Accessible Published by De Gruyter July 7, 2014

Analysis of trace elements in surface sediments, mussels, seagrass and seawater along the southeastern Adriatic coast – a chemometric approach

Slavka Stanković, Bojan Tanaskovski, Božidarka Zlatić, Milica Arsenović and Lato Pezo

Abstract

Surface sediments, mussels, seagrass, surface and bottom seawater samples were collected from the costal area of the southeastern Adriatic Sea and analyzed in order to determine the concentration and origin of the following elements: Fe, Mn, Zn, Ni, Cu, Ni, Co, As, Cd, Cr and Hg. The complexity of the obtained data was reduced by principal component analysis (PCA) and cluster analysis (CA), methods well known and accepted for the identification of the quality of marine environments. Both PCA and CA analysis were used to discriminate groups of samples according to the similarity of their chemical composition. The results revealed good diversity between the various samples, expressed by their distinctive positions of points in factor space. PCA indicated that the first two PC components explained about 73, 48, 43, 48, and 50 % of the total variance of the data for sediments, mussels, seagrass, and surface and bottom water, respectively. The results showed good discrimination capabilities between the samples taken from different locations, and also different seasons, which was especially evident in the surface and bottom water samples. Simultaneously, PCA/CA analysis of the amounts of trace elements found in the marine organisms could explain the manner of their bioaccumulation.

Introduction

The Adriatic Sea is situated between the eastern Italian and Slovenian, Croatian, Montenegrin and Albanian coasts. It is especially subject to pollution due to its enclosed character. The pollution of the Adriatic Sea is higher along the Italian coast [1]. The southeastern coastal part of the Adriatic receives large amounts of contaminants introduced by harbor, domestic, industrial, touristic and agricultural activities [2], directly, via rivers, by land erosion and through atmospheric deposition [3, 4]. Along the eastern Adriatic coast, the effects of sediment and marine pollution were investigated mainly in the Slovenian and Croatian coastal areas [5] and in the last decade some initial investigations were realized in the Albanian and Montenegrin coasts [1–4, 6–8]. Recent accelerated urbanization of the region has led to the situation in which the southeastern Adriatic marine environment is starting to suffer from pollution [1, 8]. Natural weathering and human activities are the sources contributing to trace metal contamination of the marine coastal environment through rivers and the atmospheric transport. The majority of trace elements in the coastal area of southeastern Adriatic are strongly related to river inflows [4] and the catchment geology [3, 5].

Various multivariate statistical techniques have been employed for the evaluation and characterization of environmental data [9–11]. PCA is a commonly used multivariate method mainly utilized for data reduction, (which means that the number of explanatory variables is lowered by using specific factors), but it is also employed to classify and discriminate between various samples. These factors explain the major variations within the data in order to make the components more interpretable [11, 12]. In the last decade, PCA has become accepted for the identification variations and sources of pollution in river water, groundwater, wastewater, seawater, sediment and soil [11–15], but it has seldom been used to show also the very complex influence of time.

Comprehensive applications of different chemometric methods to the analysis of marine mussels (Mytilus galloprovincialis), seagrass (Posidonia oceanica), surface/bottom water quality and surface sediments, using trace elements concentrations data are not common in the southeastern Adriatic region. In addition, identification of the sources of the metals at monitoring sites has not been fully explored in marine studies in this region. Considering their toxicity and high bioaccumulation capacity, metals are among the most common pollutants of sediments, water and biota. Sediments are preferred in regular monitoring of the environment as they represent a major sink for contaminants, contain much higher metal concentrations than seawater and have a lower temporal variability [16, 17].

Cluster analysis (CA) was also performed to classify the samples. This technique has been frequently used for fingerprinting atomic absorption spectrometry (AAS) data of observed samples. All samples were grouped in a multidimensional factor space and five dendograms were plotted, representing the similarities in the chemical composition of surface sediments, mussels, seagrass and seawater. Complete linkage was used and the distances expressed as dissimilarities (1–r, i.e., 1–correlations, which explain the degree of similarities among the samples) were calculated in the cluster analysis.

The aim of this work was to investigate the multi elemental composition of surface sediments, mussels, seagrass, surface and bottom water, including Fe, Mn, Zn, Ni, Cu, Ni, Co, As, Cd, Cr and Hg (analyzed by AAS), monitored at 10 locations, between the autumn season of 2005 and the spring season of 2007, in order to determine their origin and possible ecological risk to marine organisms. The obtained results could be used as an initial insight into the variations in the contents of trace elements in the coastal area of the southeastern Adriatic, and into the natural and anthropogenic sources of the contaminants. Within this study, pattern recognition techniques (PCA and CA) were applied on the AAS data (used as descriptors) to characterize and differentiate among the observed samples.

Materials and methods

Sampling locations and analysis

The sampling locations are situated in the proximity of different hydrological and human impacts: Sveta Stasija, Kukuljina and Herceg Novi are in the semi-enclosed Boka Kotorska Bay, which is on the World Heritage List of UNESCO. The other seven stations are located on the open Montenegrin coastline: Mamula, Žanjice, Bigova, Budva, Bar, Rt ᴆeran and Ada Bojana, Fig. 1. The locations are situated in the proximity of different geochemical, hydrological and human impacts. Co-ordinates and depths of the sampling stations and dates and type of samples are shown in Table 1.

Fig. 1 The investigated locations: 1. Sv. Stasija, 2. Kukuljina, 3. Herceg Novi, 4. Mamula, 5. Žanjice, 6. Bigova, 7. Budva, 8. Bar, 9. Rt ᴆeran, 10. Ada Bojana and 11. Kotor.

Fig. 1

The investigated locations: 1. Sv. Stasija, 2. Kukuljina, 3. Herceg Novi, 4. Mamula, 5. Žanjice, 6. Bigova, 7. Budva, 8. Bar, 9. Rt ᴆeran, 10. Ada Bojana and 11. Kotor.

Table 1

Co-ordinates and depths of the sampling stations and dates and type of samples.

