G4A (I) Atmospheric Mercury: Transport and deposition

Recent studies have reported elevated mercury (Hg) concentrations, above 0.5 ppm (higher than EPA guidelines), in fish caught around Bermuda, where the population regularly consumes fish. Elevated Hg levels in maternal cord blood of 41.3 nmol/L, approximately 8 fold above the baseline exposure levels (5 nmol/L) for North American populations have also been reported. This has raised concern on possible adverse effects that may arise from methylmercury (MeHg) toxicity. Most Hg in the ocean and other remote locations comes from atmospheric sources. Given Bermuda’s isolated location, it’s expected that fish MeHg levels would be low even though there are local sources e.g. a waste incinerator. Rain and air sampling was done on-shore and offshore on cruises, from 2008-2010. In addition, particulate samples were collected on a weekly basis. The volume weighted average concentration in rain was 4.7 ng/L, while the levels in air range from 1.2-1.8 ng/m³. On average, particulate mercury was 12.4 pg/m³. These values do not indicate that the region is receiving highly contaminated atmospheric inputs, and is reflective of the regional signal for the temperate northern Hemisphere, as has been found for other metals and pollutants. The levels in air are similar to measurements made in the open ocean. The measurements in rain are comparable to the concentration (5.6 ng/L) at Avery Point, a coastal site located in Connecticut, on the shores of Long Island Sound. The level of particulate Hg is comparable with data reported from a statewide survey carried out in Connecticut (10.5 ng/m³). Further comparison will be made to measurements made in urban and rural sites in the Southern hemisphere: Cape Point and Pretoria, South Africa. The role of air-sea exchange in atmospheric Hg cycling is supported and will be explored based on an evaluation of underway dissolved gaseous mercury (D.G.M.) measurements on the cruises. In addition, the relative importance of local versus long-range sources will also be examined and meteorological data used to establish possible sources of Hg in wet deposition collected on Bermuda.

While it is recognized that the U.S. Gulf Coast region receives relatively high mercury (Hg) wet deposition, the contribution of Hg from dry deposition is less well constrained. Previous studies suggest that reactive gaseous Hg (RGM) may adsorb onto rapidly-settling, coarse (≥2.5µm) sea salt aerosols which are abundant in coastal settings (Engle et al., 2008, Appl Geoch. 23, 419). What is unclear is the fate of this particulate Hg once it deposits onto an aquatic ecosystem where it can potentially become methylated and bioaccumulate. To address this question, in May, 2010, atmospheric particulate matter (PM) was collected along the Mississippi Gulf Coast, USA in coarse and fine (<2.5µm) fractions, for two thirty-hour intervals. The PM-bearing filters were then incubated for periods ranging from 1 hour to 1 week in aliquots of two “end member” surface waters: Grand Bay water (approx. 70% marine water, 30% fresh water); and Escatawpa River water (high DOC, acidic black water). Following incubation, the PM-filters and water samples were analyzed for particulate and methyl-Hg at the USGS mercury laboratory in Middleton, Wisconsin. In addition to the Hg analyses, a set of 10 fractions of PM, ranging in aerodynamic diameter from <0.18 to >18 µm, was collected over the course of the study and analyzed for trace elements.

Trace element enrichment factors were calculated to infer their likely source (anthropogenic vs. geogenic). When these results are combined with size segregated profiles, it appears that the fine fraction is dominated by particles formed by condensation of combustion sources whereas the coarse fraction is dominated by mechanically derived particles such as sea salt and crustal material. Mercury was leached more readily from the filters into the Escatawpa River water than into the Grand Bay water for both the fine and coarse fractions. In each case the soluble portion of the fine fraction (up to 61%) reached near steady state Hg concentrations after only one hour of incubation. Over the longer incubation (t = 1 week), however, the coarse fraction was more soluble in Grand Bay water (up to 78%) than in Escatawpa River water (as low as 28%). When the solubilization data are coupled with deposition velocities, the results suggest that large (>10µm) aerosols are the dominant source of Hg to the study area, contributing twice as much Hg as all other size fractions combined.

