1g: Mercury through time – Reconstructing Hg-deposition using natural archives

Mercury (Hg), a globally distributed air pollutant, primarily enters aquatic ecosystems through watershed runoff, industrial discharges, soil erosion, and direct atmospheric deposition. The distribution of Hg in sediment profiles, coupled with high-resolution age dating (e.g., 210Pb and 137Cs), have been broadly used to evaluate historical changes in Hg inputs. In sediment cores lacking direct contaminant discharges, enrichment factors are typically 2-4 times due to atmospheric inputs. However, in lakes with obvious local contamination histories, it is often difficult to separate atmospheric inputs versus direct point source contamination and other enriched watershed inputs. Recently, stable isotopic Hg analyses of sediment cores have been shown to be powerful tools for identifying specific Hg contamination events in rivers, lakes, and coastal regions. In this study, we investigated the deposition influx and isotopic composition of Hg in several sediment cores from lakes with a history of contamination in the Upper Midwest USA, including Lake Pepin on the Mississippi River, Lake Mendota, and Lake Michigan (open lake and within Green Bay). Mercury influx in these lake cores begin to increase in the mid-1800s, and peaked in the 1970s. In all instances, permanent enrichment of δ202Hg was observed following the onset of industrial contamination. The utility of mass independent fractionation (mainly reported as Δ199Hg) as a tracer for sources and processes became less apparent following industrial contamination. Both Lake Pepin and Lake Michigan exhibit decreases in concentration following the 1970s peak, without obvious changes in isotopic composition, suggestive of a decrease in Hg influx without changes in Hg sources. In order to identify isotopic signatures representative of single sources of Hg in the region, our ongoing studies include sediment cores from lakes with contrasting watershed to lake surface area ratios to assess atmospheric versus watershed influences. Through the comparison of lakes without a history of point source contamination, we expect to develop source attribution models for which we can apply to more complex systems in the Upper Midwest (e.g., Lake Pepin, Lake Mendota and Lake Michigan).

Mercury inputs to forest ecosystems are driven by atmospheric deposition to the forest canopy and soils, and perturbations related to clear-cut logging practices, specifically the removal of the canopy, impact Hg inputs and storage in these terrestrial ecosystems. At our study sites in the tephra fall zone of the 1980 Mt. St Helens volcanic eruption, the tephra deposits provide a clear marker that permits the study of mercury (and carbon) in surficial soils that have developed following the eruption at old-growth forest sites and adjacent sites that were clear-cut just prior to 1980. The ~15cm tephra deposition at these sites crushed the understory and left the forest overstory intact at the old-growth sites, allowing mercury deposition associated with canopy processes and litter deposition to be incorporated into overlying soils. We collected surficial soils and underlying 1980 tephra, at sites with known tree canopy species at the old growth and formerly clear-cut sites to quantify post-1980 mercury accumulation and storage in soils. At the old-growth sites, we mapped tree species and were able to investigate potential differences in the roles that specific species may play in influencing soil mercury accumulation. Overall Hg storage through the tephra layer in the soil at our sites averaged 18.60 ± 6.85 g Hg/ha at old growth and 8.86 ± 4.69 g Hg/ha at clear-cut sites. The organic horizon of the soil accounted for 42% and 16% of Hg storage in the old growth and clear-cut sites respectively. The enhanced Hg storage observed in the old-growth forest soils illustrates the significant role the canopy played in both scavenging Hg from the atmosphere (over the 33 years between sampling and eruption) and likewise in the enhanced Hg retention in underlying soils via carbon addition. A stronger positive correlation between Hg and C concentrations was observed in the old growth sites (r2 = 0.63) than in those clear-cut in the past (r2= 0.49). We estimate that annual Hg accumulation in soils at old growth and clear-cut sites were 0.51 g Hg/ha/yr and 0.22 g Hg/ha/yr respectively, such that the old growth canopy resulted in 0.29 g Hg/ha/yr more Hg stored in underlying soils.

