14
Research papers Spatial variation, speciation and sedimentary records of mercury in the Guanabara Bay (Rio de Janeiro, Brazil) Stefano Covelli a,n , Ioanna Protopsalti a , Alessandro Acquavita a,b , Marcelo Sperle c , Maurizio Bonardi d , Andrea Emili a a Dipartimento di Geoscienze, Universit a di Trieste, Via Weiss 2, 34128 Trieste, Italy b Osservatorio Alto Adriatico, Agenzia Regionale per la Protezione dell’Ambiente del Friuli Venezia Giulia (ARPA-FVG), Via Cairoli 14, 33057 Palmanova, Italy c Institute of Oceanography, Rua S ~ ao Francisco Xavier 524, S. 4018E, Universidade do Estado do Rio de Janeiro, Maracan ~ a, 20550-013 Rio de Janeiro, Brazil d ISMAR-CNR, Castello 1364/A, 30122 Venezia, Italy article info Article history: Received 23 June 2011 Received in revised form 18 November 2011 Accepted 6 December 2011 Available online 14 December 2011 Keywords: Mercury Contaminated sediments Enrichment factor Inventory Selective sequential extraction Guanabara Bay abstract As part of the ‘‘TAGUBAR’’ (TAngential GUanabara Bay Aeration Recovery) project, surface and long core sediments of the Guanabara Bay (Rio de Janeiro, Brazil) were investigated for mercury (Hg). The main, but not the only, input of Hg into the Bay’s waters is known to be a Chlor-Alkali Plant (CAP) located in the Acar ı-S ~ ao Jo ~ ao de Meritı ´ River system, on the northwestern side of the Bay. Mercury distribution in surface sediments ( o0.1–3.22 mg kg 1 , average 0.87 70.80, n ¼40) seems to be controlled by the organic component, along with sulfur rather than grain-size, where Hg concentrations are less than 1 mg kg 1 . Conversely, where the metal contents are higher than 1 mg kg 1 , accumulation in surface sediments is mostly related to the presence of nearby contamination sources, such as industrial and urban settlements in the western sector of the Bay. Although total Hg contents in surface sediments exceed the values suggested by the effects-based standard quality guidelines as potentially toxic for the benthic community, results from a sequential extraction procedure showed that the contribution of the more soluble, easily exchangeable and eventually bioavailable Hg phases was found almost negligible ( o0.1%). Most of the metal is strongly bound to the mineral lattice of the sedimentary matrix and should therefore be considered almost immobilized. The reduction in Hg accumulation in bottom sediments, expected as a consequence of the adoption of contamination control policies (i.e. Hg-free technologies in the CAP and sewage treating facilities), has not been clearly observed in the core profiles. Current estimates of Hg accumulation rates at the core top range from approximately 1 to 18 mg m 2 yr 1 . Pre-industrial bottom core samples indicate that the central and northeastern sectors of the Bay are strongly affected by Hg enrichment: concentrations exceed the estimated baseline concentration by up to 20 factors. A cumulative Hg inventory suggests that the metal content has increased with the same order of magnitude in the vicinity of potential contamination sources on the western side of the Bay, but at a different rate; this is apparently determined by local conditions. A natural attenuation of Hg concentrations to background levels is not predictable in the near future. & 2011 Elsevier Ltd. All rights reserved. 1. Introduction Coastal lagoons and estuaries are often surrounded by urban and industrial areas and are frequently loaded to a significant level with a variety of pollutants. The shallowness and limited water exchange of such coastal systems mean that the residence time of water and suspended sediments is generally much longer than in open coastal areas. This makes these environments vulnerable to even small levels of loading. Mercury (Hg) is considered to be one of the most harmful pollutants for the marine environment, and is recognized to be extremely toxic even at low concentrations. The negative effects of Hg on ecosystems and human health are mostly related to its transformation into the more toxic and mobile organic form, methyl-Hg (Ullrich et al., 2001), which can be biomagnified along the trophic chain up to human beings (Porcella, 1994). The presence of Hg in aquatic systems can be attributed to several sources, for example atmospheric deposition (Schl ¨ uter, 2000), and/or past mining and modern industrial discharge (e.g. Baldi and Bargagli, 1984; Covelli et al., 2001; Trombini et al., 2003). Mercury accumulated in sediments may be subjected to burial and/or to biogeochemical processes, which affect its speciation Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/csr Continental Shelf Research 0278-4343/$ - see front matter & 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.csr.2011.12.003 n Corresponding author. Tel.: þ39 040 5582031; fax: þ39 040 5582048. E-mail address: [email protected] (S. Covelli). Continental Shelf Research 35 (2012) 29–42

Spatial variation, speciation and sedimentary records of mercury in the Guanabara Bay (Rio de Janeiro, Brazil)

Embed Size (px)

Citation preview

Page 1: Spatial variation, speciation and sedimentary records of mercury in the Guanabara Bay (Rio de Janeiro, Brazil)

Continental Shelf Research 35 (2012) 29–42

Contents lists available at SciVerse ScienceDirect

Continental Shelf Research

0278-43

doi:10.1

n Corr

E-m

journal homepage: www.elsevier.com/locate/csr

Research papers

Spatial variation, speciation and sedimentary records of mercuryin the Guanabara Bay (Rio de Janeiro, Brazil)

Stefano Covelli a,n, Ioanna Protopsalti a, Alessandro Acquavita a,b, Marcelo Sperle c,Maurizio Bonardi d, Andrea Emili a

a Dipartimento di Geoscienze, Universit �a di Trieste, Via Weiss 2, 34128 Trieste, Italyb Osservatorio Alto Adriatico, Agenzia Regionale per la Protezione dell’Ambiente del Friuli Venezia Giulia (ARPA-FVG), Via Cairoli 14, 33057 Palmanova, Italyc Institute of Oceanography, Rua S ~ao Francisco Xavier 524, S. 4018E, Universidade do Estado do Rio de Janeiro, Maracan~a, 20550-013 Rio de Janeiro, Brazild ISMAR-CNR, Castello 1364/A, 30122 Venezia, Italy

a r t i c l e i n f o

Article history:

Received 23 June 2011

Received in revised form

18 November 2011

Accepted 6 December 2011Available online 14 December 2011

Keywords:

Mercury

Contaminated sediments

Enrichment factor

Inventory

Selective sequential extraction

Guanabara Bay

43/$ - see front matter & 2011 Elsevier Ltd. A

016/j.csr.2011.12.003

esponding author. Tel.: þ39 040 5582031; fa

ail address: [email protected] (S. Covelli).

a b s t r a c t

As part of the ‘‘TAGUBAR’’ (TAngential GUanabara Bay Aeration Recovery) project, surface and long core

sediments of the Guanabara Bay (Rio de Janeiro, Brazil) were investigated for mercury (Hg). The main,

but not the only, input of Hg into the Bay’s waters is known to be a Chlor-Alkali Plant (CAP) located in

the Acar�ı-S~ao Jo~ao de Meritı River system, on the northwestern side of the Bay. Mercury distribution

in surface sediments (o0.1–3.22 mg kg�1, average 0.8770.80, n¼40) seems to be controlled by the

organic component, along with sulfur rather than grain-size, where Hg concentrations are less than

1 mg kg�1. Conversely, where the metal contents are higher than 1 mg kg�1, accumulation in surface

sediments is mostly related to the presence of nearby contamination sources, such as industrial and

urban settlements in the western sector of the Bay. Although total Hg contents in surface sediments

exceed the values suggested by the effects-based standard quality guidelines as potentially toxic for the

benthic community, results from a sequential extraction procedure showed that the contribution of the

more soluble, easily exchangeable and eventually bioavailable Hg phases was found almost negligible

(o0.1%). Most of the metal is strongly bound to the mineral lattice of the sedimentary matrix and

should therefore be considered almost immobilized.

The reduction in Hg accumulation in bottom sediments, expected as a consequence of the

adoption of contamination control policies (i.e. Hg-free technologies in the CAP and sewage treating

facilities), has not been clearly observed in the core profiles. Current estimates of Hg accumulation

rates at the core top range from approximately 1 to 18 mg m�2 yr�1. Pre-industrial bottom core

samples indicate that the central and northeastern sectors of the Bay are strongly affected by Hg

enrichment: concentrations exceed the estimated baseline concentration by up to 20 factors. A

cumulative Hg inventory suggests that the metal content has increased with the same order of

magnitude in the vicinity of potential contamination sources on the western side of the Bay, but at a

different rate; this is apparently determined by local conditions. A natural attenuation of Hg

concentrations to background levels is not predictable in the near future.

& 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Coastal lagoons and estuaries are often surrounded by urbanand industrial areas and are frequently loaded to a significantlevel with a variety of pollutants. The shallowness and limitedwater exchange of such coastal systems mean that the residencetime of water and suspended sediments is generally much longerthan in open coastal areas. This makes these environmentsvulnerable to even small levels of loading.

ll rights reserved.

x: þ39 040 5582048.

Mercury (Hg) is considered to be one of the most harmfulpollutants for the marine environment, and is recognized to beextremely toxic even at low concentrations. The negative effectsof Hg on ecosystems and human health are mostly related to itstransformation into the more toxic and mobile organic form,methyl-Hg (Ullrich et al., 2001), which can be biomagnifiedalong the trophic chain up to human beings (Porcella, 1994).The presence of Hg in aquatic systems can be attributed to severalsources, for example atmospheric deposition (Schluter, 2000),and/or past mining and modern industrial discharge (e.g. Baldiand Bargagli, 1984; Covelli et al., 2001; Trombini et al., 2003).Mercury accumulated in sediments may be subjected to burialand/or to biogeochemical processes, which affect its speciation

Page 2: Spatial variation, speciation and sedimentary records of mercury in the Guanabara Bay (Rio de Janeiro, Brazil)

S. Covelli et al. / Continental Shelf Research 35 (2012) 29–4230

(Shi et al., 2005), remobilization (Covelli et al., 2008; Emili et al.,2011) and the bioavailability for the biota (Lacerda et al., 1992;Kehrig et al., 2002). Understanding the fate of Hg in aquaticsystems is necessary in order to identify the potential risks and toestablish appropriate remediation technologies, and is thereforeof high environmental importance (Kaplan et al., 2002).