Station Latitude (N) Longitude (E) Depth (m) Samples type a,b Sampling dates
1 42°28′04″ 18°44′99″ 6.0 Sediment, seawater, M.g., P.o. 29th September 2005
Seawater, M.g., P.o. 11th May 2006
Sediment, seawater, Mg, P.o. 23rd September 2006
Seawater, M.g., P.o. 28th May 2007
2 42°24′54″ 18°42′02″ 11 Sediment, seawater, M.g., P.o. 29th September 2005
Seawater, M.g., P.o. 11th May 2006
Sediment, seawater, Mg, P.o. 23rd September 2006
Seawater, M.g., P.o. 28th May 2007
3 42°26′93″ 18°32′15″ 7.0 Sediment, seawater, M.g., P.o. 29th September 2005
Seawater, M.g., P.o. 11th May 2006
Sediment, seawater, Mg, P.o. 23rd September 2006
Seawater, M.g., P.o. 28th May 2007
4 42°23′76″ 18°34′53″ 18 Sediment, seawater, M.g., P.o. 29th September 2005
seawater, M.g., P.o. 11th May 2006
Sediment, seawater, Mg, P.o. 23rd September 2006
Seawater, M.g., P.o. 28th May 2007
5 42°23′72″ 18°33′64″ 6.5 Sediment, seawater, M.g., P.o. 29th September 2005
Seawater, M.g., P.o. 11th May 2006
Sediment, seawater, Mg, P.o. 23rd September 2006
Seawater, M.g., P.o. 28th May 2007
6 42°21′37″ 18°41′88″ 11 Sediment, seawater, M.g., P.o. 29th September 2005
Seawater, M.g., P.o. 11th May 2006
Sediment, seawater, Mg, P.o. 23rd September 2006
Seawater, M.g., P.o. 28th May 2007
7 42°17′07″ 18°50′14″ 4.5 Sediment, seawater, M.g., P.o. 29th September 2005
Seawater, M.g., P.o. 11th May 2006
Sediment, seawater, Mg, P.o. 23rd September 2006
Seawater, M.g., P.o. 28th May 2007
8 42°07′32″ 19°04′08″ 6.5 Sediment, seawater, M.g., P.o. 29th September 2005
Seawater, M.g., P.o. 11th May 2006
Sediment, seawater, Mg, P.o. 23rd September 2006
Seawater, M.g., P.o. 28th May 2007
9 41°54′36″ 19°14′16″ 7.5 Sediment, seawater,M.g. 29th September 2005
Seawater, M.g. 11th May 2006
Sediment, seawater,M.g. 23rd September 2006
Seawater, M.g. 28th May 2007
10 41°51′97″ 19°20′29″ 7.0 Sediment, seawater 29th September 2005
Seawater 11th May 2006
Sediment, seawater 23rd September 2006
Seawater 28th May 2007
11 42°26′25″ 18°45′71″ 0.5 Mytilus galloprovincialis 29th September 2005
11th May 2006
23rd September 2006
28th May 2007

aM.g., Mytilus galloprovincialis, bP.o., Posidonia oceanica.

Boka Kotorska Bay is at the north-west of the coast of Montenegro with a mouth 2.95 km in width and an in-land length of 28.13 km; it is surrounded by high mountains, with more than 60 000 inhabitants living on its coast [18]. The Bay is naturally divided into four smaller bays: Herceg Novi, Tivat, Risan and the Kotor Bay. Sveta Stasija is located in the small Kotor bay, near tourist and harbor Kotor city, with around 20 000 inhabitants. Herceg Novi is a favored tourist city with a marine, shipyard and food industries. Herceg Novi and Mamula Island are situated close to the entrance of the Boka Kotorska Bay: Herceg Novi just into the Bay and Mamula Island just in front of the entrance of the Bay. The beach Žanjice with hotels and cottages, similar to Bigova, Budva and Bar, is in the open coastal area of Montenegro. Budva is an urban, touristic and industrial city located in the middle of the Montenegrin coastline. Bar to the south, with 15 000 inhabitants, is an important industrial harbor, the largest on the eastern side of the Adriatic, especially for the transport of crude oil and oil products. The location Rt ᴆeran is close to a channel of salt-works. Ada Bojana is located on the estuarine of the River Bojana, which is the border between Montenegro and Albania. The problem of pollution in the vicinity of these sites is high in all periods of year because wastewater is discharged directly into the sea [2].

Mussels, seagrass, surface and bottom water samples were taken between the fall of 2005 and the spring of 2007. Only sediment samples were taken in the fall of 2005 and 2006. All sampling operations were performed simultaneously, with three repetitions, within 1 day at all locations in the four seasons. The samples were collected before and after the tourist season.

Surface sediments were collected in the near of seagrass P. oceanica meadows. At all locations, the sediments are taken by a grab sampler from a boat, and prior to analysis stored in polyethylene bags and kept in a refrigerator at 4 °C. Preparation of The sediment samples were prepared for analyses as follows: after oven drying (105 °C), grinding, homogenizing, and sieving, the quantity of elements in 1 g of dry sediment samples was measured in the fraction <63 μm. Sixty sediment samples were prepared and 10 elements analyzed following a laboratory-approved QA/QC protocol.

The concentrations of the 10 elements were analyzed by the AAS technique [1]. Analytical reagent grade chemicals, Merck standards (Germany) and Milli-Q water were used. All the results of the investigated elements in sediment samples are expressed in dry weight (d.w.).

At the same time and place, about 350 g of fresh P. oceanica samples and two liters of seawater from the bottom were collected. The P. oceanica samples were washed very thoroughly, rinsed using ultra pure water, frozen, lyophilized and reduced to a powder, which was dissolved and analyzed. The mussel M. galloprovincialis and the seawater samples from the surface were collected in the vicinity of the P. oceanica meadows. At every sampling site, about 2 kg of mussels were collected, placed in nylon bags with seawater and transported to the laboratory. The biggest 25–30 mussels, of approximately the same size, were washed and cleaned out, opened raw and the flesh scraped out of the shells, which was then frozen, lyophilized, reduced to a powder, which was dissolved and analyzed.

Biota samples are freeze-dried at –40 °C for 48 h, weighed, homogenized and ground to a fine powder. The prepared samples were then kept in plastic bags. The preparation and microwave decomposition of 0.5 g samples was described previously [1, 19]. The digested samples were diluted with ultra pure water in 25 mL volumetric flasks and then transferred to 100 mL pre-cleaned polypropylene bottles. Prior to instrumental measurements, the solutions were stored at 4 °C.

To ensure the accuracy of the applied analytical methods for the determination of heavy metals in the sediments, mussels and seagrass, certified reference materials, IAEA 158 (Marine sediment), NIST 2976 (Mussel homogenate) and IAEA 140 (Fucus sample), were also analyzed [1].

Surface water samples were taken at a depth of 0.5 m. Immediately in the days following sampling, the water samples were analyzed for the elements after filtration through a 0.45 μm fiberglass membrane and acidification with nitric acid (pH ≤ 2). The filtrate was treated with ammonium pyrrolidine dithiocarbamate (APDC, Merck, Germany) to complex the investigated elements in the seawater samples. After shaking well, methyl isobutyl ketone (MIBK, Merck, Germany) was used for extraction prior to analysis by graphite furnace AAS (GF–AAS) to determine Mn, Fe, Cu, Zn, Ni, Co, Cd and Pb concentrations, while Hg and As were measured following a Cold Vapor AAS (CV-AAS) procedure [19] using a Perkin-Elmer 2000 Hydride System coupled to an AAS. The accuracy of the methods was checked by three calibration laboratory standards: standard solution of 1000 mg/L (Merck, Germany) for each element in a seawater matrix. These standards were analyzed directly after solvent extraction as described above.