Mercury (Hg) is a toxic global pollutant emitted by both natural and anthropogenic sources and dispersed through the air. Mercury deposition rates have increased in the past century as documented by lake sediment cores and other markers. Sapropels are organic-rich sediments composed primarily of dead algae, and include forms of sulfur that bind and retain Hg. Investigations into the distribution of Hg in sapropel cores are relatively few and none have been measured on Bermuda, a remote island in the Atlantic Ocean over a 1000 km from the US Coast. Here, we determined total-Hg in two sapropel cores from Mangrove Lake, Bermuda. The cores dated back ~1800 years before present (BP) based on an independent Bayesian-based age depth model. There was a sharp rise in the uppermost layers of the core starting at ~45 cm corresponding to the ~1900’s (AD). Mean concentrations in this upper region were 317 ng/g with a high of 420 ng/g (dw basis). There was also a general rise starting around the 1700’s with an anomalously high point at ~1800 AD, coinciding with ship and fortress building on the Island. Background (pre-industrialization) levels were generally less than 100 ng/g. However, the lower portion of the two cores differed, with ML3 having a relatively low & consistent background. In contrast, ML1 had added features. From the core bottom, Hg: (1) decreased during ~500 to ~1500 AD, (2) increased from ~1500 to ~1800 AD, (3) spiked at ~1800; (4) leveled out at ~70 ng/g during the 1800’s; and (5) rose sharply after. These trends are not fully explained by variations in organic matter. However, the proportion of source water (runoff vs. seawater intrusion) may have changed over time. Bermuda is often impacted hurricanes which have altered deposits to Mangrove Lake. To help interpret Hg profiles, we determined trace elements and isotope ratios using ICPMS. Heavy metals, including Pb, Cr, Cu, Ni and Zn, were elevated in the upper core. Several, but not all, elements followed the same depth profile as Hg, including a spike at ~100-cm (~1800 AD). Isotope ratios indicate gasoline as the likely source of Pb in the upper-most layers. Differences in the lower portion of the two cores is still under scrutiny, however, changes in the lower core demonstrate that early natural fluctuations in Hg must be considered.

Atmospherically mobile Hg(0) can be converted, post-deposition, to toxic methylmercury in terrestrial and aquatic environments posing significant risks to human and ecosystem health in regions far from emission sources. Accurate projections of future Hg exposure and impacts will partly depend on our ability to understand changes in the atmospheric Hg pool, but long-term knowledge of atmospheric Hg variability is limited to ~2 decades of instrumental monitoring. Natural Hg archives such as ice cores, tree-rings and lake sediments have the potential to fill this knowledge gap. Tree-rings are ideal since they are annually resolved, span multiple centuries, and cover large areas of Earth’s surface. Few studies concerning a small number of species have shown the potential of tree-rings to record local Hg emissions, but additional proof-of-concept is needed. We present an annually resolved tree-ring Hg record from a stand of white spruce ~100 m from the Bear Creek Gold Room site, Klondike Goldfields, where the Au-Hg amalgam method was used to recover fine gold from placer ore. Bear Creek operations began in 1905, but were greatly expanded in 1930 when most Klondike recovery operations were consolidated to the Bear Creek site. Results from 15 trees reveal a common pattern in tree-ring Hg over the last ~150 years (mean r = 0.60, p<0.01), extending into the pre-mining era. The lowest Hg concentrations occur prior to the Klondike Gold Rush (1896-1899). Initial Bear Creek operations mark a rise in tree-ring Hg concentrations, but this signal is small compared to a major step to peak values in the 1930s coinciding with consolidated operations. Tree-ring Hg declines rapidly after the cessation of Bear Creek operations in ca. 1966. This result suggests the principle Hg intake pathway is likely air-leaf assimilation rather than root uptake from contaminated soil, as there is no delayed decline in tree-ring Hg after the cessation of Bear Creek operations, despite a legacy of elevated soil Hg concentrations to this day. Finally, we observe that post-1966 Hg concentrations are slightly higher than pre-industrial values, likely reflecting higher global atmospheric Hg(0) concentrations. This is the first study of white spruce tree-ring Hg, and only the second to measure this variable at annual resolution. These results demonstrate that this species is a reliable recorder of local emissions, and offers great promise for longer-term reconstructions of atmospheric Hg at potentially broader spatial scales.