Uncontrolled population growth in the Rio de Janeiro metro-politan area has severely affected water quality in Guanabara Bay,which has been polluted over the last century by the additionof untreated domestic and industrial discharge. The Bay (Fig. 1)encloses the second largest industrialized region in Brazil and issurrounded by thousands of different industries (e.g. textiles,food, metallurgic, chemical) that are widely spread within thedrainage basin. The Bay also hosts several significant oil refineries,marine petroleum terminals, sanitary deposit sites, commercialports (Rio de Janeiro and Niteroi) and shipyards, and is subjectedto intense maritime traffic (Kjerfve et al., 1997). Several studieshave shown that considerable amounts of organic (e.g. Carvalhaeset al., 2002; Carreira et al., 2004) and inorganic (e.g. Rebello et al.,1986; Perin et al., 1997; Machado et al., 2002; Barbosa et al.,2004; Baptista Neto et al., 2006) pollutants entered GuanabaraBay’s eutrophic waters and were accumulated in bottom sedi-ments, significantly altering this environmental compartment.

A well-known source of Hg in the Bay waters is a Chlor-AlkaliPlant (CAP) located in the Acarı-S~ao Jo~ao de Meritı River system, onthe northwestern side of the Bay (Kehrig et al., 2003). It has beenestimated that 1.46 t of Hg were released through the plant’s liquid

Fig. 1. Location map of the study area and of the surface (grab) and core sediment sam

also indicated following information obtained from FEEMA (1990), Godoy et al. (1998)

effluents in 1975, a rate, which decreased to 20 kg yr�1 after 1979when an effluent treatment system was installed. At the beginningof the 90s, however, 160 kg yr�1 of Hg was reported to be releasedinto the Bay (Rego et al., 1993). It is evident that Hg, in commonwith other heavy metals, may derive from both point (contaminatedrivers and landfills) and diffuse sources in the Bay area (Machadoet al., 2002). Mercury accumulation, as well as the factors controllingits distribution, have been carefully investigated in bottom sedi-ments close to the main pollution source (Barrocas and Wasserman,1998; Wasserman et al., 2000), along the northern coastline of theBay (Machado et al., 2008) and in mangrove systems, which arerecognized to act as a trap for pollutants (Machado et al., 2002; Kehriget al., 2003). However, a comprehensive study of Hg contamination insediments of the whole Guanabara Bay area was still lacking and it isthe primary objective of this study.

The dispersion and accumulation patterns of this metal, relativeto the potential sources of contamination, along with the relation-ships with the sedimentary components were investigated. Remo-bilization from bottom sediments can also be a secondary sourceof Hg into the aquatic environment. A speciation technique wasapplied to investigate the main binding sites and phase associationsof Hg in the Bay’s sediments, in order to highlight the presence ofmobile and potentially available species for methylation processes.Besides, the sedimentary records of Hg from different locations wereinvestigated in order to evaluate the historical loading of this metalto the Bay’s sediments and to establish how recent sedimentsrespond to Hg inputs from human activities.

pling points. The main potential sources of mercury contamination in the Bay are

and Christensen et al. (2010).

Page 3: Spatial variation, speciation and sedimentary records of mercury in the Guanabara Bay (Rio de Janeiro, Brazil)

S. Covelli et al. / Continental Shelf Research 35 (2012) 29–42 31

2. Material and methods

2.1. Environmental setting

Guanabara Bay (Fig. 1), the second largest coastal embaymentin Brazil, covers an area of 384 km2, with a perimeter of 131 kmand a mean water volume of 1.87�109 m3. Islands constituteabout 15% of the Bay’s surface area (Kjerfve et al., 1997). The Bay ischaracterized by a central channel with a median width of 400 m,as defined by the 30 m isobaths. The deepest point of the channelhas a depth of 58 m, while the Bay’s average water depth is 5.7 m(Amador, 1997). Bottom sediments near the entrance to the Bayare principally composed of sand, reflecting wave and tidal forcing.However, sediments inside the Bay are mostly made up of mud, asa result of the Holocene transgression and rapid fluvial sedimenta-tion, accelerated by the canalization of rivers and deforestation.

Tides in the Bay are mixed and mainly semidiurnal with a meanrange of 0.7 m, mean spring range of 1.1 m and neap tidal range of0.3 m (Kjerfve et al., 2001). Tides control the circulation inside theBay, where peak tidal currents are on the order of 0.5 m s�1; at theBay’s entrance maximum current speed reaches 1.6 m s�1 duringflood-tides and 1.0 m s�1 during ebb-tides (Kjerfve et al., 2001).

The Guanabara Bay drainage basin measures 4080 km2 and isdrained by 91 rivers and stream channels. However, just six rivers areresponsible for 85% of the 100 m3 s�1 total annual freshwaterdischarge. In addition to urban and industrial areas, the drainagebasin is bordered by agricultural fields and pastures and by high reliefmountains, the Serra do Mar (Kjerfve et al., 1997; Perin et al., 1997).The Bay is an eutrophic coastal environment (Carreira et al., 2002),which is heavily affected by discharges of pollutants from the inland.

2.2. Grab and core sampling

Bottom sediment samples (n¼47) were collected using a 5 LVan Veen grab in February 2007 (Fig. 1). Samples were obtainedby scraping the upper 2–3 cm of the sediment, which was placedin plastic bags and stored in a cold room at þ4 1C until analysis.Long cores (LC), from 160 to 281 cm long, were collected by scubadiving with an hand operated piston-corer. Four cores wereconsidered in this study: LC1, LC5, LC6 and LC7 (Fig. 1). Sedimentcores were sealed and stored upright on board the boat usedduring the sampling operations, and were then cut into 70 cmlong sections, numbered consecutively and transported to thelaboratory where they were frozen.

Sediment subsampling for water content, grain-size and geo-chemical analyses was performed by cutting 2-cm thick slices,using a thin stainless steel knife. The first 10 cm of the partiallythawed cores were continuously subsampled. Slices were recov-ered at depths of 12–14, 16–18, 20–22, 24–26 and 28–30 cm, andthen at every 20 cm down to 100 cm. From a depth of 1 m down tothe bottom of the core, 2 or 3 subsamples were collected; thesewere considered to correspond to sediments accumulated beforethe industrial period. Each sediment slice was homogenized andsplit into sub-samples for subsequent analyses.

2.3. Water content, porosity, grain-size and geochemical analyses

Water content was calculated on 2–3 g subsamples as the lossof weight after drying in an oven at 105 1C for 24 h (Loring andRantala, 1992). Porosity was calculated according to Johnson et al.(1982) as follows:

f¼ðMw=rwÞ

ðMs=rsÞþðMw=rwÞ

where Mw is the weight of water lost on drying, Ms is the weightof dry sediment, rw¼1.025 g m�3 is the density of water, and

rs is the sediment density. Sediment density was measured forsome samples from each core using a Micromeritics AccuPyc 1330Picnometer.

For grain-size analysis, approximately 10 g of previouslyhomogenized wet sediment was treated with 40 mL of H2O2

(10 v/v) in a water bath for 48 h in order to remove organicmatter. Following this treatment, samples were wet-sieved withdistilled water using a 2000 mm sieve to remove coarse shellyfragments (corresponding to the gravel–sand boundary deter-mined by Wentworth (1922)). The fraction o2000 mm was thenrecovered and a Malvern Mastersizer 2000 Laser DiffractionParticle Size Analyzer, coupled with an autosampler, was usedto determine the particle sizes.

Total carbon (Ctot) and nitrogen (Ntot) content was determinedin freeze-dried and homogenized samples using a Perkin Elmer2400 CHN Elemental Analyzer at a combustion temperatureof 1020 1C. Organic Carbon (Corg) in sediments was determinedthrough progressive acidification with HCl (0.1–1.0 M), at acombustion temperature of 920 1C (Hedges and Stern, 1984).Acetanilide was used as standard compound for the calibration.

Among major and trace elements, only Al, Fe, La, Sc, Th, Ti, andV were considered in this work, since they are the potential grain-size proxies for the normalization procedure described in Section3. Concentrations of these elements, along with S content, weredetermined on 0.5 g air-dried samples, digested with 3 mL ofHCl–HNO3–H2O at 95 1C for 1 h in a water bath, and analyzedby ICP-ES and ICP-MS under contract (ACME Lab, Vancouver,Canada). The accuracy of the results was verified by analyzingStandard Reference Materials. Percentages recoveries for all theconsidered elements ranged from 92% to 107%.

Statistical analyses were performed using the SigmaPlot Systat7 software package. Data computation and plots were producedusing the Kriging interpolation method in the Golden SoftwareSurfer 8.

2.4. Mercury analyses

Total Hg content was determined on about 200 mg of lyophi-lized sediment samples, digested with a mixture of 3 mL of ‘‘aqua

regia’’ and Milli-Q water (5 mL) in a closed microwave system(Milestone, MLS 1200). The resulting solutions were analyzed byCV AAS (FIAS 100, Perkin Elmer), using NaBH4 (0.3%) in NaOH(0.1%) as a reducing agent. The accuracy of the method was testedusing a Certified Reference Material (PACS-2 Marine Sediment,NRCC, 3.0470.20 mg kg�1). On the basis of three replicates,recovery percentages ranged from 97% to 105%. The detectionlimit of the method was 0.10 mg kg�1.

