marine Pollution Bulletin xxx (2016) xxx–xxx

biological interactions, given the Ria de Aveiro food abundance (Pombo et al., 2005). A recent evaluation of the Ria de Aveiro, in light of the WFD, highlighted some environmental issues that can also influence the resident fish assemblages. Chemical contamination with tetrachloroethene was observed in the central area, comprising RIO and SJA sites, as well as LAR (MAMAOT/ARHCentro, 2012), potentially affecting the distribution of sensitive fish species. Additionally, while not a very significant problem in the majority of the lagoon, excess nutrients and eutrophication processes can still occur in some upstream areas (Lopes et al., 2007), which has resulted in elevated phytoplankton concentrations in the northern portion of the system, comprising the CAR site (MAMAOT/ ARHCentro, 2012). The elevated fish assemblage biodiversity and biomass in this site can therefore be positively correlated with this potential food source, which will be more relevant in summer time. 5. Concluding remarks In conclusion, the historical contamination in the Ria de Aveiro still seems to exert some influence on fish community structure, particularly in summer. Overall, the levels for the trace metals analyzed were within the permitted thresholds in most of the lagoon. In winter, metal contamination did not show a significant influence for the fish density distribution, but in summer, sediment contamination (Hg, Li and Zn) significantly influenced the fish community structures in terms of density. These contaminants were mainly associated with the historically and restricted contaminated area by industrial effluents, the Laranjo Basin. The species that associated the most with this contaminated area were the pelagic mullets (Liza spp.) and seabass (Dicentrarchus labrax). The elevated fish diversity associated to the uppermost area of the North Channel can be correlated with a potential food source due to the excess nutrients and eutrophication processes in this area. Acknowledgements The authors thank the fishermen's association APARA, all the local fishermen, and the colleagues who assisted in the field sampling. This work was support funded by Grupo de Ação Costeira da Região de Aveiro (GAC-RA) through PROMAR - Desenvolvimento sustentável das zonas de pesca (2007/2013) through the projects “Enguias na Ria de Aveiro, um ex-líbris a preservar: biologia, sanidade e pesca” and “Promoção dos Recursos Endógenos da Ria: Conhecer, Usufruir e Preservar”. The authors are grateful to F. Martinho for helping with the Ria de Aveiro map. C. Mieiro thanks for the postdoctoral grant SFRH/BPD/79445/2011, J. P. Coelho for the postdoctoral grant SFRH/ BPD/102870/2014, and M. Dolbeth for the postdoctoral grant SFRH/ BPD/110441/2015.

Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.marpolbul.2016.07.005.

References Abreu, S.N., Pereira, E., Vale, C., Duarte, A.C., 2000. Accumulation of mercury in sea bass from a contaminated lagoon (Ria de Aveiro, Portugal). Mar. Pollut. Bull. 40, 293–297. Abukila, A.F., 2015. Assessing the drain estuaries' water quality in response to pollution abatement. Water Science 29, 1–18. Araújo, I.B., Dias, J.M., Pugh, D.T., 2008. Model simulations of tidal changes in a coastal lagoon, the Ria de Aveiro (Portugal). Cont. Shelf Res. 28, 1010–1025. Bervoets, L., Knaepkens, G., Eens, M., Blust, R., 2005. Fish community responses to metal pollution. Environ. Pollut. 138, 338–349. Blaber, S.J.M., Blaber, T.G., 1980. Factors affecting the distribution of juvenile estuarine and inshore fish. J. Fish Biol. 17, 143–162. Blott, S.J., Pye, K., 2001. GRADISTAT: a grain size distribution and statistics package for the analysis of unconsolidated sediments. Earth Surf. Process. Landf. 26, 1237–1248.

