Toxicity bioassays in core sediments from the Bay of Santander, northern Spain

ARTICLE IN PRESS

Environmental Research 106 (2008) 304–312 www.elsevier.com/locate/envres

Toxicity bioassays in core sediments from the Bay of Santander, northern Spain A. Coza,, O. Rodrı´ guez-Obesoa, R. Alonso-Santurdea, M. A´lvarez-Guerraa, A. Andre´sa, J.R. Viguria, D. Mantzavinosb, N. Kalogerakisb a

Department of Chemical Engineering and Inorganic Chemistry, University of Cantabria, ETSII y T. Avda Los Castros s/n, 39005 Santander, Spain b Department of Environmental Engineering, Technical University of Crete, Polytechneioupolis, Chania, Crete, Greece Received 15 November 2006; received in revised form 16 May 2007; accepted 19 May 2007 Available online 9 July 2007

Abstract The use of Vibrio fischeri as luminescence bacteria is particularly effective in evaluating contaminated sediment. In this study, the ecotoxicity of five core sediments from the Bay of Santander, northern Spain, utilising V. fischeri as marine bacterium, was carried out. Different toxicity assay procedures were applied in order to study the influence of the mobility and bioavailability of the pollutants. Basic Solid Phase Test (BSPT) in whole sediment and acute toxicity test, using pore water and three leaching test procedures as liquid extracts, were applied. In addition, the study of the influence of the pH value on the toxicity results of the leaching tests was conducted. The obtained results show toxicity units (TU50) values in BSPT test ranging from 0.42 to 39.06 with a decrease with depth as general trend and TU50 values from 0.010 to 0.389 in the liquid extracts, where TU50 is calculated as the inverse of EC50 (%). The obtained data show the historical toxicity trends of the Bay of Santander and provides a technical database for the management of contaminated sediments. Moreover, these results showed evidence that each sediment test procedure provided independent and complementary ecotoxicological responses useful for a sediment classification. In order to analyse the correlations between chemical parameters (both organic and inorganic) and the toxicity results, the self-organising map (SOM) neural network and regression equations were applied. Satisfactory correlations (R ¼ 0.93) between chemical concentrations of sum of five heavy metals and 16 PAHs and BSPT toxicity were obtained. r 2007 Elsevier Inc. All rights reserved. Keywords: Sediment; Ecotoxicity; Self-organising map; Vibrio fischeri; Artificial neural networks

1. Introduction Contaminated sediments with hazardous compounds are one of the main problems in restoration activities of aquatic environment and in dredging operations. For these sediments, different guidelines have been developed in order to obtain the best assay procedures through Sediment Quality Assessment (SQA) (McCauley et al., 2000; Wenning et al., 2005; ICES, 2003; DelValls et al., 2004). In this sense, some biological tests have been developed for the characterisation of the toxicity in sediments of complex character (organic and inorganic pollutants). The advantage of the biotests is that the toxicity may be measured in different sediment compartCorresponding author. Fax: +34 942 201591.

E-mail address: [email protected] (A. Coz). 0013-9351/$ - see front matter r 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.envres.2007.05.009

ments and can be contrasted (Coz et al., 2004; Olajire et al., 2005). In addition, international trends suggest an integrated approach based on a weight of evidence sediment assessment which combines chemical characterisation with toxicity test and environmental evaluations (Riba et al., 2004; Chapman and Anderson, 2005). The use of Vibrio fischeri as luminescence bacteria test is one of the most common microbial test used for risk assessment of aquatic environment. In the literature, different biotests (fish, daphnia, algae, and luminescent bacteria) have been compared (Ferna´ndez et al., 1997; Font et al., 1998; Nendza, 2002; DelValls et al., 2004). The bacterial bioluminescence assay has been considered in this paper because it is a rapid, robust, and highly sensitive method. The results show reproducibility and they are costeffective. Furthermore, the bioluminescence assay has a worldwide application and standardisation for regulatory

