Key factors influencing Hg levels and trends in unperturbed oligotrophic temperate and boreal lakes.

Mercury (Hg) is a toxic metal that presents a major risk to ecosystems, biota, human health, and remains a priority concern. In temperate and boreal lakes Hg and methylmercury (MMHg) are expected to vary as a function of atmospheric Hg deposition, lake water chemistry, catchment characteristics and climate variables. The aim of this study was to quantify Hg and MMHg in unperturbed oligotrophic lakes and to identify the factors controlling their distribution. We first hypothesized that lake Hg (and MMHg to lesser extent) spatial variations are linked to atmospheric deposition, catchment characteristics, and terrestrial exportation of dissolved organic carbon (DOC). We secondly examined if lake Hg concentrations have followed the decrease in atmospheric Hg emission observed between the mid-1990s to the end-2010s. We found that overall, atmospheric Hg has little impact on lake Hg and MMHg concentrations, which are both primarily influenced by DOC input originating from the forest catchment. The relationship between DOC and Hg differed between the spring and the fall, with a Hg-to-DOC ratio twice as high in spring. This seems related to snowmelt input of Hg (with a relatively reduced input of DOC) or the internal lake build-up of Hg during the ice-covered period. Of the 10 lakes intensively visited over a 20-year period, only 3 showed significant lake Hg decreases despite significant negative trends in atmospheric Hg concentrations, suggesting a lag between atmospheric and surface water temporal trends. Overall, terrestrial catchments retain around 80% of atmospheric Hg implying that large Hg pools have been built up in soils in the last decades. As such, the reduction of atmospheric Hg alone will not necessarily result in Hg decreases in lakes, since the Hg concentrations may be modulated by DOC export trends and catchment characteristics. This stresses the need to improve our understanding of the processes governing Hg transfers from catchments into lakes.

How well do environmental archives of atmospheric mercury deposition in the Arctic reproduce rates and trends depicted by atmospheric models and measurements?

This review compares the reconstruction of atmospheric Hg deposition rates and historical trends over recent decades in the Arctic, inferred from Hg profiles in natural archives such as lake and marine sediments, peat bogs and glacial firn (permanent snowpack), against those predicted by three state-of-the-art atmospheric models based on global Hg emission inventories from 1990 onwards. Model veracity was first tested against atmospheric Hg measurements. Most of the natural archive and atmospheric data came from the Canadian-Greenland sectors of the Arctic, whereas spatial coverage was poor in other regions. In general, for the Canadian-Greenland Arctic, models provided good agreement with atmospheric gaseous elemental Hg (GEM) concentrations and trends measured instrumentally. However, there are few instrumented deposition data with which to test the model estimates of Hg deposition, and these data suggest models over-estimated deposition fluxes under Arctic conditions. Reconstructed GEM data from glacial firn on Greenland Summit showed the best agreement with the known decline in global Hg emissions after about 1980, and were corroborated by archived aerosol filter data from Resolute, Nunavut. The relatively stable or slowly declining firn and model GEM trends after 1990 were also corroborated by real-time instrument measurements at Alert, Nunavut, after 1995. However, Hg fluxes and trends in northern Canadian lake sediments and a southern Greenland peat bog did not exhibit good agreement with model predictions of atmospheric deposition since 1990, the Greenland firn GEM record, direct GEM measurements, or trends in global emissions since 1980. Various explanations are proposed to account for these discrepancies between atmosphere and archives, including problems with the accuracy of archive chronologies, climate-driven changes in Hg transfer rates from air to catchments, waters and subsequently into sediments, and post-depositional diagenesis in peat bogs. However, no general consensus in the scientific community has been achieved.

Using the INCA-Hg model of mercury cycling to simulate total and methyl mercury concentrations in forest streams and catchments.