The pool of total Hg in sediments is the sum of its differentchemical species, which can be distinguished on the basis of theirdifferent chemical and physical behavior (Cai et al., 1997). Thus,for Hg speciation, a five-step Selective Sequential Extraction (SSE)procedure was applied following the method developed by Bloomet al. (2003), and successively revised by Shi et al. (2005) forriver sediments. Analyses were performed as reported in Covelliet al. (2009) on selected sediment grab samples. The extractionswere carried out using approximately 2 g of lyophilized samples.Four fractions were distinguished: (1) ‘‘water soluble’’ (Hg-w);(2) ‘‘human stomach acid soluble’’ (Hg-h); (3) ‘‘organo-chelated’’(Hg-o); (4) ‘‘elemental Hg’’ (Hg-e). In the last step (5), the residuewas air dried and total Hg digestion with aqua regia wasperformed for ‘‘mercuric sulfide’’ (Hg-s) and Hg immobilized bypyrite (FeS2; Huerta-Diaz and Morse, 1992). The total Hg in eachextract was determined by CV AFS (Brooks Rand, Ltd.) after SnCl2

(30% in 10% HCl) reduction and gold trap amalgamation. Thecalculated limit of detection on the basis of three replicates was0.1 ng L�1. Quality control included Hg determinations in all the

Page 4: Spatial variation, speciation and sedimentary records of mercury in the Guanabara Bay (Rio de Janeiro, Brazil)

S. Covelli et al. / Continental Shelf Research 35 (2012) 29–4232

reagents. The method was verified for all samples, with the sumof Hg in steps 1–5 compared to total Hg concentrations.

3. Results and discussion

3.1. Surface sediments

3.1.1. Distribution of grain-size and geochemicals

Sediment grain-size was classified according to Shepard(1954). The results show the northern sector of the Bay to beenriched in the silty component (avg. 62.4730.2%), whichincrease to more than 80% in the most protected area, betweenthe Iguac-u River and the Macacu River mouths (Fig. 2). Due to theterrigenous supply of these fluvial sources, and to the limitedwater circulation in the inner part of Bay, the highest amounts ofclay (7.5% at station 22 and 7.0% at station 7) occur in this area.The southern sector of the Bay, which experiences both waveaction and tidal currents, exhibits coarser grain-size along thecentral channel and its sides. The sand content of bottom samplesranges widely, from 2.1% to 100% (avg. 30.9732.0%). However,from the Governador Island towards the inner part of the Bay, thesandy component generally constitutes less than 20% of the totalsample.

The same inner sector of the Bay shows the highest values(Table 1) of Ctot (0.12–5.78%), Ntot (0.01–0.66%,) and Corg

(0.03–5.56%) and high concentrations of these were also foundin the central area of the Bay, east of Governador Island.

Fig. 2. Map of the main textures distribution in the surface sedim

Conversely, the lowest concentrations are progressively foundapproaching southwards the entrance of the Bay (Fig. 3a).

Corg constitutes up to 99% of Ctot (about 85%, on average). BothCorg and Ntot are strongly correlated to the finest component(o16 mm) of the bottom sediments (r¼0.920 and r¼0.868,respectively, n¼47, po0.001). These two parameters also co-vary significantly (po0.001), and the linear correlation yields azero intercept, indicating that basically all N in these samples ispresent in organic form (Goni et al., 2003).

Although the d13C and d15N isotopic composition of sedimen-tary organic matter (OM) was not determined in this study, someassumptions can be made on the basis of the Corg/Ntot ratio.

This ratio can give an indication of the sources of OM in anaquatic environment. (e.g. Goni and Hedges, 1995; Hedges et al.,1997; Goni and Thomas, 2000; Goni et al., 2003), because ofthe different, origin-dependent, remineralisation efficiency (Aller,1998; Aller and Blair, 2004). In this study, Corg/Ntot values rangefrom 3.5 to 15.2 (average 10.371.7), suggesting that OM, espe-cially in the inner part of the Bay, originates mainly from terrestrialsources, such as freshwater inputs from the tributaries. Based onCorg/Ntot and d13C data from eight short sediment cores, Carreiraet al. (2002) distinguished two areas in Guanabara Bay. The innerbay sites, high in Corg/Ntot and depleted in d13C, were shown to bemainly associated with riverine/terrigenous material from theMacacu, Iguac-u and Estrela Rivers, while in the central area ofthe Bay, autochthonous sources of organic matter were seen to bepredominant. Our findings are in accordance with the results ofCarreira et al. (2002).

ents of the Bay according to Shepard’s classification (1954).

Page 5: Spatial variation, speciation and sedimentary records of mercury in the Guanabara Bay (Rio de Janeiro, Brazil)

Table 1Descriptive statistics for geochemical variables in grab and core samples.

Sample Grab LC1 core LC5 core LC6 core LC7 coren¼47 n¼16 n¼18 n¼17 n¼16

Ctot (%) Range 0.12–5.78 1.93–4.44 2.03–5.24 0.06–1.23 1.48–3.09

Avg7std 4.0071.69 3.6470.87 4.3070.97 0.6970.32 1.9970.43

Ntot (%) Range 0.01–0.66 0.20–0.43 0.20–0.55 0.01–0.17 0.11–0.26

Avg7std 0.4070.19 0.3570.08 0.4370.11 0.1770.05 0.1970.04

Corg (%) Range 0.03–5.56 1.40–4.07 1.64–4.74 0.05–1.13 1.13–2.71

Avg7std 3.6071.66 3.1870.94 3.9470.97 0.4970.29 1.6970.47

S (%) Range 0.05–3.45 0.86–2.60 0.64–1.97 0.06–0.34 0.54–1.33

Avg7std 1.4870.77 2.0670.58 1.5070.41 0.1870.06 0.7370.18

Al (%) Range 0.07–4.03 2.79–3.30 2.52–2.99 0.19–0.92 2.44–2.96

Avg7std 1.9771.03 3.0770.17 2.7670.12 0.5370.23 2.7770.14

Fe (%) Range 0.11–4.97 3.31–4.19 3.16–3.74 0.21–1.28 3.35–3.70

Avg7std 2.6971.24 3.7770.27 3.4270.18 0.7670.33 3.5470.09

Ti (%) Range 0.004–0.100 0.037–0.051 0.031–0.050 0.011–0.024 0.036–0.074

Avg7std 0.04570.022 0.04570.004 0.04570.005 0.01570.004 0.04370.009

V (mg kg�1) Range 2.0–59.0 42.0–53.0 41–55 4–18 41–48

Avg7std 34.0714.4 46.173.2 4875 1174 4472

Sc (mg kg�1) Range 0.3–8.2 6.0–7.7 5.4–6.7 0.7–2.2 6.2–7.1

Avg7std 4.772.1 6.570.4 6.170.4 1.470.5 6.670.2

La (mg kg�1) Range 2.8–60.8 35.9–54.9 24.0–32.0 2.8–10.2 28.1–33.7

Avg7std 26.0713.7 47.276.2 27.171.9 7.272.3 31.171.4

Th (mg kg�1) Range 0.4–18.7 9.5–15.2 6.3–8.6 1.6–3.0 8.4–10.3

Avg7std 7.074.4 12.971.9 7.470.6 2.470.5 9.170.5

Hg (mg kg�1) Range o0.10–3.22 o0.10–1.76 o0.10–0.70 o0.10–0.25 o0.1–1.78

Avg7std 1.0370.78 1.2070.38 0.5770.07 n.d. 0.6570.60

Sand (%) Range 2.1–100.0 1.5–20.3 4.8–10.1 55.5–95.1 6.6–17.3

Avg7std 28.4730.1 4.374.8 7.9771.5 72.3710.3 10.773.0

Silt (%) Range 0.0–91.3 71.4–89.4 85.6–91.2 4.6–41.5 74.3–84.0

Avg7std 67.8728.5 85.474.5 87.971.4 26.079.7 79.772.6

Clay (%) Range 0.0–7.9 7.5–13.2 2.9–8.2 0.3–3.0 7.9–10.6

Avg7std 3.872.0 10.371.6 4.271.4 1.670.7 9.570.8

Fig. 3. Maps of the organic carbon, Corg (a) and of total sulfur, S (b) content (%) distribution in the surface sediments of the Bay.

S. Covelli et al. / Continental Shelf Research 35 (2012) 29–42 33

The distribution pattern of S (0.05–3.45%, Fig. 3b) is also relatedto the finest component (o16 mm) of the bottom sediments(r¼0.882, n¼47, po0.001), and the highest values of S arefound along the inner coastline between the Estrela River andGuapimirim River.

Corg content showed a positive correlation with S content(r¼0.852, n¼47, po0.001), and the average C/S weight ratiocalculated for the surface sediment (2.570.9) overlaps with the2.870.8 range proposed by Berner (1982) to be characteristic

of ‘‘normal marine’’ sediments (Morse and Berner, 1995). Ingeneral, low C/S ratios are associated with higher sulfate reduction(Leventhal, 1983). C/S ratios in this study were lowest in thewestern and inner, river-influenced, areas of the Bay, as well as atthe Bay’s entrance. Riverine organic loading is responsible for theeutrophication of the estuaries, particularly in the western sector(Machado et al., 2008), leading to anoxic and even sulphidicwaters. On the other hand, the low C/S values found at the Bay’sentrance suggest that the higher reactivity of the marine-derived

Page 6: Spatial variation, speciation and sedimentary records of mercury in the Guanabara Bay (Rio de Janeiro, Brazil)

S. Covelli et al. / Continental Shelf Research 35 (2012) 29–4234

OM could enhance sulfate reduction under favorable circum-stances, although the intensity of the process is negligible whencompared to the inner Bay sediments. Although methyl-Hg occur-rence was not considered in this study, these evidences suggestthat those areas could be the most promising for investigatingin situ methyl-Hg production in the Guanabara Bay sediments, inaccordance with the general consensus (Compeau and Bartha,1985; Gilmour et al., 1998; Benoit et al., 2001) that Hg methylationin sediments is due to the activity of sulfate-reducing bacteria.

