ARTICLE IN PRESS Applied Geochemistry xxx (2005) xxx–xxx www.elsevier.com/locate/apgeochem Mercury contamination in Marano Lagoon (Northern Adriatic sea, Italy): Source identiWcation by analyses of Hg phases RaVaella Piani a, Stefano Covelli a a,¤ , Harald Biester b Department of Geological, Marine and Environmental Sciences, University of Trieste, Via Weiss, 2, 34127 Trieste, Italy b Institute of Environmental Geochemistry, INF 236, University of Heidelberg, 69120 Heidelberg, Germany Received 6 September 2004; accepted 5 April 2005 Editorial handling by W.B. Lyons Abstract Total Hg concentrations and Hg speciation were determined in bottom sediments of Marano lagoon to investigate the consequences of Hg phases on Wsh farms and shellWsh cultivation areas. Mercury phases were separated into cinnabar (HgS) and non-cinnabar compounds, via a thermo-desorption technique, in surface and core sediments; both of which had been contaminated by industrial wastes and mining activity residues. The former are due to an industrial complex, which has been producing cellulose, chlor-alkali and textile artiWcial Wbres since 1940. Processing and seepage wastewaters, which were historically discharged into the Aussa-Corno river system and therefore into the lagoon, have been signiWcantly reduced since 1984 due to the construction of wastewater treatment facilities. The second source is the Isonzo River, which has been the largest contributor of Hg into the northern Adriatic Sea since the 16th century due to Hg mining at the Idrija mine (western Slovenia). Red cinnabar (HgS) derived from the mining area is mostly stable and insoluble under current environmental conditions. In contrast, organically bound Hg, such as Hg bound to humic acids, has the potential to be transformed into bioavailable Hg compounds (for example, methylmercury). The presence of the two Hg forms permitted each Hg source to be quantiWed. It also allowed the areas with the highest risk of Hg contamination from Hg-rich sediment to be identiWed; thus potentially avoiding the transfer of Hg from the sediment into the water column and eventually into living biota. The results show that Hg Enrichment Factors in bottom sediments exceed values of 10 and cinnabar dominates the central sector near the main tidal channel where tidal Xux is more eVective. Non-cinnabar compounds were found to be enriched in Wne grained material and organic matter. In fact, up to 98% of total Hg at the Aussa-Corno river mouth and in the inner margin of the basin occurred in an organic form. This evidence, combined with the high contents of total Hg (4.1–6.6
g g¡1 and EF > 10) measured in surface sediments, suggest that Hg in Marano lagoon is involved in biogeochemical transformations (e.g., methylation).  2005 Elsevier Ltd. All rights reserved. 1. Introduction ¤ Corresponding author. Tel.: +39 040 5582031; fax: +39 040 5582048. E-mail address: covelli@univ.trieste.it (S. Covelli). 0883-2927/$ - see front matter  2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.apgeochem.2005.04.003 The biogeochemistry of Hg has become one of the most important topics in environmental sciences in the last 4 decades. This is due, in part, to Hg dispersion on a global scale, its extreme toxicity at low levels, and its ARTICLE IN PRESS 2 R. Piani et al. / Applied Geochemistry xxx (2005) xxx–xxx tendency to bioaccumulate within food chains (Mastrine et al., 1999; WHO, 1990; Porcella, 1994; Salomons, 1997). The ecological and human health eVects of Hg are generally related to the environmental transformation of inorganic Hg to the more toxic methylmercury and biomagniWcation in ecological systems (Compeau and Bartha, 1985). Moreover, several studies have demonstrated that wetlands are important areas for Hg methylation (Kierfve, 1994; King et al., 2002; Lacerda and Gonçalves, 2001) and that SO4-reducing bacteria (SRB) are the major biological mediators of Hg methylation in sediments (Gilmour et al., 1992; King et al., 1999). Also, particulate and dissolved organic C play a signiWcant role in the biogeochemical cycling of Hg in the aquatic environment (Gill and Bruland, 1990; Mason et al., 1993; Stordal et al., 1996). Possible remobilisation of heavy metal compounds, and Hg in particular, from bottom sediments into the water column of coastal lagoon ecosystems may result in variable accumulation by the lagoon biota (Lacerda et al., 1992). The presence of Hg in aquatic systems is due to several inputs, such as atmospheric deposition (Marins et al., 1996), and/or mining and industrial discharges (Baldi and Bargagli, 1984; Hines et al., 2000; Biester et al., 2002). Speciation of Hg in the water column of coastal lagoons may be of great environmental signiWcance, since it will not only control the eventual transfer of Hg to coastal waters but also the capacity of water bodies to accumulate large amounts of Hg in their sediments. As such, it is important to determine the distribution and the forms of this metal in contaminated sediments in order to establish potential risks and appropriate remediation technologies (Kaplan et al., 2002). Sequential extraction techniques have been widely used in order to diVerentiate between anthropogenic and geogenic Hg sources in soils and sediments and to evaluate the potential for remobilisation of Hg into the aqueous phase (Wallschläger et al., 1995; Bloom et al., 2003). Alternatively, Hg-thermo-desorption techniques have been successfully applied to distinguish metallic Hg from Hg(II)-binding forms; for example, between humic acids and sulphides in Xuvial and marine sediments (Biester and Scholz, 1996; Biester et al., 2000, 2002; Sladek et al., 2002). The Po River and the industrial zone of the Venice Lagoon are the primary sources of Hg contamination in coastal sediments of the Northern Adriatic sea (Fabbri et al., 2001; Donazzolo et al., 1981). Further east, the Isonzo River supplies suspended material enriched in Hg to the Gulf of Trieste from sediments and tailings of the Idrija mining area (western Slovenia). Since late 1970 several investigations on coastal sediments of the Gulf of Trieste have reported high concentrations of total Hg (23.6–47.0
