Integrative ecotoxicological assessment of sediment in Portmán Bay (southeast Spain)

ARTICLE IN PRESS Ecotoxicology and Environmental Safety 72 (2009) 1832–1841

Contents lists available at ScienceDirect

Ecotoxicology and Environmental Safety journal homepage: www.elsevier.com/locate/ecoenv

Review

Integrative ecotoxicological assessment of sediment in Portma´n Bay (southeast Spain) Augusto Cesar a,b,, Arnaldo Marı´n b, Lazaro Marin-Guirao b, Rube´n Vita b, Javier Lloret b, Toma´s Angel Del Valls c a

˜o Paulo 11045-907, Brazil Departamento de Ecotoxicologia, Universidade Santa Cecı´lia, Rua Oswaldo Cruz 266, Santos, Sa Departamento de Ecologı´a e Hidrologı´a, Facultad de Biologı´a, Universidad de Murcia, 30100 Murcia, Spain c ´tedra UNESCO/UNITWIN/WiCop, Departamento de Quı´mica Fı´sica, Facultad de Ciencias del Mar y Ambientales, Universidad de Ca ´diz, CP 11510 Puerto Real, Ca ´diz, Spain Ca b

a r t i c l e in f o

a b s t r a c t

Article history: Received 5 December 2007 Received in revised form 25 November 2008 Accepted 2 December 2008 Available online 16 July 2009

Portma´n Bay, southeast Spain, contains the most seriously metal-contaminated sediments of the Mediterranean Sea. From 1958 to 1991, approximately 50 million tons of mine tailings were dumped into the bay, completely filling up the bay and dispersing over an extensive area of the continental platform and continental slope. The objective of our study was to characterize the nature and extent of metal contamination and the responses of natural communities to it and to assess the toxicity of the sediment deposits 10 years after mining had ceased. We studied the physical and chemical characteristics of the sediments and toxicity (of the porewater and sediment–water interface) using two sea urchin species (Arbacia lixula and Paracentrotus lividus). Metal bioavailability and patterns of macroinvertebrate community composition along the contamination gradient were also studied. Univariate and multivariate analyses showed positive correlation between the sediment metal concentrations associated to the all biological effects (sea urchins toxicity tests and benthic indices). The effects of sediment contamination on the benthic community structure are visible among sampling stations. & 2008 Elsevier Inc. All rights reserved.

Keywords: Metal contamination Sediment toxicity tests Benthic index Weight of evidence Integrative assessment

Contents 1. 2.

3.

4. 5.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1833 Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1833 2.1. Sample collection and field measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1833 2.2. Sediment chemical and physical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1834 2.3. Toxicity testing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1835 2.4. Benthic community analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1835 2.5. Statistical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1835 2.6. Multivariate analysis approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1835 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1835 3.1. Sediment chemical and physical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1835 3.2. Toxicity testing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1836 3.3. Benthic community analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1837 3.4. Multivariate approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1837 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1839 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1840 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1840 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1841

 Corresponding author at: Departamento de Ecotoxicologia, Universidade Santa Cecı´lia, Rua Oswaldo Cruz 266, Santos, Sa˜o Paulo 11045-907, Brazil. Fax: +55 13 32345297. E-mail address: [email protected] (A. Cesar).

0147-6513/$ - see front matter & 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.ecoenv.2008.12.001

ARTICLE IN PRESS A. Cesar et al. / Ecotoxicology and Environmental Safety 72 (2009) 1832–1841

1. Introduction The mining of metals in the area of Portma´n (Murcia, southeast Spain) has a long history. The name Portma´n is derived from the Latin ‘‘Portus Magnus’’, because it was a natural harbor from which lead was embarked for use throughout the Roman Empire. The surrounding mountains, which are rich in metals, contain numerous old Roman lead mines. The bay itself is a metalpolluted area, where benthic communities have experienced centuries of impairment from the drainage and sedimentation associated with mining activities and their abandonment. From ˜ arroya mine pumped 3–10,000 ton of tailings 1958 to 1991, the Pen per day, first directly into the bay and later, when the bay was filled up through an emissary of more than 2 km length. In total, approximately 50 million tons of mine tailing were dumped into Portma´n Bay during this period. In the extraction process 2 m3 of water were used per ton of mineral and a ton of sodium cyanide, some 10 ton of sulfuric acid, and also, copper sulfate were used each day. All this material was poured into the sea including the remains of metals known to be toxic, such as cadmium, copper, lead, and zinc. After filling up the bay, the mining wastes dispersed over an extensive area of the continental platform and the active disposal area extended beyond the continental shelf through a submarine canyon. The spatial distribution of metal contamination (Cd, Pb, and Zn) in the water column and sediments was characterized during the 1980s (Rey and Del Rı´o, 1983; Pe´rez and Puente, 1989; De Leon et al., 1984). However, the toxicity of these sediments and the interactive effects on benthic communities have not been addressed by previous studies. We have selected two different toxicity tests, the sediment– water interface and the porewater toxicity tests, since they utilize different matrices and therefore present different ecological significance due to the different route of exposure that organisms are exposed to contaminants. Echinoderm embryo-larval development tests have been widely used to characterize a variety of toxicants, including liquid and solid phase protocols (Hunt et al., 2001). The assessment of sediment quality generally involves an evaluation of solid phase sediments, although porewater is also important, because it represents a major route of exposure to benthic organisms and substantially influences the bioavailability of contaminants (Whiteman et al., 1996; Long et al., 2003).

1833

We evaluated the structure of communities with univariate measures (Shannon–Wiener diversity, Margalef’s richness, Pielou’s evenness, RBI and EBI indices) and multivariate analyses (Multidimensional Scaling). The multi-metric RBI (relative benthic index) and EBI (exploratory benthic index) use a simple scoring system for benthic community metrics to assess benthic community health and to infer environmental quality of benthic habitats in Portma´n in a state previous to the project of restoration of the Bay. The EBI index was adapted in this study to evaluate ecological degradation of benthic communities and to identify concentrations of chemicals that are associated with biological impacts through multivariate analysis. In the present work we studied the physical and chemical characteristics of the sediments, the effect of the porewater and sediment–water interface on marine invertebrates, metal bioavailability, and patterns of macroinvertebrate community composition along the contamination gradient to evaluate the status and trends of environmental conditions in Portma´n Bay ecosystems. The goals of this study were: (a) to determine concentration of metals in sediment and relationships between contamination and biological effects; (b) to adapt and develop benthic index to evaluate effects of metal contamination; (c) to integrate chemical, toxicological, and ecological data to asses the sediment quality, aiming to categorize sampling stations for future investigation and management. The use of weigh-of-evidence (WOE) approach, improves the characterization of ‘gray areas’ of pollution and helps in the determination of the bioavailability of metals, besides being of great importance and usefulness of the integrative studies of the sediment quality assessment, before and after the application of management strategies.

2. Materials and methods 2.1. Sample collection and field measurements Samples were collected synoptically along a spatial gradient at the same depth (10–20 m) in March 2002. The spatial sampling design followed previous studies (September and December 1999, October 2000) (Cesar, et al., 2004). Sampling stations were selected at regular distances (8 km approximately) between the old mine discharge and Cabo de Palos (West–East; Fig. 1), while the reference station

N

3 Km Fig. 1. (A) Map of the study area and (B) location of sediment sampling stations, IF—Isla del Fraile; PG—Punta Galera (old emission point); PN—Punta Negra; CN—Cabo Negrete; CM—Canto de la Manceba; PL—Punta de la Loma Larga, and PE—Punta Espada.