Peat cores from filled thermokarst ponds within a subarctic permafrost peatland were used to reconstruct changes in atmospheric mercury (Hg) and lead (Pb) accumulation for the past 150 years and to distinguish between natural and anthropogenic atmospheric inputs. The study site is a permafrost peatland about 8 kilometres southeast from Kuujjuarapik, Northern Quebec. The peatland covers approximately 5 km2 at an elevation of about 110 m above sea level. The peatland is a sedge fen dotted with palsas. The study was undertaken on peat cores taken in the north-eastern part of the peatland (55.13 N, 77.41 W). The chronologies of the peat accumulation were determined using the 210Pb constant rate of supply (CRS) model in three peat cores. Temporal accumulation of mercury show up to 16 times more Hg accumulation during the mid XXth century compared to the average background mercury accumulation rate during the pre-anthropogenic period (1 g m-2 per year). Chronology of Hg accumulation is relatively similar to Pb chronology. Hg accumulation rates today still exceed the average natural background values by 4 to 5 times. Temporal variability in mercury stable isotope signatures in peat shows a clear shift between pre-industrial and industrial periods. The emission of anthropogenic Hg to the atmosphere appears to have caused variations in 202Hg and 199Hg in the uppermost layers of peatlands.

Regulatory policies enacted since the 1970s at local, state and federal levels have led to considerable reduction in anthropogenic mercury emissions in the Great Lakes region of North America. Initially, reductions in mercury concentrations in biota from the Great Lakes and smaller inland lakes were apparent. More recently, however, increasing trends in Hg concentration in fish and fish-eating birds have been reported. These trends coincide with an observed increase in watershed export of organic matter, beginning in the 1990s, that may attenuate recovery of mercury-contaminated lakes. Improved understanding of the causal mechanisms driving the rate of recovery requires knowledge of fluxes of mercury to lakes from atmospheric deposition and watershed runoff.

Sediment cores were collected from five inland seepage lakes, near the University of Michigan Biological Station in Pellston, MI, that span a range of watershed area to lake area ratios (AW:AL). Concentrations of mercury in sediment profiles for all lakes show (i) a stable pre-industrial background, (ii) a pulse event associated with land clearance (logging and agriculture), and (iii) a press event associated with industrialization. In some of the lakes, mercury concentrations decline towards the sediment surface, indicating (iv) a recovery phase. Dating of sediments is in progress and will allow for, through time, the calculation of sediment mercury fluxes and the estimation of net atmospheric mercury fluxes and the additional flux of mercury delivered from watershed to lake for each unit increase in AW:AL. Measurement of natural abundance mercury isotopes in sediments from each lake for each of the four time periods will provide additional evidence of the contributions of mercury from atmospheric deposition and watershed runoff. Finally, a decadal record (1998-2008) of total mercury in event precipitation, at Pellston, MI allows for an independent record of wet depositional flux during the industrialization and recovery periods. Mercury concentration in precipitation decreased -2.02 % year-1 during 1998-2008 (Sens slope, 95% confidence intervals -3.57, 0.22, p< 0.1), but mercury wet deposition (flux) showed no significant trend. Our findings from the geochemical archive in combination with the decadal record of atmospheric monitoring are used to lend insight into the mechanisms controlling recovery in mercury-contaminated lakes.

