Trace metal concentrations in sediments from the southwest of the Iberian Peninsula

1 Instituto de Ciencias Marinas de Andalucía (CSIC), Campus Río San Pedro, 11510 Puerto Real, Cádiz, Spain. E-mail: julian.blasco@icman.csic.es 2 CIMA, Faculty of Marine and Environmental Sciences, University of Algarve, Campus de Gambelas, 8005-139 Faro, Portugal. 3 Departamento Química y Ciencia de los Materiales, Facultad de Ciencias Experimentales, Universidad de Huelva, Avda. Fuerzas Armadas s/n, 21071 Huelva, Spain.


INTRODUCTION
Environmental contamination by metals is a worldwide problem.It is important to be aware of the possible effects of increasing levels of metal pollution on human health and the environment (Liao et al., 2006).Metals in marine sediments have a natural or anthropo-genic origin and the relative importance of these two sources depends on the metal and its location.Coastal ecosystems are often affected by anthropogenic metal inputs.The metals are transported from the water column to the sediments by adsorption to the fine particles in the seawater.These particles, with the adsorbed metals, settle on sediments, which act as both carriers and ADVANCES IN MARINE CHEMISTRY J. Blasco and J.M. Forja (eds.)potential metal sources in aquatic environments (Theofanis et al., 2001).The fate and behaviour of the metals are influenced by the characteristics of the sediment, the mineralogical composition, the redox state, adsorption/desorption processes and physical transport.Although sediments are considered to be a sink for metals, changes in the physicochemical characteristics of the sediment (redox potential, pH, dissolved oxygen) can lead to the remobilization of the metals to the overlying water.Furthermore, the burrowing activity of macroinvertebarte organisms can affect the fluxes of solutes and particles from sediment to water.Another source of metals for a deposit feeder is sediment ingestion, which can account for up to the 100% of the metal body burden in several deposit feeding invertebrates.Therefore, the metals become available to marine organisms, on which they can exert a toxic effect, and consequently humans can be affected through the food chain.One of the main problems of metals is their long biological half-life (Buccolieri et al., 2006).Research into metal contamination should focus on coastal areas, wetlands, salt marshes and estuaries, because these are highly productive and sensitive areas affected by the impact of anthropogenic contamination (Arellano et al., 1999;Cohen et al., 2001).
Although many studies on sediment metal concentrations have been carried out in the southwest of the Iberian Peninsula, the majority have focused on the area of Huelva (Fernández-Calinai et al., 1997;Ruiz et al., 1998;Grande et al., 2000;Borrego et al., 2002;Morillo et al., 2002;Santos-Bermejo et al., 2003).Few studies have been made on the Bay of Cadiz (Blasco et al., 2000;Ligero et al., 2002;Carrasco et al., 2003) and Portuguese coastal areas (Bebianno, 1995;Caetano et al., 2006).This study establishes the contamination levels in these coastal areas and determines the potential risk associated with metal contamination.The metal contamination (As, Cd, Cu, Hg, Ni, Pb and Zn) analyses of surface sediments from ecologically important ecosystems (estuaries, coastal ecosystem and salt marshes) from the southwest of the Iberian Peninsula were carried out in two seasons (spring and autumn).

Study area
The Ria Formosa is a large tidal lagoon that extends about 55 km along the south coast of Portugal and has a maximum width of 6 km.A strongly branched system of creeks and channels is connected to the ocean by six outlets.The average depth is less than 3 m.It has a water volume of approximately 31•10 6 m 3 , which is generally well mixed although occasional eutrophication problems have been reported (Newton and Mudge, 2003).The system has semidiurnal tides with 50% to 75% of the water volume exchanged during each tide.The western part of the lagoon is bounded by a heavily urbanized area surrounded by agricultural land (Ribeiro et al., 2008).
The Guadiana River estuary is located on the southern border between Portugal and Spain and is a single channel meso-tidal estuary that is 76 km long, 70 to 800 m wide and 5 to 15 m deep.The bottom consists of coarse material in the middle of the channel and finer particles near the margins (Fortunato et al., 2002).The estuary can be divided into three zones, low, middle and high, which have different salinity ranges: >25, 0.5-25 and <0.5 respectively.The lower estuary can be divided into two main habitat types: salt marshes and the main river channel.The salt marshes are important wintering grounds for many bird species and nurseries for molluscs, crustaceans and fish.The Guadiana River crosses an extensive rural area that has massive sulphide deposits (Palanques et al., 1995).The estuary receives the domestic sewage of two cities located near the mouth (Chicharo et al., 2001;Domingues et al., 2005).
The Ria of Huelva is a complex system of drainage channels that separate several areas of the salt marshes so that they function as islets in the inner estuary.This system is controlled by the tidal regime and the inputs of the Rivers Odiel and Tinto, as well as two channels which exchange water directly with the open sea.Therefore, the fate and behaviour of contaminants in the Ria of Huelva are very complex.This ecosystem has three sources of contamination: the industrial sewage from more than forty chemical industries located in three areas close to the Ría de Huelva, the urban sewage of the city of Huelva and fluvial inputs from the Rivers Tinto and Odiel, which have acidic waters and high metal levels.
The Bay of Cadiz is a littoral ecosystem with an area of 33.6 km 2 and an average depth of 3 m.It is subjected to a tidal regime and 20% of its area has intertidal characteristics.From a hydrological point of view, it can be divided into four regions: the outer bay, with oceanic characteristics; the inner bay, which is greatly affected by the tides and less exposed to wave action; the amphibious bay, of tidal marshes; and the terrestrial bay, of areas that are permanently emerged.The area around the bay has a population of 400000.The main industries located in the zone are related to shipbuilding, offshore industries and aerospace manufacturing.