ARTICLE IN PRESS 1834

A. Cesar et al. / Ecotoxicology and Environmental Safety 72 (2009) 1832–1841

was located on Fraile Island, approximately 60 km to the south of the old emission point, but is also affected by historical mining activity. The control station was select in the opposite extreme of the spatial gradient near the Cabo Palos (Punta Espada) at 20 km from the old emission point. Replicate samples (n ¼ 4) were collected from all points (n ¼ 7) along the contamination gradient on a spatial scale (kilometers) considered appropriate for examining differences along the gradient. SCUBA divers collected, capped, and sealed intact sediment cores carefully underwater and retained in the polyethylene tubes (10:15 cm diameter/height) throughout storage (4 1C in the dark). Sediment samples were divided into subsamples for the chemical analyses and toxicity testing to maximize the potential for data integration. Only the top 5 cm of the superficial sediment was used for subsamples. Sediments were stored for no longer than 7 days, prior to toxicity testing. Approximately 100 ml of porewater was extracted from each liter of sediment sampled by centrifugation (2500g) for 10 min at 4 1C. The supernatant was decanted and the process was repeated to remove any remaining particles. We kept porewater extracts for no longer than 24 h prior to toxicity testing. The control and dilution water used in the experiments consisted of natural seawater collected in unpolluted areas (where the sea urchins were also collected) and filtered through a GFC Watmans filter. Laboratory subsampling took place under strictly anaerobic conditions for acid-volatile sulfide and simultaneously extract metals (SEM-AVS), and were stored frozen (20 1C) to prevent sulfide oxidation. Four replicate samples were collected by SCUBA divers from each sampling station for benthic analysis using a 0.04 m2 metal hand grab and sieved through a 500 mm mesh. The macroinvertebrates retained on the screen were fixed with 4% buffered formalin, and later washed and transferred to 70% isopropyl alcohol prior to sorting and identifying the macrofauna. The individual taxa of each sieved sample was identified and enumerated in the laboratory by stereoscope microscopy to assess species richness and abundance. All the organisms were sorted and identified to the lowest possible taxon level and their abundance was counted. Field measurements (station coordinates and depth) were made and sediment–water interface variables (temperature, salinity, OD, pH, Eh) were measured to compare with the limits of tolerance of the species test at the time of

collection in all the sampling points (Table 1) using a field Multiline F/SET-3 (WTW-Germany) equipped with a combination of conductivity, temperature, pH, and oxygen electrodes.

2.2. Sediment chemical and physical analysis Sediment–water content was measured as a percentage of wet weight lost by drying until constant weight at 60 1C for 24 h. The dried sediments were finely ground and carefully sieved in stainless steel mesh and grain size was determined by standard mechanical dry sieve-shaker techniques to determine the sand, silt, and clay fractions (Buchanan, 1984). The total organic carbon (TOC) of each sample was measured in the fine sediment fraction (silt and clay). The TOC content was determined with Carlo Erba Instrument (EA1108), an elemental analyzer, following sample preparation with 1 N HCL to decompose the carbonate (Verdardo, et al., 1990). The percentage of organic matter (LOI) in samples was estimated by the loss of weight on ignition at 450 1C for 6 h in dried whole sediment from which the carbonates had previously been removed by acid treatment (Buchanan, 1984). The concentration of ammonia (NH3) was determined from the total ammonium (NH4) concentration, taking into account pH, temperature, and salinity of each sample (Whitfield, 1974). Sediment samples for the acid-volatile sulfide (AVS) and simultaneously extracted metals (SEM) were analyzed by a cold-acid purge-and-trap technique described in detail by Allen et al. (1993). The hydrogen sulfide was determined with an ion-selective silver/sulfide electrode (Thermo Orion, model 9616). The sulfide ion concentration in the trap solutions was measured with a combined sure-flow silver/sulfide ion-selective electrode (ISE-Orion model 9616), which offers the additional benefit of not requiring a separate reference electrode. Following digestion, simultaneously extracted metals (aluminum, arsenic, cadmium, copper, iron, mercury, nickel, lead, and zinc) were collected by filtration of the acid-sediment slurry and measured with an optical emission spectrometer (Optima 2000 DV—Perquin Elmer). All the analytical procedures were checked with reference materials (Marine Sediment References Material for Trace Metals—1, National Research Council

Table 1 Station location, depth and sediment–water interface field measurements, means7standard errors. Location

Reference

T (1C)

Salinity

OD (mg/l)

pH

Eh (mV)

Fraile Island 371240 655 N 11320 861 W 15.6 depth

IF

14.3570.06

37.7070.24

2.0270.36

7.8370.05

46.5071.55

Punta Espada 371360 417 N 01420 823 W 13.8 depth

PE

15.2570.05

36.7070.17

2.5870.02

7.5070.04

31.7572.14

P. Loma Larga 371350161 N 01470 165 W 12.8 depth

PL

14.4870.06

38.5370.06

1.9270.25

7.6070.04

33.2571.89

C. Manceba 371350 052 N 01480 376 W 15.6 depth

CM

14.7570.03

37.0070.14

2.4170.09

7.5470.01

34.5070.51

Cabo Negrete 371340 327 N 01490 326 W 12.7 depth

CN

14.5570.03

38.4070.14

2.2770.07

7.9470.03

53.0072.58

Punta Negra 371340 052 N 01500 496 W 12.8 depth

PN

14.6070.04

36.7070.04

1.7770.17

7.5670.01

33.2571.49

Punta Galera 371340 052 N 01510712 W 16.8 depth

PG

13.6370.02

37.0070.11

2.2070.17

7.5370.03

33.0071.35

ARTICLE IN PRESS A. Cesar et al. / Ecotoxicology and Environmental Safety 72 (2009) 1832–1841 (NRC), Certified Reference Material, 277 BCR, and Council National of Researches Canada, 277 BCR, for heavy metals) and allow agreement with certified values higher than 90%.

2.3. Toxicity testing Sediment toxicity tests were performed to evaluate whether metals were bioavailable to standard test organisms. The toxicity of the sediment porewater (PW) and sediment–water interface (SWI) was determined using embryo-larval development tests with two sea urchin species, Arbacia lixula and Paracentrotus lividus, following the procedures previously described (Cesar et al., 2002, 2004) and according to the accepted guidelines (Environment Canada, 1992; USEPA, 1995; CETESB, 1999; CEDEX-Spain, 2001; ABNT, 2006). For SWI system, 2 ml of the surface of an intact sediment core were introduced carefully through a syringe (5 ml) with a cut tip, and 8 ml of dilution seawater (1 sediment/4 water) was introduced carefully to minimize resuspension. New sterilized syringes were used for each sample and a circular mesh (100 mm) was placed and carefully fixed by a plastic ring to avoid displacement on the sediment–water interface and test tubes were allowed to stabilize for 24 h. The duration of subchronic tests was 28 h for P. lividus and 38 h for A. lixula, counting the number of normally developed pluteus embryos at the end of the test. After this period the number of normally developed pluteus larvae was counted and the percentage of abnormalities was determined by direct observation of 100 randomly selected individuals per vial under an inverted microscope. The sea urchins used in this study were obtained by SCUBA divers in the Fraile Island (IF).

