Mercury in the Southern Ocean

We present here the first mercury speciation study in the water column of the Southern Ocean, using a high-resolution south-to-north section (27 stations from 65.50°S to 44.00°S) with up to 15 depths (0-4440 m) between Antarctica and Tasmania (Australia) along the 140°E meridian. In addition, in order to explore the role of sea ice in Hg cycling, a study of mercury speciation in the “snow-sea ice-seawater” continuum was conducted at a coastal site, near the Australian Casey station (66.40°S; 101.14°E). In the open ocean waters, total Hg (Hg_T) concentrations varied from 0.63 to 2.76 pmol L⁻¹ with “transient-type” vertical profiles and a latitudinal distribution suggesting an atmospheric mercury source south of the Southern Polar Front (SPF) and a surface removal north of the Subantartic Front (SAF). Slightly higher mean Hg_T concentrations (1.35 ± 0.39 pmol L⁻¹) were measured in Antarctic Bottom Water (AABW) compared to Antarctic Intermediate water (AAIW) (1.15 ± 0.22 pmol L⁻¹). Labile Hg (Hg_R) concentrations varied from 0.01 to 2.28 pmol L⁻¹, with a distribution showing that the Hg_T enrichment south of the SPF consisted mainly of Hg_R (67 ± 23%), whereas, in contrast, the percentage was half that in surface waters north of PFZ (33 ± 23%). Methylated mercury species (MeHg_T) concentrations ranged from 0.02 to 0.86 pmol L⁻¹. All vertical MeHg_T profiles exhibited roughly the same pattern, with low concentrations observed in the surface layer and increasing concentrations with depth up to an intermediate depth maximum. As for Hg_T, low mean MeHg_T concentrations were associated with AAIW, and higher ones with AABW. The maximum of MeHg_T concentration at each station was systematically observed within the oxygen minimum zone, with a statistically significant MeHg_Tvs Apparent Oxygen Utilization (AOU) relationship (p < 0.001). The proportion of Hg_T as methylated species was lower than 5% in the surface waters, around 50% in deep waters below 1000 m, reaching a maximum of 78% south of the SPF. At Casey coastal station Hg_T and Hg_R concentrations found in the “snow-sea ice-seawater” continuum were one order of magnitude higher than those measured in open ocean waters. The distribution of Hg_T there suggests an atmospheric Hg deposition with snow and a fractionation process during sea ice formation, which excludes Hg from the ice with a parallel Hg enrichment of brine, probably concurring with the Hg enrichment of AABW observed in the open ocean waters. Contrastingly, MeHg_T concentrations in the sea ice environment were in the same range as in the open ocean waters, remaining below 0.45 pmol L⁻¹. The MeHg_T vertical profile through the continuum suggests different sources, including atmosphere, seawater and methylation in basal ice. Whereas Hg_T concentrations in the water samples collected between the Antarctic continent and Tasmania are comparable to recent measurements made in the other parts of the World Ocean (e.g., Soerensen et al., 2010), the Hg species distribution suggests distinct features in the Southern Ocean Hg cycle: (i) a net atmospheric Hg deposition on surface water near the ice edge, (ii) the Hg enrichment in brine during sea ice formation, and (iii) a net methylation of Hg south of the SPF.     