ARTICLE IN PRESS A. Coz et al. / Environmental Research 106 (2008) 304–312

purposes; it is preferred over other bacterial screening techniques; and it is one of the most suitable tests for sediment toxicity assessment (Guzzella, 1998; Den Besten et al., 2003; Van Beelen, 2003). Among different procedures, the Basic Solid Phase Test (BSPT) (Azur Environmental, 1998) leads to the measuring of the toxicity of the entire sediment and has been widely studied for marine sediments in the literature (Salizzato et al., 1998; Lahr et al., 2003; Vigano et al., 2003; Campisi et al., 2005). In addition, the basic test and the 100% test can be applied on pore water, on aqueous sediment elutriates, and on sediment organic extracts. Previous studies show that sediments of the Bay of Santander, northern Spain, contain significant concentrations of heavy metals and organic pollutants derived from intensive industrial, agricultural, and urban activities (Gonza´lez-Pin˜uela et al., 2006; Viguri et al., 2007). In this work, a study of the ecotoxicity in the whole core sediments from the Bay of Santander and in elutriate and extracts from these sediments was conducted in order to evaluate their possibilities of management. BSPT in the sediment matrix was carried out in order to study the total toxicity of the core sediment and the historical toxicity trends of the area. Pore water extract and two leaching tests were carried out in order to study the toxicity in different stages for dredging, landfill, or restoration. Toxicity characteristic leaching procedure (TCLP) and European Norm EN 12457 at two liquid/solid ratios, recommended for leaching tests before ecotoxicity analysis in waste and soils (Coz et al., 2004; Bednarik et al., 2005; Mantis et al., 2005), were chosen as leaching tests in order to study the influence of the extraction medium in the sediment toxicity and the bioavailability results. Bioavailability refers to the difference between the amount of pollutant and the actual dose of this pollutant the organism receives, and determines to a large extent the actual risk of contaminants in sediments. Regulatory procedures based on equilibrium leaching tests allow the characterisation of the environmental behaviour of soils and sediments in different scenarios. TCLP and EN 12457 are very common tests in USA and Europe, respectively, and both tests are used in Spanish and European regulations for the determination of the hazardousness for solid materials. In the TCLP, acetic acid is used as a leachant to simulate the organic acid biodegradation and distilled water is used in the EN 12457 test to evaluate the extractability of the pollutants. Furthermore, due to the effect of the pH medium on these bacterial metabolitic activities (ASTM, 2004; Onorati and Mecozzi, 2004), pH correction in TCLP and EN extracts was studied. Artificial neural networks (ANN) have been recently introduced as a tool for data analysis in different fields of ecology and environment, and they have mostly performed better than classical multivariate statistical methods (Lek and Gue´gan, 1999). The Kohonen self-organising map (SOM) (Kohonen, 2001), one of the most well-known ANN with unsupervised training algorithms, can be used to seek for clusters in data without prior knowledge.

305

Specifically, the SOM implements an ordered dimensionality-reducing mapping of the training data, i.e., provides projection of multidimensional data into a two-dimensional map preserving the topology of this input data space. The SOM is a useful method for data classification purposes (Vesanto, 1999), allowing an easy visualisation of multidimensional numerical data that can be utilised to obtain very useful graphic representations that help in the tasks of interpreting the data and hunting for correlations between the analysed variables. The aim of this paper is the comparison of the Microtox toxicity results in different matrices from five core estuarine sediments: sediment solid phase toxicity, obtained by BSPT; sediment pore water toxicity, obtained by Microtox 100% test previous centrifugation; and sediment water extracts toxicity obtained by Microtox 100% test, previous application of standard leaching test with acidic or nonacidic medium, with and without pH correction, all of them in order to study the toxicity of the area and the possibilities of restoration and/or dredging. The relationships between toxicity results and the chemical parameters have been examined by means of a new methodology, SOM neural network. 2. Materials and methods 2.1. Materials Five core sediments (C1–C5) from the Bay of Santander were collected. Fig. 1 indicates the location of these sediments. Samples of polluted sediments at a depth of 50 cm from the Bay of Santander were obtained using a manual core sampler. Each sample was cut in 10 cm specimens and homogenized and transferred to clean glass jars, covered with aluminium foil and kept in the dark at 4 1C until their analysis. The results of metals and organic parameters in these sediments previously characterised (Viguri et al., 2002; Gonza´lez-Pin˜uela et al., 2006; Viguri et al., 2007) P ranged from 133 to 1540 mg/kg P dw of (Cr, Cu, Ni, Pb, and Zn) and 0.010 to 23.84 mg/kg dw of 16PAHs as shown in Table 1.

2.2. Toxicity bioassays Pore water and three aqueous extracts from EN 12457 at two L/S ratios and TCLP leaching test were tested to acute toxicity with V. fischeri as luminescence bacteria. Pore water samples were obtained after centrifugation at 4800 rpm for 15 min at 4 1C. The EN leaching test was carried out in order to study the mobility of contaminants in water. For this test, both two liquid/ solid ratios described in the landfill norm (OJEC, 2003) were performed, 2 and 10 l/kg. In the TCLP leaching test, an acetic-based extraction liquid is used in order to study the behaviour of the pollutants in this medium, in particular, those of heavy metals. The liquid/solid ratio in this test is 20 l/kg. The pH values of the liquid medium for the bioluminescence assay can produce data on the effect of the bacterial metabolitic activities and can determine different toxicity responses (ASTM, 2004; Onorati and Mecozzi, 2004). The best interval of pH for this type of bioassays is 6.5–7.0 (ASTM, 2004). In order to study the pH influence in leachates, two toxicity responses were carried out for each leaching test, the first with the obtained leachate and the second after pH correction at 6.5–7.0. Two toxicity procedures, BSPT (Azur Environmental, 1998), in the whole matrix and acute toxicity 100% (ISO, 1999) in the liquid medium, were performed according to a standard procedure (ASTM, 2004) at a constant time and temperature (15 min, 15 1C). Results can be expressed as 15 min EC50 values, where EC50 (%) is expressed as the effective

