1c: Stable isotope studies of global mercury cycling and bioaccumulation

Mercury contamination in the Great Lakes region has become a prevalent concern due to elevated methylmercury (MeHg) levels in fish. While atmospheric deposition of Hg is ubiquitous, releases from legacy point-sources give rise to numerous Areas of Concern (AOCs) across the Great Lakes region. One of these AOCs is the lower St. Louis River estuary, which has MeHg concentrations in predatory fish double that of Lake Superior. Despite these highly elevated concentrations, it is difficult to infer the sources of the MeHg to these fish due to different Hg(II) inputs, as well as the complex ecology of estuarine systems. The aim of this study is to utilize Hg stable isotopes to elucidate sources of MeHg to populations of walleye and white suckers in the St. Louis River and Lake Superior region. Walleye from Lake Superior display enhanced Δ199Hg ≥ 5 ‰ and significant Δ200Hg = 0.08-0.10 ‰ for total Hg (THg) isotopes, likely from atmospheric influence and high levels of photochemical demethylation in the MeHg source. However, white suckers from Superior show depleted Δ199Hg and δ202Hg signatures and no significant Δ200Hg. This indicates that there are two distinct Hg sources for these fish in Lake Superior, presumably benthic and pelagic derived MeHg. In contrast to Lake Superior, the walleye and suckers feeding in the St. Louis River mirror the depleted signature observed in white suckers from the lake, which strongly suggests that Hg sources to these populations are obtained from the estuary and are likely sediment-based. While data supports sediment being the primary source of MeHg to fish in the St Louis River it still remains unclear whether the MeHg being produced is derived from legacy or more recent Hg (II), possibly of terrestrial or wetland origin. Preliminary data shows the presence of Δ199Hg (-0.12 ‰) and Δ200Hg (-0.05 ‰) anomalies as well as negative δ202Hg (-0.90 to -1.7) in upstream sediment THg commonly associated with precipitation and vegetation (litterfall). This signature is likely diluted downstream due to the high concentrations of legacy Hg (II) in the estuary and is not reflected in the bulk analysis of THg isotopes even though there is still a potential for methylation of this pool. In order to fully address the sources of MeHg to the St. Louis food web, further experiments examining the isotopic composition of MeHg pools in sediments and fish tissue will be performed.

An important goal of the Great Lakes Restoration Initiative is to reduce fish mercury (Hg) levels such that advisories for human consumption can be removed. However, methylmercury (MeHg), the most toxic and bioaccumulative form of Hg, is the form that predominantly occurs in edible fish tissues. MeHg has a complex biogeochemical cycle in aquatic systems, and thus it is generally very challenging to offer effective scientific results that can lead to reduced MeHg levels in fish. Of particular importance is the ability to ascertain Hg sources, the relative bioavailability of those sources, and key processes controlling bioaccumulation in fish. Previous use of Hg stable isotopes (δ202Hg, Δ199Hg, and Δ200Hg) in sediments of the Laurentian Great Lakes allowed for the quantitative identification of Hg source portfolios for each lake. Here, we apply a similar approach to identify isotopically distinct Hg signatures in Great Lakes’ piscivorous fish in order to evaluate Hg sources. Temporal variability and intra-lake differences in isotopic signatures were not evident, with the exception of Lake Erie, where fish collected in separate basins reflected distinct signatures. Additionally, concentrations of MeHg in predator fish did not correlate with isotopic Hg signatures or sedimentary total Hg concentrations. Great Lakes predatory fish span a large range of Δ199Hg (2.27 - 6.73‰) and displayed some of the highest odd mass independent fractionation (MIF) measured to date. The range in large odd MIF (Δ199Hg) may be explained by differences in euphotic depth (2-43m) among the Great Lakes, a region where Hg and MeHg may be entering the base of the aquatic food web at the lowest trophic levels. Signatures of even MIF (Δ200Hg), a potential binary tracer for precipitation, appears to be disconnected from local sedimentary sources in fish tissue, and is comparable in magnitude across five Great Lakes. This is particularly evident in Lake Ontario, where bottom sedimentary signatures suggest a significant contribution from past industrial sources, but the same signal is not evident in fish tissues. Our combined use of odd and even MIF Hg signatures allow us to evaluate the influence of atmospheric precipitation on bioaccumulation and compare the degree of photochemical processing of Hg in the Great Lakes.