Cycling of dissolved elemental mercury in Arctic Alaskan lakes

Aqueous production and water-air exchange of elemental mercury (Hg⁰) are important features of the environmental cycling of Hg. We investigated Hg⁰ cycling in ten Arctic Alaskan lakes that spanned a wide range in physicochemical characteristics. Dissolved gaseous Hg (DGM, dominated by Hg⁰) varied from 40 to 430 fM and averaged 200 fM. All surface waters were supersaturated relative to the atmosphere. DGM averaged 3 ± 2% of dissolved (i.e., filter passing) dissolved total mercury (DTM) and 15 ± 6% of dissolved labile Hg (DLM). In-lake DGM profiles generally followed the vertical distribution of light, indicating photoreduction of Hg(II) complexes as a source of Hg⁰. Additionally, DGM correlated linearly with DLM (r² = 0.82, p < 0.0001) in the lake surface, signifying that Hg complexes (mostly organic Hg associations) in dissolved phase are photoreducible and contribute to production of DGM. Further, a positive relation between DGM/DTM and both K_a (light attenuation coefficient; r² = 0.73, p < 0.02) and DOC (r² = 0.60, p = 0.02) suggests that solar radiation and dissolved organic matter control DGM production and its cycling. An average rate of DGM formation (0.6 ± 0.2% of DTM d⁻¹; range, 0.20.8) was estimated by assuming steady state with the evasional rate. In-lake DGM formation occurs at lower rates in waters with greater suspended particulate matter and dissolved organic carbon (DOC), pointing to the significant role of organic matter plays in controlling DGM formation in these aquatic systems. Estimated evasional fluxes of Hg⁰ (average, 140 ± 50 pmol m⁻² d⁻¹; range, 60-200) were comparable to those of temperate lakes (e.g., Wisconsin, Michigan). In arctic lakes, the rate of evasion during ice-free periods (7 ± 3 nmol m⁻² yr⁻¹) is similar to the atmospheric input of Hg (wet + dry) to the lakes based on levels in summertime precipitation but not including additional sources, e.g., springtime depletion.     

Mercury in lake sediments of the Precambrian Shield near Huntsville, Ontario, Canada

 Long sediment cores (>1 m) were collected from eight Precambrian Shield lakes in southern Ontario, Canada and analyzed for mercury (Hg), loss-on-ignition (LOI), and a suite of 36 other elements. Results indicated at least 100-fold variation in sediment Hg concentrations between lakes in close proximity (from 450 ppb), comparable to the variation reported for lakes across the whole of Canada. Strong areal correlations between Hg concentrations and LOI (r ² =0.77), between Hg and other trace element concentrations (Pb, Zn, Cd, Sb, As, Br), and similarities in the vertical concentration profiles of Hg and LOI, all point to the importance of organic matter in the release, transport and redistribution of metals in watershed systems. The spatial pattern of Hg concentrations in deep, precolonial sediments (>20 cm) was found to mirror the pattern of Hg concentrations in modern surface sediments, an observation that was confirmed in a follow-up survey (r ² =0.85;n=25 lakes), indicating that natural processes govern the unequal distribution of Hg among these lakes. Between-lake differences in surface sediment Hg concentrations normalized to organic carbon (Hg/C) were also reflected by Hg concentrations in smallmouth bass normalized to 35 cm length (R ² =0.63;n=15 lakes). The latter relationship suggests that smallmouth bass and lake sediment indicators provide mutually supportive information regarding Hg loading to the lacustrine environment from geological sources in the watershed system. Received: 31 October 1996 · Accepted: 27 May 1997     

The onset of anthropogenic activity recorded in lake sediments in the vicinity of the Horne smelter in Quebec,Canada: Sulfur isotope evidence

Contents and δ³⁴S values of several S compounds, enumerations of S-reducing bacteria (SRB) and Fe-reducing bacteria (IRB), and Fe, Pb and In concentrations were determined for ²¹⁰Pb-dated sediment cores from two lakes in Quebec, Canada. Both lakes are located approximately 70 km downwind of the Horne smelter and refinery in Rouyn-Noranda. Increases in Fe, Pb and In concentrations and a decrease in the δ³⁴S values of total S in both lake sediment cores coincide with the start-up of the smelter in 1927. The shift towards more negative δ³⁴S values was primarily caused by an increase in the extent of S isotope fractionation during bacterial (dissimilatory) SO₄ reduction due to SO₄ loading of the lakes after smelting began. Consequently, an enhanced accumulation of ³²S-enriched reduced inorganic S compounds is evident in the sediments. δ³⁴S values of organic S in the sediments decreased only slightly due to the smelter emissions between 1930 and 1980. Hence, due to the sulfide depositing mechanisms, S isotope ratios constitute a useful tracer recording the onset of S pollution in sediments of the two previously SO₄-limited lakes investigated. In contrast, total S concentrations alone are not reliable indicators for anthropogenic S loading in lake sediment records.     