9

Boening, D.W., 2000. Ecological effects, transport, and fate of mercury: a general review. Chemosphere 40, 1335–1351. Cabral, H.N., Fonseca, V.F., Gamito, R., Gonçalves, C.I., Costa, J.L., Erzini, K., Gonçalves, J., Martins, J., Leite, L., Andrade, J.P., Ramos, S., Bordalo, A., Amorim, E., Neto, J.M., Marques, J.C., Rebelo, J.E., Silva, C., Castro, N., Almeida, P.R., Domingos, I., Gordo, L.S., Costa, M.J., 2012. Ecological quality assessment of transitional waters based on fish assemblages in Portuguese estuaries: the Estuarine Fish Assessment Index (EFAI). Ecol. Indic. 19, 144–153. Cardoso, P.G., Marques, S.C., D'Ambrosio, M., Pereira, E., Duarte, A.C., Azeiteiro, U.M., Pardal, M.Â., 2013. Changes in zooplankton communities along a mercury contamination gradient in a coastal lagoon (Ria de Aveiro, Portugal). Mar. Pollut. Bull. 76, 170–177. Chapman, P.M., Wang, F., 2001. Assessing sediment contamination in estuaries. Environ. Toxicol. Chem. 20, 3–22. Clarke, K.R., Somerfield, P.J., Gorley, R.N., Warwick, R.M., 2014. Change in Marine Communities: An Approach to Statistical Analysis and Interpretation. third ed. PRIMER-E Ltd. Coelho, J.P., Pato, P., Henriques, B., Picado, A., Lillebø, A.I., Dias, J.M., Duarte, A.C., Pereira, M.E., Pardal, M.A., 2014. Long-term monitoring of a mercury contaminated estuary (Ria de Aveiro, Portugal): the effect of weather events and management in mercury transport. Hydrol. Process. 28, 352–360. Costa, M.J., Santos, C.I., Cabral, H.N., 2002. Comparative analysis of a temperate and a tropical seagrass bed fish assemblages in two estuarine systems: the Mira estuary (Portugal) and the Mussulo Lagoon (Angola). Cah. Biol. Mar. 43, 73–81. Costley, C.T., Mossop, K.F., Dean, J.R., Garden, L.M., Marshall, J., Carroll, J., 2000. Determination of mercury in environmental and biological samples using pyrolysis atomic absorption spectrometry with gold amalgamation. Anal. Chim. Acta 405, 179–183. da Silva, J.F., Duck, R.W., Catarino, J.B., 2004. Seagrasses and sediment response to changing physical forcing in a coastal lagoon. Hydrol. Earth Syst. Sc. 8, 151–159. Dias, J.M., Lopes, J.F., Dekeyser, I., 1999. Hydrological characterisation of Ria de Aveiro, Portugal, in early summer. Oceanol. Acta 22, 473–485. Dias, J.M., Lopes, J.F., Dekeyser, I., 2000. Tidal propagation in Ria de Aveiro lagoon, Portugal. Phys. Chem. Earth (B) 25, 369–374. Dias, J.M., Lopes, J.F., Dekeyser, I., 2001. Lagrangian transport of particles in Ria de Aveiro lagoon, Portugal. Phys. Chem. Earth (B) 26, 721–727. Dolbeth, M., Martinho, F., Freitas, V., Costa-Dias, S., Campos, J., Pardal, M.Â., 2010. Multiyear comparisons of fish recruitment, growth and production in two droughtaffected Iberian estuaries. Mar. Freshw. Res. 61, 1399–1415. Dolbeth, M., Cardoso, P.G., Grilo, T.F., Bordalo, M.D., Raffaelli, D., Pardal, M.A., 2011. Longterm changes in the production by estuarine macrobenthos affected by multiple stressors. Estuar. Coast Shelf S 92, 10–18. Dyer, S.D., White-Hull, C., Carr, G.J., Smith, E.P., Wang, X., 2000. Bottom-up and top-down approaches to assess multiple stressors over large geographic areas. Environ. Toxicol. Chem. 19, 1066–1075. Eggleton, J., Thomas, K.V., 2004. A review of factors affecting the release and bioavailability of contaminants during sediment disturbance events. Environ. Int. 30, 973–980. Fonseca, V.F., Vasconcelos, R.P., Gamito, R., Pasquaud, S., Gonçalves, C.I., Costa, J.L., Costa, M.J., Cabral, H.N., 2013. Fish community-based measures of estuarine ecological quality and pressure–impact relationships. Estuar. Coast Shelf S 134, 128–137. Gonçalves, C., Martins, M., Costa, M.H., Caeiro, S., Costa, P.M., 2013. Ecological risk assessment of impacted estuarine areas: integrating histological and biochemical endpoints in wild Senegalese sole. Ecotoxicol. Environ. Saf. 95, 202–211. Gravato, C., Guimarães, L., Santos, J., Faria, M., Alves, A., Guilhermino, L., 2010. Comparative study about the effects of pollution on glass and yellow eels (Anguilla anguilla) from the estuaries of Minho, Lima and Douro Rivers (NW Portugal). Ecotoxicol. Environ. Saf. 73, 524–533. Guilherme, S., Válega, M., Pereira, M.E., Santos, M.A., Pacheco, M., 2008. Erythrocytic nuclear abnormalities in wild and caged fish (Liza aurata) along an environmental mercury contamination gradient. Ecotoxicol. Environ. Saf. 70, 411–421. Hill, N.A., Simpson, S.L., Johnston, E.L., 2013. Beyond the bed: effects of metal contamination on recruitment to bedded sediments and overlying substrata. Environ. Pollut. 173, 182–191. Holmlund, C.M., Hammer, M., 1999. Ecosystem services generated by fish populations. Ecol. Econ. 29, 253–268. Kennish, M.J., 2002. Environmental threats and environmental future of estuaries. Environ. Conserv. 29, 78–107. Lillebø, A.I., Ameixa, O.M.C.C., Sousa, L.P., Sousa, A.I., Soares, J.A., Dolbeth, M., Alves, F.L., 2015. The physio-geographical background and ecology of Ria de Aveiro. In: Lillebø, A.I., Stålnacke, P., Gooch, G.D. (Eds.), Coastal Lagoons in Europe: Integrated Water Resource Strategies. International Water Association (IWA), UK, pp. 21–29. Lopes, C.B., Lillebø, A.I., Dias, J.M., Pereira, E., Vale, C., Duarte, A.C., 2007. Nutrient dynamics and seasonal succession of phytoplankton assemblages in a Southern European Estuary: Ria de Aveiro, Portugal. Estuar. Coast Shelf S 71, 480–490. Lopes, C.B., Lillebø, A.I., Pato, P., Dias, J.M., Rodrigues, S.M., Pereira, E., Duarte, A.C., 2008. Inputs of organic carbon from Ria de Aveiro coastal lagoon to the Atlantic Ocean. Estuar. Coast Shelf S 79, 751–757. MAMAOT/ARHCentro, 2012. Plano de Gestão das Bacias Hidrográficas dos rios Vouga, Mondego e Lis Integrados na Região Hidrográfica 4, Parte 2 – Caracterização Geral e Específica, 1.4.1 – Caracterização das Massas de Águas Superficiais. Ministério da Agricultura, Mar, Ambiente e Ordenamento do Território/Administração da Região Hidrográfica do Centro I.P. Marshall, S., Elliott, M., 1998. Environmental influences on the fish assemblage of the Humber Estuary, U.K. Estuar. Coast Shelf S 46, 175–184. Martinho, F., van der Veer, H.W., Cabral, H.N., Pardal, M.A., 2013. Juvenile nursery colonization patterns for the European flounder (Platichthys flesus): a latitudinal approach. J. Sea Res. 84, 61–69.