The Connecticut River has many narrow coves and oxbow lakes off its main channel that are partially filled with fine sediment, some displaying extremely high sedimentation rates (several cm/yr, Woodruff et al., 2013). Sediment cores from six such Connecticut River coves were analysed for total Hg, ~ 20 trace elements, organic pollutants, and dated with 210Pb-137Cs. Selected samples were analysed for Hg isotopes as a source tracer. The general spatial trend shows a typical atmospherically derived record of Hg contamination in both Barton’s Cove and Oxbow Lake in Massachusetts, but much higher Hg concentrations in Wethersfield Cove at the south end of Hartford. All coves south of Hartford show Hg enrichments, reaching 2.8 ppm Hg in Hamburg Cove in the lower CT River estuary. The onset of Hg pollution is broadly dated at the early 1900s and diminishes greatly after 1970. The records from Wethersfield Cove, also an oxbow structure that formed through a cut off associated with the 1691 floods in the CT River, show pronounced Pb, Zn and Cu pollution and the highest Hg contamination of all studied coves (~3 ppm Hg). Both the Keeney Cove and Wethersfield Cove sediments show a sharp V spike in the early 1960s. Potential sources for Hg and some of the other metals are the sewage treatment plant of Hartford situated just north of Wethersfield Cove (South Meadows). In that same area a power plant was operated between 1928 and 1961 that used Hg as the working fluid instead of steam. Mercury vapour was circulated through the turbines to generate electricity, and inspection of historic company documents at the Smithsonian Museum (DC) showed severe Hg losses during the early days of operation, especially in the mid 1930s with a >20,000 lbs Hg loss. The Wethersfield Cove records show Hg peak values in the mid 1930s. Preliminary Hg isotope data show discrete changes in d202Hg over time: more negative values (<-1 ‰) for the pre-industrial Hg and d202Hg values ~ -0.5‰ for post-industrialization. Peak Hg loadings to the sediments in Oxbow lake, Keeney Cove and Hamburg Cove are associated with the highest d202Hg values, suggesting a “pollution end member” with d202Hg close to ~ 0‰.

The value and evidence potential of the tree ring archives as a proxy for changes of atmospheric Hg pollution is still not overall accepted and appreciated. In this work, we present a proof advocating for acceptance of this archive, comparing tree ring Hg record in larch (Larix decidua) against known history of Hg emissions from Roudný gold mining site in the Czech Republic, central Europe. The period of large scale mining at Roudný started 1904 and ceased in 1930. The amalgamation technique was used for the gold ore extraction and amalgam was decomposed at an in-site laboratory with distilling equipment. Nevertheless the losses of Hg during amalgam decomposition amounted at 5-7 g per ton of processed ore. Gold extracted from 664 400 tons of processed ore amounted at 5700 kg. The annual records of gold mined and Hg lost during the extraction process are available for estimate calculation. Besides that, period 1900-2016 is covered by the life span of a currently growing larch trees.

We sampled tree ring cores of four 120-140 years old larch trees at a distance of 32, 88, 275 and 313 m from the amalgamation laboratory. The most significant record of Hg emissions was in the closest tree 32 m from Hg emission source. Its tree rings in sections corresponding to the period between 1905 and 1931 were typical with the highest Hg concentrations up to 41 µg/kg. While Hg concentration in tree rings before and after operation of laboratory averaged at 4.3 µg/kg. The relationship of the peak Hg concentrations in the tree-rings to the peak usage of Hg in amalgamation demonstrates that the Hg record within the bole has been preserved. No signs of significant mercury migration into the heartwood have been noticed as claimed by some previous studies.

The financial support was provided by the Czech Science Foundation project No. GA16-14762S.

