Marine Pollution Bulletin 96 (2015) 136–148
Contents lists available at ScienceDirect
Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul
Mercury speciation in the Adriatic Sea Jozˇe Kotnik a,⇑, Milena Horvat a, Nives Ogrinc a, Vesna Fajon a, Dušan Zˇagar b, Daniel Cossa c,1, Francesca Sprovieri d, Nicola Pirrone d a
Department of Environmental Sciences, ‘‘Jozˇef Stefan’’ Institute, Jamova 39, SI-1000 Ljubljana, Slovenia Faculty of Civil and Geodetic Engineering, University of Ljubljana, Jamova 2, SI-1000, Slovenia c Ifremer, Centre for the Mediterranean Sea, BP 330, F-83507 La Seyne-sur-Mer, France d CNR – Institute of Atmospheric Pollution Research, Rende, Italy b
a r t i c l e
i n f o
Article history: Received 15 May 2014 Revised 13 May 2015 Accepted 13 May 2015 Available online 23 May 2015 Keywords: Mercury Speciation Deep water profiles Sediments Mass balance Adriatic Sea
a b s t r a c t Mercury and its speciation were studied in surface and deep waters of the Adriatic Sea. Several mercury species (i.e. DGM – dissolved gaseous Hg, RHg – reactive Hg, THg – total Hg, MeHg – monomethyl Hg and DMeHg – dimethylmercury) together with other water parameters were measured in coastal and open sea deep water profiles. THg concentrations in the water column, as well as in sediments and pore waters, were the highest in the northern, most polluted part of the Adriatic Sea as the consequence of Hg mining in Idrija and the heavy industry of northern Italy. Certain profiles in the South Adriatic Pit exhibit an increase of DGM just over the bottom due to its diffusion from sediment as a consequence of microbial and/or tectonic activity. Furthermore, a Hg mass balance for the Adriatic Sea was calculated based on measurements and literature data. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction During recent decades special attention has been given to the mercury cycle in marine environments due to the toxicity of its organic compounds. These organic mercury compounds are produced, bioaccumulated and biomagnified in marine organisms. Of particular concern is biomagnifications in fish which are the principal route of human exposure to methyl mercury (Fitzgerald and Clarkson, 1991), thus mercury contamination is an issue of great concern globally and in the Mediterranean sea basin. It is well known that mercury is chemically, biologically and geologically active; therefore, non-conservative distributions have indeed been reported in a number of water bodies (Fitzgerald et al., 2007). The main Hg species present in the marine environment are elemental Hg (Hg0), Hg(II) and organic molecules (methyl Hg (MeHg) and dimethyl Hg (DMeHg)). The cycling of Hg in coastal marine systems is comparable to that in the open oceans, although the levels of Hg species are enhanced (Cossa et al., 1996; Fitzgerald et al., 2007). Most mercury enters marine waters by wet or dry deposition or by river discharges, with a significant fraction in oxidized form (Mason et al., 1994; Fitzgerald et al., 2007). It has been ⇑ Corresponding author. E-mail address: [email protected] (J. Kotnik). Present address: ISTerre, Université J. Fourier, BP 53, F-38041 Grenoble cedex 9, France. 1
http://dx.doi.org/10.1016/j.marpolbul.2015.05.037 0025-326X/Ó 2015 Elsevier Ltd. All rights reserved.
found that Hg0 in marine waters originates from several biotic and abiotic transformations of oxidized Hg(II) (Mason et al., 1995; Costa and Liss, 1999, 2000; Amyot et al., 1997) and decomposition of organo-mercury compounds (Mason and Fitzgerald, 1993; Mason and Sullivan, 1999). Also tectonic activity may be an important source of Hg0 (Ferrara et al., 2003; Horvat et al., 2003; Kotnik et al., 2007). Between 10% and 30% of total Hg can be present in marine waters as Hg0 (Kim and Fitzgerald, 1988; Mason and Fitzgerald, 1993). Hg0 is usually supersaturated with respect to the atmosphere, especially in surface waters, where its evasion represents an important source to the global atmosphere. Moreover, it should be highlighted that part of Hg(II) is also removed due to the processes of methylation and subsequent bioaccumulation along the food chain (Horvat et al., 2003). MeHg is present in open sea waters at very low levels, ranging from few tens of fM to few pM (Horvat et al., 1999, 2003; Mason et al., 1998; Cossa and Coquery, 2005). Methylated Hg species are primarily formed in deeper ocean waters, but are not restricted to low oxygen zones suggesting that there are additional mechanisms for methylation/demethylation processes (Horvat et al., 2003). Reducing conditions and high salinity are hypothesized to promote demethylation of MeHg (Hines et al., 2000), while it can be efficiently decomposed by photochemical reactions and microbial activity (Horvat et al., 2003). Net MeHg production in coastal marine sediments is substantial (Hammerschmidt and Fitzgerald, 2006; Heyes et al., 2006; Sunderland et al., 2006) and recent
J. Kotnik et al. / Marine Pollution Bulletin 96 (2015) 136–148
research suggests that most MeHg in marine fish may have a near-shore sedimentary origin, where it is apparent that biological methylation is more important than abiotic mechanisms (Benoit et al., 2003). The recent oceanic mass balance calculated by Mason et al. (2012) suggests that open ocean production is a greater source of MeHg that near-shore sediments. Also abiotic production of MeHg is likely to occur in hydrothermal fluids as significant levels of MeHg have been observed in hydrothermal vent fluids (Lamborg et al., 2006). DMeHg was reported to be found in deeper ocean waters (Mason and Fitzgerald, 1990; Cossa et al., 1997; Mason et al., 1995; Horvat et al., 2003; Kotnik et al., 2007). Its source at depth is thought to be linked to some heterotrophically driven production of DMeHg (e.g., Mason et al., 1998). In surface waters, it is readily lost from aquatic environments by evasion and is rapidly decomposed by photochemical degradation (Horvat et al., 2003). However, Black et al. (2009) suggest that evasion is far more important that photodecomposition as a loss term for DMeHg from marine surface water. Closed marine water systems are environments which are very sensitive to Hg pollution due to limited exchange of water between the oceans. The Adriatic Sea is a closed marine system connected to the Mediterranean Sea by the narrow Strait of Otranto and is subject to a high inflow of heavily polluted river water and other direct discharges, especially in its northern and central parts. Its most northern part, the Gulf of Trieste, is influenced by the natural and anthropogenic load of Hg from polluted River Socˇa (Isonzo), whose watershed contains the world’s second largest Hg mine in Idrija. Centuries of drainage of Hg polluted soils, cinnabar deposits, mining and smelting wastes provided the main source of Hg in the Gulf, which is one of the most Hg contaminated areas in the whole Mediterranean region. Mercury that enters the Gulf via the River Socˇa is mainly in particulate form (Širca et al., 1999). Measurements of different Hg species in the water column, suspended solids and sediments have shown several times higher concentrations than in the Central and Southern part of the Adriatic Sea (Horvat et al., 1999, 2002; Faganeli et al., 2003; Covelli et al., 2007, 2008). It was estimated, that the River Socˇa contributed approximately 2160 tonnes of Hg to the Gulf of Trieste during the 500 (1490–1995) year mining history of Idrija (Zˇagar et al., 2006). Elevated Hg levels in water and sediments were found on both western (Italian) and eastern (Croatian) coasts. However, the high Hg levels found in the Ravenna Lagoon sediments were a consequence of direct discharges of the acetaldehyde and vinyl chloride industry. More than 150 tonnes of Hg were released into the Lagoon between 1958 and 1973 (Fabbri et al., 2001; Trombini et al., 2003). The River Po delta area coastal sediments appear to be enriched in Hg as a direct consequence of very extensive industrial inputs from the river drainage basin which represents one quarter of the Italian national territory. Sources of Hg contamination were also identified in the Marghera industrial zone on the western shoreline of the Venice Lagoon. The Lagoon was extensively contaminated with Hg from chlor-alkali discharges (100–300 tonnes of Hg released between 1951 and 1988) (Bloom et al., 2004; Zonta et al., 2007). The Venice Lagoon is now a net exporter of sediments to the Adriatic Sea (1.1 million tonnes per year). The Marano and Grado Lagoons have experienced significant Hg inputs from mining (Idrija, River Socˇa discharge) and industrial sources (Aussa-Corno river, chlor-alkali plant since 1940–1984) (Piani et al., 2005). Kaštela Bay, located on the eastern Adriatic coast near Split, has been exposed to pollution by inorganic mercury derived from a chlor-alkali plant operating between 1949 and 1989. It was estimated that the total amount of elemental mercury discharged into the Bay was 39 t (Tudor, 1993). The present study was performed to obtain missing data and improve knowledge on Hg speciation and distribution in central, as well as the coastal regions of the Adriatic Sea. In this paper,