Measurement of stable mercury (Hg) isotope provides important chemical signatures for tracing the transport and transformation of mercury in the environment. However, such a measurement is challenging for natural water samples since typical Hg concentrations (0.5-100 ng/L) in these samples are at least 1,000 times lower than the needed level for isotopic measurements (>0.5 ng/ml) and therefore a pre-concentration step is required. Existing methods such as chromatographic method using ion-exchange resin or activated carbon,purge and trap method using gold trap or acid fixation. But those methods have the disadvantages such as inconsistent recovery, limited sample matrix, unsatisfactory precision and unsuitable for field deployment. In this study, we developed a pre-concentration method that uses modulated assemblies and can be deployed to field sites. The pre-concentration system include a 2.5-L bottle with a bubble diffuser, a chlorine-impregnated activated carbon (CLC) trap, a zero air filter was connected to the inlet, and a vacuum pump. Hg in aqueous phase samples is first oxidized by BrCl, then reduced with SnCl2. The produced Hg0 is purged from the aqueous phase and captured in the trap containing 500 mg of CLC. The captured Hg0 is then thermally desorbed in an argon carrying gas, followed by acid (4 M HNO3 and 1.3 M HCl) fixation. Complete purge and trap of Hg0 can be accomplished within 2 hours at 6 L/min flow rate of carrying gas. The method was verified using solutions spiked with NIST SRM 3133, UM-Almadén and BCR 482 standards at Hg concentrations of 1–200 ng /L. The results showed an analytical recovery of 98±2% (2SD, n = 45), with a deviation of d202Hg within±0.10‰ compared to the standards. The pre-concentration method represents a reliable and efficient method for determining the Hg isotope composition for natural water samples.

Anthropogenic emissions have substantially perturbed the global biogeochemical cycle of mercury (Hg) and high latitude ecosystems are particularly vulnerable to Hg pollution and climate change. We investigated temporal changes in methylmercury (MeHg) exposures of long-finned pilot whales (Globicephala melas, n=68) between 1986-2015 using Hg isotopes (δ202Hg, Δ199Hg, Δ200Hg, Δ201Hg) as tracers of the physical environment and foraging ecology. Mass-independent fraction (MIF) of Hg (Δ199Hg, Δ201Hg) is mainly driven by photochemical demethylation in seawater. Enriched δ202Hg has been shown to result from demethylation. The ranges in Δ199Hg and Δ201Hg values in whales are similar across time periods with the exception of a few years following the 2010 volcanic eruption in Iceland that may have affected light penetration in surface waters. The mean δ202Hg values of whale muscle samples are consistently ~1.5 ‰ across the study period, which is ~1 ‰ higher than their prey (squid, blue whiting, and greater argentine). This fractionation is consistent with in vivo demethylation as a detoxification mechanism in the whales. Individuals with the highest MeHg concentrations in muscle have the lowest δ202Hg values and we hypothesize that this results from more limited MeHg demethylation. We further examine the Hg speciation and isotopes in livers to test our hypothesis. We find a linear relationship between Δ200Hg anomalies (-0.1 to 0.2‰) and Δ199Hg (R2=0.76) that has not previously been reported. Variability in Δ200Hg is thought to be driven by photochemical reactions in the tropopause and may provide an effective tracer for atmospheric Hg inputs to the ocean that are methylated and accumulated in aquatic biota.