Characterization and cycling of atmospheric mercury along the central US Gulf Coast

Concentrations of atmospheric Hg species, elemental Hg (Hg^°), reactive gaseous Hg (RGM), and fine particulate Hg (Hg-PM_2.5) were measured at a coastal site near Weeks Bay, Alabama from April to August, 2005 and January to May, 2006. Mean concentrations of the species were 1.6 ± 0.3 ng m⁻³, 4.0 ± 7.5 pg m⁻³ and 2.7 ± 3.4 pg m⁻³, respectively. A strong diel pattern was observed for RGM (midday maximum concentrations were up to 92.7 pg m⁻³), but not for Hg^° or Hg-PM_2.5. Elevated RGM concentrations (>25 pg m⁻³) in April and May of 2005 correlated with elevated average daytime O₃ concentrations (>55 ppbv) and high light intensity (>500 W m⁻²). These conditions generally corresponded with mixed continental-Gulf and exclusively continental air mass trajectories. Generally lower, but still elevated, RGM peaks observed in August, 2005 and January–March, 2006 correlated significantly (p < 0.05) with peaks in SO₂ concentration and corresponded to periods of high light intensity and lower average daytime O₃ concentrations. During these times air masses were dominated by trajectories that originated over the continent. Elevated RGM concentrations likely resulted from photochemical oxidation of Hg^° by atmospheric oxidants. This process may have been enhanced in and by the near-shore environment relative to inland sites. The marine boundary layer itself was not found to be a significant source of RGM.     

Mercury concentrations in surface waters from fluvial systems draining historical precious metals mining areas ink southeastern U.S.A.

This study evaluates several southern Appalachian Piedmont mining districts for Hg contamination in surface waters and determines potential relationships between Hg discharged from historical mining operations and site-specific physical factors. Water samples were collected from 3 fluvial systems that drain areas where Hg was used to amalgamate Au from ore during the 19th century. Each of the fluvial systems exhibit similar physical characteristics such as climate, vegetation, and rock type. Total Hg (Hg_T) determinations were made using cold vapour atomic fluorescence spectroscopy techniques. Concentrations of Hg_T in the southern Appalachian Piedmont range from 1–3 ng l⁻¹ in waters of the Arbacoochee, Alabama, and South Mountains, North Carolina, Mining Districts to 13 ng l⁻¹ in waters of the Dahlonega Mining District in Georgia. The correlation between Hg_T and total suspended solids (TSS) at the southern Appalachian sites was good with a coefficient of determination (r²) of 0.82. A clear trend between environmentally-available Fe (Fe_E) and Hg_T (r²=0.86) was also evident. The correlation between Hg_T and Fe_E most likely reflects similarities in the mechanisms that control the aqueous concentrations of both metals (i.e., the particle-reactive nature of the two elements), allowing for the sorption of Hg onto Fe-oxyhydroxides. Hence, increased loads of TSS from erosional events are probably responsible for higher stream water Hg_T concentrations. Vegetation at these sites, which is heavy due to the warm, humid climate of the SE, may help reduce the total amount of Hg released from contaminated mining sites to the rivers by controlling erosion, hence, decreasing the input of contaminated particles into streams and rivers.These southern Appalachian mining sites used Hg amalgamation techniques similar to those used in other precious metals mining districts, such as the highly contaminated Comstock Au–Ag district in Nevada, yet Hg_T concentrations are orders of magnitude lower; This difference in concentration between the southern Appalachian districts and the Comstock district may correlate to the relative amounts of Hg that were used in each. However, other variables were evaluated to determine if physio-chemical differences such as climate could influence Hg_T concentrations in surface waters of the two areas.     

The distribution of total mercury and methyl mercury in a shallow hypereutrophic lake (Lake Taihu) in two seasons

