Marine Environmental Research 101 (2014) 169e183
Contents lists available at ScienceDirect
Marine Environmental Research journal homepage: www.elsevier.com/locate/marenvrev
Recent benthic foraminiferal assemblages and their relationship to environmental variables on the shoreface and inner shelf off Valencia (Western Mediterranean) pez-Belzunce a, *, Ana M. Bla zquez a, Joan Lluís Pretus b María Lo a b
Environmental and Marine Sciences Research Institute, Catholic University of Valencia, C/ Guillem de Castro 94, Valencia 46003, Spain Department of Ecology, University of Barcelona, Av. Diagonal, 643, Barcelona 08028, Spain
a r t i c l e i n f o
a b s t r a c t
Article history: Received 6 April 2014 Received in revised form 21 June 2014 Accepted 30 June 2014 Available online 9 July 2014
The environmental variables that determine the distributions of benthic foraminiferal assemblages on the shoreface and inner shelf of the north of Valencia province (Western Mediterranean) are identified. The possible influence of variables such as water depth, distance from shore, hydrodynamics, substrate type, carbonate content, organic matter content and human activity is evaluated. Multivariate cluster-Qtype analysis and redundancy analysis (RDA) are used to identify the environmental variables that have the greatest influence on the assemblage distribution. The spatial distribution of the assemblages is closely associated with water depth and substrate. The diversity and abundance of foraminifera shells increase with depth and their conservation improves. The most common species in the study area are Ammonia beccarii, Rosalina globularis, Buccella granulata, Planorbulina mediterranensis, and Lobatula lobatula. The presence of wastewater in the study area has not polluted the foraminiferal assemblages (absence of anomalous shells). The direction of the discharge plume is a potential source of nutrients for deep water. © 2014 Elsevier Ltd. All rights reserved.
Keywords: Benthic ecology Foraminifera Coastal waters SE Spain Western Mediterranean
1. Introduction Benthic foraminifera are unicellular protoctists with calcareous tests that live in all deep oceans. These organisms play a major ecological role in the ecosystem; their wide range and calcareous tests make them suitable for providing interpretation and extrapolation tools for palaeoecological and palaeoenvironmental reconstructions (Usera et al., 2002; Leorri and Cearreta, 2009; Milker zquez and Usera, 2010; Bla zquez and Ferrer, 2012). et al., 2009; Bla Recently, benthic foraminifera have been used as bioindicators for environmental studies. In shallow inner shelf environments, attempts to identify the environmental variables that could control the distribution of benthic foraminifera are difficult due to the high variability of the system. The distribution may be controlled by abiotic factors such as grain size, organic matter, salinity, dissolved oxygen, and by biotic factors such as predation, competition and reproduction (Colom, 1974; Murray, 1991, 2006; Jorissen, 1999). Therefore the analysis of benthic foraminifera has provided a proxy
* Corresponding author. pez-Belzunce), ana. E-mail addresses: [email protected] (M. Lo [email protected] (A.M. Bl azquez), [email protected] (J.L. Pretus). http://dx.doi.org/10.1016/j.marenvres.2014.06.011 0141-1136/© 2014 Elsevier Ltd. All rights reserved.
for evaluating the quality of the ecosystem. The increase in human activity on inner shelf areas has promoted a great deal of research into the effects of pollution (Martin, 2000; Sen Gupta, 2002; Bergin et al., 2006; Ferraro et al., 2006; Irabien et al., 2008). In the western Mediterranean, several studies of recent benthic foraminifera have been carried out, in Italy (Donnici and SerandreiBarbero, 2002; Frontalini and Coccioni, 2011; Magno et al., 2012), on the French Mediterranean coast (Mojtahid et al., 2008; Goineau et al., 2011; Fontanier et al., 2012; Barras et al., 2014) and off the Spanish coast (deep-sea) (Mateu, 1970; Milker et al., 2009; MateuVicens et al., 2010; Contreras-Rosales et al., 2012). At present there are few studies of the superficial distribution of benthic foramizquez, nifera on the inner shelf of the eastern seaboard of Spain (Bla zquez, 1997) and hardly any of the inner shelf or 1996; Usera and Bla n province and the north of shoreface off the coast of Castello zquez and Alca ntara-Carrio , 2009). The few studies Valencia (Bla performed to date indicate the presence of two ubiquitous species in the inner shelf of the southern part of the study area: Ammonia ) and Elphidium crispum (Linne ), and a sharp increase beccarii (Linne in the diversity and abundance of foraminifera in substrates colonised by seagrass.