Atmospheric mercury (Hg) species can be interconverted through various redox reactions, and this interconversion complicates the identification of the sources and transport pathways of Hg in the atmosphere. The Hg isotope signature is a useful tool for understanding Hg sources and the potential processes that Hg has undergone in the atmospheric environment. We examined the Hg stable isotope ratios in atmospheric Hg species (precipitation, gaseous elemental Hg, and particle-bound Hg) collected in Seoul (urban) and Taean (rural), Korea to identify the impact of local and regional transported Hg. We present both the mass-dependent fractionation (MDF, δ202Hg signature) and mass-independent fractionation (MIF) of odd (Δ199Hg signature) and even (Δ200Hg signature) Hg isotopes in precipitation, gaseous elemental Hg, and particle-bound Hg samples. Potential sources of Hg in atmospheric samples were identified by coupling the both MDF and MIF values obtained for Hg isotopes with a back-trajectory model.

Methylation and biomagnification of mercury (Hg) are well documented in most continental aquatic ecosystems. Few data exist in high altitude lake ecosystems, in particular from the Bolivian Altiplano region. This region is characterized by extreme physicochemical conditions (intense UV radiations and low dissolved oxygen contents) that were previously documented to favor in situ methylmercury production (Alanoca et al. 2016). Lake Titicaca and lake Uru-Uru are part of the endoreic Bolivian-Peruvian Titicaca catchment located at 3800m, and are connected together by the Desaguadero River. Lake Uru-Uru is located downstream of lake Titicaca, and is under the intense influence of mining activities.

Similar biological samples representative of the aquatic food webs from this region were collected in both lakes for assessing a meaningful comparison of Hg concentrations, bioaccumulation and stable isotopic composition. Biological samples included zooplankton samples, fish samples, in particular individuals of two native benthopelagic and omnivorous fish species (Orestias agassii and Orestias luteus), and of a pelagic piscivorous species (Odontesthes bonariensis). Blood samples from two endemic water bird species, Fulica ardesiaca (Andean Coot) and Rollandia roland (White-Tufted Grebe) characterized by omnivorous and piscivorous foraging behavior respectively were also collected in both lakes.

Total Hg concentrations increased significantly across the food webs in both lakes, illustrating the first evidence of Hg bioaccumulation in extreme lake ecosystems from the south American Altiplano region. Mercury concentrations were found to be 3 to 5 fold higher at each trophic level between lake Titicaca and lake Uru-Uru. This suggests that Hg levels in the food webs were directly related to the difference of in situ net production of Methylmercury (MeHg) at the base of each lake system. Lake Uru-Uru showed MeHg dissolved concentration of approximately 700±100pg/L, compared to 55±50 pg/L in lake Titicaca. This difference likely reflects a higher Hg methylation in lake Uru-Uru compared to Lake Titicaca where MeHg photodegradation is enhanced (confirmed by Hg stable isotopes analysis).

Seabirds bioaccumulate significant Hg amounts in their tissues as a consequence of their high position in marine trophic webs. Since they display contrasted foraging strategies, feeding on different trophic levels, marine birds are appropriate models to assess Hg biomagnification processes. Working on chicks is particularly interesting for tracing Hg contamination because they are fed with prey collected by their parents , who forage around the colonies during the breeding period. Therefore, chick tissues are mostly representative of Hg local contamination. Skuas are exceptional models to investigate since chicks’ diet is largely composed of seabird meat and, consequently, they generally occupy a high trophic position within their communities. This study is focused on Hg speciation and isotopic composition of blood and feathers of Antarctic (Catharacta maccormicki) and subantarctic (Catharacta lonnbergi) skua chicks from diverse breeding colonies of the Southern Ocean (Adélie Land, Crozet and Kerguelen archipelagos and Amsterdam Island). Mass dependent fractionation (MDF, δ202 Hg) values for blood and feathers clearly discriminated subtropical, subantarctic and Antarctic populations. A quite significant correlation in δ202 Hg values (R2= 0.81) was found between both tissues. Nevertheless, a lower dispersion was found in the case of blood (mean SD δ202 Hg: 0.12 ‰) than in feathers (mean SD δ202 Hg: 0.22 ‰). This is especially remarkable in the case of Adélie population with blood δ202 Hg values varying from -0.02 to 0.41‰, whereas feather displayed values from -0.56 to 0.51‰. Concerning mass independent fractionation (MIF, ∆199Hg), feathers displayed highly dispersed values whereas ∆199Hg blood values presented a geographical trend from Antarctic to the subtropics. This dispersion effect may be attributed to different integration times between the two tissues studied. In conclusion, both blood and feathers are valuable bioindicators for Hg isotopic studies, however, blood samples seem to provide more specific isotopic signature to better discriminate different populations. Differences in Hg isotopic compositions in blood and feathers presented in this study highlight the capability of Hg isotopes in seabirds to investigate the factor controlling MeHg sources and biomagnification to the food webs of the southern marine ecosystems.