137
we focus on Hg distribution and speciation in the water column of the area investigated. In addition, the concentrations of THg and MeHg in pore waters and sediment profiles down to a depth of 20 cm were determined, and a rough mass balance for THg and MeHg was calculated separately for the Northern and Southern Adriatic. 2. Materials and methods 2.1. Site description The Adriatic Sea extends northwest from 40° to 45°450 N, with an extreme length of about 770 km. Its northern part is very shallow, gently sloping with an average bottom depth of about 35 m. The site of maximum depth is south of the central area, and the average depth of the Central and South Adriatic is about 440 m, with a maximum depth of 1399 m. In the northern Adriatic the water column shows a seasonal thermal cycle. The thermocline is present in spring and summer down to 30 m depth. In winter cooling of the whole water column occurs. A freshwater surface plume was observed in spring and summer due to increased runoff and water stratification. There are two typical water masses: the seasonal layer of North Adriatic surface water (NAdSW – low salinity and high temperature in the summer), and the North Adriatic deep water (NAdDW – T = 11.35 ± 1.4 °C, S = 38.3 ± 2.8), which is cooled and renewed in winter (Artegiani et al., 1997), flowing southward and forming Central Adriatic deep water (CAdDW). In the Central Adriatic, the thermocline is formed down to 50 m in spring and summer. Below the thermocline (50 m) Levantine intermediate waters can be found (LIW – S > 38.5). The area deeper than 150 m (the Pomo Depression) is filled with CAdDW (T = 11.62 ± 0.75 °C, S = 38.47 ± 0.15). In the southern Adriatic the thermocline can be found down to 75 m. Surface waters (SAdSW) affected by river inflow exhibit a decrease in salinity. From 150 m to the bottom, Mediterranean open sea conditions were found to be modulated by LIW between 150 and 400 m (T > 13.5, S > 38.6). Bottom water masses in that area are defined as South Adriatic deep water (SAdDW – T = 13.16 ± 0.3 °C, S = 38.6 ± 0.09) (Artegiani et al., 1997). In the Strait of Otranto four water masses can be found: Adriatic surface water (AdSW) with lower salinity, flowing out of the Adriatic along the western side of the strait; Ionian surface water (ISW) flowing into the Adriatic along the eastern side; Levantine intermediate water (LIW) which flows into the Adriatic at intermediate depths; and Adriatic deep water at the bottom, which flows out of the Adriatic feeding the deep waters of the eastern Mediterranean Sea. The northern Adriatic water circulation (Fig. 1) is dominated by the North Adriatic current (NAd current), flowing southward along the W coast. During winter, it is a segment of the Po River extending only 100 km downstream, while in summer it extends farther south. The central and south Adriatic water circulation is composed of the cyclonic Central and Southern Adriatic gyres (CAd and SAd gyres), the northward flowing eastern South Adriatic current (E-SAd current) and the southward flowing western South Adriatic current (W-SAd current). These four features of the circulation strengthen during summer, while they are weaker in spring and almost absent in winter. During winter, the water circulation is dominated by smooth flow from south to north along the longitudinal centre of the basin (Artegiani et al., 1997). 2.2. Sampling and sample preparation Sampling procedures and analysis of different Hg species were performed either on board, or in the laboratory at the ‘‘Jozˇef Stefan’’ Institute (JSI), Ljubljana, Slovenia.
138
J. Kotnik et al. / Marine Pollution Bulletin 96 (2015) 136–148
Fig. 1. Sampling locations with main water circulation flows.
Water samples were collected on board the R/V Urania during two oceanographic cruises in the Adriatic Sea. The first cruise took place in autumn 2004, between 26th October and 14th November 2004. Water samples were collected at nine locations (Gargano, Pescara, Ancona, Ravenna, Venice, Trieste, Dugi Otok, Budva, Albanian coast) in the Adriatic Sea and at one in the Ionian Sea in the vicinity of the Strait of Otranto. The second cruise was performed in summer 2005, from 16th June to 5th July 2005 at seventeen locations in the Adriatic Sea (Strait of Otranto, Brindisi, Gargano, Pescara, Ancona, Cesenatico, Ravenna, Po Delta, Venice, Trieste, Pula, Premuda, Kornati, Vis, Dubrovnik, Kotor, Budva) (Fig. 1, Tables 3 and 4). At each location depth profiles were sampled depending on the salinity, temperature, oxygen and chlorophyll-a depth profile obtained prior to sampling by CTD probe. All sample containers were made of Teflon. Containers and other Teflon and glassware were acid precleaned in the laboratory up to 30 days before sampling following the procedure described by Horvat et al. (1993). Surface and deep water samples were taken by a stainless steel rosette on which 24 Niskin bottles with a volume of 10 L equipped with silicon seals and springs. Continuous monitoring of water temperature, salinity, oxygen, pressure and fluorescence were performed by CTD probe. Samples intended for THg, RHg and MeHg analysis were collected immediately after rosette boarding from the Niskin
bottles into 1 L Teflon bottles by an acid cleaned silicon tube to prevent rapid mixing of the sample and losses of volatile Hg species (DGM and DMeHg). Sample containers were rinsed three times with sample water prior to filling. Immediately after filling samples for THg and MeHg analysis were acidified by HCl (suprapur concentrated HCl; 0.5 mL of HCl per 500 mL of sample). Absorption of Hg and its species during storage was checked and was insignificant. Reactive, dissolved gaseous, total and monomethyl mercury species in nonfiltered water were analysed on board immediately after sampling. Samples for DGM and DMeHg were collected directly in 0.5 L (for DGM) and 2 L (for DMeHg) glass bubblers. Water samples (2 L) for suspended matter determination were filtered through pre-weighed, cleaned and dried Whatman GF/C glass filters (0.45 lm pore size) and stored deep frozen for further processing in the laboratory. After drying (85 °C) to a constant weight the filters were re-weighed. Sediments were taken by a stainless steel gravity box corer. The sample container was removed directly from the box corer under a N2 atmosphere in a N2 filled glovebox to prevent reactions in a more oxidative environment. Samples were cut into corresponding depths (down to 20 cm of the surface sediment profile) and then transferred to plastic and Teflon containers. Pore water was extracted by centrifugation and vacuum filtration inside the glovebox through a 0.45 lm membrane filter. Aliquots of filtered pore
J. Kotnik et al. / Marine Pollution Bulletin 96 (2015) 136–148
139
water were collected in acid precleaned Teflon containers, acidified with HCl to about 5%, and stored at 20 °C for further processing in the laboratory.
times. The results were in good agreement with the recommended and/or certified values within their uncertainties.