ARTICLE IN PRESS A. Coz et al. / Environmental Research 106 (2008) 304–312

306

N

Cantabric Sea

2 km

SANTANDER

5 Bay of Santander 1 MALIAÑO

4

Boo

2

ASTILLEO R

PEDREÑA ELECHAS

Cubas

PONTEJOS

San Salvador 3 Corer (50 cm)

Solía

Fig. 1. Location of the core sediments. Table 1 Physico-chemical characterization (heavy metals, PAHs) and Microtox bioassays results in core sediments from the Bay of Santander Chemical parameter (mg/ kg dw)

Parameter C1 core

C2 core

C3 core

C4 core

C5 core

Cr Cu Ni Pb

54 23 28 44

16 14 16 22

102 (0 cm)–174 (20 cm) 10 (50 cm)–45 (10 cm) 23 (50 cm)–43 (20 cm) 63 (50 cm)–146 (40 cm)

47 12 13 20

Zn

161 (50 cm)–1030 (30 cm) 310 (50 cm)–1540 (30 cm) 0.03 (50 cm)–0.52 (10 cm) 0.40 (50 cm)–7.70 (30 cm) 0.43 (50 cm)–7.82 (30 cm)

96 (50 cm)–232 (0 cm) 35 (50 cm)–100 (10 cm) 49 (50 cm)–77 (20 cm) 119 (20 cm)–229 (40 cm) 524 (20 cm)–956 (30 cm) 888 (20 cm)–1380 (30 cm) 0.07 (10 cm)–0.20 (40 cm) 0.61 (30 cm)–1.67 (0 cm) 0.69 (30 cm)–1.85 (50 cm)

571 (50 cm)–1160 (10 cm) 776 (50 cm)–1510 (10 cm) 0.06 (10 cm)–0.44 (0 cm) 0.80 (50 cm)–9.63 (20 cm) 0.86 (50 cm)–9.88 (20 cm)

56 (50 cm)–430 (0 cm)

P 5 metals P 6 LPAHsa P 10 HPAHsb P 16 PAHs

(50 cm)–130 (10 cm) (50 cm)–141 (40 cm) (50 cm)–41 (10 cm) (50 cm)–262 (20 cm)

(40 cm)–141 (0 cm) (40 cm)–71 (0 cm) (40 cm)–36 (10 cm) (40 cm)–220 (0 cm)

52 (40 cm)–766 (0 cm) 133 (40 cm)–1230 (0 cm) 0.04 (50 cm)–2.81 (0 cm) 0.39 (50 cm)–21.03 (0 cm) 0.44 (50 cm)–23.84 (0 cm)

(40 cm)–87 (40 cm)–27 (50 cm)–36 (50 cm)–69

(50 cm) (20 cm) (10 cm) (10 cm)

162 (40 cm)–633 (0 cm) 0.002 (30 cm)–0.015 (0 cm) 0.006 (40 cm)–0.025 (0 cm) 0.010 (50 cm)–0.040 (0 cm)

a

LPAHs: naphthalene, acenaphthylene, acenaphthene, fluorene, phenanthrene, and anthracene. HPAHs: fluoranthene, pyrene, benzo(a)anthracene, chrysene, benzo(b)fluoranthene, benzo(k) fluoranthene, benzo(a)pyrene, dibenzo(a,h)anthracene, benzo(g,h,i)perylene, and indeno(1,2,3-cd)pyrene. b

concentration of toxicant (solid or liquid) that causes a 50% decrease in light output. Another way to express toxicity is in toxicity units (TU50), which are calculated as the inverse of EC50 as a percentage, providing a comparative scale with the higher the value of TU50 the higher the toxicity, contrary to EC50 values, which decrease as toxicity increases (Salizzato et al., 1998; Coya et al., 2000; Van den Brink and Kater, 2006).

2.3. Self-organising map (SOM) In order to create and visualise the SOMs, the second version of the SOM Toolbox for Matlabs developed by the Laboratory of Computer

and Information Science at the Helsinki University of Technology (Vesanto et al., 1999) was used. A SOM consists of units called neurons that are organised on a regular two-dimensional grid. Each neuron is represented by a weight vector, which has as many components as the dimension of the input variables. The SOM is trained in an iterative way obtaining the so-called best-matching unit (BMU) or neuron whose weight vector has the minimum Euclidean distance between it and all the weight vectors of the neurons of the SOM. The weight vectors of the SOM are updated so that the BMU and its topological neighbours are moved towards the input sample vector. As a consequence, during this iterative training, the SOM behaves like a flexible net that folds onto the ‘‘cloud’’

ARTICLE IN PRESS A. Coz et al. / Environmental Research 106 (2008) 304–312 formed by input data, and after the training, BMUs of similar data samples will be close to each other on the final map grid obtained (Vesanto, 1999). In order to treat all variables as if they are of equal importance regardless of their scale of measurement, and to avoid that variables with high values dominate the map organisation just because of their greater influence on the Euclidean distances, data were normalised using a logarithmic transformation. To interpret the results, SOM, and component planes have been used. The map itself can show what unit is the BMU for each case of the input data (samples of different depths of the five core sediments). A component plane is a representation of the map that shows the values that take a same component of the weight vectors in each of the map units; the simple inspection of a component plane provides an idea of the spread of values of that variable, and by plotting all component planes and comparing them with each other, qualitative correlations between variables can be seen (Vesanto and Ahola, 1999).