g g¡1), which exponentially decrease with distance from the Xuvial source (Faganeli et al., 1991; Covelli et al., 2001). The Isonzo river inputs are dispersing westward and out of the gulf, following the main cir- culation system, and are also contributing to the Hg load in bottom sediments and, most importantly, to biota of the adjacent Grado and Marano lagoons (Brambati, 2001). Biester et al. (2000) have shown that cinnabar is the predominant form of Hg at the Isonzo River mouth, whereas organically bound Hg forms, which are mainly associated with Wne particles, are subject to long range transport. In addition, the inXux of this element into the western part of the lagoon system was signiWcantly enhanced by an industrial complex located along a stream Xowing into the lagoon. The industrial plant has been producing cellulose, chlor-alkali and textil artiWcial Wbres since the 1940s (RFVG, 1991; Marocco, 1995; Brambati, 1997). The aim of this study was to determine and quantify the proportion of cinnabar (HgS) and non-cinnabar compounds by a thermo-desorption technique in recent sediments of the Marano lagoon. As the industrial complex is a major source of non-cinnabar Hg compounds, attention was speciWcally given to locating areas of prevalent organomercury accumulation in comparison to sulphide-bound Hg. The circulation system and the relationships between Hg compounds and sediment particle characteristics were also investigated in order to recognise sites of potentially bioavailable Hg and the transfer of these Hg species from lagoon sediments into the water column and eventually into biota. Bioaccumulation, biotransfer and bioconcentration of Hg in the local environment is of direct health concern, as bivalves (clams and mussels), and a large number of Wsh farms in the lagoon are of local economic importance. 2. Environmental setting The Marano and Grado lagoons (Fig. 1(a)) are part of the extensive system of transitional environments developing along the northern Adriatic coast from the Po river delta to the Isonzo river mouth in the Gulf of Trieste. The Marano and Grado lagoons extend for about 32 km, between the Tagliamento and Isonzo River deltas, reaching up to 5 km for a total area of 160 km2. The lagoon basin is characterised by semi-diurnal tidal Xuxes (65 and 105 cm mean and spring tidal range, respectively). Small rivers Xow into the lagoon, which drain waters coming from the spring line and those from irrigation canals, which are conveyed into the water scooping machines. Particulate matter from these streams is of secondary importance, restricted to areas surrounding the spring river mouths. Conversely, the primary source of suspended sediments arrives from the sea, as the contribution of river deltas and from erosion of the barrier islands. Dispersion of sediments into the lagoon is controlled by tidal Xuxes through tidal inlets (Brambati, 1970). ARTICLE IN PRESS R. Piani et al. / Applied Geochemistry xxx (2005) xxx–xxx 3 Fig. 1. Central sector (Buso basin) of the Marano and Grado lagoons (a) and location of sampling points (b). Circles indicate bottom samples and black squares show core sites. Fish farm sites are also reported. The sampling area is located in the Buso basin (Fig. 1(b)), which occupies the central position of the lagoon system and presently has a surface area of 45 km2. The width of the Buso inlet was modiWed and reduced to 450 m after the building of jetties from 1964 to 1969. Commercial shipping through the main channel is maintained by periodical dredging, which ensures a channel depth of approximately 8 m along its axis. The most important freshwater stream Xowing into the Buso basin is the Aussa-Corno River, which has a steady Xow in the order of 25–27 m3 s¡1 (RFVG, 1983). The suspended load reaches 8–10 mg L¡1 during normal run oV periods (Piani, 1992 unpublished data). The Zellina River, which is the second and less inXuential tributary of Buso basin, has a Xow rate of just 1–2 m3 s¡1 (RFVG, 1983). The general grain-size distribution observed for the whole lagoon is also found in the Buso basin (Piani and Covelli, 2000). Medium to Wne sands are only present in front of the Buso inlet, along the main channel and the two lateral secondary channels and, subordinately, at the lower limit of the tidal Xats. The Wne pelitic component is widely dispersed in the basin and varies from 52% to 100%. The silty fraction (52–2
m) normally comprises a higher percentage than the clay fraction (<2
m) in the pelitic component. In fact, the ratio Wne-silt/clay (2–16/ <2