3.1.2. Spatial distribution of mercury

Mercury content in the surface sediments of Guanabara Bay(Fig. 4) ranges from o0.10 to 3.22 mg kg�1 (avg. 0.9170.79,n¼47). Concentrations are comparable to other Hg contaminatedcoastal environments in other parts of the world (e.g., Krom et al.,1994; Odzak et al., 2000; Trombini et al., 2003; Canario et al., 2005;Covelli et al., 2009). Values fall into the same range found byprevious research into the Bay’s surface sediments (Rebello et al.,1986; Barrocas and Wasserman, 1998) and core sediment profiles(Wasserman et al., 2000; Kehrig et al., 2003; Machado et al., 2002,2008). Both the main channel connecting the Bay with the ocean,and the eastern shoreline, show the lowest metal concentrations.

The distribution of Hg in surface sediments seems to bepartially associated with the organic fraction and, to a lesserextent, with sulfur rather than grain size. However, correlationcoefficients are not highly significant (r¼0.548 po0.001 for Corg,r¼0.462 po0.01 for S and r¼0.415 po0.01 for o16 mm, n¼47).Correlations become significant only for Hg concentrationso1 mg kg�1 (r¼0.727, po0.001 for Corg; r¼0.750, po0.001 forS; r¼0.735, po0.001 foro16 mm; n¼30, Fig. 5a–c). The similarcorrelation coefficients suggest that none of these variables playsa predominant role on Hg distribution in sediments of the leastpolluted areas, but there is a concurrent influence on it.

Fig. 4. Map of the amount of mercury (Hg, mg kg�1)

The presence of higher Hg contents (41 mg kg�1) in sedi-ments appears to be related to local factors, such as the proximityof contamination sources. In fact, some Hg ‘‘hotspots’’ were foundin the area northeast of Governador Island (stations 1, 4, 6, C13),where oil discharges have also been documented (Christensenet al., 2010), as well as along the shoreline of the urban area(stations B and C) where Rio de Janeiro Harbor and shipyards withwaste disposal sites have been indicated as potential sources ofcontamination (FEEMA, 1990).

3.1.3. Mercury speciation

One of the main concerns in Hg contaminated systems is todetermine whether Hg is mobile in the environment and its degreeof bioavailability. Guanabara Bay exhibits Hg contents whichcontinuously exceed the values suggested as potentially toxic forthe benthic community (ERL and ERM, 0.12 and 0.7 mg kg�1; Longet al., 1995). The SSE procedure was therefore applied to sixselected grab samples (1, 4, 9, B, C and C13; Fig. 1). This allowspredictions of the possible ecotoxicological effects of Hg to bemade on the basis of its mobility and bioavailability, rather thansolely from its total content within the sediment.

As reported in Table 2, Hg was mostly comprised of the Hg-efraction. However, significant levels of the Hg-o fraction were alsofound. In terms of percentage, Hg-e accounted for 96.772.2%(max. value: 98.6% in sample 9), while the OM related Hg-ofraction ranged from 1.1% to 6.7% (sample 4). The contribution ofthe more soluble species (Hg-wþHg-h) was found to be almostnegligible (0.0570.04%), with the lowest percentages associatedwith the second extraction step. The mercuric sulfide fraction(Hg-s) was always detected, with values r1.5%.

The Hg-w and Hg-h fractions represent Hg compounds (HgCl2,HgSO4 and HgO) which have a high solubility in the aqueousmedia and that are easily exchangeable with the water column.These fractions have been found to be fully dissolved in the

distribution in the surface sediments of the Bay.

Page 7: Spatial variation, speciation and sedimentary records of mercury in the Guanabara Bay (Rio de Janeiro, Brazil)

Fig. 5. The relationships between Hg and Corg (a), S (b) and o16 mm grain-size

fraction (c) in surface sediments of the Bay. Black circles emphasize samples with

Hg concentrations less than 1 mg kg�1.

Table 2Partitioning of Hg in the solid phase of the six selected bottom sediments

according to the selective extraction procedure. See text for explanation of single

fractions. Total Hg content and major geochemical parameters are also reported.

Samples

1 4 9 B C C13

Hg-w % 0.02 0.01 0.00 0.01 0.01 0.02

Hg-h % 0.09 0.02 0.00 0.00 0.01 0.01

Hg-o % 2.40 6.69 1.20 1.11 1.47 2.75

Hg-s % 1.46 0.44 0.15 0.39 0.54 1.03

Hg-e % 96.04 92.85 98.65 98.49 97.97 96.19

Hg mg kg�1 2.74 3.22 1.54 1.56 1.55 2.53

Ctot % 5.08 4.64 5.15 4.58 5.09 5.39

Corg % 4.53 4.54 4.79 4.42 4.29 5.07

N % 0.54 0.47 0.58 0.46 0.47 0.56

Clay % 3.2 5.6 3.1 3.7 4.0 2.8

Silt % 77.2 85.7 86.2 83.8 80.8 74.4

Sand % 19.6 8.7 10.7 12.5 15.3 22.8

S. Covelli et al. / Continental Shelf Research 35 (2012) 29–42 35

‘‘in vitro human stomach simulation’’ (Ruby et al., 1996; Bloomet al., 2003) and may also be considered as the main substrates forHg methylation, according to Ullrich et al. (2001). Their negligible

contribution (o0.1%) to the total Hg pool suggests that a pooravailability for the biotic compartment and a low methyl-Hgproduction should be expected. These results are significantlylower than those reported for similar estuarine environments(Covelli et al., 2009 and references therein).

The Hg-o fraction (avg. 2.672.1%), which includes Hg com-plexes with several organic compounds, including humic and fulvicacids, both living and dead biota and the relatively small fraction ofmethylated species, is positively correlated with clay (r¼0.777,n¼6, po0.07). A strict relationship between the methyl-Hg con-tent in sediments and the Hg-o fraction was reported by Bloomet al. (2003). However, due to the lack of methyl-Hg data andconsidering that each sediment type has a unique methylationpotential controlled by several factors (Ullrich et al., 2001), we canonly speculate about the importance of this fraction in the Bay’ssediments.

Hg-e represents the prevalent fraction in all the investigatedsamples. This fraction is considered a good estimate of the freeHg(0) present in the sediment matrix, although some analyticalinterferences with Hg(I), amorphous organo-sulfur and crystallineFe/Mn oxide phases could yield an overestimation (Bloom et al.,2003). The Hg-e fraction should be considered as strongly boundto the mineral lattice, and mostly immobilized in the sedimentmatrix. In previous studies, the presence of the Hg-e fraction wasfound in environments subjected to massive discharges of Hg byCAP plants (Bloom et al., 2003; Covelli et al., 2009). However, inGuanabara Bay, the only direct impact from the CAP could be site9 and other industrial sources, oil discharges and the presence ofsanitary landfills and waste disposal sites in the area probablycontribute to the total pool of Hg. There is no tool currentlyavailable to predict the biogeochemical behavior of Hg dischargedfrom these different sources.

Finally, the Hg-s fraction, which represents the less mobile andless available mineral bound Hg compounds (HgS, m-HgS, HgSeand HgAu), constitutes about 1% of total Hg in the analyzedsamples and occurred in the coarser fractions (r¼0.823, n¼6,po0.05), probably as a consequence of authigenic metacinnabarformation or the precipitation of Hg species with pyrite (Huerta-Diaz and Morse, 1992; Bower et al., 2008).

A previous study (Barrocas and Wasserman, 1998), conductedin the Acarı-S~ao Jo~ao de Meritı River system and its nearbyestuary area in front of Governador Island, found similar resultsin terms of the exchangeable fractions (Hg-w, Hg-h). On the otherhand, the same authors stated that more than 95% of total Hg wasbound to residual organic compounds. This is in contrast with ourfindings within the Bay, which highlight the predominant roleof the Hg-e fraction. Such discrepancy could be attributed to thedifferent choice of reagents employed for the SSE in the twostudies. The use of nitric acid was suggested by Bloom et al.(2003) as specific for the Hg-e phase, and not for organo-chelatedHg. Hence, the amount of Hg-o reported by Barrocas andWasserman (1998) represents the pool of Hg-e, following Bloom’sprocedure. Moreover, the presence of significant, but not elevated,percentages of Hg-s were also recorded by Barrocas andWasserman (1998), probably as a result of the fast complexationof weakly bound Hg to sulfides in marine sediments and theformation of metacinnabar.

3.2. Long sediment cores

3.2.1. Total and organic carbon, nitrogen and sulfur in the

sedimentary sequence

Surface levels in all the cores (Fig. 6) show the highest Ctot

content (about 4% at LC1, 45% at LC5 and 2–3% at LC7), whichdecreases almost exponentially with depth. The lowestCtot contents (0.18–1.23%) are found in core LC6 and they are

Page 8: Spatial variation, speciation and sedimentary records of mercury in the Guanabara Bay (Rio de Janeiro, Brazil)

Fig. 6. Vertical profiles of grain-size, total (Ctot) and organic carbon (Corg), total nitrogen (N) and sulfur (S), C/S and C/N ratios, aluminum (Al), iron (Fe), vanadium (V),

scandium (Sc) and mercury (Hg) in core LC1, LC5, LC6 and LC7.

S. Covelli et al. / Continental Shelf Research 35 (2012) 29–4236

essentially due to the coarser organic matter poor sediments,which are mainly made up of quartz.

Corg accounts for most of Ctot (470%, on average) in all thecores. Vertical profiles of Corg (Fig. 6a and b) are quite similar tothose observed for Ctot. and the decrease in concentration withdepth is probably due to the progressive degradation of organicmatter.