Sampling and sample pre-treatment
Samples were collected in three areas of the SW Iberian Peninsula (Fig. 1) during two sampling campaigns (in late autumn 2006 and spring 2007): the Ria Formosa [the Ramalhete (RS), and Ribeira de Almargem (RDA) sampling sites] and the lower estuary of the Guadiana River [Ponte International (PI)], the Ria of Huelva [Huelva (HU), Punta Sebo (PS), Punta Umbría (PU), Canal Juan Carlos I right margin (CJC-d) and left margin (CJC-i)] and the Bay of Cadiz [Puente Zuazo (PZ), Trocadero (TR) and Rio San Pedro (RSP)].Sediments were collected using PVC sediment cores that were 100 cm in length and 10 cm in diameter.The sediment cores were transported to the laboratory in a refrigerator within 3 hours of being taken and then frozen at -18ºC until analysis.Cores were sectioned into 5-cm-thick samples from the surface to 30 cm deep.To minimize the effect of grain size on metal distribution, the analyses were carried out in the <63 µm sediment fraction.The granulometric fraction was separated from the samples with the manual wet sieving method using Nylon sieves and Milli-Q water.The sediment samples were then freeze-dried.

Metal analysis
Metal concentrations in the sediment were determined using the method described by Loring and Rantala (1992).This method consists in adding a mixture of aqua regia (HNO 3 /HCl 1:3 v/v) and HF to a freezedried sediment aliquot (approx.0.1 g).The procedure was carried out in a microwave oven (CEM Mars 5) and, after acid digestion, samples were treated with boric acid to form a gelatinous precipitate of borosilicates (Bernas, 1968).Trace metal concentrations (Zn, Cd, Cu, Pb, Ni, Hg and As) were determined by ICP-OES or ICP-MS depending on the metal levels.For mercury, we used the cold-vapour technique or a LECO AMA 254 analyzer.The results were checked using reference material MESS-1 NRC.Good agreement was obtained between the analyzed and certified values.The results are expressed as µg g -1 dry weight.