170
pez-Belzunce et al. / Marine Environmental Research 101 (2014) 169e183 M. Lo
Fig. 1. Location map of the study area with samples' location. North Valencia (Spain).
The aims of this study are to characterise the main benthic foraminiferal assemblages and their spatial distribution on the inn, ner shelf zone between the provinces of Valencia and Castello including both live and dead assemblages. We assess the possible influence of variables such as depth, distance from shore, hydrodynamics, substrate type, carbonate content, organic matter content and human activity (e.g., port activity, discharges). To determine the importance of the latter factor, we evaluate the
impact of the discharge from an outfall on the distribution of these organisms. This study identifies the dominant assemblages of the different sub-environments and assesses the richness and diversity of these benthic communities by using a series of multivariate statistical analyses. The study area is located on the coast of Valencia, Spain (Fig. 1). Siliciclastic sediments come predominantly from rivers and their presence on the shelf is controlled by the balance between
pez-Belzunce et al. / Marine Environmental Research 101 (2014) 169e183 M. Lo
terrigenous input and fluctuating hydrodynamic conditions (Rey et al., 1999). On the inner shelf of the study area there are very coarse sands and pebbles and gravel deposits caused by the flooding of major rivers (Palancia, Belcaire river), characterised by a sporadic torrential regime. From the oceanographic point of view, the currents in the study area vary throughout the year in terms of velocity and distance from the coast. However, they tend to be low velocity and are predominantly westerly, west-south-westerly and south-westerly. Winds are mainly NE (9% of measures) and NNE (8%), with speeds above 8 m/s; the third component is SSE (7%). Finally, the predominant swell is easterly (27%), south-easterly (25%) and north-easterly (25%). Therefore, the longshore current in the area study has a dominant north-easterly/south-westerly direction. Other environmental variables studied show homogeneous values. Their mean values are: temperature 19.5 C, pH 8.2 (except at the mouth of the outfall, where it falls to 7 C), oxygen dissolved 8.2 mg/l and salinity 38‰. 2. Materials and methods The samples were collected in July 2009. Twenty-eight samples from the shoreface and the inner shelf were examined. The samples were collected by divers (Denebay and Fernandez, 2009; MateuVicens et al., 2010) who scraped off the surface sediments, along eight profiles perpendicular to the coast. This method of sampling was used instead of the conventional ones, as a grab e it usually washes the samples and some mixing processes may happen (Yang and Flower, 2009; Foster et al., 2012) e or a box corer e some corers do not work properly in coarsed sands-. All samples were located with GPS; the precise locations and depths of these samples are indicated in Fig. 1. During the campaign, sampling was also planned in the proximity of the outfall. Stations 12, 13, 14, 15, 16 and 17 were analysed to assess the gradient of the effects of a continued discharge of water from the treatment unit. 2.1. Study of foraminifera The decision to use the total assemblage or only the living assemblages depends on the objective of the study. Numerous ecological studies have analysed the total assemblage in order to avoid short-term impacts and to focus on long-term effects. This approach avoids the problem of the specimens' seasonal variability (Scott and Medioli, 1980; Carboni et al., 2009; Bergamin et al., 2009; Magno et al., 2012; Foster et al., 2012). However, certain recent studies published have used living assemblages as a biotic index for environmental quality (Morvan et al., 2006; Mojtahid et al., 2008; Bouchet et al., 2012; Barras et al., 2014) and a protocol was created by FOraminiferal BIoMOnitoring (FOBIMO)-group, to standardise the use of forami€nfeld et al., 2012). Dead assemnifera as an ecological index (Scho blages also provide information about post-mortem processes such as hydrodynamics and sediment inputs. We therefore decided to perform several statistical analyses with the total assemblages and with separate assemblages (living and dead) to establish whether there are significant differences. According to several authors (Buzas, 1990; Murray, 1991) when the aim is to identify foraminiferal assemblages a sample size of 300 specimens is sufficient. Fatela and Taborda (2002) consider that a sample of 100 shells is enough, because they represent more than 5% of the assemblages, and Patterson and Fishbein (1989) suggest that if the species under consideration represent almost 50% of the assemblages, then only 50 individuals are needed. In our study, 300 living and dead specimens were counted whenever possible.