To understand the geochemical cycle of Hg in hypereutrophic freshwater lake, two sampling campaigns were conducted in Lake Taihu in China during May and September of 2009. The concentrations of unfiltered total Hg (unfTHg) were in the range of 6.8–83 ng L⁻¹ (28 ± 18 ng L⁻¹) in the lake water and total Hg in the sediment was 12–470 ng g⁻¹, both of which are higher than in other background lakes. The concentration of unfTHg in ∼11% of the lake water samples exceeded the second class of the Chinese environmental standards for surface water of 50 ng L⁻¹ (GB 3838-2002), indicating that a high ecological risk is posed by the Hg in Lake Taihu. However, the concentrations of unfiltered total MeHg (unfMeHg) were relatively low in the lake water (0.14 ± 0.05 ng L⁻¹, excluding two samples with 0.81 and 1.0 ng L⁻¹). Lake sediment MeHg varied from 0.2–0.96 ng g⁻¹, with generally low ratios of MeHg/THg of <1%. The low concentrations of TMeHg in the lake water may have resulted from a strong uptake by the high primary productivity and the demethylation of MeHg in oxic conditions. In addition, contrary to the results of previous research conducted in deep-water lakes and reservoirs, the low concentrations of MeHg and low ratio of MeHg/THg in the lake sediment indicates that the net methylation of Hg was not accelerated by the elevated organic matter load created by the eutrophication of Lake Taihu. The results also showed that sediments were a source of THg and MeHg in the water. Higher diffusion fluxes of THg and MeHg may be partly responsible for the higher concentrations of THg in the lake water in May, 2009.     

Total Hg concentrations in stream and lake sediments: Discerning geospatial patterns and controls across Canada

This article presents an analysis of Geological Survey of Canada (GSC) open-file data for total Hg concentrations (THg) in stream and lake sediments at 142,028 sampling locations. This analysis was done for select survey zones across Canada, with emphasis on discerning THg-relevant geographic, geological, atmospheric and topographic controls. THg was generally highest in areas affected by mining and smelting, followed by areas with high metallogenic source locations. Background levels for THg were elevated in the more populated areas along the south, but dropped toward the remote and coldest locations in the east, north and west. This trend was correlated (R² = 0.74; P < 0.0001) with the 2005 GRAHM projections (Global/Regional Atmospheric Heavy Metals Model) for atmospheric Hg deposition (zones and locations with major geogenic sources and mining activities removed). Mean THg was higher for upland lakes (100.9 ± 0.5 SE, ppb) and streams (71.7 ± 0.6 SE, ppb) than for lowland lakes (94.4 ± 0.86 SE, ppb) and streams (64.2 ± 1.26 SE, ppb). The east-central portion of the Yukon Territory (Selwyn Basin) was analyzed in further detail. Here, THg within the sediments increased with increasing loss-on-ignition and increasing trace-element concentrations, and decreased with an increasing wet-area component per catchment above the sediment sampling locations. The characterization and quantification of these Hg trends is important for modeling and mapping health risks to ecosystems and communities across Canada and elsewhere.     

Seawater temperature proxies based on DSr, DMg, and DU from culture experiments using the branching coral Porites cylindrica

In order to investigate the incorporation of Sr, Mg, and U into coral skeletons and its temperature dependency, we performed a culture experiment in which specimens of the branching coral (Porites cylindrica) were grown for 1 month at three seawater temperatures (22, 26, and 30 °C). The results of this study showed that the linear extension rate of P. cylindrica has little effect on the skeletal Sr/Ca, Mg/Ca, and U/Ca ratios. The following temperature equations were derived: Sr/Ca (mmol/mol) = 10.214(±0.229) − 0.0642(±0.00897) × T (°C) (r² = 0.59, p < 0.05); Mg/Ca (mmol/mol) = 1.973(±0.302) + 0.1002(±0.0118) × T (°C) (r² = 0.67, p < 0.05); and U/Ca (μmol/mol) = 1.488(±0.0484) − 0.0212(±0.00189) × T (°C) (r² = 0.78, p < 0.05). We calculated the distribution coefficient (D) of Sr, Mg, and U relative to seawater temperature and compared the results with previous data from massive Porites corals. The seawater temperature proxies based on D calibrations of P. cylindrica established in this study are generally similar to those for massive Porites corals, despite a difference in the slope of D_U calibration. The calibration sensitivity of D_Sr, D_Mg, and D_U to seawater temperature change during the experiment was 0.64%/°C, 1.93%/°C, and 1.97%/°C, respectively. These results suggest that the skeletal Sr/Ca ratio (and possibly the Mg/Ca and/or U/Ca ratio) of the branching coral P. cylindrica can be used as a potential paleothermometer.     