3. Results and discussion Table 1 shows the Microtox bioassays results obtained in the solid and different liquid phases with and without pH correction. In Fig. 2, the sediment toxicity results in the BSPT according to depth were expounded. The most toxic sediments correspond to those of western area of the Bay

0

10

20

of Santander (C1–C3) as they are areas with the main anthropogenic influence, urban as well as industrial. As can be observed in Fig. 2, all samples show a low toxicity from a determined depth of the sediment. Nevertheless, in C2 core sediment the toxicity is negligible from 10 cm depth, in C3 core the toxicity from 10 cm depth is between 0.42 and 5 TU50, while C1 core shows the lowest toxicity values from a depth of 50 cm. Cores C4 and C5 show maximum toxicity of 5 TU50 at 20 cm depth and between 0 and 20 cm depth, respectively. The ecotoxicity results of the pore water and EN 12457 leaching test with a liquid/solid ratio equal to 10 l/kg show no acute toxicity (EC504100%, TU50o0.01) in all samples (Table 1). The low toxicity obtained is due to the low soluble content of contaminants in pore water and the low bioavailability of contaminants in the sediments at low L/S ratios. The ecotoxicity results, in the two leaching tests, TCLP with a L/S ratio of 20 l/kg and EN 12457 with a L/S ratio equal to 2 l/kg, with and without pH correction are shown in Fig. 3. As can be observed, the most elevated ecotoxicity values correspond to the TCLP test without pH correction, followed by the TCLP test with pH correction. The acid TU50

TU50 0

307

30

40

0

10

0

10

20

TU50 30

40

0

1964 1950

10

0

10

20

30

40

10

30

40

1901

20

30

40 1954?

50

20

C3

TU50 30

40

0

0

10

20

Depth (cm)

20

30

TU50 30

40

0.1

1

1960

1922

20

1904

30

20

30

1954? C4

5 4 3

1

40

40

50

100

10

1941

2 40

10

0

10

10 Depth (cm)

1964

C2

TU50 10

30

50

C1

0

20

40

Depth (cm)

50

0

Depth (cm)

20

Depth (cm)

Depth (cm)

1925

50

50 C5

(a)

Fig. 2. BSPT toxicity results evolution of the sediment cores: C1–C5. (a) Evolution of TU50 of all core sediments at logarithmic scale.

ARTICLE IN PRESS A. Coz et al. / Environmental Research 106 (2008) 304–312

leaching test mobilises the major quantity of contaminants increasing the leaching ecotoxicity value. The ecotoxicity values of the EN 12457 leaching test are very similar in all samples, with very low values. So that this type of leaching test for marine sediments does not mobilise large quantity of pollutants, making it necessary a more acid medium for its bioavailability. In order to study the pH influence of the TCLP leachate, the results of the toxicity in relation to the pH value are indicated in Fig. 4. As can be seen, toxicity diminishes with the pH value until values near the neutral pH, except on the C5 sediment samples for sediment depths from 30 to 50 cm. In this case, in spite of the fact that the TCLP leachate pH of these samples is close to a value of 5, low ecotoxicities are obtained. These results are according to the previous works (Ho et al., 1999; Jennings et al., 2001) where toxicity results have been conditioned by the final pH in solution. The results of ecotoxicity in the study tests are represented in Fig. 5 in logarithmic scale, with the aim to compare the results of the different procedures used in this work. In the sediments with major toxicity in the solid phase test, C1–C3, the toxicity in acid medium (TCLP leaching test) and the water medium (EN 12457 leaching test) are practically constant with the sediment depth, in spite of the fact that the solid phase test shows a marked decrease of the toxicity in the depth. Therefore, the toxicity in the whole sediments can be caused by the industrial pollutants that appeared at the last decades, in the first layers of the core sediments. However, the mobility of the pollutants is very similar in recent and old sediments of these core sediments. BSPT is recommended to give information about the toxicity of the area and TCLP and EN 12457 at 2 l/kg ratio are recommended in order to evaluate the possibilities of restoration and dredging because these assays give information about the bioavail-

ability in different water extracts. Therefore, we can recommend the restoration and/or dredging of this area based on the toxicity results in the TCLP and EN leaching tests. On the other hand, similar evolutions are obtained for the three assays (TCLP without pH correction, EN 12457 without pH correction, and BSPT) in core C5. In this core sediment, a pH correction is not necessary as can be observed in Fig. 4. In order to study the relationship among the chemical parameters in the toxicity results, SOM algorithm was applied. In the SOM algorithm, the samples of sediments were classified, using a set of five input variables: total P organic carbon (TOC), sum of five heavy metals ( 5 metals: Cr, Cu, Ni, Pb, and Zn), sum of 6 low P molecular weight PAHs—two and three rings aromatics ( 6 LPAHs: naphthalene, acenaphthylene, acenaphthene, fluorene, phenanthrene, and anthracene), sum of 10 high molecular P weight PAHs—four, five, and six aromatic rings ( 10 0.4