m) identiWes areas aVected by freshwater inputs into the lagoon: a high ratio indicates an area of fresh water input and vice versa. The highest contents of Wne silt occur where the Zellina and Aussa-Corno rivers enter the lagoon. Moreover, areas aVected by fresh water inputs and the inner part of the basin show the highest values of organic C (0.2–3.5%) and C/N ratios (up to 14). This suggests that organic matter of prevalently continental origin is partially preserved in those sectors where reducing conditions (Eh 6 ¡300 mV) in bottom sediments occur due to limited Xuvial inXuence and ARTICLE IN PRESS 4 R. Piani et al. / Applied Geochemistry xxx (2005) xxx–xxx interchange of marine waters through tidal cycles (Brambati, 1997). This area of the lagoon has high vulnerability due to the substantial human impact caused by urban settlements and industrial activities. Porto Nogaro, located 1 km north of the Aussa-Corno mouth is the most important industrial site in the region. In addition, the Torviscosa industrial complex produced cellulose from cane (Arundo donax sp.) from the 1940s until 1992. It is estimated that about 20 kg day¡1 of Hg has discharged into the Aussa River since 1949 (when the chlor-alkali plant became operative). As reported in RFVG (1991), soda production was increased from 4500 tons in 1950 to 20,000 tons during the following 10 a. The decrease of Hg discharged amounted to 6–7 kg day¡1 in 1970 and it presumably stopped in 1984 when wastewater treatment systems were installed. At the present time, industrial eZuents Xow through submarine pipes that discharge processing eZuent into the open sea; about 6.5 km oVshore from the main tidal inlet. According to the report mentioned above, an estimated total amount of 186,000 kg of Hg was released into the river. It is likely that a signiWcant volume of this Hg reached the lagoon environment. This site is currently under intense investigation and some corrective actions (e.g., soil removal and groundwater treatment) are in progress. Although high levels of Hg in bottom sediments of the Gulf of Trieste revealed a very strong supply of suspended material from the Isonzo River due to the long-term mining activity in the Idrija region (Brambati, 2001; Covelli et al., 2001), no estimation of the contributions from this source to the total amount of Hg into the lagoon has been attempted. 3. Materials and methods A total of 41 bottom sediment samples were recovered by a Van Veen grab and two cores (sampling stations 27 and 40) were hand collected from both sides of the main navigation channel (Fig. 1(b)), which connects the tidal inlet to the Aussa-Corno river mouth. Organic debris and shelly fragments were removed by sieving (<2 mm) each sample prior to grain-size analysis. Subsamples from each core and grab sample were subjected to 24 h of H2O2 treatment to remove organic matter, and subsequently wet sieved using a 52-
m sieve to separate sand from the pelitic fraction. Additionally, pelite was treated with H2O2 (10 v/v) prior to being analysed by a Micromeritics Sedigraph 5000 ET. The organic C and total N were determined on 40 °C dried and homogenised samples, after acidiWcation with progressively increasing concentration of HCl (Hedges and Stern, 1984), using a Perkin–Elmer C–H–N Elemental Analyser at a combustion temperature of 920 °C. Sediments were digested with HNO3 and H2SO4 solution (IRSA-CNR, 1985) and total Hg was measured following a modiWed form of CV AAS procedure by using a Perkin–Elmer Mercury Hydride System coupled to an AAS Perkin–Elmer mod.380. Determination of Hg phases by solid-phase-Hgthermo-desorption is based on the speciWc thermal decomposition of Hg compounds from sub-samples of dry sediment at diVerent temperatures. A detailed description of the method is found elsewhere (Biester and Zimmer, 1998; Biester et al., 2000). Results are represented by Hg-thermo-desorption curves showing release of Hg(0) versus increasing temperature. Mercury desorbed from matrix-bound Hg, such as Hg associated with humic acids, is in general normally distributed, in contrast to cinnabar which is irregularly released during thermo-desorption. Mercury peaks were then quantiWed by peak integration. Since non-cinnabar and cinnabar peaks usually partially overlap, non-cinnabar Hg was calculated by doubling the Wrst half of the corresponding peak (150– 250 °C). Therefore, Hg bound in cinnabar was obtained as the diVerence between total Hg and non-cinnabar Hg. Geochemical data for the overall Buso basin surface area were processed using the Surfer software program (Kriging algorithm). 4. Results and discussion 4.1. Mercury occurrence in the lagoon sediments The high vulnerability of coastal lagoon ecosystems is related to the shallowness and the limited water exchange of these semi-enclosed basins. Due to these features, the residence time of water, sediments and, consequently, contaminants, in the lagoon is much longer than in river mouths and in open coastal areas. Heavy metals, and in particular, Hg accumulated in sediments may be subjected to burial and/or to biogeochemical processes, which aVect their distribution, speciation and bioavailability to the lagoon biota. The high potential risk of Hg remobilisation from sediments into the water column and its subsequent bioaccumulation in the trophic chain is also enhanced by the abundance of animal and plant species in this ecosystem. Human activities such as Wshing, collection of clams and mussels and cultivation of edible Wsh species in Wsh farms increase chances of Hg uptake by living organisms and, ultimately, by human residents in the surrounding area. Mercury levels in aquatic plants and organisms from the trophic chains (Brambati, 1997) of the Marano and Grado lagoons suggest that tissue Hg is correlated with bottom sediment Hg concentrations, which show a progressive westward decrease from about 10 mg kg¡1 (Grado lagoon) to 1 mg kg¡1 (Marano lagoon). However, diVerent concentrations of Hg found in aquatic ARTICLE IN PRESS R. Piani et al. / Applied Geochemistry xxx (2005) xxx–xxx 5 Fig. 2. Distribution of total Hg (a) in