Recent paleorecords document rising mercury contamination in Lake Tanganyika

Recent Lake Tanganyika Hg deposition records were derived using ¹⁴C and excess ²¹⁰Pb geochronometers in sediment cores collected from two contrasting depositional environments: the Kalya Platform, located mid-lake and more removed from watershed impacts, and the Nyasanga/Kahama River delta region, located close to the lake’s shoreline north of Kigoma. At the Kalya Platform area, pre-industrial Hg concentrations are 23 ± 0.2 ng/g, increasing to 74 ng/g in modern surface sediment, and the Hg accumulation rate has increased from 1.0 to 7.2 μg/m²/a from pre-industrial to present, which overall represents a 6-fold increase in Hg concentration and accumulation. At the Nyasanga/Kahama delta region, pre-industrial Hg concentrations are 20 ± 3 ng/g, increasing to 46 ng/g in surface sediment. Mercury accumulation rate has increased from 30 to 70 μg/m²/a at this site, representing a 2–3-fold increase in Hg concentration and accumulation. There is a lack of correlation between charcoal abundance and Hg accumulation rate in the sediment cores, demonstrating that local biomass burning has little relationship with the observed Hg concentration or Hg accumulation rates. Examined using a sediment focusing-corrected mass accumulation rate approach, the cores have similar anthropogenic atmospheric Hg deposition profiles, suggesting that after accounting for background sediment concentrations the source of accumulating Hg is predominantly atmospheric in origin. In summary, the data document an increase of Hg flux to the Lake Tanganyika ecosystem that is consistent with increasing watershed sediment delivery with background-level Hg contamination, and regional as well as global increases in atmospheric Hg deposition.     

Distribution of the C37 tetra-unsaturated alkenone in Lake Qinghai, China: A potential lake salinity indicator

The alkenone unsaturation index U^K′₃₇ has been applied to reconstruct past temperature changes in both marine and lacustrine systems. However, few studies have addressed whether the relative abundance of the C_37:4 alkenone to the total C₃₇ production (%C_37:4) can reflect surface salinity changes in lacustrine systems. Here we present long-chain C₃₇ alkenone distribution patterns in surface sediments from Lake Qinghai, China. Surface sediments were sampled over a large range of surface salinity changes (1.7-25 g/l) within Lake Qinghai and its surrounding lakes, while temperature differences at these sampling locations should be relatively small. We have found that %C_37:4 varies from 15% to 49% as surface salinity decreases. We tentatively describe this %C_37:4-salinity link with a general linear regression: %C_37:4 = 53.4 (±7.8) − 1.73 (±0.45) × S (n = 28, r² = 0.62), although step-wise %C_37:4 changes in response to salinity variation may exist. U^K′₃₇ values vary between 0.10 and 0.16 at these sites and the inferred range of lake water temperature changes is ∼2-3 °C, suggesting that U^K′₃₇ largely reflects temperature signal across a large salinity range, consistent with previous findings that U^K′₃₇ can indicate temperature changes over a large diversity of environmental settings. We have also found that U^K′₃₇ values are correlated with salinity changes (r² = 0.4), and thus cannot exclude potential temperature effect on %C_37:4 and salinity effect on U^K′₃₇ in this study. However, even extreme estimates of temperature differences within the lake are still unable to explain the observed %C_37:4 changes. We therefore suggest that %C_37:4 could be used to infer past lake salinity changes at a regional scale.     

Atmospheric deposition of cadmium in the northeastern USA

Lake sediment cores, dated by ²¹⁰Pb, were collected from Spectacle Pond (SP), Massachusetts, and Side Pistol Lake (SPL) and Sargent Mountain Pond (SMP), Maine, USA. SP is a kettle seepage lake in granitic sand and gravel. SMP is a drainage pond on granite with little soil in the small watershed. SPL is a drainage lake in granitic till. The three cores were analyzed for total Cd. For SP and SMP, maximum concentrations of 1.7 and 3.9 mg/kg, four and eight times background concentrations, respectively, occur in the late 1960s. Accumulation rates reach maximum values concurrently with concentration and are 0.054 and 0.016 μg/cm²/a, more than 10 times background. Concentration and accumulation rate age relationships in SMP and SP are similar for background values, timing and magnitude of increase to peak values, and the decrease nearly to background values since about 1975. The chemical response to decreased atmospheric deposition lags in SPL sediment. Kettle-like lakes more clearly indicate changes in atmospheric deposition than drainage lakes.     