Data analysis

Principal component analysis (PCA) is a mathematical procedure used as a central tool in chemometrics simulations, and represents a multivariate technique in which simultaneous data reduction and classification is performed by transforming the data into orthogonal components that are linear combinations of the original variables [16]. PCA is realized by eigenvalue decomposition of a correlation matrix of the obtained data [20]. This transformation is defined in such a way that the first component contains the largest possible variance. The analysis is used to achieve maximum separation among clusters of parameters [16].

Using PCA, the source of the metals in the samples may be identified and their distributions estimated. PCA was applied to analyze the similarities at the sampling sites and seasons to identify the source and apportionment of trace elements in the surface sediments [17, 21]. The main aim of PCA is to reduce the number of variables that need to be considered into a smaller number of indices, principal components (PCs), which can be more easily interpreted [20, 22]. PCA was used to allow easy representation of the trace elements and samples in a two-dimensional diagram. The first two PCs were extracted and utilized in bivariate plots; loadings were considered to evaluate the correlations between the variables. Cluster analysis (CA) was performed to classify the samples of surface sediments, mussels, seagrass and seawater. In this study, “complete linkage” was used and the (1–r) distance was calculated. In addition, linear correlation coefficients were calculated to understand the inter-element relationships. The data were analyzed by Statistica software (Data Analysis Software System, v.10.0, StatSoft, Inc, Tulsa, OK, USA). Prior to PCA and CA, the data sets were examined for outliers [23].

Results and discussion

Descriptive analysis of the trace elements in the surface sediments

The concentration data set of the trace elements Fe, Mn, Ni, Zn, Cu, Co, As, Pb, Cd, and Hg in 20 examined sediment samples with three repetitions in two fall seasons (2005 and 2006), at 10 locations (Sv. Stasija, Kukuljina, Herceg Novi, Mamula, Žanjice, Bigova, Budva, Bar, Rt ᴆeran, and Ada Bojana) were evaluated by PCA analysis in order to determine the average concentrations of the selected trace elements, perceive their variability, define their natural or anthropogenic origin, and to identify possible sources of contamination. The descriptive statistics of the trace elements in the surface sediment samples are given in Table 2. The most frequently occurring pattern of the contents of the microelements was: Fe > Mn > Ni > Zn > Cu > Co > As > Pb > Cd > Hg.

Table 2

Descriptive statistics of the concentrations (mg/kg) of the investigated elements in the surface sediment samples.

Valid No. Mean Min. Max. Var. SD CVa
Fe 20 11 180.0 713.7 40 866.0 166 144 748 12 889.0 115.3
Mn 20 425.5 132.4 984 79 657 282.2 66.3
Zn 20 26.0 4.0 67.2 415.2 20.4 78.2
Ni 20 79.6 2.7 336 12 615 112.3 141
Cu 20 10.2 3.2 24.8 48.8 7.0 68.1
Co 20 9.3 1.0 26.2 47.9 6.9 74.5
As 20 4.8 0.0 19.8 25.9 5.1 106.7
Pb 20 3.1 0.1 9.6 6.8 2.6 83.8
Cd 20 0.3 0.05 0.9 0.07 0.27 95.5
Hg 20 0.04 0.0 0.1 0.0 0.03 79.9

aCV, coefficient of variation.

A number of geochemical and hydrological processes determine the fate of metals in sediments, which are influenced by sediment mineralogy, grain size, organic content, and anthropogenic enrichment. Different processes that contribute to trace elements storage in sediments from the water column are inorganic/organic complexation, precipitation, adsorption and mixed oxide formation [15]. The results demonstrated that Hg exhibited the lowest (0.0–0.10 mg/kg) and Fe the highest concentrations (714–40 866 mg/kg) in investigated surface sediment samples. The range of Fe concentration showed the most spatial variability among all the elements reported in the present study (Table 2): the standard deviation (SD) was the greatest for Fe (12 889).

Statistically significant correlations (p < 0.001) were observed between the contents of Fe, Mn and Zn (r = 0.748–0.959). These elements were also in strong correlation with the Cu content (p < 0.001, r = 0.874–0.899), while Ni content correlated even stronger with the Cu content (p < 0.001, r = 0.917). The Pb content correlated with the Cd content (p < 0.001, r = 0.638).

Descriptive analysis of the concentrations of trace elements in mussels

The concentration data set of the trace elements Fe, Mn, Ni, Zn, Cu, Co, Cr, As, Pb, Cd, and Hg consisted of 20 mussel samples from the five locations where they were found to grow (Sv. Stasija, Kotor, Žanjice, Bar and Rt ᴆeran), between the fall of 2005 and the spring of 2007. The descriptive statistics of the eleven trace elements in the mussel samples are given in Table 3 The most frequently occurring pattern of the micro elements content in mussels was: Fe > Zn > Mn > Pb > Ni > Cu > As > Co > Cr > Co > Hg.

Table 3

Descriptive statistics of the investigated element concentrations (mg/kg) in mussel samples.

Mussel Valid No. Mean Min. Max. Var. SD CV
Fe 20 225.6 86.7 603.3 19 685 140.3 62.2
Mn 20 20.0 1.5 85.0 466 21.6 108
Ni 20 10.0 3.4 18.9 19.3 4.4 43.9
Zn 20 167.0 82.0 345.0 5691 75.4 44.4
Cu 20 9.9 4.6 17.2 12.3 3.5 35.2
Co 20 4.5 0.1 10.0 6.3 2.5 55.3
Cr 20 3.9 2.0 6.6 1.6 1.3 32.6
As 20 7.6 1.9 20.5 25.2 5.0 66.2
Pb 20 10.5 2.3 31.2 39.0 6.2 59.4
Cd 20 2.1 1.0 3.5 0.5 0.7 32.7
Hg 20 0.4 0.03 1.0 0.1 0.4 106

The results show that Hg exhibited the lowest (0.03–1.06 mg/kg) and Fe the highest concentrations (86.7–603.3 mg/kg) in the investigated samples, the same as in the investigated sediment samples. The range of Fe concentration showed the most spatial variability among all the elements reported in the present study (Table 3): the standard deviation was the greatest for Fe (140.3). High amounts of Zn were also registered, between 82.0 and 345.0 mg/kg with a standard deviation of 75.4 mg/kg. That the highest values were found for the concentrations of Fe, Zn, and Mn in mussels is not surprising as they are essential elements for their development and growth [24].

Statistically significant correlations (p < 0.001) were observed between the contents of Fe and Mn content (r = 0.871), the contents of Ni and Co (r = 0.801) and between Fe and Zn (r = 0.733). Evaluated coefficient of correlation between Mn and Zn content was 0.689 (p < 0.01), while negative correlation was observed between Co and As (r = –0.533, p < 0.05) referring to different accumulation sources of these two elements in the mussel [1].