RESULTS
The metal concentrations (Cu, Zn, Pb, Cd, Hg, Ni and As) in surface sediments (0 to 5 cm) from eleven sampling stations for the two sampling periods are shown in Figures 2 and 3.
The highest metal concentrations were found for zinc, followed by copper, lead, arsenic (data only available for the Ría de Huelva), nickel, mercury and cadmium in decreasing order (Table 1).No seasonal pattern was observed as the concentrations were similar in both seasons at the different sampling sites.The copper levels were similar in both seasons, and although in some cases there were noticeable differences between seasons depending on the sampling station (e.g. in HU the copper levels were higher in spring than in autumn, but in PS the copper levels were lower in spring than in autumn), in general all the analyzed metals had similar values in both seasons.
The sites in the Ría de Huelva area showed higher levels of copper, zinc, lead and cadmium than the sites at the Ría Formosa, the River Guadiana estuary and the Bay of Cadiz.In general, a gradient was observed in the Ría de Huelva, with the highest values at PS, which then decreased to the outer and inner sites (HU).The PS site is affected by the River Tinto and industrial and phosphogypsum inputs.The HU station showed the second highest concentrations of the metals mentioned above due to its location in the channel where the River Odiel drains into the estuary, which is where fresh water and seawater mix.
For mercury and nickel, the highest values were found in the Bay of Cadiz and then, in decreasing order, at the sites HU≈PS, River Guadiana (PI) and Ría Formosa (RDA and RS).Levels of arsenic were only recorded in the Ría de Huelva.
The correlation matrix between metals, determined with the Spearman rank, showed high correlation coefficients between Cd, Cu, Pb and Zn, which can be related to these metals having a common origin.However, Ni was negatively correlated with these metals, as high levels were found at the sites at the Ría Formosa, River Guadiana and Bay of Cadiz, although metal levels were generally lower in these areas than in the Ría de Huelva.
A cluster analysis was carried out to classify the  sampling sites according to metal concentrations.The dendrogram (Fig. 4) shows three groups: the first group corresponds to the sites with high metal levels (HU and PS) in both seasons, both sampling stations in this group being affected by inputs from the Rivers Odiel and Tinto and chemical industries; the second group corresponds to the other stations from the Ría de Huelva, which are affected by the acid water from the two rivers and industrial chemical waste, but the outer positions of the stations means that the freshwater mixes with seawater; finally, the third group corresponds to the stations from the Ría Formosa, the River Guadiana and the Bay of Cadiz, which have a generally low or moderate level of contamination.
To calculate the risk associated with metal sediment contamination, we defined a hazard potential index (HPi) for a metal M, as: where [M] = metal concentration; F i = sequential extraction fraction expressed as the sum of the percentage for the three fractions of the BCR procedure or the first four fractions of the Tessier procedure (Tessier et al., 1979) divided by 100 (in our calculations we employed the first four fractions of the Tessier approach for these ecosystems); and SQV M = the sediment quality value for each metal.
A general hazard potential index for the metal load of the sediment can been calculated as: Table 2 shows the results for four individual metals (Zn, Cd, Cu and Pb).Due to the lack of specific information for sequential extraction of sediment samples, data from scientific literature on sediments of these areas were employed (Sáenz et al., 2003).This information is not available for the Ria Formosa and the Guadiana, so we preferred not to include these results in the table due to the high uncertainty associated with these indices.
The HPi M ranged between 1.2 and 3.6 for Pb and Cu for the Ria of Huelva and between 0.07 and 0.34 for Cu and Cd, respectively, in the Bay of Cadiz.We used SQVs for highly polluted ecosystems (Choueri et al., 2009), which ensures that there are no false negatives.Values over one indicate highly polluted ecosystems, whereas values below one, using the SQVs for nonpolluted ecosystems, show that that the ecosystem is   (Sáenz et al., 2002) and SQVs (Choueri et al., 2009).The HPi M index shows the range using SQVs corresponding to values for not polluted and highly polluted.

Metal
Average concentration  not affected by these metals.We found that the Ría de Huelva ecosystem is heavily affected by metal contamination, whereas the Bay of Cadiz is not affected.The HPi ML corresponds to the sum of all the metals, and had a value of 10 for the Ría de Huelva and 0.73 for the Bay of Cadiz.This index can be used as an initial tool for classifying areas impacted by metal contamination.