Sources of mercury to San Francisco Bay surface sediment as revealed by mercury stable isotopes

Mercury (Hg) concentrations and isotopic compositions were examined in shallow-water surface sediment (0-2 cm) from San Francisco (SF) Bay to determine the extent to which historic Hg mining contributes to current Hg contamination in SF Bay, and to assess the use of Hg isotopes to trace sources of contamination in estuaries. Inter-tidal and wetland sediment had total Hg (Hg_T) concentrations ranging from 161 to 1529 ng/g with no simple gradients of spatial variation. In contrast, inter-tidal and wetland sediment displayed a geographic gradient of δ²⁰²Hg values, ranging from −0.30‰ in the southern-most part of SF Bay (draining the New Almaden Hg District) to −0.99‰ in the northern-most part of SF Bay near the Sacramento-San Joaquin River Delta. Similar to SF Bay inter-tidal sediment, surface sediment from the Alviso Slough channel draining into South SF Bay had a δ²⁰²Hg value of −0.29‰, while surface sediment from the Cosumnes River and Sacramento-San Joaquin River Delta draining into north SF Bay had lower average δ²⁰²Hg values of −0.90‰ and −0.75‰, respectively. This isotopic trend suggests that Hg-contaminated sediment from the New Almaden Hg District mixes with Hg-contaminated sediment from a low δ²⁰²Hg source north of SF Bay. Tailings and thermally decomposed ore (calcine) from the New Idria Hg mine in the California Coast Range had average δ²⁰²Hg values of −0.37 and +0.03‰, respectively, showing that Hg calcination fractionates Hg isotopes resulting in Hg contamination from Hg(II) mine waste products with higher δ²⁰²Hg values than metallic Hg(0) produced from Hg mines. Thus, there is evidence for at least two distinct isotopic signals for Hg contamination in SF Bay: Hg associated with calcine waste materials at Hg mines in the Coast Range, such as New Almaden and New Idria; and Hg(0) produced from these mines and used in placer gold mines and/or in other industrial processes in the Sierra Nevada region and SF Bay area.     

Integrated transfers of terrigenous organic matter to lakes at their watershed level: A combined biomarker and GIS analysis

Terrigenous organic matter (TOM) transfer from a watershed to a lake plays a key role in contaminants fate and greenhouse gazes emission in these aquatic ecosystems. In this study, we linked physiographic and vegetation characteristics of a watershed with TOM nature deposited in lake sediments. TOM was characterized using lignin biomarkers as indicators of TOM sources and state of degradation. Geographical information system (GIS) also allowed us to integrate and describe the landscape morpho-edaphic characteristics of a defined drainage basin. Combining these tools we found a significant and positive relationship (R² = 0.65, p < 0.002) between mean slope of the watershed and the terrigenous fraction estimated by Λ8 in recent sediments. The mean slope also correlated with the composition of TOM in recent sediments as P/(V + S) and 3,5Bd/V ratios significantly decreased with the steepness of the watersheds (R² = 0.57, p < 0.021 and R² = 0.71, p < 0.004, respectively). More precisely, areas with slopes comprised between 4° and 10° have a major influence on TOM inputs to lakes. The vegetation composition of each watershed influenced the composition of recent sediments of the sampled lakes. The increasing presence of angiosperm trees in the watershed influenced the export of TOM to the lake as Λ8 increased significantly with the presence of this type of vegetation (R² = 0.44, p < 0.019). A similar relationship was also observed with S/V ratios, an indicator of angiosperm sources for TOM. The type of vegetation also greatly influenced the degradation state of OM. In this study, we were able to determine that low-sloped areas (0-2°) act as buffer zones for lignin inputs and by extension for TOM loading to sediments. The relative contribution of TOM from the soil organic horizons also increased in steeper watersheds. This study has significant implications in our understanding of the fate of TOM in lacustrine ecosystems.     