2.3. Analytical methods
3. Results and discussion
Reactive Hg (RHg) was determined by SnCl2 reduction, followed by gold amalgamation, thermal desorption and detection by cold vapour atomic absorption spectrometer (CVAAS). The results provided were corrected for DGM and represent ‘‘labile’’ complexes of Hg(II) calculated from the difference between DGM and measured ‘‘reactive Hg’’. The limit of detection expressed as three standard deviations of the blank was 0.05 pM. A detailed description of the method was given by Horvat et al. (1987, 1991). Total Hg (THg) in sea and pore water samples was determined after oxidation by BrCl and exposure to UV light for at least 3 h. Oxidized Hg was then reduced by SnCl2, and amalgamated, followed by thermal desorption and detection by CV AAS mercury analyser. The limit of detection was 0.5 pM calculated on the basis of three standard deviations of the reagent blank. The repeatability and reproducibility was 5% and 10%. The method used is described in detail by Horvat et al. (1987, 1991). Monomethyl mercury (MeHg) in sea and pore water samples was determined following the procedure described in Horvat et al. (1993, 2003) and Liang et al. (1994, 1996) by back extraction, ethylation, and detection by CV AFS. The limit of detection calculated on the basis of three times the standard deviation of the blank was about 0.14 pM. The repeatability and reproducibility of the method was 5% and 10%. Spike recovery was determined for each batch of analysis and ranged from 80% to 90%. The results were corrected by the recovery factors for each batch. Dissolved gaseous Hg (DGM) was determined after purging, gold amalgamation, thermal desorption by a CV AFS analyser. The system was calibrated by gas phase Hg (Hg0) kept at a defined temperature. The detection limit was 0.02 pM based on three standard deviations of the blank. The repeatability of the method was 4%. The method is described in detail by Horvat et al. (2003) and Gardfeldt et al. (2003). It should be noted that DGM concentrations reported in this study correspond to all volatile Hg species present in sea water – elemental Hg (Hg0) and dimethyl Hg ((CH3)2Hg). Dimethyl mercury (DMeHg) was determined in two steps as described by Horvat et al. (2003). Purging and trapping of DMeHg on Tenax was performed on board ship, and detection by CV AFS in laboratory. The detection limit based on three standard deviations of the baseline noise was 0.37 fM. The measurement protocol was developed by Bloom and Fitzgerald (1988) and adopted by Logar et al. (2002). Total Hg in sediments was measured following the procedure described by Horvat et al. (1991). Wet sample was digested by a mixture of HNO3/HF/HCl, reduced with SnCl2, and detected by CV AAS. The detection limit of the procedure was 2.5 pmol g1 and the reproducibility of the method was 3–5%. During analysis of each batch of samples two blanks (a reagent blank and sample processing blank) were analysed to avoid uncontrolled contamination.
3.1. Hydrology, temperature, suspended matter, oxygen, chlorophyll a
2.4. Data quality control To check the accuracy of different Hg species determinations (THg and MeHg), a certified reference material (CRM) and a reference material (RM) were used: CRM BCR 580, estuarine sediment, obtained from the Institute for Reference Materials and Measurements (IRMM) and RM IAEA 405, trace elements and methylmercury in estuarine sediment, obtained from the International Atomic Energy Agency (IAEA). All samples in a given batch, together with blanks and CRMs were prepared in duplicate and every measurement for each duplicate was repeated 2–3
Water masses were identified by their typical salinity and temperature measured during sampling by CTD probe as they were defined by Artegiani et al. (1997). Most of the northern part of the Adriatic Sea is very shallow, characterized by high freshwater input, which is reflected in the water salinity. At most northern sampling locations low salinity was found at the surface during both campaigns, but was more noticeable at locations closer to the Po and Socˇa river mouths at locations S8, and F8 and S11 respectively. The vertical salinity profile at the N Adriatic locations was characterized by low values at the surface and a maximum near the bottom. Water temperatures at the surface were generally high (19 in autumn and up to 25 °C in summer), with a decrease (12–13 °C in autumn and around 17 °C in summer) towards the bottom. A sharp thermocline was observed at locations FA and S8 at a depth of around 30 m and 10 m, respectively, while at other northern locations the thermocline was not so marked. Locations near Venice and Trieste exhibited a mixed water column with similar temperatures along the whole water column (17–19 °C in autumn and 20–25 °C in summer). Regarding the low salinities and high water temperatures in northern locations of the N Adriatic, only NAdSW were identified. Southward at locations in the central part of the Northern Adriatic, salinity and temperature profiles showed a lower impact of freshwater runoff at the surface. The influence of CAdSW could be noticed in summer as the salinities and temperatures along the whole water column were within the range defined for CAdSW (>38.0 and 11.5 °C). Below a depth of 45 m the location S6 exhibited typical CAdDW values for salinity and temperature (38.47 ± 0.15, 11.62 ± 0.75 °C). In Central Adriatic locations three water masses were found: CAdSW, LIW and CAdDW. At a location near to the W coast (S6) CAdSW were found down to a depth of 30 m, while below that, LIW were present. Two other locations in the area showed a similar distribution of water masses, while at the location in the Pomo depression CAdDW was found below a depth of 150 m. Deep water masses (SAdDW) of the Southern Adriatic Sea were identified at profiles deeper than 800 m. The depth of the SAdSW was typically down to 50 and 100 m. Below that, LIW were found at all locations to a depth of 800 m. In the Strait of Otranto warmer AdSW (Adriatic Surface Water) were present at the surface, LIW at intermediate depths and AdDW (Adriatic Deep Water) near the bottom. Oxygen and Chl-a profiles in the Adriatic showed autumn phytoplankton peaks at depths between 10 and 20 m in the N Adriatic, while summer profiles exhibited a more uniform distribution along the whole water column with a slight increase towards the bottom. All locations in the Central and Southern Adriatic had very similar distributions of Chl-a. Generally, a slight decrease towards the bottom with peaks between 30 and 80 m was observed. The waters of the Northern Adriatic are generally well oxygenized, but lower oxygen concentrations were found in autumn (155–259 lM) than in summer (415–540 lM). The distribution was similar to Chl-a, with a peak between 10 m at shallower locations and 70 m at deeper ones, and a slight decrease towards the bottom. Typically the oxygen peak occurred some ten metres above the Chl-a peak. Suspended matter was found within a wide range of concentrations (0.23–5 mg L1). Its distribution along the water column was mostly related to the Chl-a distribution with higher suspended
140
J. Kotnik et al. / Marine Pollution Bulletin 96 (2015) 136–148
matter concentrations near the bottom, or in surface waters with low salinity reflecting the influence of river water. 3.2. Concentrations and distribution of Hg and its species in water The spatial and vertical distribution of measured Hg species is presented in Figs. 2–6. Total Hg represents the sum of all Hg species (gaseous, dissolved, and bound to particulates or colloids). Total Hg (Fig. 2, Tables 3 and 4) concentrations in water were measured only during the summer cruise. The average for the whole Adriatic was 3.3 pM, ranging between 0.78 and 6.97 pM. As expected, the highest THg content along the whole water column was found at the most northern and most polluted locations along the NW coast (Venice and Trieste), and some locations near the NE coast (Premuda and Kornati Islands). The highest THg content (4.52–6.97 pM) was found in the surface water layer of the Gulf of Trieste, which exhibits low salinity, confirming the influence of River Socˇa runoff. Relatively high THg in water was also found near Venice (6.18–6.36 pM). Two locations (Po delta, Ravenna) where high Hg inputs were expected due to inputs from the chlor-alkali and other industries showed elevated THg levels (3.78–4.52 and 3.56–4.21 pM respectively). Southward along the Italian coast THg gradually decreased to values that are within the range of the Adriatic average. At locations near the N Croatian coast (Pula – S12 and Premuda Island – S13), where no large anthropogenic Hg sources are reported, high THg concentrations were found along the whole water column. In the northern part of the Croatian coast Pula and Rijeka are two large cities with busy ports with an oil refinery, and several small towns with intensive tourism during summer. THg in water column near Premuda Island (4.55–6.18 pM) was as high as those measured in the heavily Hg polluted Gulf of Trieste and in front of the Venice Lagoon. THg in the central part of N Adriatic SW of Pula was slightly lower (3.41–3.96 pM). It is typical of anthropogenically polluted waters along the N and E coast (Delta Po, Venice and Gulf of Trieste) to contain higher percentage of DGM and lower portion of MeHg, (7.6–21.7% and 4.5–12.8%, respectively) comparing to E coast of Central and South Adriatic (2.2–12.1% DGM and 3.1–71.9% MeHg). This distribution was observed at both locations on the E coast with elevated THg water levels. The nearest location where THg in sediment (Fig. 1, Table 5) was measured was close to Dugi Otok (FB). The sediment surface layer did not show any enrichment, but was rather low (95 pmol g1) and below the average for unpolluted areas of the Central and South Adriatic (596 pmol g1). The resuspension of sediment could not therefore be a reason for the high Hg level in water. As the air Hg concentrations were low during previous measurements (Sprovieri and Pirrone, 2008), it is hard to believe that elevated THg concentrations originate from some uncontrolled source inland or atmospheric deposition. As the strongly weather-dependent cyclonic NAd Gyre, flowing along the W to E coast is usually present in summer and autumn (Artegiani et al., 1997; Vilibic´ and Orlic´, 2002) it is feasible that waters from the E coast originate at the W part of the N Adriatic. Another possible source is also Hg pollution in Kaštela Bay near Split where severe Hg pollution of sediments and water, originating in a chlor-alkali plant, has been reported (Odzˇak et al., 2000; Kwokal et al., 2002). The strong A-SAd Current and the CAd Gyre could be the transport media for Hg from Kaštela Bay to more northerly locations. The presence of LIW and CAdDW in deeper layers of the water column at Premuda and Vis (S15) strengthen this hypothesis. However, previous studies performed by Horvat et al. (2003) showed that Hg enrichment in water is evident only in waters in Kaštela Bay, close to the chlor-alkali plant (7.7 pM), while in the middle and at the exit of the Bay THg concentrations decrease drastically and fall to 0.17 pM outside the Bay.