Descriptive analysis of the trace elements in seagrass

The concentration data set of the trace elements Fe, Mn, Ni, Zn, Cu, Co, Cr, As, Pb, Cd, and Hg in 32 seagrass samples from eight locations (Sv. Stasija, Kukuljina, Herceg Novi, Žanjice, Mumula, Bigova, Budva and Bar) collected between the fall of 2005 and the spring of 2007 were monitored. It was not possible to sample seagrass at Ada Bojana, because it could not survive due to input of fresh water from the Bojana River or at Rt ᴆeran, due to the presence of the salt-works in this area. The most frequently occurring pattern of the content of microelements in the seagrass samples was: Fe > Mn > Zn > Ni > Pb > Cu > Co > As > Co > Cr > Hg, Table 4. Also in the seagrass samples, the Hg content was the lowest (0.03–3.46 mg/kg) and Fe the highest (318.3–2791 mg/kg), as was the case in the surface sediment and mussel samples. The range of Fe concentration showed the most spatial variability among all the elements determined in the present study (Table 4) and the standard deviation was also the greatest (555). High amounts of Mn were also registered, between 57.0 and 385.0 mg/kg, with a standard deviation of 79.83 mg/kg. The contents of the essential heavy metals were the highest for those that have biochemical and physiological functions in plants (Fe, Mn and Zn) [24].

Table 4

Descriptive statistics of the concentrations of the investigated elements (mg/kg) in the seagrass samples.

Seagrass Valid No. Mean Min. Max. Var. SD CV
Fe 32 939.0 318.3 2791 307 898 555 59.1
Mn 32 154.3 57.0 385.0 6373 80 51.7
Ni 32 28.9 20.9 39.8 33.2 5.8 19.9
Zn 32 101.5 34.9 370.0 6213 78.8 77.6
Cu 32 7.3 1.5 19.6 16.5 4.1 55.9
Co 32 5.4 3.2 8.8 2.1 1.4 26.4
Cr 32 2.1 0.2 8.7 3.5 1.9 89.4
As 32 2.8 0.5 13.6 6.8 2.6 92.1
Pb 32 8.7 2.5 23.0 32.7 5.7 65.8
Cd 32 2.4 0.2 4.4 0.7 0.8 35.1
Hg 32 0.7 0.03 3.5 0.8 0.9 127.6

A statistically significant correlation (p < 0.001 level) was observed between the Fe and Cr content (r = 0.606) in P. oceanica. The contents of Cr was also highly correlated with the As content (r = 0.676), as was the content of Fe, while the Mn content was significantly related to the Pb content, at the p < 0.01 level with a correlation coefficient r = 0.488. The calculated correlation coefficient between the Hg and the Pb content was 0.478, statistically significant at p < 0.01 level.

Descriptive analysis of the trace elements in seawater

The concentration data set for the trace elements Fe, Mn, Ni, Zn, Cu, Co, Cr, As, Pb, Cd, and Hg in the 40 examined surface and in the 40 examined bottom seawater samples at 10 locations (Sv. Stasija, Kukuljina, H. Novi, Žanjice, Mamula, Bigova, Budva, Bar, Rt ᴆeran and Ada Bojana), between the fall of 2005 and the spring of 2007 were monitored (Tables 5 and 6). The contents of the microelements in the surface seawater samples were mostly in the following order: Fe > Co > Ni > Mn > Zn > Pb > Cu > Cu > As > Hg, while in the bottom seawater samples, the order was: Fe > Ni > Mn > Zn > Co > Pb > Cd >Cu > As > Hg.

Table 5

Descriptive statistics of the concentrations (μg/L) of the investigated elements in the surface seawater samples.

Valid No. Mean Min. Max. Var. SD CV
Fe 40 12.9 3.0 61.4 124.5 11.2 86.6
Mn 40 9.6 2.4 20.0 0.7 4.9 1.7
Ni 40 10.0 2.1 21.5 35.9 6.0 59.7
Zn 40 9.5 0.4 36.1 38.9 6.2 66.0
Cu 40 5.5 1.4 19.7 9.2 3.0 54.9
Co 40 10.5 1.8 101.0 297 17.2 163
As 40 1.4 0.5 3.1 0.6 0.8 55.6
Pb 40 7.4 1.3 10.8 6.2 2.5 33.7
Cd 40 5.9 0.0 7.4 1.7 1.3 21.6
Hg 40 1.0 0.2 2.0 0.1 0.3 28.7
Table 6

Descriptive statistics of the concentrations (μg/L) of the investigated elements investigated in bottom seawater samples.

Valid No. Mean Min. Max. Var. SD CV
Fe 40 26.2 3.8 243.0 2006 44.8 171.2
Mn 40 10.0 2.7 20.0 75.4 15.7 50.2
Ni 40 10.5 3.5 21.5 35.9 5.9 56.8
Zn 40 8.7 3.6 20.7 12.6 3.6 40.7
Cu 40 4.8 0.9 9.4 4.2 2.0 43.2
Co 40 7.5 1.8 24.0 13.1 3.6 48.4
As 40 1.4 0.4 3.2 0.6 0.8 55.9
Pb 40 7.2 2.9 10.9 5.6 2.4 32.7
Cd 40 6.1 2.1 7.7 0.9 1.0 15.8
Hg 40 1.0 0.4 2.0 0.07 0.3 27.0

The descriptive statistics of the 10 trace elements in the seawater samples are given in Tables 5 and 6 for the surface and bottom samples, respectively. The results showed that Hg exhibited the lowest concentrations (0.2–2.0 μg/L) for surface/bottom water. The highest observed concentration for Co was 101.0 μg/L in the surface water and for Fe, it was 243.0 μg/L in the bottom water. The range of Co and Fe concentrations showed the greatest spatial variability among all the elements reported in the present study (Tables 5 and Table 6): the standard deviation was the greatest for Co (17.2 μg/L) for surface water and for Fe (44.8 μg/L) for bottom water.

Statistically significant correlations between various element contents were observed in the surface water samples. The Mn content was correlated with Ni, statistically significant at the p < 0.001 level (r = 0.595), while Fe content was related to the Cu and Zn contents (r = 0.627 and r = 0.698, respectively). The content of Cu was positively correlated with the Zn content (p < 0.001, r = 0.626), while the Pb content was correlated to the As and Cd contents (p < 0.001, r = –0.765 and r = 0.599, respectively).