1835

2.5. Statistical analysis Toxicity data were checked for normality and homoscedasticity assumptions with Shapiro–Wilk’s and Bartlett’s tests, respectively. Larval development data were arcsine square root transformed prior to statistical analysis. Differences were evaluated with a parametric analysis of variance (ANOVA), followed by Tukey’s test. These analysis were carried out with the statistical package Toxstats V.3.5. The Newman–Keuls test was also applied to compare the means of normally developed larvae obtained in the sea urchin toxicity tests. Univariate measures included the Shannon–Wiener diversity indices calculated using natural logarithms (H0 ), species richness (Margalef’s d), evenness (Pielou’s J), total abundance (A), and abundance of taxa (S). The significance of differences between points was tested using one-way ANOVA. Community structure (presence or absence of pollution-sensitive and pollution-indicative species) was examined by multidimensional scaling (MDS), using the PRIMER-E (Plymouth Routines in Multivariate Ecological Research, v6) (Clark and Warwick, 2001; Clark and Gorley, 2006) suite of computer programs developed at the Plymouth Marine Laboratory, UK. Ranked lower triangular similarity matrices were constructed using a range of data transformations, the Bray–Curtis similarity measure and group-average sorting. Abundance data were fourth-root transformed in order to reduce contributions to similarity by abundant species, and thereby increasing the importance of the less-abundant species in the analysis (Clark and Green, 1988). The species contributing to dissimilarities between stations were investigated using the similarities percentages procedure (SIMPER) (Clark and Ainsworth, 1993; Somerfield, et al., 1994).

2.4. Benthic community analysis

2.6. Multivariate analysis approach

The macroinvertebrate taxonomic data was quantified using the relative benthic index adapted to Mediterranean fauna, developed by Anderson et al. (1998, 2001) and the exploratory benthic index, a new index calculated for this study which integrated different ecological community parameters, but is based on the same methodology. The original RBI was based on six categories, including total number of species (1/6), number of crustacean species (1/6), number of mollusk species (1/6), number of crustacean individuals (1/6), and the presence or absence of species indicative of sediment quality and metal pollution (2/6). The EBI was based on eight categories, including the total number of species (1/8), the number of crustacean species (1/8), number of mollusk species (1/8), number of polychaetes families (1/8), total number of individuals (1/8), and presence or absence of pollution-sensitive and pollution-indicative species (2/8), and integrating the three ecological indices of diversity, Shannon–Wiener, Pielou, and Margalef (1/8). The pollution-sensitive and pollution-indicative species were extracted from a global analysis of all samples collected in the area of study through a previous SIMPER analysis. Pollution-sensitive species were found in control stations where anthropogenic and other severe disturbances do not play a major role in structuring communities, while the pollution-indicative species are common in stressed stations and are not found in unpolluted points (Hunt et al., 2001). Each parameter value (one sixth in the RBI and one eighth in the EBI of the total indices) for each sample was the percentile at which data from that sample fit into the total range for that parameter over all samples from the Portma´n Bay data set. For the two sixths (RBI) and two eighths (EBI) of the indices represented by positive and negative indicator species, the parameter value was weighted toward presence or absence of key indicator species, with abundance given additional incremental weight by transforming the abundance of each indicator species to its double square root to compress the range of values. For each sample, the transformed abundances of the negative indicator species were summed and subtracted from the sum of the transformed abundances of the positive indicator species, and this value was converted to a percentile of the total range for all sites. The overall indices for each site was calculated by adding the values for the six (RBI) and eight (EBI) parameters together and standardizing each sum to the total range of the sums for all stations, resulting in a range of values from 0.00 (most impacted) to 1.00 (least impacted). The threshold value for a degraded benthic community was set at 0.30 since 0.00–0.30 was considered indicative of a degraded benthic community, 0.31–0.60 was considered transitional, and 0.61–1.00 was considered undegraded. These indices are based on toxicology and natural history, taking into account the responses of marine benthic communities to anthropogenic and natural disturbances, and were developed for particular areas by selecting different indicator species (Anderson et al., 1998). The selection of indicator species must be based on known responses to anthropogenic and other disturbances and related natural history, such as life history traits and abundance patterns along environmental gradients and between study stations (Anderson et al., 1998). Accordingly, the selection of indicator species along an environmental gradient or between stations can bias the results obtained with the RBI and EBI, since species are selected in relation to their presence–abundance in both extremes of the gradient or in polluted–undisturbed stations. In this sense, before selecting a species as being positive/sensitive or negative/tolerant, we must be sure that this same species has previously been cited as indicator in the same type of pollution.

The relationship amongst variables was assessed by using a multivariate analysis approach by means of a factor analysis. Principal component analysis (PCA) was used as an extraction procedure. It was based on the geochemical characteristics of the sediments (TOC, %OM, %fines, Al, As, Fe, Hg, Ni, Pb, Zn, and SEM/AVS), results of toxicity bioassays (abnormal development of sea urchin exposed to sediment PW and SWI), and the benthic indices (RBI and EBI). The concentrations of Cd and Cu were not included in the PCA, because the values were shown low and in most of the cases under the detection limit. The analysis was conducted on the matrix (varimax normalized rotation) and included any principal component axis that accounted for more than 10% of the total variance. A component loading cutoff of 0.40 was used in selecting variables for inclusion in factors. Tabachnic and Fidell (1996) suggested that a cutoff of at least 0.32 be used and that component loading of greater than 0.45 be considered fair or better. The variables were autoscaled (standardized) so as to be treated with equal importance. To confirm these relationships between chemical contamination and biological effects, the Spearman rank correlation coefficients (rho) were calculated. PCA and correlation analysis were carried out by means of the statistical packages STATISTICA software tool (Stat Soft, Inc., 2001, version 6).

3. Results 3.1. Sediment chemical and physical analysis Total organic carbon and organic matter decreased along a contamination gradient (Table 2). Most of the gradient samples did not exceed the acid-volatile sulfide values measured in control and reference stations (PE and IF). The metal concentrations in sediments showed a strong gradient despite the cessation of mining activity in Portma´n Bay approximately 20 years ago. Sediment concentrations of metals, including Zn, Al, Pb, and Fe, were low in the control and reference stations and progressively increased towards the emission point (PG), where the highest levels were recorded. These metals were found in high concentrations near the emission point and were statistically associated with the toxicity of sea urchin larvae and benthic community structures (Tables 2 and 4). Sediment metal concentrations off the Portma´n coast, PN and PG stations located on both sides of Portma´n Bay presented the highest metal concentrations, decreasing as the distance from the bay increased (PG4PN4CM4CN4PL4PE). Only IF and PL stations presented a negative molar difference between the SEM and the AVS analyzed, the rest of the stations showed a positive difference. In PN and PG

1836 Table 2 Physico-chemical measurements of grain size, un-ionized ammonia, total organic carbon (TOC), organic matter (LOI), simultaneously extracted metals (SEM), concentration of acid-volatile sulfides (AVS), total SEM (Cd/Cu/Ni/Pb/ Zn), and AVS molar ratio (AVS-SEM) in sediments of all samples, means7sd. Sampling points

PWa

SWIb

0.00570 0.00470 0.00670 0.00570 0.00770 0.00670 0.00470.1

0.00270 0.00370 0.00370 0.00470 0.00470 0.00470 0.00270

TOC (%)

LOI (%)

SEMc

Metals (mg/kg dry sed.)

Al

As

1.1770.14 0.1270.01 1.9170.14 3.1870.04 0.2470.01 0.4170.11 0.1270.01 0.7370.04 0.5670.03 0.0670.01 0.7270.07 0.2970.01 0.7470.01 1.4770.01 0.2370.01 2.9871.05 1.4870.08 2.2870.05 3.8870.29 0.2870.01 0.2370.17 0.5870.01 1.3870.08 1.3470.06 0.1970.01 4.3272.74 16.6770.27 5.1670.41 13.6270.32 0.5470.01 4.1970.11 18.3271.66 6.2270.25 11.8771.30 0.2770.06

Cd

Cu

Fe

od.l. od.l. od.l. od.l. od.l. od.l. od.l.

od.l. 18.1370.91 od.l. 7.8970.23 od.l. 13.7970.53 od.l. 49.9373.86 od.l. 16.2070.92 od.l. 197.6373.93 0.0170 165.95722.17

Hg

Ni

Pb

od.l. od.l. od.l. od.l. od.l. od.l. 0.0170

od.l. od.l. od.l. 0.0270.01 0.0570.01 0.0670.01 od.l.