In Central Adriatic slightly elevated THg levels in water were found at the location near the Kornati islands (4.43–5.37 pM, location S14). Similar DGM and MeHg percentages as more northerly Premuda and Pula were observed, suggesting anthropogenic source of pollution. The other two Central Adriatic locations (Pescara – S5 and Vis) showed much lower THg water concentrations (0.78–2.9 pM) and are within the range found for nonpoluted locations of the C and S Adriatic. It is likely that the waters sampled at Vis were the most southern part of the NAd Gyre or the most northerly part of the CAd Gyre and SAd Current, especially if we take into account that THg and MeHg in pore water of the surface sediment layer were low (33.5 and 4.68 pM, respectively). In Southern Adriatic waters along water column showed THg concentrations between 0.92 and 4.1 pM, indicating lower concentrations at locations along the Italian (0.92–2.57 pM) than the Croatian (2.63–4.08 pM) coast. The portion of DGM in those waters was relatively low (up to 10%) which excludes larger tectonic source of Hg, but rather indicates some unidentified industrial source along the Albanian and/or Montenegrian coast. However, bottom emissions of DGM, MeHg or DMeHg resulting from high tectonic activity could not be excluded in this area. These results confirm the findings from other epicontinental marine systems (Cossa et al., 1996, 1997, 2004; Cossa and Gobeil, 2000) that coastal water enrichment is only of concern for the near shore zone, where coastal sediments rapidly scavenge mercury from continental waters supplied through estuaries, and/or evolved to the atmosphere. High THg in pore waters and sediments along the Italian coast are most probably the consequence of seasonal NAd, CAd and SAd currents transporting Hg from polluted northern locations. The Strait of Otranto is characterized by lower water THg than the Adriatic average, due to mixing with open Mediterranean waters. It is well known that the concentrations of THg found in the open Mediterranean are low compared to the Adriatic. Previous studies (Horvat et al., 2003; Cossa and Coquery, 2005; Kotnik et al., 2007; Cossa et al., 2009) found levels of THg in open Mediterranean waters to be between 0.8 and 2.3 pM. At locations along the W coast of the Southern and Central Adriatic concentrations were mostly below 2.5 pM and within the range reported for open Mediterranean waters (Cossa et al., 1997; Horvat et al., 2003; Cossa and Coquery, 2005; Kotnik et al., 2007). Total Hg in Ionian waters (location S1) was within the same range. In this study reactive Hg (RHg) is defined as all Hg species that are readily available for reduction by SnCl2 solution and represents all ‘‘labile’’ complexes of Hg(II). The values reported here were corrected for DGM. In general, RHg concentrations (Fig. 3, Tables 3 and 4) ranged between 0.01 and 1.86 pM. Higher concentrations were found during the autumn cruise (average autumn RHg was 0.59 pM, summer 0.41 pM). Generally, the RHg spatial distribution follows THg patterns. Significantly higher RHg concentrations were found in waters of the most Hg polluted locations of the N Adriatic (Venice and Trieste). However, locations along the E coast (Kornati, Premuda, Pula) with elevated THg did not show an increase in RHg. It seems that in waters flowing along E coast towards the north Hg is not as reactive as along the W coast, taking into account that the RHg fraction along the whole water column was very low (2.2– 13.2%) compared to W locations (2–77.2%). The main reason for such distribution seems to be the circulation of water masses, flowing northward along the E coast and southward along the W coast, with strong summer Northern, Central and Southern Adriatic Gyres. RHg levels in locations of the Central and Southern Adriatic were slightly higher than those reported by Horvat et al. (2003) and Kotnik et al. (2007) for the Mediterranean Sea. It is difficult to compare these values with those of other studies for the Mediterranean Sea (Cossa and Coquery, 2005; Cossa et al., 2009) due to the differences in
J. Kotnik et al. / Marine Pollution Bulletin 96 (2015) 136–148
141
Fig. 2. Spatial and vertical distribution of THg in surface and bottom water layers of the Adriatic Sea (in pM).
Fig. 3. Spatial and vertical distribution of RHg in surface and bottom water layers of the Adriatic Sea (in pM).
analytical techniques employed; however, the data obtained here is also in agreement with that reported by Cossa et al. (1997). The dissolved gaseous Hg (DGM) reported here represents the sum of all volatile Hg species (elemental Hg0 and dimethyl Hg). DGM present in Adriatic waters (Fig. 4, Tables 3 and 4) had a wide range of concentrations with no significant difference in autumn (0.07–1.06, av. 0.24 pM) than summer (0.12–0.98, av. 0.31 pM). The distribution in the water column of each location was more or less within the same range. The spatial distribution of DGM mostly followed the THg distribution pattern, with the highest DGM concentrations and portions at northern locations, reflecting anthropogenic sources. It is evident that locations along the east coast did not exhibit increased DGM concentrations as for THg,
and that the DGM fraction was relatively low (2.2–11%) in comparison to locations along the west coast (7.2–45.2%). There could be several biological and/or geological factors affecting its spatial distribution. The concentrations of DGM measured in open Adriatic waters were within the range reported for the Mediterranean Sea (Cossa et al., 1997; Horvat et al., 2003; Ferrara et al., 2003; Kotnik et al., 2007; Andersson et al., 2007). The concentrations of MeHg reported here represent all methylated Hg species. MeHg (Fig. 5, Tables 3 and 4) concentrations showed no significant difference in summer (0.12–2.19, av. 0.61 pM) than in autumn (0.15–2.48, av. 0.55 pM). Interestingly MeHg spatial distribution did not follow the THg distribution. Even at the N locations, most polluted with THg, MeHg did not
142
J. Kotnik et al. / Marine Pollution Bulletin 96 (2015) 136–148
Fig. 4. Spatial and vertical distribution of DGM in surface and bottom water layers of the Adriatic Sea (in pM).
Fig. 5. Spatial and vertical distribution of MeHg in surface and bottom water layers of the Adriatic Sea (in pM).
show a significant increase either at the surface or in the near bottom water layer. In bottom water the highest MeHg was found at locations in the S Adriatic along the E coast (Budva, Kotor, Dubrovnik, Vis, Pula). The same distribution was also reflected in the surface water layer. Hg polluted northern locations did not show elevated MeHg levels, though it is evident that MeHg concentrations were higher than those in southern parts of the Adriatic. Even at locations along the southern part of the west coast, MeHg concentrations were significantly higher than the range reported for open Mediterranean waters (Cossa et al., 1997, 2009; Horvat et al., 2003; Kotnik et al., 2007). Only some locations in the Central or Southern Adriatic exhibited such values. The MeHg distribution shows that net MeHg production is much
stronger in the deeper waters of the Central or S Adriatic probably due to biological or chemical processes or there may be either a benthic source or there is less removal of MeHg from the water column (demethylation or scavenging). Dimethyl Hg was measured only during the summer cruise and represents up to 10.8% of all methylated Hg species at the location near the island of Vis (Fig. 6, Tables 3 and 4) just above the sea bottom. Elevated concentrations of DMeHg (up to 23.3 fM) were also found in deep waters of the Southern Adriatic and Strait of Otranto. The average fraction of DMeHg was 1.7%. The average concentration of DMeHg in Adriatic waters was 5.5 fM (0.2–29.3 fM). DMeHg was also detected at the surface at several locations although the concentrations were low. No significant difference
J. Kotnik et al. / Marine Pollution Bulletin 96 (2015) 136–148
143
Fig. 6. Spatial and vertical distribution of DMeHg in surface and bottom water layers of the Adriatic Sea (in fM).