Please cite this article as: García-Seoane, E., et al., Effect of historical contamination in the fish community structure of a recovering temperate coastal lagoon, Marine Pollution Bulletin (2016), http://dx.doi.org/10.1016/j.marpolbul.2016.07.005

10

E. García-Seoane et al. / Marine Pollution Bulletin xxx (2016) xxx–xxx

Mauro, J.B.N., Guimarães, J.R.D., Melamed, R., 1999. Mercury methylation in a tropical macrophyte: influence of abiotic parameters. Appl. Organomet. Chem. 13, 631–636. Mieiro, C.L., Ahmad, I., Pereira, M.E., Duarte, A.C., Pacheco, M., 2010. Antioxidant system breakdown in brain of feral golden grey mullet (Liza aurata) as an effect of mercury exposure. Ecotoxicology 19, 1034–1045. Mieiro, C.L., Duarte, A.C., Pereira, M.E., Pacheco, M., 2011a. Mercury accumulation patterns and biochemical endpoints in wild fish (Liza aurata): a multi-organ approach. Ecotoxicol. Environ. Saf. 74, 2225–2232. Mieiro, C.L., Pacheco, M., Pereira, M.E., Duarte, A.C., 2011b. Mercury organotropism in feral European sea bass (Dicentrarchus labrax). Arch. Environ. Contam. Toxicol. 61, 135–143. Mieiro, C.L., Coelho, J.P., Pacheco, M., Duarte, A.C., Pereira, M.E., 2012. Evaluation of species-specific dissimilarities in two marine fish species: mercury accumulation as a function of metal levels in consumed prey. Arch. Environ. Contam. Toxicol. 63, 125–136. Mieiro, C.L., Dolbeth, M., Marques, T.A., Duarte, A.C., Pereira, M.E., Pacheco, M., 2014. Mercury accumulation and tissue-specific antioxidant efficiency in the wild European sea bass (Dicentrarchus labrax) with emphasis on seasonality. Environ. Sci. Pollut. Res. 21, 10638–10651. Monterroso, P., Pato, P., Pereira, M.E., Millward, G.E., Vale, C., Duarte, A., 2007. Metalcontaminated sediments in a semi-closed basin: implications for recovery. Estuar. Coast Shelf S 71, 148–158. Moreira, M.H., Queiroga, H., Machado, M.M., Cunha, M.R., 1993. Environmental gradients in a southern Europe estuarine system: Ria de Aveiro, Portugal implications for soft bottom macrofauna colonization. J Aquat Ecol]–>Neth. J. Aquat. Ecol. 27, 465–482. Morelli, G., Gasparon, M., 2014. Metal contamination of estuarine intertidal sediments of Moreton Bay, Australia. Mar. Pollut. Bull. 89, 435–443. Nagato, E.G., D'eon, J.C., Lankadurai, B.P., Poirier, D.G., Reiner, E.J., Simpson, A.J., Simpson, M.J., 2013. 1H NMR-based metabolomics investigation of Daphnia magna responses to sub-lethal exposure to arsenic, copper and lithium. Chemosphere 93, 331–337. Nunes, M., Coelho, J.P., Cardoso, P.G., Pereira, M.E., Duarte, A.C., Pardal, M.A., 2008. The macrobenthic community along a mercury contamination in a temperate estuarine system (Ria de Aveiro, Portugal). Sci. Total Environ. 405, 186–194. Pacheco, M., Santos, M.A., Teles, M., Oliveira, M., Rebelo, J.E., Pombo, L., 2005. Biotransformation and genotoxic biomarkers in mullet species (Liza sp.) from a contaminated coastal lagoon (Ria de Aveiro, Portugal). Environ. Monit. Assess. 107, 133–153.