0.0370 0.0170 0.0470 0.0470 0.0870 0.1270 0.0870 0.2170 0.2970.01 0.3070.01 1.1770.09 1.4970.09 0.2070.01 0.3470.01 0.5970.03 3.1470.10 13.4770.22 16.6770.53 2.9870.15 15.3471.53 18.3371.66

AVS (mmol/g dry sed)

SEM-AVSd

0.3270.01 0.0170 0.2970.07 0.3170.03 0.2570.04 0.2870.04 0.3270.16

0.2870.01 0.1170.01 0.00570.06 1.1870.22 0.3470.03 16.3970.53 18.0173.02

Zn

PE

PE CM

PN PG

A. lixula

PG

P. lividus

PN

A. lixula

CN

P. lividus

PL CM CN Sampling points

PL

stations this positive difference was high (415 mmol/g dry sed.), which indicates the possible bioavailability of metals. The field measurements of the sediment–water interface variables (temperature, salinity, OD, pH, Eh) were measured to compare with the limits of tolerance of the species test at the time of collection (Table 1). These variables were found inside the limits of tolerance for species used in the toxicity tests and differences were not detected among the results obtained in field and in laboratory, except for OD that was inferior to the field.

3.2. Toxicity testing

IF

IF

The percentage of normally developed larvae in PW and the SWI tests were reported in Fig. 2. The results indicated adverse effects in SWI and PW tests in stations PG, PN, CN, CM, and PL, which showed statistical differences from the control station (po0.001, one-way ANOVA, post hoc Tukey test). However, station PE was not significantly different from IF and can also be

100

80

60

40

20

0

100

80

60

40

20

0 Sampling Points

Fig. 2. Comparison of mean percentage of normally developed larvae (7sd) of A. lixula and P. lividus at the different sampling points: (A) porewater tests and (B) sediment–water interface tests.

Normally Developed Larvae (%)

Normally Developed Larvae (%)

ARTICLE IN PRESS

od.l.—below detection limits. a PW—porewater. b SWI—sediment–water interface. c SEM—Cd/Cu/Ni/Pb/Zn (mmol/g dry sed.). d SEM-AVS—Cd/Cu/Ni/Pb/Zn (mmol/g dry sed.).

A. Cesar et al. / Ecotoxicology and Environmental Safety 72 (2009) 1832–1841

IF PE PL CM CN PN PG

Fines (%)

Ammonium NH3 (mg/l)

ARTICLE IN PRESS

The RBI and EBI threshold value was at 0.30 since 0.00–0.30 was considered indicative of a degraded benthic community, 0.31–0.60 was considered transitional, and 0.61–1.00 was considered undegraded. a Relative benthic index by Hunt et al. (2001), six parameters (RBI1) and exploratory benthic index, ten parameters (EBI2).

b

43.0079.46 11.2570.75 98.25725.16 26.2572.84 2.9570.07 6.0070.25 0.8970.01 32.2575.12 9.0070.81 77.5075.17 25.7571.18 2.9370.07 5.8870.24 0.8970.01 59.7579.60 9.0070.71 88.75720.24 18.7572.25 2.4370.12 4.2570.43 0.8270.01 115.50725.51 8.7570.48 128.00727.11 18.2571.11 2.1370.11 3.8670.18 0.7270.04 9.2571.97 4.0070.71 17.2572.87 8.5071.04 1.9570.07 2.8770.19 0.8870.02 59.00711.67 10.0071.08 94.75713.54 15.5071.50 2.0470.13 3.5170.41 0.7370.03 7.0070.71 0.8070.11 1.3870.16 0.3970.04 24.2576.70 2.0070.41 117.75715.99 1.7570.25 1.5070.64 1.0070.71 1.2570.63 0.2570.25 0.7570.48 2.0070.57 7.2571.43 40.00712.35 12.2571.18 11.5071.19 43.5074.97 14.7571.60 6.2571.03 28.25711.73 9.0071.15 5.5070.50 11.7571.65 9.0071.29 1.7570.48 6.0071.92 3.7570.48 3.5070.65 35.50711.56 6.0071.08 1.2570.63 91.50716.07 3.2570.63 IF PE PL CM CN PN PG

25.50711.02 37.2574.60 21.75710.48 7.7571.03 2.2570.95 7.5071.66 1.2570.63

4.5070.96 1.5070.64 1.2570.94 1.2570.63 0.5070.50 2.0071.22 2.5070.96

(J0 ) d H0 (loge) Indiv Indiv

Spp Spp Indiv

Indiv

Spp

Indiv

Spp

Total Polychaeta Mollusca All Crustacea Sampling points Amphipods

Table 3 Individual and species abundances of infaunal benthic macroinvertebrates and values of the relative benthic indices (RBI1 and EBI2)a, means7sd.

Spp

S. Wiener

Margalef’s

Pielou’s

RBI1

EBI2

0.6470.07 0.7070.11 0.6570.06 0.6770.06 0.4670.05 0.5470.11 0.4270.04 0.5470.04 0.2670.03 0.3070.05 0.3170.03 0.4270.09 0.3170.01 0.2770.06

A. Cesar et al. / Ecotoxicology and Environmental Safety 72 (2009) 1832–1841

1837

considered as a reference station (p40.05, one-way ANOVA, post hoc Tukey test). Toxicity tests with PW and SWI presented a similar pattern of response although the percentage of normally developed larvae were lower in the first one. Results of the analysis of variance and post hoc tests pointed to significant differences in the percentage of normally developed larvae among the samples in both PW and SWI tests (ANOVA, post hoc Tukey test, po0.001, Table 5). For both treatments, the concentrations of un-ionized ammonium (NH3) were low. Evidently, the highest concentration of NH3 was found in PW, but below the effect threshold for the used species (Table 2). 3.3. Benthic community analysis Univariate metrics of the community structure for all the sampling stations are shown in Table 3, where they point to a significant variation along the metal contamination gradient. Stations PG, PN, and CN (with a degraded benthos) had fewer species, lower abundance, lower diversity, and lower benthic index scores than stations with a healthy benthos (PE, IF). Results of the analysis of variance and post hoc tests pointed to significant differences among sampling stations for both benthic indices (RBI and EBI) (ANOVA, post hoc Tukey test, po0.001, Table 5). A similar pattern of disturbance was indicated by the RBI and EBI values. In both indices, all the samples from PG, PN, and CN showed signs of disturbed assemblages and presented the lowest values. However, a more effective assessment of the gradual changes in the benthic community structure along the Portma´n gradient was obtained with the EBI values than with the original RBI (Table 3). In addition, the EBI index explained better the ordination of the sampling stations established by multidimensional scaling (MDS) (Fig. 4). The benthic community of Portma´n Bay (PG and PN) was characterized by the general reduction in invertebrate species and the presence of relatively high numbers of a few species including the crustacean (Apseudes latreilli), polychaetes (family Sabellidae), and the bivalves (Chamelea gallina and Macoma cumana). CM and PL stations were separated from the metal-contaminated stations by the presence of a relatively high abundance of polychaetes (Cirratullidae, Sabellidae, Spionidae, Syllidae, Orbinidae, and Capitellidae). Dissimilarities between PE and IF resulted from changes in the relative abundance of crustaceans (Bathyporeia guilliamsoniana, Corophium acutum, Phascolion strombus, Phoxocephalus aquosus, and Anapagurus leavis). The species that contributed most to the dissimilarity between control stations and metal-contaminated sampling stations were classified as pollution-sensitive species; these were Phoxocephalus aquosus, Urothoe grimaldii (Amphipoda) and Branchiostoma lanceolatum (Branchiostomida). The pollution-indicative species were Apseudes latreilli, Chamelea gallina and Macoma cumana, which contributed to the dissimilarity between the metalcontaminated stations and the control (Tables 4 and 5). 3.4. Multivariate approach The factor analysis reorganized the data of the original data set in two principal factors, which together explained 88.15% of the total variance in the original data set. The loadings of the variables and percentage of total variance for these two factors are represented in Table 6. The predominant factor (F1) accounted for 70.00% of the total variance and combines the chemical concentrations of metals (Al, As, Fe, Hg, Pb, Zn, and SEM-AVS), fines, TOC, and organic matter associated to the all biological effects (sea urchins toxicity tests and benthic indices). Second