0.3 TU50

308

0.2

0.1

0

C5 at 30-50 cm 5

4

6

7

pH

Fig. 4. Influence of the pH in the TCLP leaching test on the toxicity results. The rectangle area highlights the samples that show low toxicity at lower pH values.

0.4

TU50

0.3

0.2

0.1

0 10 20 30 40 50 0 10 20 30 40 0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50

0 C1-Depth (cm)

C2-Depth (cm) TCLP

C3-Depth (cm) TCLP pH 7

EN

C4-Depth (cm)

C5-Depth (cm)

EN pH 7

Fig. 3. Results of toxicity units (TU50) in TCLP and EN leaching tests with or without pH correction. (’) Toxicity results in the TCLP leaching test (at 20 l/kg) without pH correction. (&) Toxicity results in the TCLP leaching test (at 20 l/kg) with pH correction. ( ) Toxicity results in the EN leaching test (at 2 l/kg) without pH correction. ( ) Toxicity results in the EN leaching test (at 2 l/kg) with pH correction.

ARTICLE IN PRESS A. Coz et al. / Environmental Research 106 (2008) 304–312

0.1

TU50

TU50

TU50 0.01 0

309

1

10

100

0

0.01

0.1

1

10

100

0.01 0

0.1

1

10

100

1964 10

1950

10

10

30

40

1901

20

30

40 1954?

50

50

1964

50 C3

TU50 1

10

0.01 0

0.1

1

1960

10 Depth (cm)

10 Depth (cm)

30

C2

TU50 0.1

20

40

C1

0.01 0

Depth (cm)

20

Depth (cm)

Depth (cm)

1925

20

30

1941 1922

20

1904

10

BSPT TCLP TCLP with pH correction EN

30

EN with pH correction 40

40 1954?

50

50

C4

C5

Fig. 5. Ecotoxicity results (TU50) in all assays represented in logarithmic scale: toxicity in solid phase through BSPT test and toxicity in the leaching extracts TCLP and EN with or without pH correction.

HPAHs: fluoranthene, pyrene, benzo(a)anthracene, chrysene, benzo(b)fluoranthene, benzo(k)fluoranthene, benzo (a)pyrene, dibenzo(a,h)anthracene, benzo(g,h,i)perylene, and indeno(1,2,3-cd)pyrene) and the ecotoxicity results in the BSTP expressed as TU50. Fig. 6 shows the SOM of 8  3 units that was obtained, where the labels added to the hexagons of the map indicate the BMU corresponding to each sediment sample. A k-means algorithm was then applied to cluster the trained map, helping to classify the units of the SOM into three different groups (clusters I–III, Fig. 6) based on the minimum Davies–Bouldin index (Vesanto and Alhoniemi, 2000). The component planes of the five input variables (Fig. 7) are an excellent aid for the interpretation of the clusters obtained. Cluster I includes deep samples of the cores C5 (depths 30–50 cm), C2 (20–40 cm), and C1 (50 cm). As shown in the component planes (Fig. 7), all these sediments are characterised by having low toxicity, with TU50 values less than 1%, as well as low levels of metals and PAHs. Cluster II corresponds to samples with toxicity values between 1.5% and 4% in TU50: that is the case of all C3 core from 10 to 50 cm, core C5 from 1 to 20 cm, C4 at 30

and 40 cm, and the sample of C2 at a depth of 10 cm. Finally, the superficial samples of cores C2 and C3, and the whole C1 core sediment (except the deepest sample at 50 cm), which have the common feature of exhibiting high values of TU50, are grouped in cluster III, which therefore can be interpreted as the cluster of the most toxic samples. Moreover, and according to the component planes of Fig. 7, these samples with the highest ecotoxicity values also correspond generally to the highest levels of metals and PAHs. These correlations between toxicity and the levels of heavy metals and PAHs that could be guessed from the component planes of the SOM were mathematically quantified. In this sense, based on the results of the SOM algorithm, regression mathematical modelling between the P chemical parameters ( 6LPAHs, S10HPAHs, and SMetals) and the ecotoxicity as BSPT results of all studied cores was developed. The software package STATGRAPHICS Plus 5.1 was used to perform both simple and multiple regression analysis of the variables and their logarithms. For all the studied cases, the best fitting corresponds to the following equation, where all chemical concentrations are