We present a new, catchment-scale, process-based dynamic model for simulating mercury (Hg) in soils and surface waters. The Integrated Catchments Model for Mercury (INCA-Hg) simulates transport of gaseous, dissolved and solid Hg and transformations between elemental (Hg(0)), ionic (Hg(II)) and methyl (MeHg) Hg in natural and semi-natural landscapes. The mathematical description represents the model as a series of linked, first-order differential equations describing chemical and hydrological processes in catchment soils and waters which we believe control surface water Hg dynamics. The model simulates daily time series between one and 100 years long and can be applied to catchments ranging in size from <1 to ~10,000 km(2). Here we present applications of the model to two boreal forest headwater catchments in central Canada where we were able to reproduce observed patterns of stream water total mercury (THg) and MeHg fluxes and concentrations. Model performance was assessed using Monte Carlo techniques. Simulated in-stream THg and MeHg concentrations were sensitive to hydrologic controls and terrestrial and aquatic process rates.

Assessment of modeled mercury dry deposition over the Great Lakes region.

Three sets of model predicted values for speciated mercury concentrations and dry deposition fluxes over the Great Lakes region were assessed using field measurements and model intercomparisons. The model predicted values were produced by the Community Multiscale Air Quality Modeling System for the year 2002 (CMAQ2002) and for the year 2005 (CMAQ2005) and by the Global/Regional Atmospheric Heavy Metals Model for the year 2005 (GRAHM2005). Median values of the surface layer ambient concentration of gaseous elemental mercury (GEM) from all three models were generally within 30% of measurements. However, all three models overpredicted surface-layer concentrations of gaseous oxidized mercury (GOM) and particulate bound mercury (PBM) by a factor of 2-10 at the majority of the 15 monitoring locations. For dry deposition of GOM plus PBM, CMAQ2005 showed a clear gradient with the highest deposition in Pennsylvania and its surrounding areas while GRAHM2005 showed no such gradient in this region; however, GRAHM2005 had more hot spots than those of CMAQ2005. Predicted dry deposition of GOM plus PBM from these models should be treated as upper-end estimates over some land surfaces in this region based on the tendencies of all the models to overpredict GOM and PBM concentrations when compared to field measurements. Model predicted GEM dry deposition was found to be as important as GOM plus PBM dry deposition as a contributor to total dry deposition. Predicted total annual mercury dry deposition were mostly lower than 5 µg m(-2) to the surface of the Great lakes, between 5 and 15 µg m(-2) to the land surface north of the US/Canada border, and between 5 and 40 µg m(-2) to the land surface south of the US/Canada border. Predicted dry deposition from different models differed from each other by as much as a factor of 2 at regional scales and by a greater extent at local scales.

Spatial trends and historical deposition of mercury in eastern and northern Canada inferred from lake sediment cores.

Recent and historical deposition of mercury (Hg) was examined over a broad geographic area from southwestern Northwest Territories to Labrador and from the U.S. Northeast to northern Ellesmere Island using dated sediment cores from 50 lakes (18 in midlatitudes (41-50 degrees N), 14 subarctic (51-64 degrees N) and 18 in the Arctic (65-83 degrees N)). Distinct increases of Hg overtime were observed in 76% of Arctic, 86% of subarctic and 100% of midlatitude cores. Subsurface maxima in Hg depositional fluxes (microg m(-2) y(-1)) were observed in only 28% of midlatitude lakes and 18% of arctic lakes, indicating little recent reduction of inputs. Anthropogenic Hg fluxes adjusted for sediment focusing and changes in sedimentation rates (deltaF(adj,F)) ranged from -22.9 to 61 microg m(-2) y(-1) and were negatively correlated (r = -0.57, P < 0.001) with latitude. Hg flux ratios (FRs; post-1990)/pre-1850) ranged from 0.5 to 7.7. The latitudinal trend for Hg deltaF(adj,F) values showed excellent agreement with predictions of the global mercury model, GRAHM for the geographic location of each lake (r = 0.933, P < 0.001). The results are consistent with a scenario of slow atmospheric oxidation of mercury, and slow deposition of reactive mercury emissions, declining with increasing latitude away from emission sources in the midlatitudes, and support the view that there are significant anthropogenic Hg inputs in the Arctic.