ARTICLE IN PRESS 1838

A. Cesar et al. / Ecotoxicology and Environmental Safety 72 (2009) 1832–1841

Table 4 Spearman rank correlation coefficients for selected chemicals significantly correlated with sea urchin (Arbacia lixula and Paracentrotus lividus), normal development (PW n ¼ 28/SWI n ¼ 28)a, RBI (n ¼ 28), and EBI (n ¼ 28). Chemicals identified by principal components analysis (PCA) as covarying with inhibited sea urchin development or degraded benthos are denoted by ‘‘S’’ ( ¼ significant-component loading X0.40)b. Chemical

Arbacia lixula PW

Al Fe Pb Ni Zn NH3 Fines LOI TOC AVS SEMc SEM-AVS a b c

Paracentrotus lividus SWI

PW

SWI

Spearm

PCA

Spearm

PCA

Spearm

PCA

Spearm

PCA

0.478** 0.774*** 0.782*** 0.561** 0.817*** 0.318 0.287 0.458* 0.254 0.114 0.797*** 0.650***

S S S S S

0.686*** 0.812*** 0.905*** 0.510** 0.920*** 0.208 0.444* 0.725*** 0.287 0.175 0.904*** 0.700***

S S S S S

0.552** 0.781*** 0.844*** 0.587*** 0.880*** 0.307 0.276 0.587*** 0.272 0.082 0.854*** 0.718***

S S S S S

0.670*** 0.752*** 0.911*** 0.387* 0.910*** 0.176 0.480** 0.751*** 0.320 0.200 0.910*** 0.689***

S S S S S

S

S S

S S

S S

S

S S

S S

S S

Relative benthic index (RBI)

Exploratory benthic index (EBI)

Spearm

PCA

Spearm

PCA

0.688*** 0.720*** 0.793*** 0.277 0.795*** 0.180 0.550** 0.706*** 0.159 0.237 0.758*** 0.591***

S S S

0.521** 0.663*** 0.612*** 0.504** 0.617*** 0.118 0.429** 0.597*** 0.290 0.001 0.619*** 0.617***

S S S S S

S S S

S S

S S

S S

PW ¼ pore water; SWI ¼ sediment–water interface. Indicates significance at pp0.05*; indicates significance at pp0.01**; indicates significance at pp0.001***. SEM (Cd/Cu/Ni/Pb/Zn).

Table 5 Analyses of variances and post hoc test for toxicity tests and relative benthic index.

3

Variables

Univariant measures

F

p-level

Post hoc test Tukey HSD

2

Benthic Index Toxicity Tests

RBI EBI A. lixula (PW) A. lixula (SWI) P. lividus (PW) P. lividus (SWI)

12.5583 17.0134 676.451 410.292 325.814 320.624

*** *** *** *** *** ***

IF IF IF IF IF IF

PL PL PL PL PL PL

CM CM CM CM CM CM

CN CN CN CN CN CN

PN PN PN PN PN PN

PG PG PG PG PG PG

***po0,001.

Table 6 Sorted rotated factor loadings (pattern) of the original 17 variables and percentage of total variance for two principal factors. Variable

A. lixula (PW) A. lixula (SWI) P. lividus (PW) P. lividus (PW) RBI EBI Fines (%) TOC LOI Al As Fe Hg Ni Pb Zn SEM-AVS Variance (%)

Factor loadings principal components marked loadings are 40.4 F1

F2

0.356310 0.611704 0.366612 0.718389 0.321346 0.581360 0.689289 0.653864 0.919018 0.932088 0.862845 0.432841 0.848928 0.766481 0.080625 0.881927 0.926777 70.00299

0.891831 0.784946 0.893788 0.659636 0.910823 0.681004 0.694675 0.716911 0.342341 0.307204 0.368814 0.580471 0.419319 0.076453 0.870510 0.392856 0.328999 18.14706

Only loadings greater than 0.40 are shown in bold format.

factor (F2) represented 18.14% of the variance and combines the chemical concentrations of metals (As and Ni) and fines correlated to the all biological effects. The representation of estimated factor

1 Factor scores

PE PE PE PE PE PE

F2 F1

0 -1 -2 -3 IF

PE

PL CM CN Sampling points

PN

PG

Fig. 3. Estimated factor scores from each of five sampling stations to the centroid of cases for the original data. The factor scores quantify to the prevalence of every component for each station and are used to confirm the factor description.

scores from each station is represented in Fig. 3. Factor 1 scores were negative for stations IF, PE, PL, CM, and CN. On the other hand, the positive scores of factor 2 that were measured at stations CM, CN, and PN, confirmed that the two factors were related to the association of biological effect with metals concentrations. These factors indicate environmental degradation caused by the related metals, since toxicity (sea urchins tests) is correlated in both factors, in addition to in situ alterations (RBI and EBI) (Fig. 4). Spearman rank correlations between SWI and PW toxicity tests and bulk chemical concentrations were examined in the two sea urchin species. In the PW toxicity tests with A. lixula and P. lividus, the analysis showed highly significant positive correlations between abnormally developed larvae and zinc, lead, iron, and SEM-AVS concentrations (Table 4). For its part, Spearman rank correlations indicated significant negative correlations between the benthic community structure (as represented by the RBI and EBI indices) and several metals (Zn, Al, Pb, Fe). In addition, there

ARTICLE IN PRESS A. Cesar et al. / Ecotoxicology and Environmental Safety 72 (2009) 1832–1841 Stress: 0.12

IF-2

IF-1

PE-1 PE-2 PE-4 PE-3 PLL-4

IF-4 CN-1

IF-3

PLL-3CM-1 PLL-1 CM-3 PLL-2 CM-2 CM-4

PN-3

CN-4 CN-2

PN-1 CN-3

PN-4

PN-2

PG-4 PG-1

PG-3 PG-2

Fig. 4. Multidimensional scaling ordinations for fourth root transformed total fauna abundance (stress ¼ 0.12) in sampling stations.

were significant negative correlations between RBI and EBI values, and organic matter and AVS-SEM (Table 4).