The analysis of the bottom water also showed some relatively good correlations between the concentrations of the elements. The content of Mn was correlated to the Ni content, statistically significant at p < 0.001, with r = 0.639, as well as to the As and Pb content (p < 0.01 level, r = –0.462 and r = 0.517, respectively). Fe was correlated to Co and As at the p < 0.01 level, r = 0.408 and r = 0.518, respectively. The As content was negatively correlated to the Hg content at the p < 0.01 level, r = –0.416 and to the Pb content, p < 0.001, r = –0.758. The correlation coefficient between the Cd content and the Pb content was 0.416, statistically significant at the p < 0.01 level.

The ranges of the concentrations of the observed elements in the surface sediment, mussel, seagrass and seawater were comparable to those reported for the Adriatic coastal area in the neighboring countries – Albania and Croatia, [3–5, 25–31], Table 7. The seagrass P. oceanica is a widely distributed species in the southern, but not in the northern part of the Adriatic Sea. This grass is very sensitive to pollutants and disturbances, and requires a slightly warmer sea, which are probably the reasons no data exists for P. oceanica from the Croatian coastal area.

Table 7

Ranges of the concentrations of the studied metals in the surface sediments, mussels, seagrass (mg/kg) and seawater (μg/L) of neighboring countries.

Element Seawater (μg/L) P. oceanica (mg/kg d.w.) M. galloprovincialis (mg/kg d.w.) Sediment (mg/kg d.w.)
Range

[27, 29–31]
Range

[27]
Range

[27, 28]
Range

[3–5, 25–27, 31]
Fe 0.60–12.1 NA 53.4–719 4.7–63.5a
Mn 0.0–4.8 NA 2.0–13.0 2.0–2344.0
Zn 0.1–14.1 NA 59.0–273.0 9.6–135.0
Ni 0.1–1.4 NA 0.8–5.0 15.3–462.0
Cu 9.3–72.0 6.3–16.5 3.7–88.4 0.1–336.0
Co NA NA NA NA
Cr 0.7–3.9 1.29–37.1 0.98–2.9 4.1–253.0
As NA NA 4.0–30.0 NA
Pb 1.9–4.4 2.21–19.1 1.15–7.0 2.5–80.8
Cd 0.0–3.4 0.22–1.86 1.97–4.10 0.0–5.7
Hg 0.0–0.24 NA NA 0.1–2.0

NA, not analyzed; aFe (%)-in the sediments.

According to literature data (Table 7), Cu and As were present in higher concentrations in mussels from Albania and Croatia, respectively, and Pb in lower concentrations [27, 28] compared to mussels from Montenegro. The range of Cd was lower in mussels from Montenegro than in mussels from Albania [27]. There is no data about heavy metals in P. oceanica in the Croatian coastal area. In P. oceanica from Montenegro, there were higher concentrations of Cu, Pb and Cd, and much lower Cr concentration than in P. oceanica from Albania. In the case of seawater the highest concentration of Cu was measured in the Albanian coastal area [27].

According to the literature data (Table 7), the concentrations of Fe, Mn, Ni, Cu, Zn, and Pb are much higher in the surface sediments of the Albanian and lower in the Croatian coastal area, in relation to those found in the Montenegrin coastal area. The content of Mn is similar in Croatian and Montenegrin surface sediment, but in Albanian, it is higher. The Ni concentration was the highest in the Albanian, lower in the Montenegrin and the lowest in the Croatian coastal areas. There is no data related to the Co concentration in the surface sediments of the Albanian and Croatian coasts. Generally, the ranges of the concentrations of Cd and Hg in the surface sediments of the Albanian coastal area were slightly higher than those in the Montenegrin coastal area [4, 31].

If the conclusions derived from Table 7 and the geographical distribution of metals concentrations in Montenegrin coastal area, Fig. 2, are compared, trends could be observed among the obtained data, i.e., a decreasing trend of the Fe, Mn, Ni, Cu, Co and Zn concentrations in the surface sediment from the south (Albanian coastal area) to the northern part of Montenegro. The contents of Cd, Hg and Pb in the surface sediment were the highest in the semi-closed Boka Kotorska Bay when the Montenegrin coast is considered, Fig. 2. The geographical distributions of the concentrations of pollutants in the surface sediments in the study area are shown in Fig. 2.

Fig. 2 Geographical distribution of pollutants concentrations in the surface sediment in the study area.

Fig. 2

Geographical distribution of pollutants concentrations in the surface sediment in the study area.

The decreasing distribution pattern of the Fe, Mn, Ni, Cu, Co, and Zn contents from the south (Albanian coastal area) to the northern part of Montenegro and Croatia is influenced by the counter-clockwise system of the southeastern Adriatic current, which carries these elements from their sources [5]. The majority of trace elements in the coastal area of the southeastern Adriatic are strongly related to riverine inflows and catchment geology [3–5]. The input of Ni, Cu, and Mn near the Montenegrin coastal area was observed as a result of the inflow of the Drina River in Albania [3], and of the Bojana River, which is the natural border of Montenegro and Albania. The highest levels of Fe, Mn and Ni in the surface sediments are closely correlated to the geological background of the study area [3, 4] and hydro-geological impact [5].

Beside hydro geological impact and atmospheric inputs [2], the main toxic pollutant loads in the Montenegrin coastal area include agriculture, and domestic and industrial waste, especially in the semi-closed Boka Kotorska Bay.

Principal component analysis (PCA)

Principal component analysis was performed to analyze the relationships between environmental impacts and the concentrations of the elements in the surface sediment, mussels, seagrass and water (surface and bottom) samples along the Montenegrin coastal area between the autumn of 2005 and the spring of 2007. The rotation of the principal components was executed by the varimax method with Kaiser Normalization.

PCA analysis of element composition found in the sediments showed that the first two principal components (PCs) explained 72.90 % of the total variance in the original data, Fig. 3a. The first component (PC1) contributed 52.02 %, and the second component (PC2) 20.88 % to the total variance, Fig. 3a. Fe, Zn, Ni, and Cu contributed the most in the PC1 calculation (16.7, 16.2, 16.0, and 17.8 %, respectively), while Pb, Cd, and Hg contributed the most in the PC2 evaluation (26.9, 31.9, and 21.7 %, respectively). The main contributions to PC1 are natural to geological structure, as well as some anthropogenic sources. Most of the relatively high concentrations these ions present, especially Fe, Ni, Mn, and Cu, could have resulted from soil leaching and hydro-geological processes [4]. which include rivers and rainfall inflows, and simultaneously represent a type of surface sediment below the water column [4, 7]. This relationship showed that trace metals were mainly supplied from the drainage area of the rivers of the southeastern Adriatic [5], which confirms further that the soil of the coastal area must be rich in Fe, Mn, Ni, and Cu [3]. This indicates a high possibility of erosion of the coastal soil by rainfalls and tidal waves, and geological catchment [5]. A recent input of Ni, Cu, and Mn near the Montenegrin coastal area was observed as result of the input of the Drina River [4]. PC1 could be considered as being formed mostly from terrestrial sources.