was found between W and E locations of the Central and Southern Adriatic. The average DMeHg concentration for northern locations was significantly lower. The concentrations reported here are within the same range as found in open Mediterranean waters (
Contents lists available at ScienceDirect
Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul
Mercury speciation in the Adriatic Sea Jozˇe Kotnik a,⇑, Milena Horvat a, Nives Ogrinc a, Vesna Fajon a, Dušan Zˇagar b, Daniel Cossa c,1, Francesca Sprovieri d, Nicola Pirrone d a
Department of Environmental Sciences, ‘‘Jozˇef Stefan’’ Institute, Jamova 39, SI-1000 Ljubljana, Slovenia Faculty of Civil and Geodetic Engineering, University of Ljubljana, Jamova 2, SI-1000, Slovenia c Ifremer, Centre for the Mediterranean Sea, BP 330, F-83507 La Seyne-sur-Mer, France d CNR – Institute of Atmospheric Pollution Research, Rende, Italy b
a r t i c l e
i n f o
Article history: Received 15 May 2014 Revised 13 May 2015 Accepted 13 May 2015 Available online 23 May 2015 Keywords: Mercury Speciation Deep water profiles Sediments Mass balance Adriatic Sea
a b s t r a c t Mercury and its speciation were studied in surface and deep waters of the Adriatic Sea. Several mercury species (i.e. DGM – dissolved gaseous Hg, RHg – reactive Hg, THg – total Hg, MeHg – monomethyl Hg and DMeHg – dimethylmercury) together with other water parameters were measured in coastal and open sea deep water profiles. THg concentrations in the water column, as well as in sediments and pore waters, were the highest in the northern, most polluted part of the Adriatic Sea as the consequence of Hg mining in Idrija and the heavy industry of northern Italy. Certain profiles in the South Adriatic Pit exhibit an increase of DGM just over the bottom due to its diffusion from sediment as a consequence of microbial and/or tectonic activity. Furthermore, a Hg mass balance for the Adriatic Sea was calculated based on measurements and literature data. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction During recent decades special attention has been given to the mercury cycle in marine environments due to the toxicity of its organic compounds. These organic mercury compounds are produced, bioaccumulated and biomagnified in marine organisms. Of particular concern is biomagnifications in fish which are the principal route of human exposure to methyl mercury (Fitzgerald and Clarkson, 1991), thus mercury contamination is an issue of great concern globally and in the Mediterranean sea basin. It is well known that mercury is chemically, biologically and geologically active; therefore, non-conservative distributions have indeed been reported in a number of water bodies (Fitzgerald et al., 2007). The main Hg species present in the marine environment are elemental Hg (Hg0), Hg(II) and organic molecules (methyl Hg (MeHg) and dimethyl Hg (DMeHg)). The cycling of Hg in coastal marine systems is comparable to that in the open oceans, although the levels of Hg species are enhanced (Cossa et al., 1996; Fitzgerald et al., 2007). Most mercury enters marine waters by wet or dry deposition or by river discharges, with a significant fraction in oxidized form (Mason et al., 1994; Fitzgerald et al., 2007). It has been ⇑ Corresponding author. E-mail address: [email protected] (J. Kotnik). Present address: ISTerre, Université J. Fourier, BP 53, F-38041 Grenoble cedex 9, France. 1
http://dx.doi.org/10.1016/j.marpolbul.2015.05.037 0025-326X/Ó 2015 Elsevier Ltd. All rights reserved.
found that Hg0 in marine waters originates from several biotic and abiotic transformations of oxidized Hg(II) (Mason et al., 1995; Costa and Liss, 1999, 2000; Amyot et al., 1997) and decomposition of organo-mercury compounds (Mason and Fitzgerald, 1993; Mason and Sullivan, 1999). Also tectonic activity may be an important source of Hg0 (Ferrara et al., 2003; Horvat et al., 2003; Kotnik et al., 2007). Between 10% and 30% of total Hg can be present in marine waters as Hg0 (Kim and Fitzgerald, 1988; Mason and Fitzgerald, 1993). Hg0 is usually supersaturated with respect to the atmosphere, especially in surface waters, where its evasion represents an important source to the global atmosphere. Moreover, it should be highlighted that part of Hg(II) is also removed due to the processes of methylation and subsequent bioaccumulation along the food chain (Horvat et al., 2003). MeHg is present in open sea waters at very low levels, ranging from few tens of fM to few pM (Horvat et al., 1999, 2003; Mason et al., 1998; Cossa and Coquery, 2005). Methylated Hg species are primarily formed in deeper ocean waters, but are not restricted to low oxygen zones suggesting that there are additional mechanisms for methylation/demethylation processes (Horvat et al., 2003). Reducing conditions and high salinity are hypothesized to promote demethylation of MeHg (Hines et al., 2000), while it can be efficiently decomposed by photochemical reactions and microbial activity (Horvat et al., 2003). Net MeHg production in coastal marine sediments is substantial (Hammerschmidt and Fitzgerald, 2006; Heyes et al., 2006; Sunderland et al., 2006) and recent
J. Kotnik et al. / Marine Pollution Bulletin 96 (2015) 136–148
research suggests that most MeHg in marine fish may have a near-shore sedimentary origin, where it is apparent that biological methylation is more important than abiotic mechanisms (Benoit et al., 2003). The recent oceanic mass balance calculated by Mason et al. (2012) suggests that open ocean production is a greater source of MeHg that near-shore sediments. Also abiotic production of MeHg is likely to occur in hydrothermal fluids as significant levels of MeHg have been observed in hydrothermal vent fluids (Lamborg et al., 2006). DMeHg was reported to be found in deeper ocean waters (Mason and Fitzgerald, 1990; Cossa et al., 1997; Mason et al., 1995; Horvat et al., 2003; Kotnik et al., 2007). Its source at depth is thought to be linked to some heterotrophically driven production of DMeHg (e.g., Mason et al., 1998). In surface waters, it is readily lost from aquatic environments by evasion and is rapidly decomposed by photochemical degradation (Horvat et al., 2003). However, Black et al. (2009) suggest that evasion is far more important that photodecomposition as a loss term for DMeHg from marine surface water. Closed marine water systems are environments which are very sensitive to Hg pollution due to limited exchange of water between the oceans. The Adriatic Sea is a closed marine system connected to the Mediterranean Sea by the narrow Strait of Otranto and is subject to a high inflow of heavily polluted river water and other direct discharges, especially in its northern and central parts. Its most northern part, the Gulf of Trieste, is influenced by the natural and anthropogenic load of Hg from polluted River Socˇa (Isonzo), whose watershed contains the world’s second largest Hg mine in Idrija. Centuries of drainage of Hg polluted soils, cinnabar deposits, mining and smelting wastes provided the main source of Hg in the Gulf, which is one of the most Hg contaminated areas in the whole Mediterranean region. Mercury that enters the Gulf via the River Socˇa is mainly in particulate form (Širca et al., 1999). Measurements of different Hg species in the water column, suspended solids and sediments have shown several times higher concentrations than in the Central and Southern part of the Adriatic Sea (Horvat et al., 1999, 2002; Faganeli et al., 2003; Covelli et al., 2007, 2008). It was estimated, that the River Socˇa contributed approximately 2160 tonnes of Hg to the Gulf of Trieste during the 500 (1490–1995) year mining history of Idrija (Zˇagar et al., 2006). Elevated Hg levels in water and sediments were found on both western (Italian) and eastern (Croatian) coasts. However, the high Hg levels found in the Ravenna Lagoon sediments were a consequence of direct discharges of the acetaldehyde and vinyl chloride industry. More than 150 tonnes of Hg were released into the Lagoon between 1958 and 1973 (Fabbri et al., 2001; Trombini et al., 2003). The River Po delta area coastal sediments appear to be enriched in Hg as a direct consequence of very extensive industrial inputs from the river drainage basin which represents one quarter of the Italian national territory. Sources of Hg contamination were also identified in the Marghera industrial zone on the western shoreline of the Venice Lagoon. The Lagoon was extensively contaminated with Hg from chlor-alkali discharges (100–300 tonnes of Hg released between 1951 and 1988) (Bloom et al., 2004; Zonta et al., 2007). The Venice Lagoon is now a net exporter of sediments to the Adriatic Sea (1.1 million tonnes per year). The Marano and Grado Lagoons have experienced significant Hg inputs from mining (Idrija, River Socˇa discharge) and industrial sources (Aussa-Corno river, chlor-alkali plant since 1940–1984) (Piani et al., 2005). Kaštela Bay, located on the eastern Adriatic coast near Split, has been exposed to pollution by inorganic mercury derived from a chlor-alkali plant operating between 1949 and 1989. It was estimated that the total amount of elemental mercury discharged into the Bay was 39 t (Tudor, 1993). The present study was performed to obtain missing data and improve knowledge on Hg speciation and distribution in central, as well as the coastal regions of the Adriatic Sea. In this paper,