4. Discussion Historically, the Sierra of Cartagena-Portma´n (Murcia, southeast Spain) was exploited to extract pyrite and lead sulfide. During the 20th century, the mineral laundries used floating techniques to extract metal, producing great quantities of mining wastes. These muddy wastes were discharged into the bay, producing a high degree of sediment metal contamination. In total, approximately 50 million tons of mine tailings were dumped into the ocean, including metals known to be toxic, such as cadmium, copper, lead, and zinc (Marin-Guirao et al., 2005). The bay has received mining effluents during three decades which resulted in the fulfilling of the bay with mining wastes and therefore the loss of its natural coastal line and conditions. Since all mining activities ceased in the beginning of the 1990s and the surrounding areas possess high ecological values, there exits a great interest in restoring the bay during the last few years. The restoration plan has recently been approved and the works will start at the end of 2007. As a first step in the restoration program it is necessary to assess the contamination of the bay and for such purposes it has been argued that the best way must be a weightof-evidence approach (Burton et al., 2002; Chapman et al., 2002; Riba et al., 2004a, b; Cesar et al., 2007) where complementary environmental tools must be integrated in a correct manner. The development of this approach may serve also for environmental monitoring during the different restoration stages. The chemical analysis of the contaminants present in a sediments sample is one of the first approaches employed historically for pollution assessment. In this sense, our results indicate that sediments from the Portma´n Bay are heavily polluted by metals; metal concentrations decrease as the distance to the bay increases. According to the classification proposed by Long et al. (1995), PG and PN stations were highly polluted by Zn and Pb, whereas the rest of the stations were classified as less polluted by the four metals, except CM station, which was classified as moderately polluted by Pb. There is a considerable uncertainty regarding the concentration of metals that may pose significant ecological risks due to metal bioavailability as it is determined by the concentration of

1839

acid-volatile substances formed in anoxic conditions (Ankley, 1996). In marine sediments, sulfides can be responsible for metal bioavailability in pore and overlying water (USEPA, 1995) and affect the distribution of benthic invertebrates. The determination of acid-volatile sulfides (AVS) is widely used as a measure to reduce sulfur species in sediments. Di Toro et al. (1992) proposed that if the SEM/AVS ratio is less than 1, there will be no toxic effect from Cd, Cu, Hg, Ni, Pb, or Zn. Short-term sediment bioassays showed that the molar ratio of AVS/SEM determines the activities of at least some metals in porewater (Di Toro et al., 1990; Ankley et al., 1991). Metal activities were reduced to very low levels at ratios o1, because of the high stability of the metal sulfide. Toxicity and bioaccumulation were correlated with porewater metal activities in the toxicity tests (Luoma and Fisher, 1997). Sediment toxicity tests are widely used and accepted environmental tools to assess the toxicity of the metal content as well as the bioavailability of contaminants in marine sediments. They are technically well developed (USEPA, 1994; ASTM, 1997) and are widely accepted as useful tools for a wide variety of research and regulatory purposes (Swartz, 1989; Luoma and Ho, 1993; Burton, 1991). For example, they are used to determine the sediment toxicity of single chemicals and mixtures, chemical bioavailability, the potential adverse effects of dredged material on benthic marine organisms, and the magnitude and spatial and temporal distribution of pollution impacts in the field (Ferraro and Cole, 2002). Ecotoxicological monitoring requires simple, rapid and sensitive methods which can be used to measure the potential risk of sediment metal concentrations and their toxicity in marine invertebrates. Predicting the bioavailability and toxicity of metals in aquatic sediments is a critical component in the development of sediment quality criteria. The exposure of developing sea urchin embryos to the interface between sediment and water (SWI) provides a more ecologically relevant bioassay for this species (Anderson et al., 1998), and the results of the laboratory toxicity tests could be considered predictive of ecological change on a station-by-station basis because these are subchronic toxicity tests and may reflect chronic impacts in individual stations. The toxicity test employing sea urchin embryos identified as toxic the sediments from the Portma´n Bay. As observed with the chemical analysis the toxicity also decreased with the distance from the bay. Both PW and SWI tests presented a similar pattern of toxicity along the studied gradient. Furthermore, the toxicity results are in accordance with the assumption that toxicity does not exist in those sediments where the molar concentration of sulfides is higher than the molar concentration of divalent metals. Benthic faunal communities, as living components of the sediments, represents the integrate response of the biological effects of pollutants content in a sediment sample. The classical descriptive parameters (Margalef richness, Shannon–Wiener diversity, Pielou evenness, and Simpson dominance) showed a progressive variation in benthic communities along the metal contamination gradient of the Portma´n coast. Those stations with a degraded benthos (PG, PN, and CN) had fewer species, lower abundance, lower diversity, and lower benthic index scores than stations with a healthy benthos (PE and PL). The development of quantitative indices of benthic community health as indicators of environmental quality of estuarine and coastal water is a critical task for management of coastal ecosystems. The present study has evaluated a multi-metric benthic index known as relative benthic index, which is used in USA regional water quality controls (Anderson et al., 1998, 2001; Hunt et al., 2001) and exploratory benthic index a new index that introduces other descriptive parameters of the fauna in an integrated way. This latter index was advantageous for quantifying and estimating the cumulative effects of multiple stressors on benthic biota between the lowand high-impact zones. The index scores calculated for each

ARTICLE IN PRESS 1840

A. Cesar et al. / Ecotoxicology and Environmental Safety 72 (2009) 1832–1841

station provided a good and practical overall discrimination capacity. The metrics, pollution-indicative and pollution-sensitive species, were established from SIMPER analysis. Thus, the species indicators of sediment quality and metal pollution were selected for their contribution to the dissimilarity between the control and the polluted zones. There is evidence that in areas of high metal concentration, the effects of metals may be ameliorated by the development of tolerance mechanisms in some species and the evolution of tolerant strains in others (Grant et al., 1989). Our data suggest that the Tanaidacea Apseudes latreilli, the Sabellidae polychaetes of the Family and the bivalves Chamelea gallina and Macoma cumana can tolerate the whole range of metal concentrations found in Portma´n Bay. The Tanaidacea Apseudes latreilli is common in Portma´n Bay and also in metal-polluted areas of the nearby Mar Menor, a hypersaline lagoon. This euryhaline specie inhabits streams close to deserts that transport mining wastes from the mountains of Portma´n area. The crustaceans Bathyporeia guilliamisoniana, Coropium acutum, Phoxocephalus aquosus, and Anapagurus leaves, were absent from the metal-polluted sampling stations (PG, PN, and CN), suggesting that they were incapable of adapting to such extreme levels of metals. The sampling stations used in the present study may be classified using EBI scores with values at or above 5 used as the breaking point between ‘‘control’’ and ‘‘degraded’’ sites. The multi-metric EBI index showed a sensitivity and resolution for distinguishing differences in habitat quality. Analyses of variances and post hoc test for toxicity tests and relative benthic index indicates that the communities of the nearest stations to the poured of mine sterile (PN, PG and CN) they are negatively affected and present different populations from the others studied stations. Since the plan of restoration of the bay is centered in self-hardly affecting to the stations located in the bay (PN and PG), the others stations (except impacted CN) will serve as reference during and after the restoration and they will allow the comparison to evaluate the evolution and the success of the restoration process. CN that will not be restored and that is contained as one of the impacted stations can be used as positive control during the restoration process. Since the restoration program is proposed to remove a huge volume of sediments that fulfill the bay, it is important to take into account the importance of sediment resuspension in the metal liberation, and therefore to incorporate appropriate techniques which must reduce the sediment resuspension or mitigate the effects of the metal liberation. In this context, it is necessary to incorporate environmental tools such as water toxicity tests to be applied in the environmental monitoring during the restoration operations to assure that the metals content in sediments are not released to the water column, and consequently affecting the environmental quality in the surrounding areas. Furthermore, since the present sediments in the bay showed high toxicity, evidenced through the sea urchin embryo-larval test here applied, it is important to incorporate sediment toxicity test not only during the restoration operations, but also later, in order to ensure that the restoration was successful. In the present study we demonstrated that mine wastes are spread at least 7300 m along the Mediterranean coast of Murcia. The multivariate analysis and the benthic indices have shown that community structure changes along this metal pollution gradient. The marine sediments of Portma´n Bay continue to show high toxicity due to the high metal concentrations they contain and any sediment resuspension in this toxic hot spot of the Mediterranean Sea must be treated with caution. Sediment quality values need to be a develop to help protect public health and the environment (DelValls and Chapman, 1998; DelValls et al., 1998). The sediment quality triad (SQT) tools were used in this study, and the