g g¡1 and organic C (b) in % for the surface sediments of the lagoon basin. species are related to their trophic characteristics. For example, mussel (Mytilus) and common cockle (Cardium edule) which are Wlter feeders show low concentrations; higher concentrations were found in prawn (Palaemon) a coarse detritus feeder; whereas very high levels (average 0.5–0.7 mg kg¡1, maximum 2–3 mg kg¡1) were detected in the carnivorous goby (Zosterissessor ophiocephalus) and silverside (Atherina boyeri). The same high Hg anomalies were observed in the Wsh farms where edible Wsh (e.g., gilt-head bream, Sparus auratus and bass, Dicentrarchus labrax), which are at the top of the food chain, showed levels far higher (up to about 5 mg kg¡1) than the same species collected in the natural environment. Unfortunately, Hg biomagniWcation increases the total Hg levels in marine species of commercial value (Brambati, 2001), and the magniWcation process seems to be more eVective in the Wsh farms. Brambati (2001) also found high Hg concentrations in the hair of people having a Wsh-based ARTICLE IN PRESS 6 R. Piani et al. / Applied Geochemistry xxx (2005) xxx–xxx Table 1 Relative abundances of main textural and geochemical parameters in surface sediment samples Sample Sand (%) Silt (%) Clay (%) 2–16
m (%) C org. (%) Hg (
g g¡1) RP1 RP2 RP3 RP4 RP5 RP6 RP7 RP8 RP9 RP10 RP11 RP12 RP13 RP14 RP15 RP16 RP17 RP18 RP19 RP20 RP21 RP22 RP23 RP24 RP25 RP26 RP27 RP28 RP29 RP30 RP31 RP32 RP33 RP34 RP35 RP36 RP37 RP38 RP39 RP40 RP4A 0.7 2.5 16.4 0.4 1.5 3.3 1.0 3.0 19.2 13.4 11.2 5.6 98.3 47.8 30.5 29.3 8.7 0.9 0.4 0.5 5.0 30.8 4.4 6.1 5.6 7.4 9.6 18.0 16.9 12.0 16.2 3.1 2.1 8.7 21.1 18.6 5.0 7.2 13.6 5.3 0.2 71.3 70.5 62.6 70.1 80.5 75.7 72.0 72.0 55.8 64.1 71.8 74.4 1.7 33.7 36.0 36.2 67.3 62.1 61.6 65.5 71.5 42.2 70.6 67.9 71.4 74.1 65.4 57.0 59.1 66.0 62.8 74.9 79.4 73.8 60.9 56.4 63.0 76.8 57.4 71.2 66.8 28.0 27.0 21.0 29.5 18.0 21.0 27.0 25.0 25.0 22.5 17.0 20.0 0 18.5 33.5 34.5 24.0 37.0 38.0 34.0 23.5 27.0 25.0 26.0 23.0 18.5 25.0 25.0 24.0 22.0 21.0 22.0 18.5 17.5 18.0 25.0 32.0 16.0 29.0 23.5 33.0 42.0 51.0 45.5 51.5 49.5 31.5 46.0 33.0 36.5 35.5 24.0 28.0 0 23.5 39.5 42.0 36.0 48.5 39.5 40.0 28.5 32.0 36.5 32.5 32.0 43.5 30.0 32.0 31.0 27.5 27.0 31.0 33.5 22.5 24.0 33.0 41.0 36.0 38.0 44.5 58.5 1.92 2.18 1.81 3.06 2.51 1.61 2.62 1.81 1.29 1.40 1.28 1.33 0.21 0.75 1.17 1.22 1.68 3.48 2.06 2.08 1.29 1.66 1.53 1.41 1.59 1.64 1.36 1.60 1.30 1.27 1.23 1.42 1.60 1.04 1.07 1.41 2.21 1.64 1.63 2.03 3.20 3.61 3.52 3.51 4.70 4.77 5.48 4.62 3.98 3.03 3.66 3.69 3.57 0.13 3.07 3.53 3.89 5.73 4.89 5.24 5.62 4.14 5.63 6.19 4.95 4.56 4.09 3.45 4.57 2.74 4.11 3.20 6.58 3.21 3.55 4.81 2.14 4.50 3.57 3.57 4.69 5.50 diet (up to 20 mg kg¡1) and Hg levels showed a direct positive correlation with age. Furthermore, high levels of Hg were also found in birds, demonstrating that the top consumers in the area may bioaccumulate Hg. Results of textural and geochemical parameters of the study area reported by Piani and Covelli (2000) revealed that major and trace elements (Al, Fe, K and Zn, Cr and Ni) as well as organic C, are signiWcantly correlated with Wne particles (<16
m) and they reXect the distribution pattern of this textural component in the lagoon basin. The anomalously high concentrations of Hg occur near the Zellina and Aussa-Corno river mouths and on tidal Xats in the eastern part of the basin (Fig. 2(a) and Table 1). Due to the double source of contamination mentioned above, Hg concentrations in sediments, ranging between 0.1 and 6.6
g g¡1, are on average (4 § 1.2
g g¡1) far higher than natural background values proposed in previous studies for the northern Adriatic sea (0.1
g g¡1, Faganeli et al., 1991; 0.2
g g¡1, Covelli et al., 2001). Higher concentrations of organic C are found in the inner and western parts of the basin where bottom sediments are mostly pelites rich with Wne silt (Fig. 2(b)). The two short cores do not show noticeable grainsize variations, although a slight increase in the coarse sandy fraction was detected at the top of core 27. This core was collected in the western part of the basin where salt marshes are absent and tidal Xuxes are more eYcient ARTICLE IN PRESS 7 R. Piani et al. / Applied Geochemistry xxx (2005) xxx–xxx Table 2 Relative abundances of main textural and geochemical parameters in core samples Depth (cm) Sand (%) Silt (%) Clay (%) 2–16
m (%) C org. (%) Hg (
g g¡1) Core 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0–3 3–5 5–7 7–9 9–12 12–14 14–16 16–18 18–20 20–22 24–26 28–30 33–35 37–39 41–43 44–46 48–50 54–55 58–60 62–64 1.1 0.8 0.5 0.4 1.0 0.0 0.1 0.5 0.5 0.4 0.4 1.2 0.2 1.1 0.3 0.3 0.3 1.2 0.8 0.6 74.5 68.8 73.4 65.8 67.2 60.2 62.8 63.4 64.4 63.0 65.3 70.0 67.3 76.9 75.9 71.6 68.7 75.2 73.0 73.2 24.4 30.4 26.1 33.8 31.8 39.8 37.1 36.1 35.1 36.6 34.3 28.8 32.5 22.0 23.8 28.1 31.0 23.6 26.2 26.2 45.2 40.6 44.5 44.8 42.4 41.1 44.4 44.5 43.5 42.5 43.4 41.3 48.9 42.2 45.3 47.0 51.2 39.1 42.1 42.7 2.07 1.98 1.91 1.75 1.66 2.03 1.88 1.84 1.96 1.80 1.62 1.63 1.45 1.76 1.53 2.12 2.87 1.41 – – 5.80 7.19 7.43 8.41 6.55 9.84 5.99 4.23 4.13 3.58 3.55 5.13 3.40 3.42 2.09 2.14 1.10 0.52 0.63 0.41 Core 27 1 2 3 4 5 6 7 8 9 10 12 13 0–3 5–7 9–11 13–15 17–19 21–23 25–27 29–31 33–35 37–39 41–43 45–47 0.7 1.8 2.3 1.9 0.1 0.4 0.6 0.9 0.4 0.1 1.3 0.4 70.3 73.4 76.7 74.7 69.7 70.6 72.3 71.2 72.0 68.1 71.1 64.9 29.0 24.8 21.0 23.4 30.2 29.0 27.1 27.9 27.6 31.8 27.6 34.7 41.9 37.0 34.3 35.9 45.1 44.7 44.4 45.1 45.0 46.3 44.9 46.9 