ARTICLE IN PRESS A. Coz et al. / Environmental Research 106 (2008) 304–312

310

C2-40

C5-40

C5-30 C5-50 C1-50

C2-20 C2-30

C5-10 C5-20 C3-20 C5-1

C2-10

C4-40

C3-10 C3-30 C3-50 C3-40 C4-30

C4-20 C2-1 C1-10 C1-20 C1-30

C1-1 C1-40

C3-1

Fig. 6. Distribution of the samples of different depths in the five core sediments on the self-organising map (SOM). The codes stand for the core and the depth of the sample (e.g., C2-40: sample of core sediment C2 at a depth of 40 cm). The three clusters (I–III) have been derived from the kmeans algorithm applied to the trained SOM.

expressed as mg/kg dw and ecotoxicity is expressed as EC50 as percentage: X ecotoxicity ¼ 6:07  1:96  log½ 5 metals X  0:30  log½ 10HPAHs X  0:13  log½ 6LPAHs; R ¼ 0:93. As can be observed, the heavy metal concentration gives the highest ecotoxicity influence on the sediments and the high molecular weight PAHs gives more toxicity than the low molecular weight PAHs with parameters equal to 1.96, 0.30, and 0.13, respectively. In previous works, the quantitative agreement of toxicity and chemical concentrations in sediments was poor for BSPT assay (Munawar et al., 2000; Mowat and Bundy, 2001). Other authors show significant correlations between concentration of heavy metals and toxic response with BSPT (Tay et al., 1992) and toxic effects attributable mainly to organic compounds (Salizzato et al., 1998; Van den Brink and Kater, 2006) and sulphur in sediments (Ricking et al., 2002; Calace et al., 2005). 4. Conclusions Different toxicity procedures with V. fischeri as marine bacteria were carried out in short core sediments (50 cm

P P P Fig. 7. Component planes of the SOM for the input variables ( 5 metals, TOC, 6 LPAHs, 10 HPAHs, and TU50). Each map corresponding to one variable (component) is to compare to the map representing the distribution of the sediment samples presented in Fig. 6; hexagons in the same place on different component planes correspond to the same map unit. Colours indicate the value of the component in the weight vector of each unit of the map, according to the colourbars on the right.

ARTICLE IN PRESS A. Coz et al. / Environmental Research 106 (2008) 304–312

depth) in order to study the toxicity of the area and the possibilities of these sediments as dredging, landfilling, and/or restoration. BSPT, acute toxicity in pore water and acute toxicity in two leaching tests, EN 12457 at two L/S ratios and TCLP, with and without correction of the pH value have been conducted. The BSPT for the analysis of toxicity in marine sediments gives useful information about the toxicity of the sediments in relation to the historical evolution in core sediments. The obtained results show toxicity units (TU50) values from 0.42 to 39.06 with a decrease with depth as general trend. The most toxic sediments correspond to those of western area of the Bay of Santander (C1, C2, and C3) with the highest urban and industrial activity in their coastal zone. The results of ecotoxicity in the pore water test and EN 12457 leaching test with a 10 l/kg L/S ratio show no acute toxicity. However, TCLP and EN 12457 with a 2 l/kg ratio leaching tests show higher ecotoxicity results. These leaching tests are recommended in the case of evaluating the possibilities of restoration and dredging, in order to study the bioavailavility in acid (TCLP) and neutral mediums (EN 12457 at 2 l/kg ratio). Based on the low results of toxicity in TCLP and EN 12457 leaching tests, restoration, and/or dredging are recommended in this area. In many cases, the pH value of the TCLP leachate has a great influence on the ecotoxicity results and in these cases, a neutral pH value is recommended. However, few samples have a very low toxicity results in spite of having a pH value close to 5. The interpretation of the SOM algorithm makes it possible to obtain visual correlations between toxicity and the concentration of heavy metals and PAHs that then were satisfactory quantified applying regression models (R ¼ 0.93). Taking into account the obtained correlation, the BSPT toxic effects are mainly under the influence of the Cr, Cu, Ni, Pb, and Zn studied metal concentration in the sediments (coefficient of 1.96) rather than high molecular weight PAHs concentration (coefficient of 0.3) and low molecular weight PAHs concentration (coefficient 0.13). Acknowledgments This research project was supported by the financial help of the Spanish Science and Technology Ministry Project CTM 2005-07282-C03-03. AC wishes to thank the Marcelino Botı´ n Foundation for granting him a research fellowship. MA-G was funded by the Spanish Ministry of Education and Science under an FPU fellowship. References ASTM, 2004. Standard Test Method for Assessing the Microbial Detoxification of Chemically Contaminated Water and Soil Using a Toxicity Test with a Luminescent Marine Bacterium. ASTM D 5660-96. Azur Environmental, 1998. Microtox Basic Solid-Phase Test (Basic SPT). Carlsbad, CA, USA.