171
We used the relation between dead and living assemblages to assess the representativeness of the data. We tested the similarity between the living and dead assemblages using Sanders' similarity index (Sanders, 1960). Values generally above 70% indicate a good correlation between the two associations (Rodríguez-Lazaro et al., 2013). These values are normal for estuarine or lagoon areas; however, in the conditions that prevail in the study area, with an open coast without landforms and intense coastal dynamics, the ranges are expected to vary more widely. In our case, the species found in the living assemblages were also the most abundant in the dead assemblages. Live organisms represented 1.2% of the total of shells analysed in previous studies zquez, 1997), of which over 90% correspond to the (Usera and Bla Rotaliina suborder. In our study, this ratio oscillated between 5% and 40% and was considered statistically representative. We stress the difficulties in identifying living suborders like the Miliolids or agglutinated foraminifera, due to their shells (Le Calvez and Cesana, 1972; Martin and Steinker, 1973). The samples were wet sieved through a 63 mm screen. Foraminifera were fixed with alcohol and stained with Rose Bengal dye following the procedure described by Walton (1952) to differentiate between living and dead assemblages. The washed and dried fraction was concentrated by flotation with a dense liquid (trichloroethylene). Foraminifera shells were picked (under a binocular stereomicroscope) using reflected light until a representative number of 300 individuals per sample was obtained. In the superficial samples, in which shells were less abundant,100 individuals were collected. The foraminifera were classified following Loeblich and Tappan (1987). The data obtained were quantitatively analysed, and the diversity index (Shannon and Wiener, 1949), equitability and Fisher's alpha (Fisher et al., 1943) were calculated. The alpha index is unreliable in samples of fewer than 100 individuals (Murray, 2006). It was estimated in order to establish the composition and proportional abundance of the species identified, and the dominance of the species in the samples. Diversity values were computed with the PAST (Palaeontological statistics) software (Hammer et al., 2001, 2008). 2.2. Sedimentary and chemical analysis The variations in temperature, salinity, pH and dissolved oxygen were very small and so they were not included in the statistical analysis. After drying the material, sub-samples of 100 g were treated with sodium hydroxide and hydrogen peroxide to disaggregate and washed through three sieves with mesh diameters of 0.2 mm, 0.125 mm and 0.063 mm. Three size fractions for each sample were thus available to provide sedimentological data. Sand grain size was analysed, differentiating between bioclastic and lithoclastic. Organic matter was determined using the WalkleyeBlack method (Walkley and Black, 1934). The proportion of calcium carbonate was calculated only in clay and silt, using a Bernard calcimeter. 2.3. Statistical analysis Numerous methods have been developed to quantitatively reconstruct palaeoenvironmental variables. These methods differ in terms of the numerical assumptions made regarding the data used: for instance, whether the taxoneenvironment response is unimodal (Gaussian) or linear (Sejrup et al., 2004). First, Principal Component Analysis (PCA) was carried out to extract the most important foraminiferal assemblages. Due the high variability of the study area, with a mean value of 27
172
pez-Belzunce et al. / Marine Environmental Research 101 (2014) 169e183 M. Lo
Table 1 Statistical results of redundancy analyses (RDAs). RDA results Axis 1 (Living and dead) assemblages Eigenvalues Specieseenvironment correlations Cumulative percentage variance of Species data of Specieseenvironment relation Correlation Water depth Bioclastic sands Organic matter Marine vegetation Dead assemblages Eigenvalues Specieseenvironment correlations Cumulative percentage variance of Species data of Specieseenvironment relation Correlation Water depth Bioclastic sands Organic matter Marine vegetation Living assemblages Eigenvalues Specieseenvironment correlations Cumulative percentage variance of Species data of Specieseenvironment relation Correlation Bioclastic sands Carbonates Organic matter
Contents lists available at ScienceDirect
Marine Environmental Research journal homepage: www.elsevier.com/locate/marenvrev