combination of analytical chemistry, toxicity tests and benthic community structure, proved to be useful in providing a full picture of the extent of metal pollution along the coast of Murcia. This was the first sediment quality assessment along the Mediterranean coast of Murcia using the weight-of-evidence approach. Other sediment assessments were carried out in this area, but only focusing on sediment chemistry and such results were not published and are not available. The use of three lines of evidence, i.e. sediment physical–chemical characteristics, sediment toxicity, and benthic community analysis, integrated by multivariate analysis, was useful to assess the quality of the sediments of Portma´n Bay, giving an insight about the bioavailability of contaminants as well as in situ alterations. Such information is valuable to support dredged material management in this area. Portma´n Bay represents the hottest spot of metal contamination in the whole Mediterranean basin. This study has special importance for this ecosystem because of its fragile nature and the high amounts of metals it receives. The use of the WOE approach, integrating sediment toxicity, sediment chemical concentration, and infaunal community structure data was fundamental to determine the extent and the environmental significance of sediment contamination in Portma´n Bay and consequently to support the management actions which has been taking place in this ecosystem.

5. Conclusions Univariate and multivariate analyses showed positive correlation between the sediment metal concentrations associated to the all biological effects (sea urchins toxicity tests and benthic indices). The multi-metric EBI index showed a sensitivity and resolution for distinguishing differences in habitat quality. Analyses of variances and post hoc test for toxicity tests and benthic index indicates that the communities of the nearest stations to the poured of mine sterile (PN, PG, and CN) are altered and present different populations from the rest of the stations. The ordination of environmental sediment data by PCA suggested that metals (Zn, Pb, and Fe), fines, TOC, and LOI were associated with biological effects along the contamination gradient. In addition, Spearman rank correlations indicated significant negative correlations between benthic community structure and the metals, Zn, Pb, Al, and LOI. The use of the WOE approach, integrating sediment toxicity, sediment chemical concentration, and infaunal community structure data was fundamental to determine the extent and the environmental significance of metal sediment contamination and consequently to support the management actions which has been taking place in this ecosystem. This study provided a good insight about the bioavailability of contaminants as well as in situ alterations in this valuable estuarine ecosystem.

Acknowledgments We are grateful to the anonymous referee for his useful comments and constructive suggestions. The first author thanks ˜ ola de Cooperacio´n Internacional) of MUTIS-AECI (Agencia Espan the Spanish Government and CAPES/MEC-Brazil (BEX-3238/06-7) for the doctoral and postdoctoral scholarships. The work was partially funded by the Brazilian–Spanish joint project (CAPESBrazil #099/06 and MEC-Spain PHB 2005-0100-PC). The authors declare that this study was conducted in accordance with the national and institutional guidelines for the protection of human subjects and animal welfare.

ARTICLE IN PRESS A. Cesar et al. / Ecotoxicology and Environmental Safety 72 (2009) 1832–1841

References Allen, H.E., Gongmin, F., Deng, B., 1993. Analysis of acid-volatile-sulfide (AVS) and simultaneously extracted metals (SEM) for estimation of potential toxicity in aquatic sediments. Environ. Toxicol. Chem. 12, 1441–1453. American Society for Testing and Materials, 1997. Standard guide for conducting 10-day static sediment toxicity tests with marine and estuarine amphipods. E1367-92. In: Annual Book of ASTM Standards, vol. 11.05. Philadelphia, PA, pp. 731–756. Anderson, B.S., Hunt, J.W., Phillips, B.M., Fairey, R., Roberts, C.A., Oakden, J.M., Puckett, H.M., Stephenson, M., Tjeerdema, R.S., Long, E.R., Wilson, C.J., Lyons, J.M., 1998. Chemistry, toxicity and benthic community conditions in selected sediments of the Los Angeles Region. Final Report. State Water Resources Control Board, Sacramento, CA, USA. Anderson, B.S., Hunt, J.W., Phillips, B.M., Fairey, R., Roberts, C.A., Oakden, J.M., Puckett, H.M., Stephenson, M., Tjeerdema, R.S., Long, E.R., Wilson, C.J., Lyons, J.M., 2001. Sediment quality in Los Angeles Harbor, USA: a triad assessment. Environ. Toxicol. Chem. 20, 359–370. Ankley, G.T., Phipps, G.L., Leonard, E.N., Benoit, D.A., Mattson, V.R., Kosian, P.A., Cotter, A.M., Dierkes, J.R., Hansen, D.J., Mahoney, J.D., 1991. Acid-volatile sulfide as a factor mediating cadmium and nikel bioavailability in contaminated sediments. Environ. Toxicol. Chem. 10, 1299–1307. Ankley, G.T., 1996. Evaluation of metal/acid-volatile sulfide relationships in the prediction of metal bioaccumulation by benthic macroinvertebrates. Environ. Toxicol. Chem. 15, 2138–2146. Associac- a˜o Brasileira de Normas Te´cnicas, 2006. Ecotoxicologia aqua´tica— ˆ nica de curta durac- a˜o—Me´todo de ensaio com ouric- o do mar Toxicidade cro (Echinodermata: Echinoidea), NBR 15350. Buchanan, J.B., 1984. Sediment analysis. In: Holme, N.A., Mcintyre, A.D. (Eds.), Methods for the Study of Marine Benthos. Blackwell Scientific Publications, Oxford, pp. 41–65. Burton Jr., G.A., 1991. Assessing the toxicity of freshwater sediments. Environ. Toxicol. Chem. 10, 1585–1627. Burton Jr., G.A., Batley, G.E., Chapman, P.M., Forbes, V.E., Smith, E.P., Reynoldson, T., Schlekat, C.E., den Besten, P.J., Bailer, A.J., Green, A.S., Dwyer, R.L.A., 2002. Weight-of-evidence framework for assessing sediment (or other) contamination: improving certainty in the decision-making process. Hum. Ecol. Risk Assess. 8 (7), 1675–1696. Cesar, A., Marı´n, A., Marı´n-Guirao, L., Vita, R., 2002. Sensitivity of Mediterranean amphipods and sea urchins to reference toxicants. Cienc. Mar. 12, 407–417. Cesar, A., Marı´n, A., Marı´n-Guirao, L., Vita, R., 2004. Amphipod and sea urchin tests to assess the toxicity of Mediterranean sediments: the case of Portma´n Bay. Sci. Mar. 68 (1), 205–213. Cesar, A., Choueri, R.B., Riba, I., Moralles-Caselles, C., Pereira, C.D.S., Santos, A.R., Abessa, D.M.S., Delvalls, T.A., 2007. Comparative sediment quality assessment in different littoral ecosystems from Spain (Gulf of Cadiz) and Brazil (Santos and Sa˜o Vicente estuarine system). Environ. Int. 33, 429–435. CETESB, 1999. Me´todo de ensaio: A´gua do mar-Teste de toxicidade croˆnica de curta durac- a˜o com Lytechinus variegatus, Lamark,1816 (Echinodermata: Echinoidea). Cia. de Tecnologia de Saneamento Ambiental do Estado de Sa˜o Paulo, Brasil, L5.250. CEDEX, 1994. DelValls, T.A., Casado-Martinez, M.C., Riba, I., Martı´n-Dı´az, M.L., Forja, J.M., Garcı´a-Luque, E., Go´mez-Parra, A., 2001. Investigacio´n conjunta sobre la viabilidad de utilizar ensayos ecotoxicolo´gicos para la evaluacio´n de la calidad ambiental del material de dragado. Technical Report OT/060/01. Centro de Estudios y Experimentacio´n de Obras Pu´blicas, Ca´diz, Spain. Clark, K.R., Ainsworth, M., 1993. A method of linking multivariate community structure to environmental variables. Mar. Ecol. Prog. Ser. 92, 205–219. Clark, K.R., Green, R.H., 1988. Statistical design and analysis for a biological effects study. Mar. Ecol. Prog. Ser. 46, 213–226. Clark, K.R., Warwick, R.M., 2001. Change in Marine Communities: An Approach to Statistical Analysis and Interpretation, second ed. Plymouth Marine Laboratory, UK. Clark, K.R., Gorley, R.N., 2006. PRIMER v6: User Manual/Tutorial. Plymouth, United Kingdom. Chapman, P.M., McDonald, B.G., Lawrence, G.S., 2002. Weight-of-evidence issues and frameworks for sediment quality (and other) assessments. Hum. Ecol. Risk. Assess. 8, 1489–1515. De Leon, A.R., Mas, J., Guerrero, J., Jornet, A., 1984. Monitoring of heavy metals in superficial sediment and some marine organisms from the western Mediterranean coast. VIIes Journe´es E´tud. Pollutions, Lucerne, CIESM, 321–326. DelValls, T.A., Chapman, P.M., 1998. The use of multivariate analysis to link the sediment quality triad components to site-specific sediment quality values in