311

Bednarik, V., Vondruska, M., Koutny, M., 2005. Stabilization/solidification of galvanic sludges by asphalt emulsions. J. Hazard. Mater. B122, 139–145. Calace, N., Ciardullo, S., Petronio, B., Pietrantonio, M., Abbodanzi, F., Campisi, T., Cardellicchio, N., 2005. Influence of chemical parameters (heavy metals, organic matter, sulphur and nitrogen) on yoxicity of sediments from the Mar Piccolo (Taranto, Ionian Sea, Italy). Microchem. J. 79, 243–248. Campisi, T., Abbondanzi, F., Casado-Martinez, C., DelValls, T.A., Guerra, R., Iacondini, A., 2005. Effect of sediment turbidity and color on light output measurement for Microtoxs Basic Solid-Phase Test. Chemosphere 60, 9–15. Chapman, P.M., Anderson, J., 2005. A decision-making framework for sediment contamination. Integr. Environ. Manage. 1 (3), 163–173. Coya, B., Maran˜o´n, E., Sastre, H., 2000. Ecotoxicity assessment of slag generated in the process of recycling lead from waste batteries. Resour. Conserv. Recycl. 29, 291–300. Coz, A., Andre´s, A., Irabien, A., 2004. Ecotoxicity assessment of stabilized/solidified foundry sludge. Environ. Sci. Technol. 38, 1897–1900. DelValls, T.A., Andres, A., Belzunce, M.J., Buceta, J.L., CasadoMartı´ nez, M.C., Castro, R., Riba, I., Viguri, J., Blasco, J., 2004. Chemical and ecotoxicological guidelines used for the management of dredged material disposal. TrAC-Trends Anal. Chem. 23, 819–828. Den Besten, P.J., De Deckere, E., Babut, M.P., Power, B., DelValls, T.A., Zago, Ch., Oen, A.M.P., Heise, S., 2003. Biological effects-based sediment quality in ecological risk assessment for European waters. J. Soils Sed. 3, 144–162. Ferna´ndez, J., Barrueso, M.L., Font, R., Sabater, M.C., 1997. Characterization of tannery wastes comparison of three leachability tests. J. Hazard. Mater. 54, 31–45. Font, R., Gomis, V., Fernandez, J., Sabater, M.C., 1998. Physicochemical characterization and leaching of tannery wastes. Waste Manage. Res. 16 (2), 139–149. Gonza´lez-Pin˜uela, C., Coz, A., Andre´s, A., Irabien, M.J., Yusta, I., Viguri, J., 2006. PAHs compositions and sources in surface-core sediments from the Santander Bay (northern Spain). Fresenius Environ. Bull. 9a, 1031–1036. Guzzella, L., 1998. Comparison of test procedures for sediment toxicity evaluation with Vibrio fischeri bacteria. Chemosphere 37, 2895–2909. Ho, K.T., Kuhn, A., Pelletier, M.C., Hendricks, T.L., Helmstetter, A., 1999. pH dependent toxicity of five metals to three marine organisms. Environ. Toxicol. 14 (2), 235–240. ICES, 2003. Report of the Working Group on Marine Sediments in Relation to Pollution, March. International Council for the Exploration of the Sea. Akvaplan-Niva, Tromsø, Norway. ISO, 1999. International Organization for Standardization. Water Quality. Determination of the Inhibitory Effect of Water Samples on the Light Emission of Vibrio fischeri (luminescent bacteria test). Part 3: Method Using Freezedried Bacteria. European Standard ISO 11348-3: 1998. European Committee for Standardization, Bruxelles. Jennings, V.L.K., Rayner-Brandes, M.H., Bird, D.J., 2001. Assessing chemical toxicity with the bioluminescent photobacterium (Vibrio fischeri): a comparison of three commercial systems. Water Res. 35 (14), 3448–3456. Kohonen, T., 2001. Self-Organizing Maps, third ed. Springer, Berlin. Lahr, J., Maas-Diepeveen, J.L., Stuijfzand, S.C., Leonards, P.E.G., Druke, J.M., Lucker, S., Espeldoorn, A., Kerkum, L.C.M., van Stee, L.L.P., Hendriks, A.J., 2003. Responses in sediment bioassays used in the Netherlands: can observed toxicity be explained by routinely monitored priority pollutants? Water Res. 37, 1691–1710. Lek, S., Gue´gan, J.F., 1999. Artificial neural networks as a tool in ecological modelling, an introduction. Ecol. Model. 120, 65–73. Mantis, I., Voutsa, D., Samara, C., 2005. Assessment of the environmental hazard from municipal and industrial wastewater treatment sludge by employing chemical and biological methods. Ecotoxicol. Environ. Saf. 62, 397–407.