137
we focus on Hg distribution and speciation in the water column of the area investigated. In addition, the concentrations of THg and MeHg in pore waters and sediment profiles down to a depth of 20 cm were determined, and a rough mass balance for THg and MeHg was calculated separately for the Northern and Southern Adriatic. 2. Materials and methods 2.1. Site description The Adriatic Sea extends northwest from 40° to 45°450 N, with an extreme length of about 770 km. Its northern part is very shallow, gently sloping with an average bottom depth of about 35 m. The site of maximum depth is south of the central area, and the average depth of the Central and South Adriatic is about 440 m, with a maximum depth of 1399 m. In the northern Adriatic the water column shows a seasonal thermal cycle. The thermocline is present in spring and summer down to 30 m depth. In winter cooling of the whole water column occurs. A freshwater surface plume was observed in spring and summer due to increased runoff and water stratification. There are two typical water masses: the seasonal layer of North Adriatic surface water (NAdSW – low salinity and high temperature in the summer), and the North Adriatic deep water (NAdDW – T = 11.35 ± 1.4 °C, S = 38.3 ± 2.8), which is cooled and renewed in winter (Artegiani et al., 1997), flowing southward and forming Central Adriatic deep water (CAdDW). In the Central Adriatic, the thermocline is formed down to 50 m in spring and summer. Below the thermocline (50 m) Levantine intermediate waters can be found (LIW – S > 38.5). The area deeper than 150 m (the Pomo Depression) is filled with CAdDW (T = 11.62 ± 0.75 °C, S = 38.47 ± 0.15). In the southern Adriatic the thermocline can be found down to 75 m. Surface waters (SAdSW) affected by river inflow exhibit a decrease in salinity. From 150 m to the bottom, Mediterranean open sea conditions were found to be modulated by LIW between 150 and 400 m (T > 13.5, S > 38.6). Bottom water masses in that area are defined as South Adriatic deep water (SAdDW – T = 13.16 ± 0.3 °C, S = 38.6 ± 0.09) (Artegiani et al., 1997). In the Strait of Otranto four water masses can be found: Adriatic surface water (AdSW) with lower salinity, flowing out of the Adriatic along the western side of the strait; Ionian surface water (ISW) flowing into the Adriatic along the eastern side; Levantine intermediate water (LIW) which flows into the Adriatic at intermediate depths; and Adriatic deep water at the bottom, which flows out of the Adriatic feeding the deep waters of the eastern Mediterranean Sea. The northern Adriatic water circulation (Fig. 1) is dominated by the North Adriatic current (NAd current), flowing southward along the W coast. During winter, it is a segment of the Po River extending only 100 km downstream, while in summer it extends farther south. The central and south Adriatic water circulation is composed of the cyclonic Central and Southern Adriatic gyres (CAd and SAd gyres), the northward flowing eastern South Adriatic current (E-SAd current) and the southward flowing western South Adriatic current (W-SAd current). These four features of the circulation strengthen during summer, while they are weaker in spring and almost absent in winter. During winter, the water circulation is dominated by smooth flow from south to north along the longitudinal centre of the basin (Artegiani et al., 1997). 2.2. Sampling and sample preparation Sampling procedures and analysis of different Hg species were performed either on board, or in the laboratory at the ‘‘Jozˇef Stefan’’ Institute (JSI), Ljubljana, Slovenia.
138
J. Kotnik et al. / Marine Pollution Bulletin 96 (2015) 136–148
Fig. 1. Sampling locations with main water circulation flows.
Water samples were collected on board the R/V Urania during two oceanographic cruises in the Adriatic Sea. The first cruise took place in autumn 2004, between 26th October and 14th November 2004. Water samples were collected at nine locations (Gargano, Pescara, Ancona, Ravenna, Venice, Trieste, Dugi Otok, Budva, Albanian coast) in the Adriatic Sea and at one in the Ionian Sea in the vicinity of the Strait of Otranto. The second cruise was performed in summer 2005, from 16th June to 5th July 2005 at seventeen locations in the Adriatic Sea (Strait of Otranto, Brindisi, Gargano, Pescara, Ancona, Cesenatico, Ravenna, Po Delta, Venice, Trieste, Pula, Premuda, Kornati, Vis, Dubrovnik, Kotor, Budva) (Fig. 1, Tables 3 and 4). At each location depth profiles were sampled depending on the salinity, temperature, oxygen and chlorophyll-a depth profile obtained prior to sampling by CTD probe. All sample containers were made of Teflon. Containers and other Teflon and glassware were acid precleaned in the laboratory up to 30 days before sampling following the procedure described by Horvat et al. (1993). Surface and deep water samples were taken by a stainless steel rosette on which 24 Niskin bottles with a volume of 10 L equipped with silicon seals and springs. Continuous monitoring of water temperature, salinity, oxygen, pressure and fluorescence were performed by CTD probe. Samples intended for THg, RHg and MeHg analysis were collected immediately after rosette boarding from the Niskin
bottles into 1 L Teflon bottles by an acid cleaned silicon tube to prevent rapid mixing of the sample and losses of volatile Hg species (DGM and DMeHg). Sample containers were rinsed three times with sample water prior to filling. Immediately after filling samples for THg and MeHg analysis were acidified by HCl (suprapur concentrated HCl; 0.5 mL of HCl per 500 mL of sample). Absorption of Hg and its species during storage was checked and was insignificant. Reactive, dissolved gaseous, total and monomethyl mercury species in nonfiltered water were analysed on board immediately after sampling. Samples for DGM and DMeHg were collected directly in 0.5 L (for DGM) and 2 L (for DMeHg) glass bubblers. Water samples (2 L) for suspended matter determination were filtered through pre-weighed, cleaned and dried Whatman GF/C glass filters (0.45 lm pore size) and stored deep frozen for further processing in the laboratory. After drying (85 °C) to a constant weight the filters were re-weighed. Sediments were taken by a stainless steel gravity box corer. The sample container was removed directly from the box corer under a N2 atmosphere in a N2 filled glovebox to prevent reactions in a more oxidative environment. Samples were cut into corresponding depths (down to 20 cm of the surface sediment profile) and then transferred to plastic and Teflon containers. Pore water was extracted by centrifugation and vacuum filtration inside the glovebox through a 0.45 lm membrane filter. Aliquots of filtered pore
J. Kotnik et al. / Marine Pollution Bulletin 96 (2015) 136–148
139
water were collected in acid precleaned Teflon containers, acidified with HCl to about 5%, and stored at 20 °C for further processing in the laboratory.
times. The results were in good agreement with the recommended and/or certified values within their uncertainties.