50-year record and solid state speciation of mercury in natural and contaminated reservoir sediment

Contaminated fluvial sediments represent both temporary sinks for river-borne pollutants and potential sources in case of natural and/or anthropogenic resuspension. Reservoir lakes play a very important role in sediment dynamics of watersheds and may offer great opportunities to study historical records of river-borne particles and associated elements transported in the past. The fate and potential environmental impact of Hg depends on its abundance, its carrier phases and its chemical speciation. Historical Hg records and solid state Hg speciation were compared in sediments from two contrasting reservoirs of the Lot River (France) upstream and downstream from a major polymetallic pollution (e.g. Cd, Zn) source. Natural (geochemical background) and anthropogenic Hg concentrations and their relationships with predominant carrier phases were determined. The results reveal important historical Hg contamination (up to 35 mg kg⁻¹) of the downstream sediment, reflecting the historical evolution of industrial activity at the point source, i.e. former coal mining, Zn ore treatment and post-industrial remediation work. Single chemical extractions (ascorbate, H₂O₂, KOH) suggest that at both sites most (∼75%) of the Hg is bound to organic and/or reactive sulphide phases. Organo-chelated (KOH-extracted) Hg, representing an important fraction in the uncontaminated sediment, shows similar concentrations (∼0.02 mg kg⁻¹) at both sites and may be mainly attributed to natural inputs and/or processes. Although, total Hg concentrations in recent surface sediments at both sites are still very different, similar mono-methylmercury concentrations (up to 4 μg kg⁻¹) and vertical distributions were observed, suggesting comparable methylation-demethylation processes. High mono-methylmercury concentrations (4–15 μg kg⁻¹) in 10–40 a-old, sulphide-rich, contaminated sediment suggest long-term persistence of mono-methylmercury. Beyond historical records of total concentrations, the studied reservoir sediments provided new insights in solid state speciation and carrier phases of natural and anthropogenic Hg. In case of sediment resuspension, the major part of the Hg historically stored in the Lot River sediments will be accessible to biogeochemical recycling in the downstream fluvial-estuarine environment.     

Phytoreduction and volatilization of mercury by ascorbate in Arabidopsis thaliana, European beech and Norway spruce

Mercury vapor (Hg⁰) emission from plants contributes to the atmospheric Hg cycle. Young barley (Hordeum vulgare L.) plants grown on a hydroponic cultivation medium containing Hg(II) have previously been shown to increase their Hg⁰ emission significantly by reduction of Hg(II) with endogenous ascorbic acid. Regarding the potential contribution to the Hg cycle from the vast forest-covered areas, it was important to investigate this mechanism in trees. The increase in Hg⁰ emission from young European beech plants cultivated on a HgCl₂ medium exceeded that from controls by ca. tenfold and was proportional to the Hg(II) concentration. From these experiments, a flux of 12.8 μg Hg⁰/h/m² was estimated at an exposure of the roots to 20 μM Hg(II). Mercury vapor release from homogenates of Norway spruce needles exceeded that from European beech leaves by a factor of 2.3–4, i.e. in proportion to the reported AA concentrations; the reduction was maximal at alkaline pH which is typical for AA. The 8.4-fold difference in Hg⁰ release between homogenates from wild-type Arabidopsis thaliana and from its AA-deficient mutant vtc 1-1 also paralleled the reported difference in AA levels of both species. It is concluded that the phytoreduction and vaporization of Hg by AA is an important mechanism as much for Hg detoxification in trees as for Hg emission to the atmosphere. The efficiency of this process seems to result from the optimal coordination of transfer and biochemical transformation of mercuric ions and Hg vapor. There is no evidence for a relevant difference in the mechanisms of biogenic Hg(II) reduction between grass plants and trees.     