ARTICLE IN PRESS 312

A. Coz et al. / Environmental Research 106 (2008) 304–312

McCauley, D.J., DeGraeve, G.M., Linton, T.K., 2000. Sediment quality guidelines and assessment: overview and research needs. Environ. Sci. Policy 3, 133–144. Mowat, F.S., Bundy, K.J., 2001. Correlation of field-measured toxicity with chemical concentration and pollutant availability. Environ. Int. 27, 479–489. Munawar, M., Munawar, J.F., Sergeant, D., Wenghofer, C., 2000. A preliminary bioassessment of Lake Baikal sediment toxicity in the vicinity of a pulp and paper mill. Aquat. Ecosyst. Health Manage. 3, 249–257. Nendza, M., 2002. Inventory of marine biotest methods for the evaluation of dredged material and sediments. Chemosphere 48, 865–883. OJEC, 2003. Council Decision of 19 December 2002 establishing criteria and procedures for the acceptance of waste at landfills pursuant to article 16 of and Annex II to Directive 1999/31/EC. Official Journal of the European Communities 16.01.2003, 2003/33/EC. Olajire, A.A., Altenburger, R., Kqster, E., Brack, W., 2005. Chemical and ecotoxicological assessment of polycyclic aromatic hydrocarbon— contaminated sediments of the Niger Delta, Southern Nigeria. Sci. Total Environ. 340, 123–136. Onorati, F., Mecozzi, M., 2004. Effects of two diluents in the Microtoxs toxicity bioassay with marine sediments. Chemosphere 54, 679–687. Riba, I., Forja, J., Go´mez-Parra, A., DelValls, T.A., 2004. Sediment quality in littoral regions of the Gulf of Ca´diz: a triad approach to address the influence of mining activities. Environ. Pollut. 132, 341–353. Ricking, M., Beckman, E., Svenson, A., 2002. PAHS and Microtox acute toxicity in contaminated sediments in Swede. J. Soils Sed. 2 (3), 129–136. Salizzato, M., Pavoni, B., Ghirardini, A.V., Ghetti, P.F., 1998. Sediment toxicity measured using Vibrio fischeri related to the concentrations of organic (PCBs, PAHs) and inorganic (metals, sulphur) pollutants. Chemosphere 36, 2949–2968. Tay, K.-L., Doe, K.G., Wade, S.J., Vaughan, D.A., Berrigan, R.E., Moore, M.J., 1992. Sediment bioassessment in Halifax harbour. Environ. Toxicol. Chem. 11 (11), 1567–1581.

Van Beelen, P., 2003. A review of the application of microbial toxicity test for deriving sediment quality guidelines. Chemosphere 53, 795–808. Van den Brink, P.J., Kater, B.J., 2006. Chemical and biological evaluation of sediments from the Wadden Sea, The Netherlands. Ecotoxicology 15, 451–460. Vesanto, J., 1999. SOM-based data visualization methods. Intell. Data Anal. 3, 111–126. Vesanto, J., Ahola, J., 1999. Hunting for correlations in data using the self-organizing map. In: Proceeding of the International ICSC Congress on Computational Intelligence Methods and Applications (CIMA’99), ICSC Academic Press, pp. 279–285. Vesanto, J., Alhoniemi, E., 2000. Clustering of the self-organizing map. IEEE Trans. Neural Netw. 11, 586–600. Vesanto J., Himberg, J., Alhoniemi, E., Parhankangas, J., 1999. Selforganizing map in Matlab: the SOM Toolbox. In: Proceedings of the Matlab DSP Conference 1999. Comsol Oy, Espoo, Finland, pp. 35–40. Vigano, L., Arillo, A., Buffagni, A., Camusso, M., Ciannarella, R., Crosa, G., Falugi, C., Galassi, S., Guzzella, L., Lopez, A., Mingazzini, M., Pagnotta, R., Patrolecco, L., Tartari, G., Valsecchi, S., 2003. Quality assessment of bed sediments of the Po River (Italy). Water Res. 37, 501–518. Viguri, J.R., Verde, J., Irabien, J.A., 2002. Environmental assessment of polycyclic aromatic hydrocarbons (PAHs) in surface sediments of the Santander Bay, Northern Spain. Chemosphere 48, 157–165. Viguri, J., Irabien, M.J., Yusta, I., Soto, J., Go´mez, J., Rodrı´ guez, P., Martı´ nez, M., Irabien, J.A., Coz, A., 2007. Physico-chemical and toxicological characterisation of the historic estuarine sediments: a multidisciplinary approach. Environ. Int. 33, 436–444. Wenning, R.J., Adams, W.J., Batley, G.E., Berry, W.J., Birge, W.J., Bridges, T.S., Burton, G.A., Chapman, P.M., Douglas, W.S., Engler, R.M., Ingersoll, C.G., Moore, D.W., Stahl, R.G., Word, J.Q., 2005. Executive summary. In: Wenning, R.J., Batley, G.E., Ingersoll, C.G., Moore, D.W. (Eds.), Use of Sediment Quality Guidelines and Related Tools for the Assessment of Contaminated Sediments. SETAC Press (Society of Environmental Toxicology and Chemistry), Pensacola, FL, USA, pp. 11–38.