2.3. Analytical methods
3. Results and discussion
Reactive Hg (RHg) was determined by SnCl2 reduction, followed by gold amalgamation, thermal desorption and detection by cold vapour atomic absorption spectrometer (CVAAS). The results provided were corrected for DGM and represent ‘‘labile’’ complexes of Hg(II) calculated from the difference between DGM and measured ‘‘reactive Hg’’. The limit of detection expressed as three standard deviations of the blank was 0.05 pM. A detailed description of the method was given by Horvat et al. (1987, 1991). Total Hg (THg) in sea and pore water samples was determined after oxidation by BrCl and exposure to UV light for at least 3 h. Oxidized Hg was then reduced by SnCl2, and amalgamated, followed by thermal desorption and detection by CV AAS mercury analyser. The limit of detection was 0.5 pM calculated on the basis of three standard deviations of the reagent blank. The repeatability and reproducibility was 5% and 10%. The method used is described in detail by Horvat et al. (1987, 1991). Monomethyl mercury (MeHg) in sea and pore water samples was determined following the procedure described in Horvat et al. (1993, 2003) and Liang et al. (1994, 1996) by back extraction, ethylation, and detection by CV AFS. The limit of detection calculated on the basis of three times the standard deviation of the blank was about 0.14 pM. The repeatability and reproducibility of the method was 5% and 10%. Spike recovery was determined for each batch of analysis and ranged from 80% to 90%. The results were corrected by the recovery factors for each batch. Dissolved gaseous Hg (DGM) was determined after purging, gold amalgamation, thermal desorption by a CV AFS analyser. The system was calibrated by gas phase Hg (Hg0) kept at a defined temperature. The detection limit was 0.02 pM based on three standard deviations of the blank. The repeatability of the method was 4%. The method is described in detail by Horvat et al. (2003) and Gardfeldt et al. (2003). It should be noted that DGM concentrations reported in this study correspond to all volatile Hg species present in sea water – elemental Hg (Hg0) and dimethyl Hg ((CH3)2Hg). Dimethyl mercury (DMeHg) was determined in two steps as described by Horvat et al. (2003). Purging and trapping of DMeHg on Tenax was performed on board ship, and detection by CV AFS in laboratory. The detection limit based on three standard deviations of the baseline noise was 0.37 fM. The measurement protocol was developed by Bloom and Fitzgerald (1988) and adopted by Logar et al. (2002). Total Hg in sediments was measured following the procedure described by Horvat et al. (1991). Wet sample was digested by a mixture of HNO3/HF/HCl, reduced with SnCl2, and detected by CV AAS. The detection limit of the procedure was 2.5 pmol g1 and the reproducibility of the method was 3–5%. During analysis of each batch of samples two blanks (a reagent blank and sample processing blank) were analysed to avoid uncontrolled contamination.
3.1. Hydrology, temperature, suspended matter, oxygen, chlorophyll a
2.4. Data quality control To check the accuracy of different Hg species determinations (THg and MeHg), a certified reference material (CRM) and a reference material (RM) were used: CRM BCR 580, estuarine sediment, obtained from the Institute for Reference Materials and Measurements (IRMM) and RM IAEA 405, trace elements and methylmercury in estuarine sediment, obtained from the International Atomic Energy Agency (IAEA). All samples in a given batch, together with blanks and CRMs were prepared in duplicate and every measurement for each duplicate was repeated 2–3
Water masses were identified by their typical salinity and temperature measured during sampling by CTD probe as they were defined by Artegiani et al. (1997). Most of the northern part of the Adriatic Sea is very shallow, characterized by high freshwater input, which is reflected in the water salinity. At most northern sampling locations low salinity was found at the surface during both campaigns, but was more noticeable at locations closer to the Po and Socˇa river mouths at locations S8, and F8 and S11 respectively. The vertical salinity profile at the N Adriatic locations was characterized by low values at the surface and a maximum near the bottom. Water temperatures at the surface were generally high (19 in autumn and up to 25 °C in summer), with a decrease (12–13 °C in autumn and around 17 °C in summer) towards the bottom. A sharp thermocline was observed at locations FA and S8 at a depth of around 30 m and 10 m, respectively, while at other northern locations the thermocline was not so marked. Locations near Venice and Trieste exhibited a mixed water column with similar temperatures along the whole water column (17–19 °C in autumn and 20–25 °C in summer). Regarding the low salinities and high water temperatures in northern locations of the N Adriatic, only NAdSW were identified. Southward at locations in the central part of the Northern Adriatic, salinity and temperature profiles showed a lower impact of freshwater runoff at the surface. The influence of CAdSW could be noticed in summer as the salinities and temperatures along the whole water column were within the range defined for CAdSW (>38.0 and 11.5 °C). Below a depth of 45 m the location S6 exhibited typical CAdDW values for salinity and temperature (38.47 ± 0.15, 11.62 ± 0.75 °C). In Central Adriatic locations three water masses were found: CAdSW, LIW and CAdDW. At a location near to the W coast (S6) CAdSW were found down to a depth of 30 m, while below that, LIW were present. Two other locations in the area showed a similar distribution of water masses, while at the location in the Pomo depression CAdDW was found below a depth of 150 m. Deep water masses (SAdDW) of the Southern Adriatic Sea were identified at profiles deeper than 800 m. The depth of the SAdSW was typically down to 50 and 100 m. Below that, LIW were found at all locations to a depth of 800 m. In the Strait of Otranto warmer AdSW (Adriatic Surface Water) were present at the surface, LIW at intermediate depths and AdDW (Adriatic Deep Water) near the bottom. Oxygen and Chl-a profiles in the Adriatic showed autumn phytoplankton peaks at depths between 10 and 20 m in the N Adriatic, while summer profiles exhibited a more uniform distribution along the whole water column with a slight increase towards the bottom. All locations in the Central and Southern Adriatic had very similar distributions of Chl-a. Generally, a slight decrease towards the bottom with peaks between 30 and 80 m was observed. The waters of the Northern Adriatic are generally well oxygenized, but lower oxygen concentrations were found in autumn (155–259 lM) than in summer (415–540 lM). The distribution was similar to Chl-a, with a peak between 10 m at shallower locations and 70 m at deeper ones, and a slight decrease towards the bottom. Typically the oxygen peak occurred some ten metres above the Chl-a peak. Suspended matter was found within a wide range of concentrations (0.23–5 mg L1). Its distribution along the water column was mostly related to the Chl-a distribution with higher suspended
140
J. Kotnik et al. / Marine Pollution Bulletin 96 (2015) 136–148
matter concentrations near the bottom, or in surface waters with low salinity reflecting the influence of river water. 3.2. Concentrations and distribution of Hg and its species in water The spatial and vertical distribution of measured Hg species is presented in Figs. 2–6. Total Hg represents the sum of all Hg species (gaseous, dissolved, and bound to particulates or colloids). Total Hg (Fig. 2, Tables 3 and 4) concentrations in water were measured only during the summer cruise. The average for the whole Adriatic was 3.3 pM, ranging between 0.78 and 6.97 pM. As expected, the highest THg content along the whole water column was found at the most northern and most polluted locations along the NW coast (Venice and Trieste), and some locations near the NE coast (Premuda and Kornati Islands). The highest THg content (4.52–6.97 pM) was found in the surface water layer of the Gulf of Trieste, which exhibits low salinity, confirming the influence of River Socˇa runoff. Relatively high THg in water was also found near Venice (6.18–6.36 pM). Two locations (Po delta, Ravenna) where high Hg inputs were expected due to inputs from the chlor-alkali and other industries showed elevated THg levels (3.78–4.52 and 3.56–4.21 pM respectively). Southward along the Italian coast THg gradually decreased to values that are within the range of the Adriatic average. At locations near the N Croatian coast (Pula – S12 and Premuda Island – S13), where no large anthropogenic Hg sources are reported, high THg concentrations were found along the whole water column. In the northern part of the Croatian coast Pula and Rijeka are two large cities with busy ports with an oil refinery, and several small towns with intensive tourism during summer. THg in water column near Premuda Island (4.55–6.18 pM) was as high as those measured in the heavily Hg polluted Gulf of Trieste and in front of the Venice Lagoon. THg in the central part of N Adriatic SW of Pula was slightly lower (3.41–3.96 pM). It is typical of anthropogenically polluted waters along the N and E coast (Delta Po, Venice and Gulf of Trieste) to contain higher percentage of DGM and lower portion of MeHg, (7.6–21.7% and 4.5–12.8%, respectively) comparing to E coast of Central and South Adriatic (2.2–12.1% DGM and 3.1–71.9% MeHg). This distribution was observed at both locations on the E coast with elevated THg water levels. The nearest location where THg in sediment (Fig. 1, Table 5) was measured was close to Dugi Otok (FB). The sediment surface layer did not show any enrichment, but was rather low (95 pmol g1) and below the average for unpolluted areas of the Central and South Adriatic (596 pmol g1). The resuspension of sediment could not therefore be a reason for the high Hg level in water. As the air Hg concentrations were low during previous measurements (Sprovieri and Pirrone, 2008), it is hard to believe that elevated THg concentrations originate from some uncontrolled source inland or atmospheric deposition. As the strongly weather-dependent cyclonic NAd Gyre, flowing along the W to E coast is usually present in summer and autumn (Artegiani et al., 1997; Vilibic´ and Orlic´, 2002) it is feasible that waters from the E coast originate at the W part of the N Adriatic. Another possible source is also Hg pollution in Kaštela Bay near Split where severe Hg pollution of sediments and water, originating in a chlor-alkali plant, has been reported (Odzˇak et al., 2000; Kwokal et al., 2002). The strong A-SAd Current and the CAd Gyre could be the transport media for Hg from Kaštela Bay to more northerly locations. The presence of LIW and CAdDW in deeper layers of the water column at Premuda and Vis (S15) strengthen this hypothesis. However, previous studies performed by Horvat et al. (2003) showed that Hg enrichment in water is evident only in waters in Kaštela Bay, close to the chlor-alkali plant (7.7 pM), while in the middle and at the exit of the Bay THg concentrations decrease drastically and fall to 0.17 pM outside the Bay.