Regional Hg atmospheric deposition records in southern New England are often impacted by local sources. Block Island, RI, is a small glacial relic island just south of the Rhode Island coastline. Its many seepage lakes and ponds have almost no drainage basins, and Block Island has no industry and sparse population. We therefore expected Block Island sediment records would capture local atmospheric Hg deposition with little interference from other sources or watershed sediment focusing. We analysed 14 sediment cores for Hg and then focussed on Fresh Pond and a freshwater marsh near West Beach. These cores were analysed for total Hg, dated with 210Pb-137Cs, 14C, and pollen methods. From sedimentation rates and bulk dry densities we calculated Hg accumulation rates and standing inventories for Hg and 210Pb. Pre-industrial concentrations were 30-40 ppb Hg, whereas concentrations in the mid 20th century were up to 320 ppb Hg. The Hg accumulation rates increased from 2 ng Hg/cm2 yr to 18 ng Hg/cm2 yr in the mid 20th century. Some cores show a clear decrease in Hg over the last 35 years. The Hg - 210Pb inventory ratios vary slightly for the two locations, indicating some sediment focussing, largely from aeolian transport. Both records show gentle increases in the mid to late 1800s, but then a strong increase starting around 1940. The modern Hg accumulation rates are of the same order as measured regional bulk Hg deposition rates. Long Island Sound sediment cores show the start of strong Hg contamination in the mid 1800s, and no clear change in slope at 1940. These contamination records are dominated by local Hg sources, especially Hg pollution from the Western Connecticut hat making industry, and also have higher Hg-210Pb inventory ratios. The early part of the Block Island record (1400-1600CE) is characterized by natural background values only. Evidence for atmospheric Hg pollution associated with the early colonial era silver and gold mining by the Spanish in S-America is absent. The magnitude of Hg releases from these colonial mining operations would lead us to expect anomalous Hg values during this era (1500-1700 CE). This missing Hg is either the result of local retention of the mining Hg or emission of Hg species that were rapidly removed from the atmosphere.

Forest vegetation plays a key role of the cycling of mercury (Hg) in terrestrial ecosystems and litterfall has been indicated to be the major transport vector of atmospheric Hg to forest soils. Soil organic matter (OM) is the dominant carrier of Hg from catchment soil to lake sediments. Hence, it is important to understand how changes in forest vegetation affect Hg in soil and its biogeochemical cycling in lake systems. To evaluate whether long-term shifts in forest vegetation induced by climate- or land-use changes influenced Hg-accumulation in lakes we investigated a sediment record from a cirque lake (Glaswaldsee) located in the Black Forest (Southern Germany). We were particularly interested if coniferous vegetation leads to a larger export of Hg to aquatic systems than deciduous vegetation. The radiocarbon-dated sediment core shows a well-defined Holocene vegetation history based on pollen-analyses. The forest vegetation during the Mesolithic at the Glaswaldsee was dominated by Corylus avellana (hazel) until ~ 7800 year BP. Hazel was replaced by Quercus robur (oak), which was replaced at ~ 5400 years BP by a mixed forest of Abies alba (fir) and Fagus sylvatica (beech). When oak replaced hazel, there was neither a significant change in mean Hg-concentration (~ 117 ng g-1) nor a shift in Hg:C ratio. However, Hg-concentrations increase to 198 ng g-1 directly after the appearance of fir and beech. Nevertheless, there was no substantial increase in Hg-accumulation rates because the increased Hg concentrations were caused by decreasing input of OM through litterfall in coniferous forest resulting in higher Hg:C ratios. At around 1270 years BP an additional increase in Hg concentrations in the Glaswaldsee sediments was observed, which might be due to an increasing in atmospheric fluxes attributed to increasing human activities e.g., forest clearance or mining. Hg concentrations peak (640 ng g-1) around 650 years BP when deforestation was highest. Our results contrast earlier studies that suggest an inevitable higher release of Hg from coniferous than deciduous forests.

Acknowledgement: This research is funded by the Deutsche Forschungsgemeinschaft (DFG) by a grant to H. Biester.