Abstract not available.

Mercury (Hg) stable isotope ratio analysis (IRA) is nowadays widely used for determination of Hg sources and its biogeochemical cycling. Hg has seven stable isotopes, which fractionate during different biogeochemical processes. As a result of fractionation, specific isotopic signatures can be found in the end products, which can be further used for tracing Hg sources. Modern methods based on Multicollector Inductively Coupled Plasma Mass Spectrometry (MC-ICP-MS) enable total and species specific Hg isotopic analysis with high precision. As for all analytical techniques, a metrological approach is needed in order to assess and improve the quality of the results. Hg isotopic compositions are commonly reported as delta values relatively to the standard reference material NIST 3133. However, in IRA use of the secondary isotopic standard is very important. Nowadays, only few standards are used for quality control of isotope analysis. Due to the lack of Hg IRA secondary standards, current research is directed in producing ‘in-house’ Hg IRA secondary standard from cinnabar ore from Idrija mercury mine (Idrija, Slovenia). To get more information of elemental composition of that ore, neutron activation analysis has been done. In this work, storing conditions, different digestion ways (acid digestions, using 3 different acid mixtures, and high pressure asher), matrix effects and Hg stability were studied on Cold Vapour Generator (CVG) coupled with MC-ICP-MS. Results were compared in two laboratories using two different instruments. Negative mass-dependent fractionation (MDF) expressed as δ202Hg values of -0.90 ± 0.20 ‰ was obtained.

Atmospheric mercury (Hg) monitoring programs such as GMOS , AMNet, CAMNet and the Chinese Hg program make observations of atmospheric Hg species concentrations, but do not provide direct information on Hg emission sources. Hg stable isotope studies of emission sources (coal, mining, natural, re-emission) and atmospheric GEM have over the past decade shown great potential for better understanding Hg emission sources, and transformation and exchange processes in the atmosphere. The objective of this work was to develop a Hg isotope monitoring tool that is compatible with current mercury monitoring programs. The tool is based on the coupling of the new automated plug-and-play Model 1115i multi-valve module with the Tekran 2537X analyzer. The valve box, operated and programmed via a 2537X plug-in, routes analyzed Hg(0) at the 2537X cell vent into different channels where the Hg(0) is accumulated on traps. The traps are then manually desorbed for Hg isotope analysis by MC-ICPMS. We will present RD results on the use of gold traps and oxidizing solution traps, and illustrate results for diurnal Hg isotope monitoring. Potential applications other than diurnal monitoring are pollution-event (GEM, CO threshold) based valve switching, TGM/GEM Hg isotope monitoring using a pyrolizer, DGM/GEM isotope monitoring using an equilibrator, or gradient Hg isotopes using soils wells, snow or atmosphere towers, etc.