The same vertical trends can also be seen for Ntot (Fig. 6). Thehighest values are evident in the first 30 cm of all the cores (about0.5% at LC1 and LC5, 0.12% at LC6 and 0.22% at LC7). Corg/Ntot ratios(6.4–13.5; Fig. 6a and b) do not show a clear trend in core LC1, LC5and LC7 whereas an increase with depth is evident in core LC6.Corg/Ntot ratios can be modified during diagenesis because OMwithin the sediments is remineralized with different efficiencies

Page 9: Spatial variation, speciation and sedimentary records of mercury in the Guanabara Bay (Rio de Janeiro, Brazil)

S. Covelli et al. / Continental Shelf Research 35 (2012) 29–42 37

depending on its origin and the lability of the organic compounds(Aller, 1998; Aller and Blair, 2004; Hedges et al., 1997). However,the observed Corg/Ntot increase in core LC6 could be simply due tothe combined effect of increasing sediment grain size and thecorresponding decrease in Ntot.

Total S contents (Fig. 6) are generally higher in the upper coresections. Considering the C/S ratio in the sediment core profiles(Fig. 6), core LC1 shows rather constant values with depth, withan average of 1.5670.23. On the other hand, average values forthe other cores (LC5, LC6 and LC7) agree well, and fall in the 2.34–2.70 range, with higher values generally being observed in the topsediment layers, most evidently in core LC6. C/S ratios in thesedimentary sequence can give an indication on the paleo condi-tions of the depositional environment. Although ‘‘normal marinesediments’’ show C/S ratios generally close to the 2.870.8 valueproposed by Berner (1982), alternating freshwater and marinedeposits or highly variable inputs of organic matter (Morse andBerner, 1995) can yield considerably different C/S values. In coresLC5, LC6 and LC7, C/S values approach the previously mentioned‘‘normal marine’’ value, although LC6 does not fit this description,being characterized by a higher content of sand. Core LC1 showsmuch lower values, suggesting that rather different conditions arepresent at this site. Raiswell and Berner (1986) suggested that alowering of the C/S value occurs in euxinic basins, i.e. insediments overlain by anoxic bottom waters containing H2S,since organic carbon is not a limiting factor and more pyrite isformed per unit of buried Corg. The vertical variability of the C/Sratio in sediments of cores LC5, LC6 and LC7 could therefore beascribed to the resulting balance of autochthonous and allochtho-nous alternating OM sources, with core LC6 showing signs of aprobable shift towards more refractory OM in the more recentsediments. On the other hand, the very low C/S values found incore LC1 are indicative of a less efficient preservation of OM,probably as a result of a higher Corg lability and more anoxicconditions.

3.2.2. Mercury in the sedimentary sequence

Mercury contamination has been well recorded in fine bottomsediments sampled from the western sector of the Bay. Hg contentprofiles in LC1 and LC7 show a characteristic trend (Fig. 6) which isclearly due to anthropogenic inputs in the more recent history of thiscoastal environment, in contrast to cores LC5 and LC6 which aresampled from the eastern area of the Bay. Hg concentrations are quiteconstant (about 1.4 mg kg�1) in the first 30 cm of core LC1, showinga peak (1.76 mg kg�1) at a depth of 35 cm, and a sharp decreasedown to background values (o0.10 mg kg�1) in the basal coresection. An exponential decrease in the Hg concentration from thesubsurficial peak (1.78 mg kg�1 at a depth of 2–4 cm) is evident incore LC7. Background values are reached at similar core depths towhere they are found in LC1 (34–36 cm), probably reflecting smalldifferences in sedimentation rates between the two areas. A Hgprofile similar to that in core LC7 was found in one of the short cores(S1) collected by Wassermann et al. (2000) in front of the Acar�ı-S~aoJo~ao de Meritı River mouth. A peak concentration of less than4 mg kg�1 was found at a depth of 3 cm.

The Hg profile in core LC5 is rather peculiar (Fig. 6). Hgconcentration ranges from 0.44 to 0.70 mg kg�1 down to a depthof 170 cm, before reaching the background level in the basal coresection. Taking into account the sampling location, it can behypothesized that this profile is due to active physical reworkingof the sediment. No information can be obtained from the coarsersediment of core LC6, where Hg content is lower than the limitof detection (0.10 mg kg�1) along the whole profile, except fortwo levels, at depths between 6 and 10 cm, where Hg content is0.24–0.25 mg kg�1.

No correlation is observed between the concentrations of Hgand fine grained particles (o63 and o16 mm), due to the scarcegrain-size variability within the sedimentary sequence in coresLC1, LC5 and LC7. Corg content does, however, show a highcorrelation with Hg distribution in cores LC1 (r¼0.894, n¼12,po0.001) and LC7 (r¼0.872, n¼14, po0.001), probably as aconsequence of the greater OM deposition in the upper sedimentlayers of these cores. Total S showed a relatively good correlationwith Hg only in cores LC1 (r¼0.783, n¼12, po0.001) and LC7(r¼0.731, n¼14, po0.01).

3.3. Normalization procedure and mercury enrichment factor

A common approach to assessing whether anomalous metalcontributions are present in the sediment, is to normalize metalconcentrations for grain size and mineralogy effects. The normal-izing element is usually an important constituent of one or moreof the major trace metal carriers, and reflects their grain-sizevariability in the sediments such as Al (Covelli and Fontolan,1997; Rubio et al., 2000), Li (Loring, 1990) or Fe (Sinex and Wright1988; Rule 1986; Baptista Neto et al., 2005).

In the entire set of grab and core samples, the best correlationwas observed between Fe and the o16 mm component (r¼0.953;po0.001), but a close relationship is also seen for the other proxyelements (Sc, r¼0.952; Al, r¼0.937; Ti, r¼0.735; po0.001),reflecting their coexistence in the crystalline structure of alumi-nosilicate minerals. Vanadium was highly correlated to theo63 mm content (r¼0.942; po0.001), whereas La and Thshowed the strongest correlation with the o2 mm fraction ofthe sediment (r¼0.842 and 0.825, respectively, with po0.001).

Based on these observations, Fe was chosen as the mostsuitable proxy for the calculation of the Hg enrichment factor(EF) in surface samples. The EF was obtained by dividing the ‘‘Hgto normalizing element’’ ratio in the sample by the same ratiofound in the baseline, as follows:

EF¼(Hg/Fe)sample/(Hg/Fe)baseline

The EF ratio denotes no enrichment or depletion relative to thebaseline when it has a value close to 1, but indicates an anthro-pogenic contribution when it has a value greater than 1 (e.g. Covelliand Fontolan, 1997; Rubio et al., 2000; Devesa-Rey et al., 2010).Five degrees of contamination are commonly defined (Sutherland,2000): EFo2, deficiency to low enrichment; EF 2–5, moderateenrichment; EF 5–20, significant enrichment; EF 20–40, very highenrichment; and EF440, extremely high enrichment.

The Hg/Fe ratio calculated for the local baseline (0.041) wasassumed to be the average found in the basal sections of the longcores, due to the natural concentrations of the two elements.These basal sections, at depths ranging from 184 to 262 cm,consist of muds, which were not dated, but are assumed tocorrespond to sediment deposition in the pre-industrial period.In fact, Baptista Neto et al. (1996) dated muds from the basal partof some cores collected in the Jurujuba Embayement (harbor areaof Niteroi) as in excess of 3000 years. Although EF valuesgenerally reflect the spatial trend of Hg concentrations, severalsites of anomalous Hg accumulation due to local sources (Fig. 7)were highlighted by the EF plot compared to the metal contents.

Surface sediments reveal the highest EF (420) at samplingpoints close to Governador Island (stations 1, C13 and 4). Most ofthe area surrounding this island (stations 9, 10, U, V, C34)corresponds to high EF values relative to the rest of the Bay.This is unsurprising, as the main polluted rivers supply theirwater discharge and associated sewage to the Bay in this sector.In fact, the largest input of Hg in Guanabara Bay was the CAP in theAcarı-S~ao Jo~ao de Meritı River drainage basin, which began

Page 10: Spatial variation, speciation and sedimentary records of mercury in the Guanabara Bay (Rio de Janeiro, Brazil)

Fig. 7. Distribution of the mercury Enrichment Factor (Hg EF) in surface sediments of the Bay.

S. Covelli et al. / Continental Shelf Research 35 (2012) 29–4238

operation in the 50s with the use of Hg amalgam technology. Adrastic decline in Hg pollution followed the adoption of a Hg-freemembrane production process in the 70s (Godoy et al., 1998).However, lower but significant contributions of Hg were also reportedfor the Iguac-u River (Machado et al., 2002) and for the Estrela River(Rebello et al., 1986). The oil terminal located close to station 1 hasbeen cited in previous research as a potential source of other heavymetals—mostly Pb but also Zn and Cu (Baptista Neto et al., 2006). Asexpected, high EF values (10–20) are evident in front of the Rio deJaneiro Harbor area (stations B, 40, C and D), due to the presence ofcontaminated sediments influenced by anthropogenic activities.

Another hot spot is located close to the entrance of JurujubaEmbayment and the harbor area of Niteroi (station C3), where thehigh enrichments of Pb, Zn, Cr and Ni in surface muds have beenattributed to the intense urbanization of the catchment basin(Baptista Neto et al., 2000). Only the central (stations 13, 14, 16,31, 32, G) and northeastern (stations S, R, P, N, 22) sectors of theBay appear not to be affected by high Hg enrichment (EFo5). Thehigh energy of tidal currents, and the prevailing coarser sedi-ments (not suitable for accumulation of contaminants), explainsthe low Hg enrichment seen in the central area of the Bay.Freshwater inputs from the Macacu River system, whose drainagebasin is affected by rural activities (Faria and Sanchez, 2001), donot seem to have supplied Hg to the sediments accumulating attheir mouths (stations P, N, 22), although a low contaminationdegree for the Environmental Protection Area of Guapimirim waspreviously reported (Rebello et al., 1986).