1.38 1.27 1.06 1.27 1.22 1.24 1.28 1.28 1.19 1.38 1.41 1.57 2.15 2.74 2.24 2.66 2.46 0.85 0.39 0.42 0.41 0.63 0.32 0.30 Fig. 3. Calculated contribution of cinnabar and non-cinnabar Hg compounds in surface sediment samples as estimated by peak integration. The cumulative bar charts provide the amount of both fractions related to total Hg content. ARTICLE IN PRESS 8 R. Piani et al. / Applied Geochemistry xxx (2005) xxx–xxx Fig. 5. Areal distribution of the amount of non-cinnabar compounds, expressed as percentage of the total Hg content in the surface sediments of the lagoon. Fig. 4. Typical Hg-thermo-desorption curves show surface sediment samples where Hg is predominantly present as non-cinnabar (sample 18) and cinnabar (sample 14) compounds, and where cinnabar and non-cinnabar compounds are almost equally mixed (sample 24). on the mudXats. In contrast, core 40 was sampled on a tidal Xat near a salt marsh area, which is typically a low energy environment. Major and trace elements do not show a signiWcant vertical trend along the core proWles (Piani and Covelli, 2000). The only exception is Hg that is detected at a constant value of about 2
g g¡1 in the Wrst 20 cm of core 27. Higher concentrations are found in core 40 where subsurface maximum is reached at 13 cm depth (9.9
g g¡1) and Hg decreases downcore approaching background values at 50 cm depth (Table 2). Many trace elements are known to be preferentially associated with Wne particles in aquatic systems. Therefore, a common approach is to normalise metal concentrations to account for grain-size variability prior to assessing the degree of contamination (Loring and Rantala, 1992). The degree of enrichment of a metal is then calculated by dividing its ratio to the normalising element by the same ratio found in the baseline (e.g., Middleton and Grant, 1990). A non-dimensional enrichment factor EF D (Me/N)sample/(Me/N)baseline is obtained, where Me is the concentration of the potentially enriched metal and N is the concentration of the normalising element. A value of unity denotes no enrichment or depletion relative to the pre-industrial datum. Iron was chosen as the normalising element due to its signiWcant correlation with the <16
m fraction (r D 0.87; p < 0.001) and the local baseline was assumed to be the average concentration of the two elements in the basal levels of the two cores (Febaseline D 2.11%, from Piani and Covelli, 2000). Mercury enrichment factors indicate that the centralsouthern sector of the basin along with the inner coastline extending from the Aussa-Corno river mouth westwards are preferential accumulation areas (EF > 10). In the latter, the higher Hg enrichment (maximum 18) can be explained by freshwater input, whereas in the Wrst ARTICLE IN PRESS R. Piani et al. / Applied Geochemistry xxx (2005) xxx–xxx 9 Fig. 6. Calculated contribution of cinnabar and non-cinnabar Hg compounds in core 27 and 40 sediment samples as estimated by peak integration. The cumulative bar charts provide the amount of both fractions related to the total Hg content downcore. area periodic dredging operations in the main channel could be considered as a factor that contributed to increase metal concentration in sediments. Also, dumping of excavated material in the mudXats, immediately adjacent to tidal channels, was a common procedure followed in the past. These operations were recently ceased due to introduction of severe national legislation rules. Therefore, subsequent reworking and dispersion of dredged sediments through tidal Xuxes cannot be excluded and EF areal distributions may reXect this complex situation. The constant Hg concentrations in the upper part of core 27 (Table 2) appear to be related to mixing eVects due to bioturbation in the Wrst 20 cm, which would have homogenised recent metal supplies. This evidence was conWrmed by X-ray radiography of the core, which shows concentrations of small tube-like burrows at the core top. Conversely, contamination in the upper section of core 40, extending downcore to about 50 cm, shows two apparent maxima (Table 2). The relevant diVerence between the two cores, in terms of thickness of contaminated sediment suggests two hypotheses. The core 40 proWle could be explained as the original sedimentary sequence, which was altered by anthropogenic activities, although dumping of excavated muds seems to be improbable due to the distance of the coring site from the main navigation channels. A second hypothesis, supported by 137Cs proWles in the same cores (Piani and Covelli, 2000), suggests that the two sampling sites are characterised by diVerent local sedimentation rates. This has also been reported for example by Pavoni et al. (1987) and Donazzolo et al. (1981) for the Venice lagoon and depends on variable energy hydrodynamic conditions from the lower to the upper mudXats. 