With the 2011 promulgation of the Mercury and Air Toxics Standards by the U.S. Environmental Protection Agency, and the successful negotiation by United Nations Environment Programme of the Minamata Convention, global emissions of mercury (Hg) to the atmosphere from anthropogenic sources are expected to decline. Recent reports suggest regional gaseous Hg declines have already begun well before they were anticipated; however, providing independent evidence for the drivers of such declines is difficult. To address this challenge, the U.S. Geological Survey and National Atmospheric Deposition (NADP) program have initiated a national-scale effort to establish a baseline of total gaseous mercury (TGM) and Hg stable isotopic compositions at 31 sites distributed across North America. Samples were collected on gold traps and integrated over a 2-week period every other month using a simple static air pump. Collaboration with national scale air quality monitoring networks, such as the Mercury Deposition Network (MDN) and the Atmospheric Mercury Network (AMNet) has provided the backbone for linking the baseline to long-term wet Hg deposition and gaseous Hg monitoring sites, with approximately two thirds of the bulk air samplers operating at MDN or AMNet sites. The network includes a highly diverse set of sites ranging from remote locations (Denali National Park, Alaska) to highly urbanized sites (Bronx, New York and Boston, Massachusetts). For all sites, the average odd isotope mass independent fractionation (MIF) was -0.23‰ (D199Hg) and had a range of -0.38‰ to 0.01‰. We observed a positive correlation between D199Hg and D200Hg (r2 = 0.68). Mass dependent fractionation (MDF, d202Hg) was generally positive with a mean value of 0.27‰, although negative MDF was also observed at some sites. For example, near the highly contaminated Y12 Plant at Oak Ridge National Lab, Tennessee, d202Hg was consistently very light (-0.40‰), similar to co-located sediment. A comparison of urban vs background sites showed no difference in D199Hg, but there was a difference in d202Hg of 0.28‰ (urban sites consistently lighter). Differences in odd MIF and MDF among regions were small; however, there were interesting trends within some regions. For example, in eastern North America D199Hg was negatively correlated with d202Hg (r2 = 0.86). The long-term goal of this effort is to help provide a better understanding of anticipated future changes in atmospheric Hg concentrations and isotope abundances to evaluate the success of these regulations.

Natural gas (NG) represents an important and rapidly growing global energy source, and some commercially relevant reserves of NG are known to contain mercury (Hg) at concentrations between 0.01 and 5000μg/m3. The amount of gaseous elemental mercury released to the atmosphere from gas production and combustion is largely unknown, but is likely an increasing contribution to the global atmospheric Hg pool. Some models of changing sources of Hg to the atmosphere rely on Hg stable isotope compositions of atmospheric sources as baseline parameters. However, no Hg isotopic compositions have been published for Hg entering the atmosphere from NG. In an effort to characterize the isotopic composition of Hg released from NG, we analyzed the stable isotopic compositions of used mercury removal unit (MRU) catalysts (PURASPEC) supplied by Johnson Matthey Inc. MRU catalyst samples were analyzed from production facilities that process NG from different gas fields (Malaysia, North Sea, Brazil, N. Europe, Australia). Catalyst samples had a wide range of δ202Hg values (4.5‰) and a narrower but still large range of ∆199Hg values (0.74‰). We suggest that the bulk of Hg adsorbed to catalyst near the inlet of each MRU reactor is representative of the Hg isotopic composition of the NG source. Different gas fields range in δ202Hg and ∆199Hg from -3.75‰ to -0.68‰ and -0.09‰ to 0.65‰, respectively. Analysis of four samples from different positions within a single MRU reactor demonstrates significant fractionation of a small fraction of Hg that is not removed at the entrance to the MRU. We suggest that this fractionation is due to sorption of Hg to the catalyst surface from the gas phase, and that this process follows a Rayleigh fractionation model with ε ≈ -0.40‰. The ∆199Hg values of these samples decrease from inlet (0.25‰) to outlet (0.08‰), which suggests that the Hg sorption to catalyst surfaces may be causing mass independent fractionation related to nuclear volume effects of the sorption reaction. These results suggest that Hg isotopic analysis may be a feasible monitoring tool for Hg emissions from NG production in some gas fields. With further analyses of NG from around the world, a global average Hg isotopic composition could be estimated (as has been done for coal) to characterize this input to atmospheric Hg isotope models.