It is widely acknowledged that there are large uncertainties in our understanding of the global mercury cycle. However, uncertainty analysis of current global-scale models for mercury (e.g., CTM-Hg, GRAHM, ECHMERIT, GEOS-Chem) has not been conducted. This is likely because of the heavy computational costs associated with modeling the atmospheric chemistry and transport of mercury using these models. In this work we use a multimedia modeling approach to estimate the uncertainties in the global mass balance (fluxes, concentrations and inventories) of mercury in active circulation in air, oceans, soil and vegetation. A new spatially-resolved, descriptive multimedia model was developed to explicitly describe mercury cycling in the environment and a Monte Carlo uncertainty analysis conducted on its unit-world variant which has high computational efficiency (25,000 simulations in < 8 hours on a desktop computer). This efficiency is achieved by representing all processes as pseudo-first order, and by modeling atmospheric chemistry of mercury using species concentration ratios. Results from our modeling agree well with observations and with results from the above listed chemistry-transport models. We find that the 95% confidence level in the modeled atmospheric deposition varies within a factor ~36-45; modeled residence time of mercury in the atmosphere varies within a factor of about 10. We further identify a limited number of model input parameters whose uncertainties contribute most to the uncertainties in the global mercury cycling. Processes described by these parameters are: (i) reduction and oxidation reactions in surface oceans and the amount of reducible mercury present in surface oceans (up to 50% or more contribution to the uncertainty in many model outputs); (ii) mercury interconversion reactions in the atmosphere (up to 20% contribution to the uncertainty in some outputs); (iii) mercury mass transfer processes (up to 10-30% contribution in some outputs); and (iv) vegetation characteristics (mostly influencing the mercury exchange between air and land surface). Uncertainties in these four processes contribute more than uncertainties in anthropogenic emissions, which only contribute 10% or less to the uncertainties in many model outputs. We therefore recommend that more research is directed towards a better representation of the above identified four processes.

Recently, the Arctic Monitoring and Assessment Programme (AMAP) re-analyzed 1990-2005 anthropogenic global mercury emissions and developed a compatible set of historical global emission inventories that are suitable for modeling the impact of changing mercury emissions on global distribution of atmospheric mercury. The re-analyzed inventories use a common methodology and a more consistent information base for estimating certain emissions. Details are found in a report on this work available at AMAP website. Environment Canada’s mercury model GRAHM (Global/Regional Atmospheric Heavy Metals Model) was applied to assess the impact of changing anthropogenic emissions and meteorology on the mercury concentrations and deposition globally from 1990 to 2005. The changes in meteorology reflect the inter-annual variability and climate change during this period.

GRAHM is an Eulerian multi-scale atmospheric model which simulates all meteorological processes and Hg related physico-chemical processes in a single system. The model simulates transport, transformation and surface exchange of Hg fluxes. The sources of Hg emissions in the model include anthropogenic, natural and reemission of previously deposited Hg from land and oceans. GRAHM was previously shown to give satisfactory results. Continuous run from 1990 to 2005 was performed to develop a comprehensive historical picture of the global distribution of mercury. The Japanese 25-year Reanalysis Project (JRA-25) data were used to provide meteorological and climatological fields to the model.

Globally mercury deposition shows very small trend, however regionally the changes are large. Deposition significantly increased for the given period in East and South Asia, 26% and 8% correspondingly. Whereas deposition in East Asia steadily rises, deposition in South Asia peaked in 1997. Besides these regions, there is only one other region, Australia & Oceania, which has noticeable positive trend of 2.5%. Large decrease of the deposition occurred in Europe, 35%, and to less extent in Central Asia and North America, 18% and 14% correspondingly. South America and Africa have no noticeable tendency. Deposition to Arctic exhibits large variability, however without obvious trend, which shows larger impact of meteorology and global sources of mercury to this region. Complete analysis of the results will be presented at the conference making use of the observed trends where possible.