In the modern world, primary and legacy anthropogenic Hg emissions account for most of the Hg released to Earths atmosphere and dwarf the contribution of Hg from natural sources (e.g. volcanoes). However, there are intervals in the geologic past where natural processes, such as massive volcanism, likely emitted of large amounts of Hg to Earths atmosphere at rates comparable to human activities. Evidence for enhanced Hg emissions in the geologic past is recorded in marine sedimentary rocks, which contain global-scale spike(s) in their Hg contents and/or Hg/TOC ratios (i.e., Hg anomalies) coincident with periods of prolonged, large-scale volcanism. Here, we examine the Hg stable isotopic composition of marine sediments deposited during times of massive volcanism, drawing both from our own work on Triassic-Jurassic Hg anomalies and the broader literature. In particular, we provide evidence that, over the past several hundred million years, the isotopic signature of near-shore sediments has been more sensitive to changes in Hg-loading and Hg cycle perturbations than deeper-water sediments. Investigating the isotopic composition of marine sediments deposited before, during, and after these ancient Hg-loading events provides a geologic context for understanding the sources and pathways of anomalous Hg deposition to the oceans and supplies insight into the ways that anthropogenic Hg emissions may alter the Hg isotopic ratios of modern marine sediments.

The monitoring record of atmospheric Hg content and deposition is both brief and natural as well as anthropogenic Hg emission estimates have considerable uncertainties. Paleo-reconstructive methods are needed to determine the historical and natural deposition rates of Hg and a wide variety of other contaminants as well as help constrain modeling of contemporary atmospheric processes of Hg deposition and cycling between reservoirs. As part of the Government of Canadas Clean Air Regulatory Agenda (now Climate Change and Air Pollutants) program we have been trying fill in knowledge gaps regarding Hg deposition history and inventories on the Canadian landscape near major point sources and in regions remote from population and industrial centres using multi-lake sediment core reconstructions. The aim of which is to produce quantitative spatial and temporal Hg deposition data spanning pre-industrial times through to the modern day, tracking the onset of Hg emitting industries and the post regulation period comparing with contemporary monitoring data where possible. Multiple lake coring sites within a region serve as multiple receptors producing a coherent landscape signal after accounting for site specific natural Hg background conditions, changes in sedimentation rates and sediment focusing. The combined signal from multiple lake sediment cores helps remove residual site specific biases and can be used to develop a temporal-spatial distribution of Hg inventory deposited on the landscape with respect to a major point source which can be compared directly with long-term estimated emission inventories from the point source. Here we present some preliminary data collected from lake sediment cores in a region isolated from industry (Kejimkujik National Park Nova Scotia, Canada) and lake sediment cores collected various distances from the Pb-Zn smelter in Trail British Columbia, Canada which has been in operation for more than a century. As Trail is situated on the Columbia River 11 km from the US-Canada border it has given rise to in some of the earliest transborder pollution disputes in North America.

The Tibetan Plateau (TP) is located in the Central Asia and represents the highest and remotest landform on the earth. It is distant from point sources of mercury pollution and is relatively isolated. It is endowed with various environmental matrices such as glaciers, rivers and lakes, serving as a unique place for understanding the cycling of mercury and its history. Reconstructed time series of Hg deposition from ice core and lake sediments showed consistent patterns along with the Asias mercury emission. However, the rough resolution of lake sediments and the loss of recent glacial records preclude detailed insights into recent changes in mercury over the TP. Active monitoring of atmospheric mercury started from ~2010 at sporadic sites (e.g., Waliguan, Nam Co, and Everest) reported TGM of~1-2 ng/m3, which agrees well with the background of Northern Hemisphere. A combination of passive air sampling and biomonitoring of a native plant (Androsace tapete) in the central TP revealed a general decline of atmospheric mercury since ~2010.