Normalization to Fe and EF values produced little change inthe shape of Hg concentration profiles with depth (Fig. 6). Peaksand troughs are not severely affected by grain-size variability,since the cores are lithologically very homogenous throughouttheir length. The degree of Hg enrichment at the core top appearsto depend on the location of the sampling sites, relative to thepunctual contamination sources (LC1, EF¼9.2; LC5, EF¼3.4; LC6,EF¼4.8; LC7, EF¼11). Maximum enrichments were achieved inthe recent past, since they are recorded in the sub-surface levels: at depths of 7, 17 and 35 cm for LC1 (EF¼9.3), at65 cm for LC5 (EF¼4.9) and only at 3 cm for LC6 (EF¼6.5) andLC7 (EF¼12).

3.4. Accumulation rate and mercury inventory

Depositional total Hg flux calculations for the sedimentarysequence are carried out in order to understand how thedynamics of metal accumulation in sediments varies with timeand in order to predict, if possible, the more recent trends. The Hgaccumulation rate (HgAR) in core sediments can be obtained fromthe following equation, taking into account the rate at whichsediment is accumulating at each site:

HgAR (mg m�2 yr�1)¼o � (Hg)s with o¼(1�j) �v �r

where o is the mass sedimentation rate (g m�2 yr�1), (Hg)s is theHg concentration (mg kg�1mg kg�1) in the sediment, j is

Page 11: Spatial variation, speciation and sedimentary records of mercury in the Guanabara Bay (Rio de Janeiro, Brazil)

Fig. 8. Estimated accumulation rate of mercury (Hg total flux) in LC1 and LC7 core

sediments with time (a) and mercury inventory (Hg Inv) with time (b) in the same

cores, where contribution calculated for each sediment core level is reported

divided for the corresponding time interval.

S. Covelli et al. / Continental Shelf Research 35 (2012) 29–42 39

porosity, v is the sedimentation rate (mm yr�1) and r is thesediment density (g cm�3).

On the basis of 210Pb activity, Wilken et al. (1986) estimated asedimentation rate of 2 cm yr�1 for the north–west region ofGuanabara Bay. More recently, Godoy et al. (1998) determinedpresent sedimentation rates for several locations in the Bay andreported values ranging from 0.86 to 2.2 cm yr�1 (1.6 cm yr�1 onaverage). In agreement with the first sedimentation rates assessedfor the Bay (Amador, 1980), Godoy et al. (1998) found the highestrates to occur in the top section of the sedimentary sequencerecovered close to the fluvial sources of the Acarı-S~ao Jo~ao deMeritı and Iguac-u Rivers. Amador (1980) stated that the highestrates correspond to the most recent period (beginning in thelate 50s, when industrial development significantly increased),and that they are due to modifications in the water circula-tion pattern, which occurred as a consequence of landfilling ofchannels and small inlets.

In our HgAR determination, the previously mentioned averagesedimentation rate was used only for core LC7, whereas for theother cores, rates were applied according to the location of ourcores relative to the sampling points reported by Godoy et al.(1998; i.e. LC1¼P2, 22 mm yr�1, LC5¼P4, 1.5 mm yr�1 andLC6¼P5, 22 mm yr�1).

Present accumulation rates of Hg estimated for the Bay fromthe sampled cores range from about 1 mg m�2 yr�1 (LC5) to18 mg m�2 yr�1, found in core LC7, close to the Rio de JaneiroHarbor (Table 3). An intermediate HgAR of 8 mg m�2 yr�1 wasrecorded in LC1, north of Governador Island and in front of theIguac-u and Estrela rivers mouths. The minimum rate of HgAR isless than 1 mg m�2 yr�1 in LC6, at the western side of the Bay.

Not considering LC5, which, as previously mentioned, wasaffected by physical reworking, and LC6, temporal variations inHg loading increase towards the present day in LC1 and LC7(Fig. 8a). The onset of the increase of Hg flux dates to 1963 in coreLC1 and to ca. 1980 in core LC7, respectively. However, whereastrends are similar for the two sites from 1950s to ca. 2000, afterthat date they diverge (Fig. 8a). HgAR increases sharply in LC7,whereas it does not show the same apparent trend in LC1. HighestHg loadings found at the top of the two cores suggest ongoingaccumulation of Hg into the Bay, which accounts for more thandouble near Rio de Janeiro Harbor. Mercury binding particleswhich deriving from the contaminated S~ao Jo~ao de Merit�ı Riverare transported by strong tidal currents in the channel betweenGovernador Island and the mainland during ebb tides (Kjerfveet al., 1997). However, considering the location of core LC7 withrespect to the S~ao Jo~ao de Meritı River mouth and the supposedHg pollution decline, due to both an effluent treatment systemand the adoption of a Hg-free process from the CAP, results seemto indicate the contribution of additional inputs. Mercury mayderive from other sources such as emissions from industrial andurban settlements and/or from contaminated sediment resuspen-sion during frequent dredging operations. For instance, the

Table 3Hg concentration peak (mg kg�1) and surface, peak and average (1950–2007) H

for each core in the period 1950–2007 are also reported for each sediment cor

Peak (mg kg�1) Estimated age (year) Hg total accumulation rate

Surface (mg m�2 y�1) Pea

LC1 1.43 1977 8 9

LC5 0.67 n.d. 1 2

LC7 1.78 2005 18 18

LC6 0.26 2003 1 1 ?

n.d.¼not determined.

reclamation works in the 1950s for the construction of the Fund~aoIsland affected irreversibly the water flow in the Fund~ao Channel,which received domestic and industrial effluents without ade-quate treatment. Barbosa et al. (2004), evaluating the proposeddredging strategy for the revitalization of the channel, reportedHg concentrations (2.3 mg g�1) for the contaminated channelsediments comparable to what observed in core LC7. The poorwater quality of this area of the Bay was also confirmed byhighest Hg concentrations (380 mg kg�1) found in the suspendedparticulate matter (Kehrig et al., 2002).

We cannot exclude a significant contribution associated withevents of high atmospheric emissions from the industrial park ofRio de Janeiro metropolitan area. Lacerda et al. (2002) measuredhigh concentrations of Hg in bulk atmospheric deposition inNiteroi, since many industries (oil refineries, iron and steel plants)have Hg in their atmospheric effluents. In addition, the landfillrunoff (diluting and transporting leachate) and the resultingwastewater overflow from landfills to the Bay probably affectedthe metal input in recent times (Machado et al., 2002). This lastcase along with metal contributions from the Iguac-u and Estrelarivers, mainly affected by domestic wastes (da Silveira et al.,2011), would explain the relatively constant Hg flux observed inLC1 core for the last decade.

Since the cumulative Hg inventory quantifies the total amountof the metal loaded to the sediments over time (Kolak et al.,

g accumulation rates in the sediment cores; cumulative inventories of Hg

e sampling point.

Hg inventory (1950–2007)

k (mg m�2 y�1) Medium (mg m�2 y�1) mg m�2

5 91

2 26

8 86

? 15 ?

Page 12: Spatial variation, speciation and sedimentary records of mercury in the Guanabara Bay (Rio de Janeiro, Brazil)

S. Covelli et al. / Continental Shelf Research 35 (2012) 29–4240

1998), an assessment of the total Hg buried in the sedimentarycolumn from the beginning of the industrial period wasattempted as follows:

Hg inventory (mg m�2)¼S[(Hg)s � (1�j) �r � d]

where d is the thickness of sediment between two consecutiveanalyzed core levels.

The inventory represents the total loading of Hg, includingnatural background input, that was deposited on a 1 m2 surfacearea over the period of industrialization and, presumably, relatedcontamination (1950-present). In spite of the different currentaccumulation rates and the evolution of Hg flux with time, theresults show a slight difference between cumulative Hg inventoryin core LC1 (91 mg m�2) and in core LC7 (86 mg m�2). Due to thefeatures of the other two cores, only a rough estimate of the Hginventory can be made: approximately 26 mg m�2 for LC5 coreand approximately 15 mg m�2 for LC6.

The amount of Hg buried at LC1 site is comparable to thatobtained by Machado et al. (2008) in front of the Iguac-u Rivermouth, but it is almost ten-folds lower than the site closest to theS~ao Jo~ao de Merit�ı River investigated by the same authors. Thisfact confirms that, on the whole, the western side of the Bay is asink for the metal. Mercury has accumulated with the same orderof magnitude but following different time intervals (Fig. 8b) as aconsequence of the several existing contamination sources pre-viously discussed. This can also be seen if the Hg inventorycalculated for each sediment core level is reported, divided forthe corresponding time interval (Fig. 8b). In core LC1 there is ageneral constant contribution along the core followed by a slightdecrease towards the core top from the year 2005. Conversely, incore LC7, the increase in the Hg inventory is evident for the lastdecade, which accounts for up to 10 times the amount of Hgburied in the previous fifty years.

4. Conclusions

Mercury distribution in Guanabara Bay seems to be dependenton different factors. Where Hg content is lower than 1 mg kg�1,the organic component, along with sulfur rather than grain-size,seems to affect Hg occurrence in sediments. Conversely, wherethe metal contents are higher, accumulation in surface sedimentsis mostly related to the presence of nearby contaminationsources, such as industries and urban settlements in the westernsector of the Bay and around Governador Island. The reduction inHg accumulation in bottom sediments expected as a consequenceof the adoption of contamination control policies (i.e. Hg-freetechnologies in the CAP and sewage treating facilities), has notbeen clearly observed in the core profiles. Given the scale of Hgcontamination in the present sediments and considering themultiple potential direct (industries) and indirect (dredging,landfilling) sources of Hg for the Bay, a natural attenuation ofHg concentrations close to background levels in the near future isnot predictable.

Although total Hg contents in surface sediments exceed thevalues suggested by the effects-based standard quality guide-lines as potentially toxic for the benthic community, the con-tribution of the more soluble, easily exchangeable and,eventually bioavailable, species of Hg was found to be almostnegligible. Similarly, the contribution of Hg associated with theorganic fraction, reported in the literature to be positivelycorrelated to methyl-Hg, was also small. Most of the metal isbound to the Hg-e fraction, which is normally due to CAPdischarges and should be considered almost immobilized sincethese compounds are strongly bound to the mineral lattice of thesedimentary matrix.