Mercury emissions from two cement plants in China with pre-calcine technique were measured. Flue gas samples were taken with Ontario Hydro Method (OHM) and Hg within were detected with CVAAS (F732). Total Hg in solid samples such as limestone, Fe material, clay, coal, feed meal, ESP ash, clinker, cement products were determined with Lumex CVAAS. Hg isotope signatures in the flue gas and solid samples were detected with MC-ICP-MS (Nu Instruments, Nu plasma model II). Total Hg in the stack flue gas varied distinctively for two plants, with about 15 µg/m3 for plant #1, while <0.5 µg/m3 for plant #2, and most Hg were in divalent form. Hg emission factors were 76 mg Hg/ t clinker for plant #1 and 1.8 mg Hg/ t clinker for plant #2. Hg was introduced into the production system mainly by the raw material, such as limestone and Fe material, and coal constitute for another lesser portion (< 30%). Hg levels in the raw material and Hg cycling and accumulation ability in the system determined the Hg emission levels. Clinker contained no mercury, and Hg in the cement were introduced by additives, such gypsum and fly ash. No clear mass-dependent fraction (MDF) of Hg was found between raw material and flue gas samples, with δ202Hg in a range of about -1.3~-2.8‰, which were much lighter than the stack flue gas emitted from coal-fired power plants. Also, there were no mass-independent fraction (MIF) in different solid and flue gas samples.

Mercury has attracted a great concern because of its accumulative toxic effect to the organisms and human beings. Deposition of Hg from the atmosphere is regarded as an important source for aquatic and terrestrial ecosystem. Before that, mercury is subject to a series of physical and chemical processes in the atmosphere. Total mercury (THg) in daily precipitation was determined at four sites in a coastal city in Southeast China from June 2012 to May 2013. In addition, the precipitation samples during May-October (rainy season) and the rest of months (dry season) were respectively combined to examine the Hg isotopic composition. The volume-weighted mean concentration of THg ranged from 1.0 to 59.4 ng/L, with an annual value of 12.3 ng/L, and the wet deposition flux was 14.0 µg/m2/y. No significant difference of THg concentrations was observed between the rainy and dry seasons, which was inconsistent with the seasonal variation presenting higher THg in dry season due to increasing combustion in inland cities of China. A total of 8 mixed precipitation samples were first concentrated and then measured by MC-ICP-MS. All precipitation samples displayed the significant negative MDF (202Hg ranged from -2.2 to -1.5‰), the positive MIF of odd isotopes (∆199Hg and ∆201Hg ranged from 0.08 to 0.22‰ and from 0.02 to 0.12‰ respectively), as well as the positive but not evident MDF of even isotope (∆200Hg ranged from 0.01 to 0.07‰). The ratio of ∆199Hg/∆201Hg was 1.10, being close to the ratio of MIF produced by magnetic isotope effect (MIE) but different from the value caused by nuclear volume effect (NVE), which suggested that MIE could be the main mechanisms during the aqueous Hg photoreduction. The ∆199Hg value in precipitation samples was quite low, and the correlation of δ202Hg and ∆199Hg was significantly negative, which indicated that the influence of long-range transport was small. The δ202Hg and ∆199Hg in precipitation and in some environmental samples from other studies were compared. The δ202Hg and ∆199Hg of all precipitation samples fell in the ranges in coal, but whether the coal combustion source could be considered producing the fractionation of Hg isotope was not sure because the fractionation of Hg isotope would be occurred during the process of coal combustion.