4.2. Mercury phases in bottom sediments Results of Hg thermo-desorption measurements performed on bottom samples demonstrate that both cinnabar and non-cinnabar compounds are present in lagoon sediments (Fig. 3) and that the relative contribution to total Hg contents can be assessed. The speciWc Hg phase in the sediments is the most important factor in terms of potential bioavailability of Hg to the aquatic food chain ARTICLE IN PRESS 10 R. Piani et al. / Applied Geochemistry xxx (2005) xxx–xxx Fig. 7. Distribution of non-cinnabar compounds, expressed as percentage of the total Hg content, in the sedimentary sequence of core 40 and 27. Hg-thermo-desorption curves for each core level are also reported. rather than the total Hg concentration. Presence and abundance of the non-cinnabar component in bottom sediments of some sampling sites may identify those areas in the lagoon where Hg is present in chemical forms that may be more mobile and/or involved in methylation processes. Almost all samples show curves with two peaks. A Wrst maximum was detected between 230 and 260 °C, which represents Hg released from non-cinnabar compounds such as humic acids, according to the curves of standard Hg compounds reported by Biester et al. (2000). Similarly, the second peak (300–350°C) is related to red cinnabar. As noted by the same authors, several high narrow Hg peaks usually appear in the curves due to slight diVerences in the crystallinity and grain-size of cinnabar, which inXuences Hg release temperatures. Three types of Hg-thermo-desorption curves were obtained for surface sediments (Fig. 4). Just one sample taken upstream in the Aussa-Corno River is almost entirely characterised by the non-cinnabar peak. Due to the low temperature of this Wrst peak, Hg could be associated with organosulphides or occurs as metacinnabar as a result of absorption of Hg(II) or Hg(0), by the prevailing silty-clay fraction with higher speciWc surface area, and subsequent transformation. In contrast, thermo-desorption curves of samples (e.g., 14 and 36) collected near the lagoon inlet and in the eastern part of the basin show Hg release temperatures mostly related to red crystalline cinnabar. All other samples from the lagoon show curves with two peaks, thus representing an evident mixing of both Hg compounds in surface sediments as a result of tidal dynamics. The spatial distribution of non-cinnabar compounds (Fig. 5 and also Fig. 3) is clearly inXuenced by the freshwater inputs (AussaCorno and Zellina rivers). Non-cinnabar Hg is relatively more abundant, especially in the lower course of the tributary streams and in the lagoon just in front of the river mouths. This evidence conWrms that most of the Hg dis- ARTICLE IN PRESS R. Piani et al. / Applied Geochemistry xxx (2005) xxx–xxx 11 Fig. 8. Relations between non-cinnabar Hg and organic carbon, and the 2–16
m fraction. charged into the Aussa River from the industrial complex preferentially accumulates in the river itself and in the inner part of the lagoon. In contrast, cinnabar, although present at all locations, is mostly concentrated near the southern and eastern sectors of the basin. This fact testiWes to the role of tidal Xuxes in controlling dispersion and accumulation of cinnabar bound to suspended matter from Isonzo River and entering the lagoon through at least 3 of the 4 inlets. 4.3. Mercury phases in core sediments Occurrence of Hg phases in core 27 and 40 (Fig. 6) shows an almost equal distribution of cinnabar and noncinnabar compounds in the upper section of the sedimentary sequence. In contrast, Hg sulphides decrease downcore. Only non-cinnabar compounds are present in the basal part of the cores where Hg concentrations become very low (Fig. 7). The sediment core from the Gulf of Trieste (Covelli et al., 2001) was mostly composed of Wne grained sediments, which have recorded the historical evolution of Hg production at Idrija mine through an exponential increase of metal concentrations since 1800 (although mining activity began in 1496). Between 1850 and 1970 the amount of excavated ore increased as much as the Hg production although the ore grade decreased from several tens to less than 1%. This activity produced an increasing amount of mining residues along with an improved eYciency of the roasting processes, including intensive crushing of the ore, which would have signiWcantly increased the amount of Wne grained material in the mining residues (Biester et al., 2000). The large amount of Hg enriched Wne particles (silt-clay) transported by freshwaters of the Isonzo river into the Gulf of Trieste would ARTICLE IN PRESS 12 R. Piani et al. / Applied Geochemistry xxx (2005) xxx–xxx have also entered the lagoons following the anticlockwise circulation system characteristic of this part of the northern Adriatic sea and Xood tidal Xuxes. Comparing Hg compounds with other textural and geochemical parameters (Fig. 8), the best correlations were found between Hg and organic C (r D 0.83) and the 2–16
m fraction (r D 0.86), suggesting that both the Wne particles and organic matter play an important role in transferring and accumulating the metal in the lagoon sediments. Although Hg behaviour in this system depends on its chemical form introduced into the aquatic environment, Hg associated with industrial eZuents has been shown to be positively correlated with Wne grain sizes such as <63
m (Rae and Aston, 1981) or <20
m (Barghigiani et al., 1996); Hg also has a strong aYnity for organic matter (Smith and Loring, 1981; Baldi and Bargagli, 1984; Gagnon et al., 1997). In areas where chlor-alkali plants are responsible for Hg contamination of sediments, up to 70–90% of the total metal in sediments can be associated with easily oxidizable organic matter, which is easily degraded and may then be released from the substrate (Smith and Loring, 1981; Gagnon et al., 1997; Biester et al., 2002). Distinction between detrital Hg, i.e. cinnabar particles, settling in deltaic sandy sediments in front of the Isonzo river mouth and non-cinnabar compounds accumulated in the central sector of the Gulf of Trieste was made on the basis of correlation with the <16