It appears that the risk of Hg remobilization from sediments intothe water column under undisturbed conditions is low. However,the possibility of accidental resuspension events, which could havestrong effects on the redox conditions important to the biogeo-chemical fate of Hg, cannot be ruled out. Additional research, suchas direct measurement of methylation–demethylation rates, couldbe important to the consideration of euthropication events, whichmainly occur in the western sector of the Bay, and are oftenassociated with the hypoxic/anoxic conditions that enhance bothsulfate reduction and methylation rates. Experimental investigationinto the effects of physical resuspension should be promoted in thefuture, since sediments still remain a potential contamination sourcefor the whole aquatic trophic chain.

Acknowledgments

This study was conducted as part of the ‘‘TAGUBAR’’ (TAngentialGUanabara Bay Aeration Recovery) project framework, an inter-disciplinary cooperation project between Italy (Ministry for ForeignAffairs of Italy and General Direction for Development Coopera-tion) and Brazil (Agencia Brasileira de Cooperac- ~ao Governo Federaldo Brasil) coordinated by the University ‘‘Ca’ Foscari’’ of Venice(resp. G. Perin) and the UERJ, Universidade do Estado do Rio deJaneiro (resp. L. Verc-osa Carvalheira). The authors are gratefulto Daniela Bottos, Andrea Mao, Silvia De Pieri, Giancarlo Arcarifor their valuable help in field and subsampling operations.Special thanks to Mauro Bussi for grain-size analyses and CristianoLanducci for C and N analyses and graphics support. The authorsare also indebted to the personnel of the Institute of Oceanographyof the UERJ for kind assistance during lab work.

References

Aller, R.C., 1998. Mobile deltaic and continental shelf muds as suboxic, fluidizedbed reactors. Marine Chemistry 61, 143–155.

Aller, R.C., Blair, N.E., 2004. Early diagenetic remineralization of sedimentaryorganic C in the Gulf of Papua deltaic complex (Papua New Guinea): net loss ofterrestrial C and diagenetic fractionation of C isotopes. Geochimica etCosmochimica Acta 68, 1815–1825.

Amador, E.S., 1980. Assoreamento da Baia de Guanabara—Taxas de Sedimentac- ~ao.Anais da Academia Brasileira de Ciencias 52, 723–742.

Amador, E.S., 1997. Baia de Guanabara e ecosistemas perifericos—Homen eNaturaleza. Reproarte Grafica e Editora, Rio de Janeiro, p. 529.

Baldi, F., Bargagli, R., 1984. Mercury pollution in marine sediments near a chlor-alkali plant: distribution and availability of the metal. Science of the TotalEnvironment 39, 15–26.

Baptista Neto, J.A., Smith, B.J., McAllister, J.J., 1996. Sedimentological evidence ofhuman impact on a nearshore environment: Jurujuba Sound, Rio de JaneiroState, Brazil. Applied Geography 19, 153–177.

Baptista Neto, J.A., Smith, B.J., McAllister, J.J., 2000. Heavy metal concentrations insurficial sediments in a nearshore environment, Jurujuba Sound, SoutheastBrazil. Environmental Pollution 109, 1–9.

Baptista Neto, J.A., Crapez, M., McAllister, J.J., Vilela, C.G., 2005. Concentration andbioavailability of heavy in sediments from Niteroi Bay (Guanabara Bay/S.E.Brazil). The Journal of Coastal Research 21, 811–817.

Baptista Neto, J.A., Gingele, F.X., Leipe, T., Brehme, I., 2006. Spatial distribution ofheavy metals in surficial sediments from Guanabara Bay: Rio de Janeiro, Brazil.Environmental Geology 49, 1051–1063.

Barbosa, M.C., Almeida, M., Mariz, D.F., Almeida, J.L.D., 2004. Studies of channelsediments contaminated with organic and heavy metals. Journal of HazardousMaterials 110, 29–38.

Barrocas, P.R.G., Wasserman, J.C., 1998. Mercury behaviour in sediments from asub-tropical coastal environment in SE Brazil. In: Wasserman, J.C., Silva-Filho,E.V., Villas-Boas, R. (Eds.), Environmental Geochemistry in the Tropics, vol. 72,Springer-Verlag, Heidelberg, Lecture Notes in Earth Sciences, pp. 171–184.

Benoit, J.M., Gilmour, C.C., Mason, R.P., 2001. The influence of sulfide on solid-phase mercury bioavailability for methylation by pure cultures of desulfobul-bus propionicus (1pr3). Environmental Science and Technology 35, 127–132.

Berner, R.A., 1982. Burial of organic carbon and pyrite sulfur in the modern ocean:its geochemical and environmental significance. American Journal of Science282, 451–473.

Bloom, N.S., Preus, E., Katon, J., Hiltner, M., 2003. Selective extractions to assessbiogeochemically relevant fraction of inorganic mercury in sediments andsoils. Analytica Chimica Acta 479, 233–248.

Page 13: Spatial variation, speciation and sedimentary records of mercury in the Guanabara Bay (Rio de Janeiro, Brazil)

S. Covelli et al. / Continental Shelf Research 35 (2012) 29–42 41

Bower, J., Savage, K.S., Weinman, B., Barnett, M.O., Hamilton, W.P., Harper, W.F.,2008. Immobilization of mercury by pyrite (FeS2). Environmental Pollution156, 504–514.

Cai, Y., Jaffe, R., Jones, R., 1997. Ethylmercury in the soils and sediments of theFlorida Everglades. Environmental Science and Technology 31, 302–305.

Canario, J., Vale, C., Caetano, M., 2005. Distribution of monomethylmercury andmercury in surface sediments of the Tagus Estuary (Portugal). Marine Pollu-tion Bulletin 50, 1121–1145.

Carreira, R.S., Wagener, A.L.R., Readman, J.W., Fileman, T.W., Macko, S.A., Veiga, A.,2002. Changes in the sedimentary organic carbon pool of a fertilized tropicalestuary, Guanabara Bay, Brazil: an elemental, isotopic and molecular markerapproach. Marine Chemistry 79, 207–227.

Carreira, R.S., Wagener, A.L.R., Readman, J.W., 2004. Sterols as markers of sewagecontamination in a tropical urban estuary (Guanabara Bay, Brazil): space-timevariation. Estuarine, Coastal and Shelf Science 60, 587–598.

Carvalhaes, G.K., Brooks, P., Azevedo, G.C., Azevedo, J.A.T., Machado, M.C.S.,Stringer, R., 2002. The status of sediments samples from Guanabara Bay Riode Janeiro: a continuous study considering PCDD/F and PCB’S contamination.Organohalogen Compounds 57, 357–360.

Christensen, J.H., Tomasi, G., de Lemos Scofield, A., Guadalupe Meniconi, M.F.,2010. A novel approach for characterization of polycyclic aromatic hydro-carbon (PAH) pollution patterns in sediments from Guanabara Bay, Rio deJaneiro, Brazil. Environmental Pollution 158, 3290–3297.

Compeau, G.C., Bartha, R., 1985. Sulfate-reducing bacteria: principal methylatorsof mercury in anoxic estuarine sediment. Applied and Environmental Micro-biology 50, 498–502.

Covelli, S., Fontolan, G., 1997. Application of a normalization procedure indetermining regional geochemical baselines: Gulf of Trieste, Italy. Environ-mental Geology 30, 34–45.

Covelli, S., Faganeli, J., Horvat, M., Brambati, A., 2001. Mercury contamination ofcoastal sediments as the result of long-term cinnabar activity (Gulf of Trieste,northern Adriatic Sea). Applied Geochemistry 16, 541–558.

Covelli, S., Faganeli, J., De Vittor, C., Predonzani, S., Acquavita, A., Horvat, M., 2008.Benthic fluxes of mercury species in a lagoon environment (Grado Lagoon,Northern Adriatic Sea, Italy). Applied Geochemistry 23, 529–546.

Covelli, S., Acquavita, A., Piani, R., Predonzani, S., De Vittor, C., 2009. Recentand past contamination of mercury in an estuarine environment (Maranolagoon, Northern Adriatic, Italy). Estuarine, Coastal and Shelf Science 82,273–284.

da Silveira, R.P., Rodrigues, A.P.d.C., Santelli, R.E., Cordeiro, R.C., Bidone, E.D., 2011. Massbalance in the monitoring of pollutants in tidal rivers of the Guanabara Bay, Rio deJaneiro, Brazil. Environmental Monitoring and Assessment 181, 165–173.

Devesa-Rey, R., Barral, M.T., Jouanneau, J.-M., Diaz-Fierros, F., 2010. Analyses of thedegree of contamination and evolution in the last 100 years of the compositionof the bed sediments of the Anllons Basin. Environmental Earth Sciences 61,1401–1417.

Emili, A., Koron, N., Covelli, S., Faganeli, J., Acquavita, A., Predonzani, S., De Vittor,C., 2011. Does anoxia affect mercury cycling at the sediment-water interface inthe Gulf of Trieste (northern Adriatic Sea)? Incubation experiments usingbenthic flux chambers. Applied Geochemistry 26, 194–204.

Faria, M.M., Sanchez, B.A., 2001. Geochemistry and mineralogy of recent sedi-ments of Guanabara Bay (NE sector) and its major rivers—Rio de JaneiroState—Brazil. Anais da Academia Brasileira de Ciencias 73, 121–133.

FEEMA, 1990. Projeto de recuperac- ~ao gradual do ecossistema da Baıa de Guana-bara. Fundac- ~ao Estadual de Engenharia do Meio Ambiente—FEEMA, Rio deJaneiro, vol. 2, p. 203.