DISCUSSION
Sediments are recognized as sinks for many contaminants discharged into surface waters due to the adsorption processes of particulate matter, which then settles on the sediment.The southwest of the Iberian Peninsula is affected by metal contamination processes due to mining, industry and urban activities, and coastal ecosystems from these zones have metal inputs from a variety of sources.
The recent impact of anthropogenic activities can be assessed by measuring the metal concentrations in surface sediment, although the background values should be subtracted from the total metal concentration.Using average shale is common (Turekian and Wedephol, 1961) for calculating a geocumulation index (I geo ) (Müller, 1981).However, as the natural geology can differ from the average shale, these calculations can produce misleading results.Different background values have been reported for the ecosystems in this study (Table 3), which means that the enrichment factor (EF M = C sM /C bM ), defined as the ratio of the surface sediment metal concentration (C sM ) to the background values (C bM ), can vary greatly depending on the background values selected.It is clear that the use of average shale can only be considered as an initial approach when the lack of data means that sitespecific values cannot be used.For the sediments collected in the Ría Formosa and the River Guadiana no site-specific values were found and the use of average shale can lead to biased results and underestimations of the contaminant inputs.However, when several backgrounds are available, the most appropriate option is to use the most conservative data to avoid underestimating the contamination events.Therefore, to calculate the EF M in the Ría de Huelva we used the background values reported by Borrego et al. (2002).According to these, the EF M values ranged between 1.3 for nickel and 71 for copper.The low EF Ni value for nickel is because there is no pollution source for this metal (Morillo et al., 2002).In the Bay of Cadiz, the EF M ranged between 2.3 for zinc and 10 for mercury, although the background values were not available for several metals.If background concentrations of the Bay of Cadiz were used to calculate the EF M for the Ría Formosa and the River Guadiana, these ranged between 1.0 for cadmium and 5 for mercury.
The high metal contamination in the Ría de Huelva is caused by sediment contributions from the Rivers Tinto and Odiel and from the industrial activity on the estuary banks (Borrego et al., 2002).The sediment inputs correspond to fine-grained sediment enriched in sulphide-associated metals (Fernández-Caliani et al., 1997).This pattern is evident for the HU and PS sites located close to the mouth of the two rivers.Cu and Zn are of fluvial origin, whereas As is associated with industrial wastes (Grande et al., 2000).Sediment metal transport has been reported to be a source of metal for the inner shelf (Fernández-Caliani et al., 1997).
In the Bay of Cadiz, the EF Hg values close to 10 evidence the mercury input in the ecosystem, which has been observed from the beginning of the twentieth century, when this metal began to increase in the ecosystem (Ligero et al., 2002).Although cadmium had an EF Cd that indicates a recent input, it is difficult to establish the temporal evolution of this metal because diffusion processes and affinity with fine-grained sediment and organic matter can disturb the sediment profile concentrations (Ligero et al., 2002).
For the Ría Formosa and River Guadiana, the values recorded for metals in the sediments are in the same range as those reported by Bebianno (1995) and Caetano et al. (2006), indicating that there are similar levels of contamination.The lack of temporal series of metal concentrations or dated sediment profiles makes it difficult to establish the evolution of contamination.In the Guadiana River the process of sediment transport can affect metal levels in sediments, so during the peak flood process the metal content decreases significantly (Caetano et al., 2006).Although the River Guadiana sediments are less contaminated than the River Tinto-Odiel sediments, they have been reported to be the main source of sediment and metal contamination in the northern Gulf of Cadiz (Gonzalez et al., 2007).Metal partitioning can be useful as an initial tool for modelling metal bioavailability and the potential risk for ecosystems.Several authors have determined the metal distribution pattern using sequential extraction procedures for the Ría de Huelva, salt marsh and Bay of Cadiz areas (Izquierdo et al., 1997;Usero et al., 1998;Morillo et al., 2002;Sáenz et al., 2003).In the Bay of Cadiz, metals are generally associated with the residual fraction, with the exception of lead.However, in the Ría de Huelva area, copper, cadmium and zinc are in the mobile phase and lead and nickel in the residual fraction (Morillo et al., 2002).The diffusive flux of Cu in this area showed the high concentration of this metal and its extractability and therefore presumed bioavailability (Blasco et al., 2000).
We mainly used sediment quality values (SQVs) to assess the status of contaminated sediment, but the results can vary greatly depending on the values used (Chapman et al., 1999).Although SQVs have this drawback, they are employed in the context of a tiered strategy for ecological risk assessment; they should be applied with care and cannot be employed for regulatory purposes (Chapman and Mann, 1999).Further development of this approach should take into account the factors that affect bioavailability and toxicity.
Using only those metals associated with fractions which can be remobilized and enter the aquatic ecosystems is an improvement over using total sediment metal concentrations (Riba et al., 2002;Sáenz et al., 2003), including metals belonging to the lattice structure of the sedimentary material that are not available for organisms.The HPi M allowed zones to be classified according to the effect associated with each metal or the total metal load (Hpi ML ).Although these indices do not distinguish the origin of the metal because the natural source and anthropogenic or post-deposition redistribution are not separated, the total metal load, index should be used as an initial tool because the same index can be obtained when the index of a metal is very high and the other metals have values below one, or if all the metals or some of them have values above one.Both indices can be employed using the average concentration of one area or specifically for each sampling station.However, SQVs are obtained for the correlation between effect and total metal concentration (DelValls and Chapman, 1998) and if bioavailability is not taken into account the index can have higher values.SQVs are affected by uncertainty because benchmark values for large coastal ecosystems involve considering that metal bioavailability is similar for areas with different sediment geochemistry.Recently, Hewitt et al., (2009) pointed out the need to increase the ecological significance of sediment contamination guidelines by taking into account community analysis measurements, such as field-based species sensitivity distributions and multivariate models, to ensure environmental protection.

Fig. 2 .
Fig.2.-Concentrations of copper, zinc, lead and cadmium (mg kg -1 dry sediment) at the sampling stations for spring and autumn.

Fig. 3 .
Fig. 3. -Concentrations of mercury and nickel (mg kg -1 dry sediment) at the sampling stations for spring and autumn.

Table 1 .
-Average and range concentration for the analyzed metal at the three locations in the two sampling campaigns.Results are expressed as µg g -1 dry weight.

Table 2 .
-Average concentration of Zn, Cd, Cu and Pb in the Ría de Huelva (RHU) and the Bay of Cadiz (BC) and HPi M calculated using the sum of the first four fractions (∑ n i=1 Fi) of the Tessier approach

Table 3 .
-Average concentration in sediment from the three coastal zones and background values for the Ría de Huelva, Bay of Cadiz and average shale composition.The results are expressed as µg•g -1 dry weight.