Pato, P., Otero, M., Válega, M., Lopes, C.B., Pereira, M.E., Duarte, A.C., 2010. Mercury partition in the interface between a contaminated lagoon and the ocean: the role of particulate load and composition. Mar. Pollut. Bull. 60, 1658–1666. Pereira, M.E., Lillebø, A.I., Pato, P., Válega, M., Coelho, J.P., Lopes, C.B., Rodrigues, S., Cachada, A., Otero, M., Pardal, M.A., Duarte, A.C., 2009. Mercury pollution in Ria de Aveiro (Portugal): a review of the system assessment. Environ. Monit. Assess. 155, 39–49. Pérez-Domínguez, R., Maci, S., Courrat, A., Lepage, M., Borja, A., Uriarte, A., Neto, J.M., Cabral, H., Raykov, V. St., Franco, A., Alvarez, M.C., Elliott, M., 2012. Current developments on fish-based indices to assess ecological-quality status of estuaries and lagoons. Ecol. Indic. 23, 34–45. Pombo, L., Elliott, M., Rebelo, J.E., 2005. Environmental influences on fish assemblage distribution of an estuarine coastal lagoon, Ria de Aveiro (Portugal). Sci. Mar. 69, 143–159. R Development Core Team, 2012. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Viena, Austriahttp://www.R-project.org. Rebelo, J.E., 1992. The ichthyofauna and abiotic hydrological environment of the Ria de Aveiro, Portugal. Estuaries 15, 403–413. Souid, G., Souayed, N., Yaktiti, F., Maaroufi, K., 2013. Effect of acute cadmium exposure on metal accumulation and oxidative stress biomarkers of Sparus aurata. Ecotoxicol. Environ. Saf. 89, 1–7. Ter Braak, C.J.F., Smilauer, P., 2002. CANOCO reference manual and CanoDraw for Windows user's guide: software for canonical community ordination (version 4.5). Tkatcheva, V., Poirier, D., Chong-Kit, R., Furdui, V.I., Burr, C., Leger, R., Parmar, J., Switzer, T., Maedler, S., Reiner, E.J., Sherry, J.P., Simmons, D.B.D., 2015. Lithium an emerging contaminant: bioavailability, effects on protein expression, and homeostasis disruption in short-term exposure of rainbow trout. Aquat. Toxicol. 161, 85–93. Whitfield, A.K., Elliott, M., 2002. Fishes as indicators of environmental and ecological changes within estuaries: a review of progress and some suggestions for the future. J. Fish Biol. 61, 229–250. Wołowicz, M., Sokołowski, A., Lasota, R., 2007. Estuaries—a biological point of view. Oceanol. Hydrobiol. St. 36, 113–130. Zuur, A.F., Ieno, E.N., Elphick, C.S., 2010. A protocol for data exploration to avoid common statistical problems. Methods Ecol. Evol. 1, 3–14.

Please cite this article as: García-Seoane, E., et al., Effect of historical contamination in the fish community structure of a recovering temperate coastal lagoon, Marine Pollution Bulletin (2016), http://dx.doi.org/10.1016/j.marpolbul.2016.07.005