In Central Adriatic slightly elevated THg levels in water were found at the location near the Kornati islands (4.43–5.37 pM, location S14). Similar DGM and MeHg percentages as more northerly Premuda and Pula were observed, suggesting anthropogenic source of pollution. The other two Central Adriatic locations (Pescara – S5 and Vis) showed much lower THg water concentrations (0.78–2.9 pM) and are within the range found for nonpoluted locations of the C and S Adriatic. It is likely that the waters sampled at Vis were the most southern part of the NAd Gyre or the most northerly part of the CAd Gyre and SAd Current, especially if we take into account that THg and MeHg in pore water of the surface sediment layer were low (33.5 and 4.68 pM, respectively). In Southern Adriatic waters along water column showed THg concentrations between 0.92 and 4.1 pM, indicating lower concentrations at locations along the Italian (0.92–2.57 pM) than the Croatian (2.63–4.08 pM) coast. The portion of DGM in those waters was relatively low (up to 10%) which excludes larger tectonic source of Hg, but rather indicates some unidentified industrial source along the Albanian and/or Montenegrian coast. However, bottom emissions of DGM, MeHg or DMeHg resulting from high tectonic activity could not be excluded in this area. These results confirm the findings from other epicontinental marine systems (Cossa et al., 1996, 1997, 2004; Cossa and Gobeil, 2000) that coastal water enrichment is only of concern for the near shore zone, where coastal sediments rapidly scavenge mercury from continental waters supplied through estuaries, and/or evolved to the atmosphere. High THg in pore waters and sediments along the Italian coast are most probably the consequence of seasonal NAd, CAd and SAd currents transporting Hg from polluted northern locations. The Strait of Otranto is characterized by lower water THg than the Adriatic average, due to mixing with open Mediterranean waters. It is well known that the concentrations of THg found in the open Mediterranean are low compared to the Adriatic. Previous studies (Horvat et al., 2003; Cossa and Coquery, 2005; Kotnik et al., 2007; Cossa et al., 2009) found levels of THg in open Mediterranean waters to be between 0.8 and 2.3 pM. At locations along the W coast of the Southern and Central Adriatic concentrations were mostly below 2.5 pM and within the range reported for open Mediterranean waters (Cossa et al., 1997; Horvat et al., 2003; Cossa and Coquery, 2005; Kotnik et al., 2007). Total Hg in Ionian waters (location S1) was within the same range. In this study reactive Hg (RHg) is defined as all Hg species that are readily available for reduction by SnCl2 solution and represents all ‘‘labile’’ complexes of Hg(II). The values reported here were corrected for DGM. In general, RHg concentrations (Fig. 3, Tables 3 and 4) ranged between 0.01 and 1.86 pM. Higher concentrations were found during the autumn cruise (average autumn RHg was 0.59 pM, summer 0.41 pM). Generally, the RHg spatial distribution follows THg patterns. Significantly higher RHg concentrations were found in waters of the most Hg polluted locations of the N Adriatic (Venice and Trieste). However, locations along the E coast (Kornati, Premuda, Pula) with elevated THg did not show an increase in RHg. It seems that in waters flowing along E coast towards the north Hg is not as reactive as along the W coast, taking into account that the RHg fraction along the whole water column was very low (2.2– 13.2%) compared to W locations (2–77.2%). The main reason for such distribution seems to be the circulation of water masses, flowing northward along the E coast and southward along the W coast, with strong summer Northern, Central and Southern Adriatic Gyres. RHg levels in locations of the Central and Southern Adriatic were slightly higher than those reported by Horvat et al. (2003) and Kotnik et al. (2007) for the Mediterranean Sea. It is difficult to compare these values with those of other studies for the Mediterranean Sea (Cossa and Coquery, 2005; Cossa et al., 2009) due to the differences in
J. Kotnik et al. / Marine Pollution Bulletin 96 (2015) 136–148
141
Fig. 2. Spatial and vertical distribution of THg in surface and bottom water layers of the Adriatic Sea (in pM).
Fig. 3. Spatial and vertical distribution of RHg in surface and bottom water layers of the Adriatic Sea (in pM).
analytical techniques employed; however, the data obtained here is also in agreement with that reported by Cossa et al. (1997). The dissolved gaseous Hg (DGM) reported here represents the sum of all volatile Hg species (elemental Hg0 and dimethyl Hg). DGM present in Adriatic waters (Fig. 4, Tables 3 and 4) had a wide range of concentrations with no significant difference in autumn (0.07–1.06, av. 0.24 pM) than summer (0.12–0.98, av. 0.31 pM). The distribution in the water column of each location was more or less within the same range. The spatial distribution of DGM mostly followed the THg distribution pattern, with the highest DGM concentrations and portions at northern locations, reflecting anthropogenic sources. It is evident that locations along the east coast did not exhibit increased DGM concentrations as for THg,
and that the DGM fraction was relatively low (2.2–11%) in comparison to locations along the west coast (7.2–45.2%). There could be several biological and/or geological factors affecting its spatial distribution. The concentrations of DGM measured in open Adriatic waters were within the range reported for the Mediterranean Sea (Cossa et al., 1997; Horvat et al., 2003; Ferrara et al., 2003; Kotnik et al., 2007; Andersson et al., 2007). The concentrations of MeHg reported here represent all methylated Hg species. MeHg (Fig. 5, Tables 3 and 4) concentrations showed no significant difference in summer (0.12–2.19, av. 0.61 pM) than in autumn (0.15–2.48, av. 0.55 pM). Interestingly MeHg spatial distribution did not follow the THg distribution. Even at the N locations, most polluted with THg, MeHg did not
142
J. Kotnik et al. / Marine Pollution Bulletin 96 (2015) 136–148
Fig. 4. Spatial and vertical distribution of DGM in surface and bottom water layers of the Adriatic Sea (in pM).
Fig. 5. Spatial and vertical distribution of MeHg in surface and bottom water layers of the Adriatic Sea (in pM).
show a significant increase either at the surface or in the near bottom water layer. In bottom water the highest MeHg was found at locations in the S Adriatic along the E coast (Budva, Kotor, Dubrovnik, Vis, Pula). The same distribution was also reflected in the surface water layer. Hg polluted northern locations did not show elevated MeHg levels, though it is evident that MeHg concentrations were higher than those in southern parts of the Adriatic. Even at locations along the southern part of the west coast, MeHg concentrations were significantly higher than the range reported for open Mediterranean waters (Cossa et al., 1997, 2009; Horvat et al., 2003; Kotnik et al., 2007). Only some locations in the Central or Southern Adriatic exhibited such values. The MeHg distribution shows that net MeHg production is much
stronger in the deeper waters of the Central or S Adriatic probably due to biological or chemical processes or there may be either a benthic source or there is less removal of MeHg from the water column (demethylation or scavenging). Dimethyl Hg was measured only during the summer cruise and represents up to 10.8% of all methylated Hg species at the location near the island of Vis (Fig. 6, Tables 3 and 4) just above the sea bottom. Elevated concentrations of DMeHg (up to 23.3 fM) were also found in deep waters of the Southern Adriatic and Strait of Otranto. The average fraction of DMeHg was 1.7%. The average concentration of DMeHg in Adriatic waters was 5.5 fM (0.2–29.3 fM). DMeHg was also detected at the surface at several locations although the concentrations were low. No significant difference
J. Kotnik et al. / Marine Pollution Bulletin 96 (2015) 136–148
143
Fig. 6. Spatial and vertical distribution of DMeHg in surface and bottom water layers of the Adriatic Sea (in fM).
was found between W and E locations of the Central and Southern Adriatic. The average DMeHg concentration for northern locations was significantly lower. The concentrations reported here are within the same range as found in open Mediterranean waters (