1841

the Gulf of Ca´diz (Spain) and in San Francisco Bay (USA). Cien. Mar. 24, 313–336. DelValls, T.A., Forja, J.M., Gonza´lez-Mazo, E., Go´mez-Parra, A., 1998. Determining contamination sources in marine sediments using multivariate analysis. TracTrend. Anal. Chem. 17, 181–192. Di Toro, D.M., Mahony, J.D., Hansen, D.J., Scott, K.J., Hicks, M.B., Mayr, S.M., Redmond, M.S., 1990. Toxicity of cadmium in sediments: the role of acid volatile sulfide. Environ. Toxicol. Chem. 9, 1487–1502. Di Toro, D.M., Mahony, J.D., Hansen, D.J., Scott, K.J., Carlson, A.R., Ankley, G.T., 1992. Acid volatile sulfide predicts the acute toxicity of cadmium and nikel in sediments. Environ. Sci. Technol. 26, 96–101. Environment Canada, 1992. Biological test method: Fertilization assay using echinoids (sea urchins and sand dollars), amended November 1997. EPS 1/RM/ 27. North Vancouver, BC. Ferraro, S.P., Cole, F.A., 2002. A field validation of two sediment-amphipod toxicity tests. Environ. Toxicol. Chem. 21, 1423–1437. Grant, A., Hately, J.G., Jones, N.V., 1989. Mapping the ecological impact of heavy metals in the estuarine polychaete Nereis diversicolor using inherited metal tolerance. Mar. Pollut. Bull. 20, 235–238. Hunt, J.W., Anderson, B.S., Phillips, B.M., Tjeerdema, R.S., Taberski, K.M., Wilson, C.J., Puckett, H.M., Stephenson, M., Fairey, R., Oakden, J., 2001. A large-scale categorization of sites in San Francisco Bay, USA, based on the sediment quality triad, toxicity identification evaluations, and gradient studies. Environ. Toxicol. Chem. 20, 1252–1265. Long, E.R., MacDonald, D.D., Smith, S.L., Calder, F.D., 1995. Incidence of adverse biological effects within ranges of chemical concentrations in marine and estuarine sediments. Environ. Manage. 19, 81–97. Long, E.R., Carr, R.S., Montagua, P.A., 2003. Porewater toxicity tests: Values as a component of sediment quality triad assessments. In: Carr, R.S., Nipper, M. (Eds.), Porewater Toxicity Testing: Biological, Chemical, and Ecological Considerations. SETAC Press, Pensacola, FL, pp. 163–200. Luoma, S.N., Ho, K.T., 1993. Appropriate uses of marine and estuarine sediment bioassays. In: Calow, P. (Ed.), Handbook of Ecotoxicology. Blackweel Scientific, Oxford, pp. 193–226. Luoma, S.N., Fisher, N., 1997. Uncertainties in assessing contaminant exposure from sediments. In: Ingersoll, C.G., Dillon, T., Biddinger, G.R. (Eds.), Ecological Risk Assessment of Contaminated Sediments, first ed. SETAC Press, Pensacola, FL, USA, pp. 211–237. Marin-Guirao, L., Cesar, A., Marı´n, A., Vita, R., 2005. Establishing the ecological quality status of soft-bottom mining-impacted coastal water bodies in the scope of the Water Framework Directive. Mar. Pollut. Bull. 50, 374–387. Pe´rez, J.G., Puente, C.R., 1989. Estudio de la contaminacio´n marina entre Cabo de ˜ oso (SE-Espan ˜ a). Concentraciones de cadmio, plomo y cinc en Palos y Cabo Tin ˜ ol de Oceanografı´a. sedimentos superficiales. Informes Te´cnicos Instituto Espan MAPA. Rey, J., Del Rı´o, V.D., 1983. La plataforma continental Mediterra´nea, entre cabo de ˜ oso: morfologı´a y estudios sı´smicos de la cobertura Palos y cabo Tin ˜ ol de Oceanografı´a. sedimentarı´a. Informes Te´cnicos Instituto Espan Riba, I., Forja, J.M., Go´mez-Parra, A., DelValls, T.A., 2004a. 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. Riba, I., Casado-Martı´nez, M.C., Forja, J.M., DelValls, T.A., 2004b. Sediment quality in Atlantic Coast of Spain. Environ. Toxicol. Chem. 23 (2), 271–282. Somerfield, P.J., Gee, J.M., Warwicck, R.M., 1994. Soft sediment meiofaunal community structure in relation to a long-term heavy metal gradient in the Fal estuary system. Mar. Ecol. Prog. Ser. 105, 79–88. Swartz, R.C., 1989. Marine sediment toxicity tests. In: Contaminated Marine Sediments—Assessment and Remediation. National Academy Press, Washington, DC, pp. 115–129. Tabachnic, B.G., Fidell, L.S., 1996. Using Multivariate Statistics. Harper Collins, College Publishers, New York, NY, USA. US Environmental Protection Agency, 1994. Methods for assessing the toxicity of sediment-associated contaminants with estuarine and marine amphipods. EPA/600/-94/025, Narragansett, RI 02882. US Environmental Protection Agency, 1995. Short-term methods for estimating the chronic toxicity of effluents and receiving waters to west coast marine and estuarine organisms. EPA/600/R-95-136, Cincinnati, OH. Verdardo, D.J., Forelich, P.N., Mc Intyre, A., 1990. Determination of organic carbon and nitrogen in marine sediments using the Carlo Erba NA-1500 analyzer. Deep-Sea Res. 37, 157–165. Whitfield, M., 1974. The hydrolysis of ammonia ions in sea water—a theoretical study. J. Mar. Biol. Assoc. UK 54, 565–580. Whiteman, F.M., Ankley, G.T., Kahl, M.D., Rau, D.M., Bacer, M.D., 1996. Evaluation of interstitial water as route of exposure for ammonia in sediment tests with benthic macroinvertebrates. Environ. Toxicol. Chem. 15, 794–801.