Fig. 3 Biplot of the correlated trace elements and locations in surface sediment samples of the southeastern Adriatic: (a) sediments, (b) mussels, (c) seagrass, (d) surface seawater, and (e) bottom seawater.

Fig. 3

Biplot of the correlated trace elements and locations in surface sediment samples of the southeastern Adriatic: (a) sediments, (b) mussels, (c) seagrass, (d) surface seawater, and (e) bottom seawater.

PC2 could be considered as being formed mostly from anthropogenic sources (Hg, Cd, Pb, and partly As), i.e., implying the impact of manmade sources. The factor loading of As was of no great burden to any one component, which means that As originates from both natural and anthropogenic sources.

Wastewater and run-off from agriculture, phosphate fertilizers and pesticides, could be responsible for the Cd and As concentrations in surface sediments in this area [2], while the Pb and Hg distributions in the surface sediments were from a common main source, atmospheric deposition [32]. In addition, many studies have described increases in Cd, Pb, Hg, and As concentrations in coastal sediments as been the result of particulates produced during the burning of coal and oil fuel [33].

However, in the investigated area, a trend of decreasing Hg, Cd, Pb and As contents was observed over the studied period, and also a decreasing trend in the Co, Zn, Cu, Mn, Fe, and Ni contents, as could be seen from Fig. 3a, was observed at almost all the sampled locations (Sv. Stasija, Kukuljina, Herceg Novi, Mamula, Žanjice, Bigova, Budva, and Bar), while the samples found at Rt ᴆeran and Ada Bojana showed increases in the contents of Co, Zn, Cu, Mn, Fe, and Ni between the fall of 2005 and 2006. The increasing content of these elements in the surface sediment samples found in Rt ᴆeran and Ada Bojana was not surprising, as Rt ᴆeran is a location close to salt-works and Ada Bojana is located at the mouth of the River Bojana.

PCA analysis of the element composition found in mussels showed that first two principal components (PCs) explained 48.15 % of the total variance in the original data, Fig. 3b. The first component (PC1) contributed 27.52 % and the second component (PC2) of 20.63 % to the total variance, Fig. 3b. Mn, Fe, Zn, Ni, Co, and Cu contributed the most in the calculation of PC1 (18.9, 21.5, 16.4, 18.5, and 19.8 %, respectively), while As contributed the most in the evaluation of PC2 (18.0 %). In this case, a marked separation between the accumulation patterns of metals in the mussel species was evident [34]. The grouping of the elements could be explained by the nature of the mussel’s diet: since mussels are filter feeders, this supports the view that particulate arsenic is the most important source of metal accumulation in bivalves, as the filter-feeding bivalves accumulate arsenic from ingested living and dead particles. Marine mussels have a limited ability to bioconcentrate inorganic arsenic from seawater, but can bioaccumulate organo-arsenic compounds from their food, depending on the salinity and phosphorus content [35]. Among the studied trace metals, for Fe, Zn, and Mn, a close relationship seems to exist between their contents in mussel tissue and their contents in the corresponding water [36, 37]. Co, Cu, and Ni could arise from organic and inorganic sources and their entries into mussel are from both suspended particles and water [1, 19]. Increasing trends in the concentrations of Pb, Cd, and Cr, and partly As, were observed over the investigation years, as can be seen from Fig. 3b, at all observed locations (Sv. Stasija, Kotor, Žanjice, Bar, and Rt ᴆeran), which could be attributed to human activity.

PCA analysis of the element composition found in seagrass showed that the first two principal components (PCs) explained 43.4 % of the total variance in the original data. The first component (PC1) contributed 23.95 %, and the second component (PC2) contributed 19.39 % to the total variance, Fig. 3c. Cr, The content of Fe and As contributed the most in the PC1 calculation (23.0, 20.6, and 17.5 %, respectively), while the contents of Mn, Hg and Pb contributed the most in the evaluation of PC2 (18.8, 16.3, and 32.5 %, respectively). In the case of marine flowering, the PCA grouping of the elements into the main components could be considered as the bioaccumulation route of these elements into the aquatic plants: Cr, Fe, and As probably accumulated directly from the sediment through the roots, but Mn, Hg, and Pb are assumed to come from the water column [1, 38]. Trace element uptake in sea grasses differs depending on the trace element and the plant tissue. This specificity depends on the chemical properties of each trace element [38]. Researchers found that some groups of trace elements have similar accumulation patterns [38]. Therefore, the PCA of P. oceanica plotted the same elements very close together, indicating very similar accumulation patterns by this plant. The content of Cu, and also of Cd and Ni, tended to increase in the seagrass over the observed seasons, for almost all the observed locations (Sv. Stasija, Herceg Novi, Žanjice, Mumula, Bigova, Budva, and Bar). The exception was Kukuljina, where increased contents of Fe, Cr, and As were registered in all samples, for all the observed seasons.

The PCA analysis of the element composition found in the surface water samples showed that the first two principal components (PCs) explained 48.35 % of the total variance in the original data. The first component (PC1) contributed 30.68 %, and the second component (PC2) contributed 17.67 % of the total variance, Fig. 3d. The contents of Mn, As and Pb contributed the most in the calculation of PC1 (18.6, 23.8, and 22.9 %, respectively), while the contents of Fe and Co contributed the most in the evaluation of PC2 (28.0 and 29.2 %, respectively). The measurements of the contents of the elements obtained from observed locations (Sv. Stasija, Kukuljina, H. Novi, Žanjice, Mamula, Bigova, Budva, Bar, Rt ᴆeran, and Ada Bojana) showed that lower contents of As and Fe, as well as of Pb, Mn, and Ni, were observed in the spring seasons. The areas marked by circling on the diagram represent the specific seasonal behavior of the tested samples: they contained the highest levels of Fe and As and the lowest contents of Hg and Pb in the fall of 2005 until the fall of 2006, when the situation was completely the opposite. There was less variation in the concentrations of studied trace elements in the spring seasons.

PCA analysis of the composition of the elements found in the bottom water samples showed that the first two principal components (PCs) explained 50.29 % of the total variance in the original data. The first component (PC1) contributed 32.11 %, and the second component (PC2) contributed 18.18 %. The contents of Fe, As and Pb contributed the most in the calculation of PC1 (22.0 %, 18.9 % and 17.4 %, respectively), while the contents of Cu, Zn and Pb contributed the most in the evaluation of PC2 (26.1, 19.2, and 15.5 %, respectively). The measurement of the contents of elements obtained from the observed locations (Sv. Stasija, Kukuljina, H. Novi, Žanjice, Mamula, Bigova, Budva, Bar, Rt ᴆeran and Ada Bojana) showed that lower contents of As, Fe, Zn and Cu, as well as of Ni, Mn, Pb, and Cd, were observed in the spring seasons. The areas marked by circling on the diagrams represent the specific seasonal behavior of the tested samples: in the fall of 2005, they contained the highest concentrations of Fe, As, Zn, and Cu, and lower contents of the other elements. In the fall of 2006, almost all samples contained Hg and Cd, and again, there was less variation in the contents of the studied trace elements in the spring season.