Recent benthic foraminiferal assemblages and their relationship to environmental variables on the shoreface and inner shelf off Valencia (Western Mediterranean) pez-Belzunce a, *, Ana M. Bla zquez a, Joan Lluís Pretus b María Lo a b
Environmental and Marine Sciences Research Institute, Catholic University of Valencia, C/ Guillem de Castro 94, Valencia 46003, Spain Department of Ecology, University of Barcelona, Av. Diagonal, 643, Barcelona 08028, Spain
a r t i c l e i n f o
a b s t r a c t
Article history: Received 6 April 2014 Received in revised form 21 June 2014 Accepted 30 June 2014 Available online 9 July 2014
The environmental variables that determine the distributions of benthic foraminiferal assemblages on the shoreface and inner shelf of the north of Valencia province (Western Mediterranean) are identified. The possible influence of variables such as water depth, distance from shore, hydrodynamics, substrate type, carbonate content, organic matter content and human activity is evaluated. Multivariate cluster-Qtype analysis and redundancy analysis (RDA) are used to identify the environmental variables that have the greatest influence on the assemblage distribution. The spatial distribution of the assemblages is closely associated with water depth and substrate. The diversity and abundance of foraminifera shells increase with depth and their conservation improves. The most common species in the study area are Ammonia beccarii, Rosalina globularis, Buccella granulata, Planorbulina mediterranensis, and Lobatula lobatula. The presence of wastewater in the study area has not polluted the foraminiferal assemblages (absence of anomalous shells). The direction of the discharge plume is a potential source of nutrients for deep water. © 2014 Elsevier Ltd. All rights reserved.
Keywords: Benthic ecology Foraminifera Coastal waters SE Spain Western Mediterranean
1. Introduction Benthic foraminifera are unicellular protoctists with calcareous tests that live in all deep oceans. These organisms play a major ecological role in the ecosystem; their wide range and calcareous tests make them suitable for providing interpretation and extrapolation tools for palaeoecological and palaeoenvironmental reconstructions (Usera et al., 2002; Leorri and Cearreta, 2009; Milker zquez and Usera, 2010; Bla zquez and Ferrer, 2012). et al., 2009; Bla Recently, benthic foraminifera have been used as bioindicators for environmental studies. In shallow inner shelf environments, attempts to identify the environmental variables that could control the distribution of benthic foraminifera are difficult due to the high variability of the system. The distribution may be controlled by abiotic factors such as grain size, organic matter, salinity, dissolved oxygen, and by biotic factors such as predation, competition and reproduction (Colom, 1974; Murray, 1991, 2006; Jorissen, 1999). Therefore the analysis of benthic foraminifera has provided a proxy
* Corresponding author. pez-Belzunce), ana. E-mail addresses: [email protected] (M. Lo [email protected] (A.M. Bl azquez), [email protected] (J.L. Pretus). http://dx.doi.org/10.1016/j.marenvres.2014.06.011 0141-1136/© 2014 Elsevier Ltd. All rights reserved.
for evaluating the quality of the ecosystem. The increase in human activity on inner shelf areas has promoted a great deal of research into the effects of pollution (Martin, 2000; Sen Gupta, 2002; Bergin et al., 2006; Ferraro et al., 2006; Irabien et al., 2008). In the western Mediterranean, several studies of recent benthic foraminifera have been carried out, in Italy (Donnici and SerandreiBarbero, 2002; Frontalini and Coccioni, 2011; Magno et al., 2012), on the French Mediterranean coast (Mojtahid et al., 2008; Goineau et al., 2011; Fontanier et al., 2012; Barras et al., 2014) and off the Spanish coast (deep-sea) (Mateu, 1970; Milker et al., 2009; MateuVicens et al., 2010; Contreras-Rosales et al., 2012). At present there are few studies of the superficial distribution of benthic foramizquez, nifera on the inner shelf of the eastern seaboard of Spain (Bla zquez, 1997) and hardly any of the inner shelf or 1996; Usera and Bla n province and the north of shoreface off the coast of Castello zquez and Alca ntara-Carrio , 2009). The few studies Valencia (Bla performed to date indicate the presence of two ubiquitous species in the inner shelf of the southern part of the study area: Ammonia ) and Elphidium crispum (Linne ), and a sharp increase beccarii (Linne in the diversity and abundance of foraminifera in substrates colonised by seagrass.