Variations in the isotopic composition of molybdenum in freshwater lake systems

Variations in molybdenum isotopic composition, spanning the range of ∼ 2.3‰ in the terms of ⁹⁷Mo/⁹⁵Mo ratio, have been measured in sediment cores from three lakes in northern Sweden and north-western Russia. These variations have been produced by both isotopically variable input of Mo into the lakes due to Mo isotopic heterogeneity of bedrock in the drainage basins and fractionation in the lake systems due to temporal variations in limnological conditions. Mo isotope abundances of bedrock in the lake drainage basins have been documented by analysis of Mo isotope ratios of a suite of molybdenite occurrences collected in the studied area and of detrital fractions of the lake sediment cores. The median δ⁹⁷Mo value of the investigated molybdenites is 0.26‰ with standard deviation of 0.43‰ (n = 19), whereas the median δ⁹⁷Mo value of detrital sediment fractions from two lakes is − 0.40‰ with standard deviation of 0.36‰ (n = 15).The isotopic composition of Mo in the sediment cores has been found to be dependent on redox conditions of the water columns and the dominant type of scavenging phases. Hydrous Fe oxides have been shown to be an efficient scavenger of Mo from porewater under oxic conditions. Oxidative precipitation of Fe(II) in the sediments resulted in co-precipitation of Mo and significant authigenic enrichment at the redox boundary. In spite of a pronounced increase in Mo concentration associated with Fe oxides at the redox boundary the isotopic composition of Mo in this zone varies insignificantly, suggesting little or no isotope fractionation during scavenging of Mo by hydrous Fe oxides. In a lake with anoxic bottom water a chironomid-inferred reconstruction of O₂ conditions in the bottom water through the Holocene indicates that increased O₂ concentrations are generally associated with low δ⁹⁷Mo/⁹⁵Mo values of the sediments, whereas lowered O₂ contents of the bottom water are accompanied by relatively high δ⁹⁷Mo/⁹⁵Mo values, thus confirming the potential of Mo isotope data to be a proxy for redox conditions of overlying waters. However, it is pointed out that other processes including input of isotopically heterogeneous Mo and Mn cycling in the redox-stratified water column can be a primary cause of variations in Mo isotopic compositions of lake sediments.     

Influence of temperate mixed and deciduous tree covers on Hg concentrations and photoredox transformations in snow

Mercury dynamics in snowpacks under forested canopy are currently unknown, even though these snowpacks may represent important Hg pools eventually released towards lakes at snowmelt. We followed Hg distribution and partitioning in snowpacks under different temperate canopy types over space and time, and conducted short-term experiments on Hg redox behaviour in these snowpacks. Hg concentrations were ca. two times higher in snow deposited under coniferous than deciduous canopies; the lowest concentrations were observed in snow over a frozen lake in the same watershed. In snow on the ground, up to 80% of the Hg was bound to particles between 10 and 70 μm. Incubations of snow in situ showed that (i) Hg photoreduction and evasion was significant in open areas (lake surface) but was greatly hampered by light attenuation under winter canopies and (ii) oxidation of newly produced Hg⁰ was a significant process in boreal snow, affecting Hg evasion to the atmosphere. We used a mass balance approach to compare Hg pools in snowpacks with wet deposition measured by precipitation collectors. A net gain of Hg was observed in snow under mixed canopies whereas, under a deciduous canopy, the pool of Hg stored at the end of the winter was comparable to that of wet deposition. Snow over lake acted as a winter source of Hg. Whereas most Hg deposited by snow on lakes is lost before snowmelt, Hg deposited on the forested watershed is largely retained in snowpacks, presenting a threat to systems receiving meltwaters.     

Distribution and significance of long-chain alkenones as salinity and temperature indicators in Spanish saline lake sediments

We investigated relationships between sedimentary solvent-extractable long-chain alkenone (LCA) concentration and composition and environmental factors in a suite of endorheic lakes from inland Spain. LCAs were found in 14 of the 54 lakes examined, with concentrations comparable with those from previously published lacustrine settings. The composition of LCAs in our sites, however, contrast from the majority of those previously reported from lake environments; in our study the tri-unsaturated component is the most abundant component at most sites where LCAs are detected, and C_38:3 is the most abundant LCA in the majority of sites. LCA occurrence appears to be restricted to brackish-hypersaline sites and C₃₇ LCAs are absent above a salinity of ∼40 g L⁻¹ suggesting a salinity control on LCA-producing organisms in these sites. Low concentrations of C₃₇ LCA components means and temperature indices are generally not applicable. Instead we find good relationships between C₃₈ components and (in particular mean autumn) temperature and the strongest LCA-temperature relationships are found when using a combination of all C₃₇ and C₃₈ compounds. We propose a new alkenone temperature index for lakes with elevated salinity and where the C₃₈ components dominate the LCA distributions. This is expressed as (r² = 0.80, n = 13). In this paper, we provide the first account of sedimentary LCA distributions from lakes in inland Spain, extending the range of environments within which these compounds have been found and highlighting their significance as indicators of both salinity and temperature in saline, endorheic lake environments. This has important implications for extending the potential role of LCAs as palaeoclimatic indicators in lacustrine environments.