Lake sediments provide an important record of changing mercury (Hg) deposition and isotopic composition. However, there are few records that preserve a record of direct deposition from the global Hg pool. A remote lake in northwestern Wyoming was studied because of its small watershed to lake surface ratio of ~2, and a lack of regional upwind Hg sources. The lake was cored and sectioned at 0.5 cm intervals. Organic soil and foliage were also collected from the watershed. The lake core samples were dated using 210Pb and 137Cs and analyzed for THg and Hg isotopic compositions including mass dependent (MDF) and mass independent (MIF) fractionation. THg concentrations ranged from 16.2 to 20.8 ng/g in pre-1850 samples before increasing to a maximum of 103 ng/g in 2005. Hg isotopic compositions of the lake sediments were relatively constant before 1850 with average values for δ202Hg (MDF), Δ199Hg (odd-MIF) and Δ204Hg (even-MIF) of -0.86‰, -0.02‰and -0.19‰ respectively. Beginning in 1850, isotope values shifted toward recent sediment values for δ202Hg, Δ199Hg and Δ204Hg of -0.45‰, 0.47‰ and -0.41‰ respectively, which are similar to values measured in precipitation at other locations in North America. Soil isotopic values for δ202Hg, Δ199Hg and Δ204Hg are -1.69‰, -0.32‰ and -0.08‰ respectively, suggesting a negligible input of Hg from soil to lake sediments. Previous studies have suggested that the mechanisms responsible for MIF in precipitation are predominantly photochemical reduction (odd-MIF) and photochemical oxidation (even-MIF). Thus, the isotopic shifts that we observe are more consistent with changes in photochemical processes in the atmosphere rather than simply additions of anthropogenic Hg to the atmosphere.

Boreal peatland ecosystems are well known as an important sink of atmospheric mercury (Hg) as well as a source of highly toxic methylmercury (MeHg) to downstream ecosystems. Therefore, a thorough understanding of the complex Hg cycling in these northern ecosystems is warranted. Stable Hg isotopes were evaluated as a tool for investigating Hg cycling in an upland-peatland ecosystem. Specifically, we examined the variability of stable Hg isotopes in exported waters in a relatively well characterized small peatland-upland watershed (~0.1 km2, Watershed S2) in Marcell Experimental Forest in northern Minnesota. Exported water samples at the watershed outlet were collected biweekly when the stream flowed in years 2014 and 2015. Additionally, we collected surface and subsurface runoff from upland soils and porewater from the lagg zones, bog cores, upland soil, and litter. We determined total Hg concentrations and stable Hg isotope compositions for these sample types. Due to the high dissolved organic matter (DOM) in the exported waters, we completely oxidized the DOM by digesting the samples with 1-5% of an acidic mixture of KMnO4 and K2S2O8 at 95oC, before we purged and trapped the sample Hg by slowly adding SnCl2, and used 1% KMnO4 traps to collect reduced Hg. Data on exported waters over time (2014- 2015) showed a large range of mass-dependent fractionation (as δ202Hg, ranging from -2.1 to -1.3 ‰; n=21) but only a small range of mass-independent fractionation (as D199Hg, ranging from -0.35 to -0.10 ‰; n=21), with D199Hg resembling the isotopic values of Hg in bog and vegetation samples. Samples with more positive δ202Hg in the exported waters were collected during the spring snow-melt period, and these δ202Hg values were similar to the values of surface and subsurface runoff from the upland (i.e., -1.4 to -1.2 ‰; n=4). Our results suggest that the temporal variation of stable Hg isotope compositions in the exported waters is relatively large in this small upland-peatland watershed, which was largely driven by the hydrology of the watershed. The differences in δ202Hg values may allow us to distinguish two specific sources of Hg in the exported waters as derived from upland forest vs. peatland bog. Therefore, stable Hg isotopes could potentially be used in studying Hg cycling in DOM-rich surface waters, provided that effective methods are used to remove the influence of DOM on subsequent Hg extraction process (e.g., purge and trap) for high-precision isotopic analysis.