170
pez-Belzunce et al. / Marine Environmental Research 101 (2014) 169e183 M. Lo
Fig. 1. Location map of the study area with samples' location. North Valencia (Spain).
The aims of this study are to characterise the main benthic foraminiferal assemblages and their spatial distribution on the inn, ner shelf zone between the provinces of Valencia and Castello including both live and dead assemblages. We assess the possible influence of variables such as depth, distance from shore, hydrodynamics, substrate type, carbonate content, organic matter content and human activity (e.g., port activity, discharges). To determine the importance of the latter factor, we evaluate the
impact of the discharge from an outfall on the distribution of these organisms. This study identifies the dominant assemblages of the different sub-environments and assesses the richness and diversity of these benthic communities by using a series of multivariate statistical analyses. The study area is located on the coast of Valencia, Spain (Fig. 1). Siliciclastic sediments come predominantly from rivers and their presence on the shelf is controlled by the balance between
pez-Belzunce et al. / Marine Environmental Research 101 (2014) 169e183 M. Lo
terrigenous input and fluctuating hydrodynamic conditions (Rey et al., 1999). On the inner shelf of the study area there are very coarse sands and pebbles and gravel deposits caused by the flooding of major rivers (Palancia, Belcaire river), characterised by a sporadic torrential regime. From the oceanographic point of view, the currents in the study area vary throughout the year in terms of velocity and distance from the coast. However, they tend to be low velocity and are predominantly westerly, west-south-westerly and south-westerly. Winds are mainly NE (9% of measures) and NNE (8%), with speeds above 8 m/s; the third component is SSE (7%). Finally, the predominant swell is easterly (27%), south-easterly (25%) and north-easterly (25%). Therefore, the longshore current in the area study has a dominant north-easterly/south-westerly direction. Other environmental variables studied show homogeneous values. Their mean values are: temperature 19.5 C, pH 8.2 (except at the mouth of the outfall, where it falls to 7 C), oxygen dissolved 8.2 mg/l and salinity 38‰. 2. Materials and methods The samples were collected in July 2009. Twenty-eight samples from the shoreface and the inner shelf were examined. The samples were collected by divers (Denebay and Fernandez, 2009; MateuVicens et al., 2010) who scraped off the surface sediments, along eight profiles perpendicular to the coast. This method of sampling was used instead of the conventional ones, as a grab e it usually washes the samples and some mixing processes may happen (Yang and Flower, 2009; Foster et al., 2012) e or a box corer e some corers do not work properly in coarsed sands-. All samples were located with GPS; the precise locations and depths of these samples are indicated in Fig. 1. During the campaign, sampling was also planned in the proximity of the outfall. Stations 12, 13, 14, 15, 16 and 17 were analysed to assess the gradient of the effects of a continued discharge of water from the treatment unit. 2.1. Study of foraminifera The decision to use the total assemblage or only the living assemblages depends on the objective of the study. Numerous ecological studies have analysed the total assemblage in order to avoid short-term impacts and to focus on long-term effects. This approach avoids the problem of the specimens' seasonal variability (Scott and Medioli, 1980; Carboni et al., 2009; Bergamin et al., 2009; Magno et al., 2012; Foster et al., 2012). However, certain recent studies published have used living assemblages as a biotic index for environmental quality (Morvan et al., 2006; Mojtahid et al., 2008; Bouchet et al., 2012; Barras et al., 2014) and a protocol was created by FOraminiferal BIoMOnitoring (FOBIMO)-group, to standardise the use of forami€nfeld et al., 2012). Dead assemnifera as an ecological index (Scho blages also provide information about post-mortem processes such as hydrodynamics and sediment inputs. We therefore decided to perform several statistical analyses with the total assemblages and with separate assemblages (living and dead) to establish whether there are significant differences. According to several authors (Buzas, 1990; Murray, 1991) when the aim is to identify foraminiferal assemblages a sample size of 300 specimens is sufficient. Fatela and Taborda (2002) consider that a sample of 100 shells is enough, because they represent more than 5% of the assemblages, and Patterson and Fishbein (1989) suggest that if the species under consideration represent almost 50% of the assemblages, then only 50 individuals are needed. In our study, 300 living and dead specimens were counted whenever possible.