Single environmental medium or monitoring technique can record mercury changes with respective advantages and limitations. For example, lake sediments can yield long term records (hundreds to tens of thousands of years) but has low resolution (5-10 years). Glacial ice cores span thousands of years with seasonal to annual resolution, however, ice core mercury records are sometimes questioned due to active mercury mobility in the interface of snow-atmosphere. Active mercury monitoring is accurate but only covers very recent years and is deficient. We suggest undertaking intercomparison, integration, and reconciliation of historical mercury records and measurements from diverse natural archives and monitoring techniques, to reveal a larger and more detailed view of mercury changes over the TP.

Lake Chungar is the largest and deepest lacustrine ecosystem in the Chilean Altiplano located in the active volcanic setting of Parinacota (6348 m a.s.l.). The late Holocene period is characterized by intense volcanic activities resulting in the deposition of numerous tephra layers followed by a gradual decrease of volcanic activity which stopped around ~500 cal yr BP. Two sediment cores collected in Lake Chungar (4500 m a.s.l.) were analyzed in high-resolution (mm to cm scale) to provide mercury concentration (Hg) profiles together with X-ray fluorescence (XRF) geochemical parameters (e.g., K, Fe, Si) and molecular composition of organic matter (OM) determined by pyrolysis-gas chromatography-mass spectrometry (Py-GC/MS). Our high-throughput Py-GC/MS method yielded semi-quantitative data on 86 organic compounds belonging to different biochemical classes of OM (e.g., carbohydrate, lipid, chlorophyll). The combination of inorganic and organic geochemical properties allowed us reconstructing the Hg deposition history and mechanisms with regard to the volcanic, limno-ecological and anthropogenic history of the region during the last 3,000 cal. yr BP.

Results highlight that volcaniclastic layers (~ 12) are characterized by abrupt rises in dark mafic mineral enriched in Fe and K accompanied with a strong decrease in Hg concentrations and in the abundance of organic compounds associated to plant and/or algal production (e.g., carbohydrates, proteins and chlorophylls). This is followed by abrupt rises in Hg and in abundances of carbohydrates, proteins and/or chlorophylls. Such feature demonstrates the rapid deposition of coarse volcanic minerals impoverished in Hg followed by a post-eruption deposition of Hg due to atmospheric oxidation of Hg and its recycling by lake biota (i.e., mostly phytoplankton) which recovers from the eruption.

Finally, for the last 500 years of our record, Hg exhibited a sharp rise associated with Ag mining activities during the Colonial era (15541900 AD), followed by a Hg steady decrease illustrating lower Hg inputs during the twentieth century although they remain elevated relative to the pre-industrial background.

Mercury (Hg) highly accumulates in aquatic biota and uptake by algae has been found to be its entry-point into the aquatic foodchain, but the contribution of Hg uptake by algae under changing solar irradiance and climatic conditions has not yet been investigated for pre-industrial times. We analyzed the link between Hg accumulation, cyclic changes in total solar insolation (TSI), related changes in productivity and climate during the past 5 kyrs in sediments of a small highly productive lake located in Southern Patagonia (53°S). Analyzes encompass proxies for solar irradiance (TSI based on 10Be), sediment geochemical composition and lake productivity (FTIR spectra, and hydrogen-index (HI)).

The sediment record shows high accumulation of organic matter (median 70 %) and strong vari­ations in Hg accumulation which correspond to changes in TSI and productivity. Accumulation of Hg was highest during drier periods when insolation and lake productivity was high and erosion fluxes from the catchment were low. This indicates that sediment Hg accu­mulation and potential methylation in this highly productive lake are to a large extent controlled by insolation and related algae production and to a lesser extent by direct atmospheric deposition and Hg fluxes supplied by erosion from catchment soils. We suggested that the high Hg uptake by algae is attributed to oxic water-column methylation in settling patticles in anaerobic micro-niches at times of high productivity. Hg accumulation during periods of high TSI and lake productivity were on an avergae threefold higher compared to periods of lower TSI and productivity and wetter conditions. Our findings suggest that sediment Hg accumulation releated to increased TSI, water phase methylation and algae uptake can surpass Hg fluxes from the catchment in productive lakes.