m grain size fraction (Covelli et al., 2001). According to these authors, Hg is preferentially adsorbed to the Wnest particles in the form of Hg(II), Hg(I) and Hg(0), due the higher speciWc surface area of each particle; although the presence of micro-crystalline cinnabar cannot be excluded. forms as a consequence of recent industrial inputs. The distinction between the two Hg forms allows the areas at most risk from Hg remobilisation from bottom sediments to the water column and, possible interaction with lagoon biota to be identiWed. The main risk areas occur near some Wsh farms, which are protected from high energy regimes. For example, the areas of most concern are in front of the Zellina river mouth and the eastern part of the basin, as well as the whole Aussa-Corno River mouth. The relevant, although extremely variable, thickness of sediments contaminated by Hg, the high percentage of non-cinnabar compounds, and the extension of the basin make it diYcult to devise an appropriate, in situ, remediation strategy. Due to the importance of local Wsh farm production, the most suitable choice would be to concentrate the eVorts on the bottom sediments and remediation of restricted pools. For example, excavation and removal of the contaminated layer and a reduction of new Hg supplies bound to suspended matter through tidal Xuxes Xowing periodically into the Wsh farms could be the most Wtting procedures. Acknowledgements This work has been supported by Regione Autonoma Friuli-Venezia Giulia (Direzione Regionale dell’Istruzione e Cultura) in the project “Rimobilizzazione e biodisponibilità del mercurio in ambienti lagunari e Xuviali regionali” (resp. Prof. A.Brambati). A special thanks to Ben Gildfedder for his valuable editing contribution. References 5. Conclusions Marano and Grado lagoons have experienced signiWcant historic Hg inputs from mining and industrial point sources. The Buso basin, located in the central sector of this lagoon, has been severely aVected both by long-term Hg contamination through Isonzo river suspended load and by untreated eZuents discharged into the AussaCorno river from a chlor-alkali plant since 1950. This study has shown that the Hg-thermo-desorption technique is a useful method to discriminate between organically bound Hg and sulphide (principally cinnabar) bound Hg. Using this method, the relative contribution of each river system to the inXux of Hg into the lagoon can be determined. Cinnabar from Isonzo River dominates the central sector of the lagoon, near the main tidal channel where tidal Xux is more eVective. Non-cinnabar compounds (avg. 49 § 22%) were found to be enriched in Wne grained material and organic matter in the inner margin of the basin, particularly at the Aussa-Corno River mouth where up to 98% of total Hg is bound in non-cinnbar Baldi, F., Bargagli, R., 1984. Mercury pollution in marine sediments near a chlor-alkali plant: distribution and availability of the metal. Sci. Total Environ. 39, 15–26. Barghigiani, C., Ristori, T., Lopez Arena, J., 1996. Mercury in marine sediment from a contaminated area of the northern Tyrrhenian Sea: <20
m grain-size fraction and total sample analysis. Sci. Total Environ. 192, 63–73. Biester, H., Scholz, C., 1996. Determination of mercury binding forms in contaminated soils. Mercury pyrolysis versus sequential extractions. Environ. Sci. Technol. 31, 233–239. Biester, H., Zimmer, H., 1998. Solubility and changes of mercury binding forms in contaminated soils after immobilization treatment. Environ. Sci. Technol. 32, 2755–2762. Biester, H., Gosar, M., Covelli, S., 2000. Mercury speciation in sediments aVected by dumped mining residues in the drainage area of the Idrija mercury mine, Slovenia. Environ. Sci. Technol. 34, 3330–3336. Biester, H., Müller, G., Schöler, H.F., 2002. Binding and mobility of mercury in solid contaminated by emissions from chlor-alkali plants. Sci. Total Environ. 284, 191–203. Bloom, N.S., Preus, E., Katon, J., Hiltner, M., 2003. Selective extraction to assess the biogeochemically relevant fraction- ARTICLE IN PRESS R. Piani et al. / Applied Geochemistry xxx (2005) xxx–xxx ation of inorganic mercury in sediments and soils. Anal. Chim. Acta 479, 233–248. Brambati, A., 1970. Provenienza, trasporto e accumulo dei sedimenti recenti nelle lagune di Marano e di Grado e nei litorali tra i Wumi Isonzo e Tagliamento. Mem. Soc. Geol. It. 9, 281–329. Brambati, A., 1997. Metalli pesanti nelle lagune di Marano e Grado. Piano di studi Wnalizzato all’accertamento di sostanze persistenti nelle Lagune di Marano e Grado ed al loro risanamento. (RFVG) Direzione Regionale dell’Ambiente, Servizio dell’Idraulica, Trieste. Brambati, A., 2001. Coastal sediments and biota as indicators of Hg contamination in the Marano and Grado Lagoons. RMZ – Material Geoenviron. 48, 165–171. Compeau, G.C., Bartha, R., 1985. Sulfate reducing bacteria: principle methylators of mercury in anoxic estuarine sediments. Appl. Environ. Microbiol. 50, 498–502. Covelli, S., Faganeli, J., Horvat, M., Brambati, A., 2001. Mercury contamination of coastal sediments as the result of long-term cinnabar activity (Gulf of Trieste, northern Adriatic Sea). Appl. Geochem. 16, 541–558. Donazzolo, R., Hieke Merlin, O., Menegazzo Vitturi, L., Orio, A.A., Pavoni, B., Perin, G., Rabitti, S., 1981. Heavy metal contamination in surface sediments from the Gulf of Venice, Italy. Mar. Pollut. Bull. 12, 417–425. Fabbri, D., Gabbianelli, G., Locatelli, C., Lubrano, D., Trombini, C., Vassura, I., 2001. Distribution of mercury and other heavy metals in core sediments of the northern Adriatic sea. Water Air Soil Pollut. 