Montenegro is one of the richest areas in the hydrological world and is one of the rainiest parts of Europe [18, 39]. Rainfalls, riverine inflows and hydrodynamic forces may contribute to the same spatial distribution of the dissolved metals in the Montenegrin coastal area. The higher level of the toxic elements in the previous investigated seasons in seawater at all the investigated locations (Fig. 3d, e) could be the result of the regional climate [39], and implementation of EU regulations, especially the coastal wastewater regulation. At the same time, the greater variation in the concentrations of the studied trace elements in the fall seasons in seawater samples as compared to smaller variation of the concentrations of the studied trace elements in the spring seasons could be explained as a result of the summer tourist seasons.

Cluster analysis

The separate dendrograms of CA for the tested samples of surface sediments, mussels, seagrass and seawater are shown in Fig. 4. A complete linkage algorithm and (1–r) distances were used as the measure of similarity among the samples. The (1–r) distances (shown on the ordinate axis) are measured as the average difference across dimensions of the tested samples. This distance measure yields results similar to the Euclidean distance, but in this measuring technique, the effect of single large differences (outliers) is dampened (since they are not squared). Cluster analysis was used for the grouping of the trace elements in the analyzed samples. A hierarchical agglomerative procedure was employed to the scaled data using a single linkage with (1–r) distances to establish clusters. The results obtained as dendrograms are presented in Figs. 4 and 5. A shorter linkage distance indicates good correlation between the assays and a longer distance indicates a less indicative correlation between the assays (Figs. 4 and 5).

Fig. 4 Dendrograms of the cluster analysis for the grouping of the trace elements in: (a) sediments, (b) mussels, (c) seagrass, (d) surface seawater, and (e) bottom seawater of the southeastern Adriatic.

Fig. 4

Dendrograms of the cluster analysis for the grouping of the trace elements in: (a) sediments, (b) mussels, (c) seagrass, (d) surface seawater, and (e) bottom seawater of the southeastern Adriatic.

Fig. 5 Dendrograms of the cluster analysis for the grouping of sampling stations depending on the investigated elements in: (a) sediments, (b) mussels, (c) seagrass, (d) surface seawater, and (e) bottom seawater of the southeastern Adriatic.

Fig. 5

Dendrograms of the cluster analysis for the grouping of sampling stations depending on the investigated elements in: (a) sediments, (b) mussels, (c) seagrass, (d) surface seawater, and (e) bottom seawater of the southeastern Adriatic.

Fig. 4a represents two distinct groups (clusters) for the surface sediments, defined by two branches in the dendogram: a first branch (Co, Mn, Zn, Cu, Ni, and Fe) and a second branch (As, Hg, Cd, and Pb). The calculated values of the scores allowed the identification of two groups of trace elements assays in mussels, Fig. 4b. The first group of trace elements assays was divided into two branches: the consisting of Cu, Ni and Co, and the second of Zn, Fe and Mn, while the second group also consisted of two branches: the first included Pb, Hg, and As, and second Cd and Cr.

The grouping of the assays in the trace elements found in seagrass (Fig. 4c) corresponded to the PCA analysis. The first cluster assembled the Cd, Ni, Co, Pb, and Mn assays, while the second cluster was composed of the contents Zn, Cu, Hg, Fe, As, and Cr.

The PCA results were in agreement with the results of cluster analysis, almost coinciding, whereas the samples showed good discrimination capabilities.

Figure 5 represents the grouping of the sample stations according to CA for the tested samples of surface sediments, mussels, seagrass and seawater. As before, the complete linkage algorithm and (1–r) distances were used as the measure of similarity among the sample stations. Figure 5 showed quite complex, distinctive graphs of the measured concentrations of the trace elements, which correspond to those of the PCA analysis.

Figure 5a represents two distinct groups (clusters) of the surface sediments, defined by two dendogram branches: in the first branch are locations 9 and 10 from the fall 2005 and 2006. In the second brunch are all the other locations. This is not surprising, as site 9 is near the salt-works and 10 is at the mouth of the Bojana River. In Fig. 5b, the sampling locations of the mussels are also grouped into two cluster branches: the first branch is the mussels from the fall 2005 and 2006 and the second branch is mussels mainly from the spring 2006 and 2007. The mussels from the locations grouped in the second cluster branch have concentrations of the analyzed elements that are similar to each other. Deviations from this rule are related to the sampling locations of seagrass and seawater: these clusters are too complex for explanation (Fig. 5b, c, and d).

Conclusions

  • Principal component analyses (PCA) and cluster analyses (CA) were applied to determine the origin and distribution of 10 trace elements in surface sediments, mussels, seagrass, and surface and bottom seawater of the southeastern Adriatic coast, Montenegro.

  • Two principal components were recognized in the studied area using PCA, including natural and anthropogenic factors. The associations between the trace metals, i.e., as Fe, Mn, Ni, Cu, Zn, and Co, were considered to be mainly functions of natural background, whereas As was a function of natural and anthropogenic sources in the observed region of the Adriatic Sea.

  • The concentrations of the trace metals Fe, Mn, Ni, Cu, Zn, and Co decreased from the southeastern to the northern part of the studied Adriatic, i.e., from the Albanian to the Montenegrin coastal area.

  • Moreover, the contribution from rock weathering seems not to cause a meaningful increase in the levels of heavy metals (Hg, Pb, and Cd) along the southeastern Adriatic coast. This was not the case for As. The highest concentrations of Cd, Hg and Pb were in the Bay.

  • PCA analysis showed a good separation between samples collected before and during the spring season of 2007. It was found that contents of Fe, Mn, Ni, As, and Pb in seawater (surface/bottom) were lower in the spring than in the fall seasons at all the investigated locations.

  • Based on PCA analysis, the trends of the Hg, Cd, Pb, and As contents were decreasing over two investigated years in seawater in almost all the observed locations.

  • CA analysis related to the locations in the case of seawater exhibited differences in their distribution in dependence on the hydro-geological and anthropogenic input of the metals.

  • PCA analysis of the concentration of trace elements found in mussel and seagrass could be explained by their bioaccumulation. Thus, conclusions could be draw concerning the sources of trace elements in seagrass (seawater and/or sediments) and the mussels (seawater and/or sediment particles).

  • The temporal and spatial distribution of the contents of trace elements in the surface sediments and mussels could be monitored with the help of PCA and CA analyses.

Acknowledgments

This work was supported by project III 43009 of the Ministry of Education, Science and Technological Development of the Republic of Serbia.