171
We used the relation between dead and living assemblages to assess the representativeness of the data. We tested the similarity between the living and dead assemblages using Sanders' similarity index (Sanders, 1960). Values generally above 70% indicate a good correlation between the two associations (Rodríguez-Lazaro et al., 2013). These values are normal for estuarine or lagoon areas; however, in the conditions that prevail in the study area, with an open coast without landforms and intense coastal dynamics, the ranges are expected to vary more widely. In our case, the species found in the living assemblages were also the most abundant in the dead assemblages. Live organisms represented 1.2% of the total of shells analysed in previous studies zquez, 1997), of which over 90% correspond to the (Usera and Bla Rotaliina suborder. In our study, this ratio oscillated between 5% and 40% and was considered statistically representative. We stress the difficulties in identifying living suborders like the Miliolids or agglutinated foraminifera, due to their shells (Le Calvez and Cesana, 1972; Martin and Steinker, 1973). The samples were wet sieved through a 63 mm screen. Foraminifera were fixed with alcohol and stained with Rose Bengal dye following the procedure described by Walton (1952) to differentiate between living and dead assemblages. The washed and dried fraction was concentrated by flotation with a dense liquid (trichloroethylene). Foraminifera shells were picked (under a binocular stereomicroscope) using reflected light until a representative number of 300 individuals per sample was obtained. In the superficial samples, in which shells were less abundant,100 individuals were collected. The foraminifera were classified following Loeblich and Tappan (1987). The data obtained were quantitatively analysed, and the diversity index (Shannon and Wiener, 1949), equitability and Fisher's alpha (Fisher et al., 1943) were calculated. The alpha index is unreliable in samples of fewer than 100 individuals (Murray, 2006). It was estimated in order to establish the composition and proportional abundance of the species identified, and the dominance of the species in the samples. Diversity values were computed with the PAST (Palaeontological statistics) software (Hammer et al., 2001, 2008). 2.2. Sedimentary and chemical analysis The variations in temperature, salinity, pH and dissolved oxygen were very small and so they were not included in the statistical analysis. After drying the material, sub-samples of 100 g were treated with sodium hydroxide and hydrogen peroxide to disaggregate and washed through three sieves with mesh diameters of 0.2 mm, 0.125 mm and 0.063 mm. Three size fractions for each sample were thus available to provide sedimentological data. Sand grain size was analysed, differentiating between bioclastic and lithoclastic. Organic matter was determined using the WalkleyeBlack method (Walkley and Black, 1934). The proportion of calcium carbonate was calculated only in clay and silt, using a Bernard calcimeter. 2.3. Statistical analysis Numerous methods have been developed to quantitatively reconstruct palaeoenvironmental variables. These methods differ in terms of the numerical assumptions made regarding the data used: for instance, whether the taxoneenvironment response is unimodal (Gaussian) or linear (Sejrup et al., 2004). First, Principal Component Analysis (PCA) was carried out to extract the most important foraminiferal assemblages. Due the high variability of the study area, with a mean value of 27
172
pez-Belzunce et al. / Marine Environmental Research 101 (2014) 169e183 M. Lo
Table 1 Statistical results of redundancy analyses (RDAs). RDA results Axis 1 (Living and dead) assemblages Eigenvalues Specieseenvironment correlations Cumulative percentage variance of Species data of Specieseenvironment relation Correlation Water depth Bioclastic sands Organic matter Marine vegetation Dead assemblages Eigenvalues Specieseenvironment correlations Cumulative percentage variance of Species data of Specieseenvironment relation Correlation Water depth Bioclastic sands Organic matter Marine vegetation Living assemblages Eigenvalues Specieseenvironment correlations Cumulative percentage variance of Species data of Specieseenvironment relation Correlation Bioclastic sands Carbonates Organic matter