Biomass burning is a significant source of mercury to the global atmosphere. The boreal forest, in particular, is believed to harbour a large amount of mercury that is released during wildfires – fires that are expected to increase in magnitude and frequency with climate change. The impact of these releases on the mercury cycle in the Arctic region is of particular interest, given ongoing concerns with mercury levels in local wildlife. However, mercury emissions from boreal forest fires are not well constrained. In order to assess the relative contributions of natural and anthropogenic sources of mercury to the Canadian atmosphere before any regulations are implemented, this gap of knowledge must be understood. Continuous monitoring of mercury (Hg) and other pollutants such as carbon monoxide (CO) at two sites in western Canada provides the opportunity to capture emission signatures such as Hg/CO ratios from boreal forest burning in the region and thereby better constrain emissions from this particular source.

Atmospheric mercury concentrations have been monitored at Little Fox Lake, Yukon, since 2007 and Whistler, British Columbia, since 2008, under the International Polar Year project INCATPA (Intercontinental Atmospheric Transport of anthropogenic Pollutants to the Arctic), the Clean Air Regulatory Agenda (CARAII) mercury program and the Northern Contaminants Program (NCP). Early results include observations of plumes at Whistler with Hg/CO ratios typical of (a) industrial emissions and (b) biomass burning emissions, and the identification of several large forest fires as sources of elevated Hg at Little Fox Lake. Current results are presented along with ongoing mercury research plans at these sites.

The fate of mercury deposited onto snow- and ice-covered surfaces is of critical importance for atmospheric mercury models. At high-latitudes, springtime Atmospheric Mercury Depletion Events (AMDEs) are accompanied by important deposition of oxidized mercury to the cryosphere. A significant portion of the deposited mercury may revolatilise from the cryosphere rapidly. However, a combination of physical and chemical environmental factors may trap a substantial portion of the deposited mercury within the cryosphere for an extended period of time. During snowmelt, the cryospheric mercury may enter the snowpack’s meltwater. The subsequent emission of mercury from the snowpack’s meltwater can be an important source of atmospheric mercury. Indeed, it is possible that the high-latitude summertime increase in the concentration of surface-level atmospheric gaseous elemental mercury is caused by significant emission of mercury from cryospheric meltwater. To simulate these various processes, a new atmosphere/cryosphere parameterisation has been developed for atmospheric mercury models. This parameterisation predicts the concentrations of elemental and oxidised mercury in the snowpack surface layer, the underlying snowpack, and the snowpack meltwater. Emission of elemental mercury from the snowpack and its meltwater to the atmosphere is represented.

In this presentation, we will describe the developed snowpack/meltwater mercury model. This parameterisation has been incorporated into Environment Canada’s atmospheric mercury model, the Global/Regional Atmospheric Heavy Metals model (GRAHM). We will discuss the performance of the updated GRAHM. We find that the addition of the snowpack/meltwater model significantly improves the seasonality of GRAHM’s simulated surface-level atmospheric gaseous elemental mercury concentrations.

Gaseous Elemental Mercury (GEM) was investigated in the troposphere and in the interstitial air extracted from the snow at depths ranging from the surface to 1.6 meters at Dome C Concordia Station, (central Antarctica,75°06S, 123°20E, 3220 m above sea level) located 1100 km away from the nearest coast of East Antarctica during January 2009.

Measurements in the boundary layer combined with modeling of vertical turbulent mixing revealed evidence of fast GEM oxidation processes inside this layer during low irradiation periods. However, unexpected high GEM concentrations for remote places (> 3 ng/m3) were measured under higher solar radiations. GEM variations were well anti-correlated with ozone levels resulting in strong diel variations of GEM. Photoreduction at the air-snow interface and fast oxidation inside the snow pack were also observed. Chemical mechanisms involving bromine compounds can be a realistic explanation for these phenomena. Given the size of the Plateau, the impact of this specific reactivity may have consequences at a larger scale on the global cycle of mercury.