Gilmour, C.C., Riedel, G.S., Ederington, M.C., Bell, J.T., Gill, G.A., Stordal, M.C., 1998.Methylmercury concentrations and production rates across a trophic gradientin the northern Everglades. Biogeochemistry 40, 327–345.

Godoy, J.M., Moreira, I., Branganc-a, M.J., Wanderley, C., Mendes, L.B., 1998. A studyof Guanabara Bay sedimentation rates. Journal of Radioanalytical and NuclearChemistry 227, 157–160.

Goni, M.A., Hedges, J.I., 1995. Sources and reactivities of marine-derived organicmatter in coastal sediments as determined by alkaline CuO oxidation.Geochimica et Cosmochimica Acta 59, 2965–2981.

Goni, M.A., Thomas, K.A., 2000. Sources and transformations of organic matter insurface soils and sediments from a tidal estuary, North Inlet, South Carolina,USA. Estuaries 23, 548–564.

Goni, M.A., Teixeira, M.J., Perkey, D.W., 2003. Sources and distribution of organicmatter in a river-dominated estuary (Winyah Bay, SC, USA). Estuarine, Coastaland Shelf Science 57, 1023–1048.

Hedges, J.I., Stern, J.H., 1984. Carbon and nitrogen determinations in carbonate-containing solids. Limnology and Oceanography 29, 657–663.

Hedges, J.I., Keil, R.G., Benner, R., 1997. What happens to terrestrial organic matterin the ocean? Organic Geochemistry 27, 195–212.

Huerta-Diaz, M.A., Morse, J.W., 1992. Pyritization of trace metals in anoxic marinesediments. Geochimica et Cosmochimica Acta 56, 2681–2702.

Johnson, T.C., Evans, J.E., Eisenreich, S.J., 1982. Total organic carbon in LakeSuperior sediments: comparison with hemipelagic and pelagic marine envir-onments. Limnology and Oceanography 15, 958–960.

Kaplan, D.I., Knox, A.S., Myers, J., 2002. Mercury geochemistry in wetland and itsimplications for in situ remediation. Journal of Environmental Engineering128, 723–732.

Kehrig, H.A., Costa, M., Moreira, I., Malm, O., 2002. Total and methylmercuryin a Brazilian estuary, Rio de Janeiro. Marine Pollution Bulletin 44,1018–1023.

Kehrig, H.A., Pinto, F.N., Moreira, I., Malm, O., 2003. Heavy metals and methylmer-cury in a tropical coastal estuary and mangrove in Brazil. Organic Geochem-istry 34, 661–669.

Kjerfve, B., Ribeiro, C.H.A., Dias, G.T.M., Filippo, A.M., Da Silva Quaresme, V., 1997.Oceanographic characteristics of an impacted coastal bay: Ba�ıa de Guanabara,Rio de Janeiro, Brazil. Continental Shelf Research 17, 1609–1643.

Kjerfve, B., de Lacerda, L.D., Dias, G.T.M., 2001. Baia de Guanabara, Rio de Janeiro,Brasil. In: Seeliger, U., Kjerfve, B. (Eds.), Coastal Marine Ecosystemof Latin America. Ecological Studies, vol. 144. , Springer-Verlag, Berlin,pp. 107–117.

Kolak, J.J., Long, D.T., Beals, T.M., 1998. Anthropogenic inventories and historicaland present acuumulation rates of copper in Great Lakes sediments. AppliedGeochemistry 13, 59–75.

Krom, M.D., Kaufman, A., Hornung, H., 1994. Industrial mercury in combinationwith natural 210Pb as time-dependent tracers of sedimentation and mercuryremoval from Haifa Bay, Israel. Estuarine, Coastal and Shelf Science 38,625–642.

Lacerda, L.D., Fernandez, M.A., Tanizaki, K.F., Calazans, C.F., 1992. Bioavailability ofheavy metals in sediments of two coastal lagoon in Rio de Janeiro, Brazil.Hydrobiologia 228, 65–70.

Lacerda, L.D., Paraquetti, H.H.M., Rezende, C.E., Silva, L.F.F., Filho, E.V.S., Marins,R.V., Ribeiro, M.G., 2002. Mercury concentrations in bulk atmospheric deposi-tion over the coast of Rio de Janeiro, Southeast, Brazil. Journal of BrazilianChemical Society 13, 165–169.

Leventhal, J.S., 1983. An interpretation of carbon and sulfur relationships in BlackSea sediments as indicators of environments of deposition. Geochimica etCosmochimica Acta 47, 133–138.

Long, E.R., MacDonald, D.D., Smith, S.L., Calder, F.D., 1995. Incidence of adversebiological effects within ranges of chemical concentrations in marine andestuarine sediments. Environmental Management 19, 81–97.

Loring, D.H., 1990. Lithium—a new approach for the granulometric normalizationof trace metal data. Marine Chemistry 29, 155–168.

Loring, D.H., Rantala, R.T.T., 1992. Manual for the geochemical analyses of marinesediments and suspended particulate matter. Earth-Science Reviews 32,235–283.

Machado, W., Moscatelli, M., Rezende, L.G., Lacerda, L.D., 2002. Mercury, zinc, andcopper accumulation in mangrove sediments surrounding a large landfill inSoutheast Brazil. Environmental Pollution 120, 455–461.

Machado, W., Santelli, R.E., Loureiro, D.D., Oliveira, E.P., Borges, A.C., Ma, V.K.,Lacerda, L.D., 2008. Mercury accumulation in sediments along an eutrophica-tion gradient in Guanabara Bay, Southeast Brazil. Journal of Brazilian ChemicalSociety 19, 569–575.

Morse, J.W., Berner, R.A., 1995. What determines sedimentary C/S ratios? Geochi-mica et Cosmochimica Acta 59, 1073–1077.

Odzak, N., Zvonaric, T., Kljakovic Gaspic, Z., Horvat, M., Baric, A., 2000. Biomonitor-ing of mercury in the Kastela Bay using transplanted mussels. Science of theTotal Environment 261, 61–68.

Perin, G., Fabris, R., Manente, S., Rebello Wagener, A., Hamacher, C., Scotto, S.,1997. A five-year study on the heavy-metal pollution of Guanabara Baysediments (Rio de Janeiro, Brazil) and evaluation of the metal bioavailabilityby means of geochemical speciation. Water Research 31, 3017–3028.

Porcella, D.B., 1994. Mercury in the environment: biogeochemistry. In: Watras,C.J., Huckabee, J.W. (Eds.), Mercury Pollution: Integration and Synthesis, LewisPublishers, Boca Raton, Florida (p. 727).

Raiswell, R., Berner, R.A., 1986. Pyrite and organic matter in Phanerozoic normalmarine shales. Geochimica et Cosmochimica Acta 50, 1967–1976.

Rebello, A. deL., Haekel, W., Moreira, I., Santelli, R., Schroeder, F., 1986. The fateof heavy metals in an estuarine tropical system. Marine Chemistry 18,215–255.

Rego, V.S., Pfeiffer, W.C., Barcellos, C.C., Rezende, C.E., Malm, O., Souza, C.M.M.,1993. Heavy metal transport in the Acarı-S~ao Jo~ao de Meritı River System,Brazil. Environmental Technology 14, 167–174.

Rubio, R., Nombela, M.A., Vilas, F., 2000. Geochemistry of major and trace elementsin sediments of the Ria de Vigo (NW Spain): an assessment of metal pollution.Marine Pollution Bulletin 40, 968–980.

Ruby, M.W., Davis, A., Link, T.E., Schoff, R., Eberle, S., Sellstone, C.M., 1996.Estimation of lead and arsenic bioavailability using a physiologically basedextraction test. Environmental Science and Technology 30, 422–430.

Rule, J.P., 1986. Assessment of trace element geochemistry of Hampton RoadsHarbor and lower Cheseapeake Bay area sediments. Environmental Geologyand Water Sciences 8, 209–219.

Schluter, K., 2000. Review: evaporation of mercury from soils. An integration andsynthesis of current knowledge. Environmental Geology 39, 249–271.

Shepard, F.P., 1954. Nomenclature based on sand–silt–clay ratios. Journal ofSedimentary Petrology 24, 151–158.

Shi, J., Liang, L., Jiang, G., Jin, X., 2005. The speciation and bioavailability ofmercury in sediments of Haihe River, China. Environment International 31,357–365.

Sinex, S.A., Wright, D.A., 1988. Distribution of trace metals in the sediments andbiota of Chesapeake Bay. Marine Pollution Bulletin 19, 425–431.

Sutherland, R.A., 2000. Bed sediment-associated trace metals in an urban stream,Oahu, Hawaii. Environmental Geology 39, 611–627.

Trombini, C., Fabbri, D., Lombardo, M., Vassura, I., Zavoli, E., Horvat, M., 2003.Mercury and methylmercury contamination in surficial sediments and clamsof a coastal lagoon (Pialassa Baiona, Ravenna, Italy). Continental Shelf Research23, 1821–1831.

Page 14: Spatial variation, speciation and sedimentary records of mercury in the Guanabara Bay (Rio de Janeiro, Brazil)

S. Covelli et al. / Continental Shelf Research 35 (2012) 29–4242

Ullrich, S.M., Tanton, T.W., Abdrashitova, S.A., 2001. Mercury in the aquaticenvironment: a review of factors affecting methylation. Critical Reviews inEnvironmental Science and Technology 31, 241–293.

Wasserman, J.C., Freitas-Pinto, A.A.P., Amoroux, D., 2000. Mercury concentrationsin sediments profiles of a degraded tropical coastal environment. Environ-mental Technology 22, 297–305.

Wentworth, C.K., 1922. A scale of grade and class terms for clastic sediments.Journal of Geology 30, 377–392.

Wilken, R.D., Moreira, I., Rebello, A., 1986. 210Pb and 137Cs fluxes in a sediment

core from Guanabara Bay, Brazil. Science of the Total Environment 58,195–198.