References

[1] D. Joksimovic. S. Stankovic. J. Serb. Chem. Soc.77, 105 (2012). Search in Google Scholar

[2] M. Jovic, R. A. Stankovic, L. Slavkovic-Beskoski, I. Tomic, S. Degetto, S. Stankovic. J. Serb. Chem. Soc.76, 933 (2011). Search in Google Scholar

[3] P. Rivaro. C. Ianni. S. Massolo. N. Ruggieri. R. Frache. Toxicol. Environ. Chem.86, 85 (2004). Search in Google Scholar

[4] P. Rivaro, A. Cullaj, R. Frache, C. Lagomarsino, S. Massolo, C. M. De Mattia, N. Ungaro. J. Coastal Res.58, 54 (2011). Search in Google Scholar

[5] T. Dolenc, J. Faganeli, S. Pirc. Geol. Croat.1, 59 (1998). Search in Google Scholar

[6] P. Lazo. A. Cullaj. B. Baraj. J. Phys. IV107, 715 (2003). Search in Google Scholar

[7] B. Tanaskovski, M. Petrovic, Z. Kljajic, S. Degetto, S. Stankovic. Maced J. Chem. Chem. En.33, 139 (2014). Search in Google Scholar

[8] J. Markovic, D. Joksimovic, S. Stankovic. Arch. Biol. Sci. 64, 265 (2012). Search in Google Scholar

[9] H. Juahir, T. Mohd Ekhwan, M. S. Zain, B. M. Mokhtar, Z. Jalaludin, M. K. L. J. Jan. American-Euroasian J. Agric. Environ. Sci.4, 258 (2008). Search in Google Scholar

[10] R. R. R. Kannan, R. Arugumam, P. Anantharaman. Biol. Trace Elem Res.143, 1149 (2011). Search in Google Scholar

[11] F. Akbal, L. Gürel, T. Bahadır, I. Güler, G. Bakan. H. Büyükgüngör. Water Air Soil Poll.216, 21 (2011). Search in Google Scholar

[12] F. Zhou, H. Guo, Y. Liu, Z. Jiang. Marine Poll. Bull.54, 745 (2007). Search in Google Scholar

[13] K. Manikandan, P. Kannan, M. Sankar. Earth Sci. India, 5, 1 (2012). Search in Google Scholar

[14] L. Slavkovic, B. Skrbic, N. Miljevic, A. Onjia. Environ. Chem. Lett.2, 105 (2004). Search in Google Scholar

[15] M. Erdogan. Monitoring and Statistical Assessment of Heavy Metal Pollution in Sediments Along the Izmir Bay Using ICP-MS. PhD thesis, the Graduate School of Engineering and Sciences of Izmir Institute of Technology, Izmir, Turkey (2009). Search in Google Scholar

[16] J. Hlavay, T. Prohaska. M. Weisz, W. W. Wenzel, G. J. Stingeder. Pure Appl. Chem.76, 415 (2004). Search in Google Scholar

[17] M. Dassenakis. V. Paraskevopoulou, C. Cartalis, N. Adaktilou, K. Katsiabani. Pure Appl. Chem.84, 335 (2011). Search in Google Scholar

[18] D. Magas. Geoadria7, 51 (2002). Search in Google Scholar

[19] C. Lafabrie, C. Pergent-Martini. G. Pergent. Env. Poll.151, 262 (2008). Search in Google Scholar

[20] M. Arsenović, Z. Radojević, S. Stanković, Ž. Lalić, L. Pezo. Ceram. Int.39, 1667 (2013). Search in Google Scholar

[21] M. R. Rodríguez-Barroso, L. J. García-Morales, D. M. Coello Oviedo, M. J. Quiroga Alonso. Env. Mon. Assess163, 489 (2010). Search in Google Scholar

[22] C. Mendiguchia, C. Moreno, M. Garcia-Vargas. Chemosphere69, 1509 (2007). Search in Google Scholar

[23] F. Grubbs. Technometrics1, 1 (1969). Search in Google Scholar

[24] C. P. Nagajyoti, D. K. Lee, M. V. T. Sreekanth. Environ. Chem. Lett. 8, 199 (2010). Search in Google Scholar

[25] P. Ž. Rožič, T. Dolenec, B. Baždarić, V. Karamarko, G. Kniewald, M. Dolenec. Env. Sci.Poll. Res.19, 2708 (2012). Search in Google Scholar

[26] A. Popadić, J. Vidović, V. Cosović, D. Medaković, M. Dolenec, I. Felja. Marine Poll. Bull. 76, 333 (2013). Search in Google Scholar

[27] J. Malltezi, V. Dini, A. Bani, S. Sulçe. Fresen. Env. Bull.21, 3249 (2012). Search in Google Scholar

[28] V. Orescanin, I. Lovrencic, L. Mikelic, D. Barisic, Z. Matasin, S. Lulic, D. Pezelj. Nucl. Instr. Meth. Phys. Res. B 245, 495 (2006). Search in Google Scholar

[29] L. Manfra, A. Accornero. Marine Poll. Bull.50, 682 (2005). Search in Google Scholar

[30] V. Cuculic, N. Cukrov, Z. Kwokal, M. Mlakar. Est. Coast. Shelf Sci.95, 253 (2011). Search in Google Scholar

[31] J. Abshi, L. Dhaskali, M. Adhami, E. Canaj, Z. Rada. Nat. Montenegrina7, 475 (2008). Search in Google Scholar

[32] D. Romic. M. Romic. M. Zovko. H. Bakic. G. Ondrasek. Envron. Geochem. Health34, 399 (2012). Search in Google Scholar

[33] M. E. García, J. J. Cruz-Motta, O. Farina, C. Bastidas. Cont. Shelf Res.28, 2757 (2008). Search in Google Scholar

[34] L. Campanella, M. E. Conti, Principal Component Analysis Applied to Biomonitoring Studies: A Case Study, SIS 2009, Pescara, September 23–25, Italy, pp. 297–300. Search in Google Scholar

[35] J.M. Neef. Environ. Toxol. Chem.16, 917 (1997). Search in Google Scholar

[36] C.U. Rivonker, A.H. Parulekar. Indian J. Mar. Sci.27, 411 (1998). Search in Google Scholar

[37] C. K. Yap, S. G. Tan. Indian J Mar. Sci.36, 227 (2007). Search in Google Scholar

[38] C. Sanz-Lazaro, P. Malea, E. T. Apostolaki, I. Kalantzi, A. Marın, I. Karakassis. Biogeosciences9, 2497 (2012). Search in Google Scholar

[39] V. Ducić, J. Luković, D. Burić, G. Stanojević, S. Mustafić. Nat.Hazard Earth Syst. Sci.12, 687 (2012). Search in Google Scholar

Published Online: 2014-7-7
Published in Print: 2014-7-22

©2014 IUPAC & De Gruyter Berlin/Boston