129, 143–153. Faganeli, J., Planinc, R., Pezdic, J., Smodis, B., Stegnar, P., Ogorelec, B., 1991. Marine geology of the Gulf of Trieste (northern Adriatic): geochemical aspects. Mar. Geol. 99, 93–108. Gagnon, C., Pelletier, E., Mucci, A., 1997. Behaviour of anthropogenic mercury in coastal marine sediments. Mar. Chem. 59, 159–176. Gill, G.A., Bruland, K.W., 1990. Mercury speciation in 1994, distribution and Xux of particulate mercury in four stratiWed seepage lakes, surface freshwater system in California and other areas. Environ. Sci. Technol. 31, 489–495. Gilmour, C.C., Henry, E.A., Mitchell, R., 1992. Sulfate stimulation of mercury methylation in freshwater sediments. Environ. Sci. Technol. 26, 2281–2287. Hedges, J.I., Stern, J.H., 1984. Carbon and nitrogen determinations in carbonate-containing solids. Limnol. Oceanog. 29, 657–663. Hines, M.E., Horvat, M., Faganeli, J., Bonzongo, J.C.J., Barkay, T., Major, E.B., Scott, K.J., Bailey, E.A., Warwick, J.J., Berry Lyons, W., 2000. Mercury biogeochemistry in the Idrija River, Slovenia, from above the Mine into the Gulf of Trieste. Environ. Res. A 83, 129–139. IRSA-CNR, 1985. Istituto di Ricerca sulle Acque. Metodi analitici per i fanghi: parametri chimico-Wsici, vol. 3, pp. 1–5. Kaplan, D.I., Knox, A.S., Myers, J., 2002. Mercury geochemistry in wetland and its implications for in situ remediation. J. Environ. Engin. 128, 723–732. Kierfve, B., 1994. Coastal lagoons. In: Kjerfve, B. (Ed.), Coastal Lagoon Processes. Elsevier, Amsterdam, pp. 1–8. King, J.K., Harmon, S.M., Fu, T.T., Gladden, J.B., 2002. Mercury removal, methylmercury formation, and sulfate-reducing bacteria proWles in wetland mesocosms. Chemosphere 46, 859–870. 13 King, J.K., Saunders, F.M., Lee, R.F., Jahnke, R.A., 1999. Coupling mercury methylation rates to sulfate reduction rates in marine sediments. Environ. Toxicol. Chem. 18, 1362–1369. Lacerda, L.D., Gonçalves, G.O., 2001. Mercury distribution and speciation in waters of the coastal lagoons of Rio de Janeiro, SE Brazil. Mar. Chem. 76, 47–58. Lacerda, L.D., Fernandez, M.A., Tanizaki, K.F., Calazans, C.F., 1992. Bioavailability of heavy metals in sediments of two coastal lagoon in Rio de Janeiro, Brazil. Hydrobiol. 228, 65–70. Loring, D.H., Rantala, R.T.T., 1992. Manual for the geochemical analyses of marine sediments and suspended particulate matter. Earth-Sci. Rev. 32, 235–283. Marins, R.V., Lacerda, L.D., Silva Filho, E.V., 1996. Atmospheric mercury deposition over Sepetiba Bay, Brazil. J. Braz. Chem. Soc. 9, 77–88. Marocco, R., 1995. Sediment distribution and dispersal in northern Adriatic lagoons (Marano and Grado paralyc system). Geologia 57, 77–89. Mason, R.P., Fitzgerald, W.F., Hurley, J., Hanson Jr., A.K., Donaghay, P.L., Sieburth, J.M., 1993. Mercury biogeochemical cycling in a stratiWed estuary. Limnol. Oceanog. 38, 1227–1241. Mastrine, J.A., Bonzongo, J.C.J., Lyons Berry, W., 1999. Mercury concentrations in surface waters from Xuvial systems draining historical precious mining areas in southeastern USA. Appl. Geochem. 14, 147–158. Middleton, R., Grant, A., 1990. Heavy metals in the Humber estuary: scrobicularia clay as a pre-industrial datum. Proc. Yorkshire Geol. Soc. 48, 75–80. Pavoni, B., Donazzolo, R., Marcomini, A., Degobbis, D., Orio, A.A., 1987. Historical development of the Venice lagoon: contamination as recorded in radio-dated sediment cores. Mar. Pollut. Bull. 18, 18–24. Piani, R., 1992. Il particellato sospeso nella Laguna di Grado e Marano: aspetti tessiturali e composizionali. Università di Trieste, Istituto di Geologia e Paleontologia, Tesi di Laurea in Sedimentologia (unpublished). Piani, R., Covelli, S., 2000. Contributo antropico di metalli pesanti e 137Cs nei sedimenti del bacino di Buso (Laguna di Marano e Grado, Italia settentrionale). Studi Trentini di Scienze Naturali – Acta Geolog. 77, 169–177. Porcella, D.B., 1994. Mercury in the environment: biogeochemistry. In: Watras, C.J., Huckabee, J.W. (Eds.), Mercury Pollution: Integration and Synthesis. Lewis Publishers, Boca Raton, FL. Rae, J.E., Aston, S.R., 1981. Mercury in coastal and estuarine sediments of the northeastern Irish Sea. Mar. Pollut. Bull. 12, 367–371. RFVG Regione Autonoma Friuli Venezia Giulia – USL n. 8 Bassa Friulana, 1991. La qualità delle acque della Laguna di Marano. Regional Technical Report. RFVG Regione Autonoma Friuli Venezia Giulia, 1983. Progetto del piano di risanamento del bacino idrograWco della Laguna di Marano e Grado. Rapporto conclusivo. Regional Technical Report, Trieste. Salomons, W., 1997. Biogeochemical of contaminated sediments and soils: perspective and future research. J. Geochem. Explor. 62, 37–40. Sladek, C., Gustin-Sexauer, M., Biester, H., Kim, C., 2002. Assessment of three methods for determining mercury speciation in mine wastes. Geochem. Explor. Environ. Anal. 4, 369–375. ARTICLE IN PRESS 14 R. Piani et al. / Applied Geochemistry xxx (2005) xxx–xxx Smith, J.N., Loring, D.H., 1981. Geochronology of mercury pollution in the sediments of the Saguenay Fjord, Quebec. Environ. Sci. Technol. 15, 944–951. Stordal, M.C., Gill, G.A., Wen, L.S., Santschi, P.H., 1996. Mercury phase speciation in the surface waters of three Texas estuaries: importance of colloidal forms. Limnol. Oceanog. 41, 52–61. Wallschläger, D., Desai, M.V.M., Spengler, M., Wilken, R.-D., 1995. Mercury in contaminated sediments and Xoodplains of the river Elbe – assessment of major species and mobilization behaviour. In: Wilken, R.-D., Forstner, U., Knochel, A. (Eds.), International Conference Heavy Metals in the Environment, vol. 2, Hamburg, September 1995, pp. 52–56. WHO (World Health Organization), 1990. Environmental Criteria 101: Methylmercury. World Health Organization. Geneva, Switzerland.