1a-2: Atmospheric mercury cycling and transformations: Insights from measurements and models

The transport and transformation of mercury in the atmosphere has been a topic of discussion in the Arctic for over two decades. While considerable information has been published from Alert, Canada, less has been discussed in Canada’s western Arctic. It’s been shown that 95% of the anthropogenic Hg deposited in Canada comes from external sources and that ~27% of the anthropogenic mercury deposited in the Canadian Arctic is contributed from Asia. Little Fox Lake, situated in the Yukon Territory in Western Canada, has been identified as a good location to monitor mercury plumes from Asian long range transport into Canada. Since 2007, atmospheric mercury has been measured at Little Fox Lake where annual median concentration levels of total gaseous mercury (TGM) are 1.34 ± 0.15 ng m-3. This annual median concentration is lower than the average Canadian background levels (mean 1.47 ± 0.22 ng m-3). As well, there is a distinct seasonal/monthly median signature in the TGM levels that ranges between 1.18-1.50 ± 0.09 ng m-3). Trends of atmospheric mercury are on the decline in the northern hemisphere and across Canada, but a recent trend analysis from Little Fox Lake (2007-2014) shows a significant increase in the TGM concentration levels in all months but January and February (where the trends were not statistically significant). The MannKendall trend analysis of the Little Fox Lake data showed a monthly median increasing trend of +1.7 ± 0.6% per year (not including January and February) with a range of +0.7 to +2.5% per year. The same trend analysis was performed on data Alert, Nunavut over a similar time period (2007-2013) and showed a monthly median trend of –2.7 ± 1.6 % per year with a range of -7 to-0.8% per year. The increasing trend in the TGM concentration level may be explained by increasing emissions in Asia but also may be a result of regional forest fires or other regional input. An investigation into the atmospheric mercury concentrations at these Arctic locations will be presented.

Volcanic systems are a poorly-constrained, natural contributor to the global mercury cycle. Lack of comprehensive data has led to uncertainties in 1) the total amount of mercury emitted from volcanoes; 2) the speciation of mercury upon release from volcanic plumes; and 3) the pathways and timing of deposition following plume release. Three volcanoes (Masaya volcano, Nicaragua; Poas and Turrialba volcanoes, Costa Rica) were studied extensively to determine these uncertainties. These volcanoes represent a diverse variety of magmatic systems within the Central American Volcanic Arc. At each volcanic site, air and soil samples were collected along a distance gradient up to several kilometers downwind from the emission source. Air samples were collected using iodated-carbon traps with a portable pump either by walking through the plume or via drone for the analysis of gaseous and particulate mercury. Soil samples at each site consisted of a depth profile for the analysis of total mercury to establish a timeline of mercury deposition. A suite of aqueous samples was also collected, including rainwater, condensates from a hydrothermal vent, and water from hyper-acidic crater lakes; in situ filtration was used to differentiate dissolved versus particulate mercury. Analysis of the samples was carried out at the Class 100 Ultra-Clean Trace Elements Laboratory (UCTEL) at the University of Manitoba. In addition, real-time atmospheric mercury speciation measurements (gaseous elemental mercury, gaseous oxidized mercury and particulate-bound mercury) were carried out intermittently at locations that are adjacent to Poas and Turrialba volcanos. Preliminary results show that the total gaseous mercury concentrations in the air reach near ambient levels within a short distance from the plume center. Total mercury concentrations in all the water samples are in the order of tens of ng/L. Methylated forms of mercury have also been detected within two volcanic crater lakes and in the condensate from a hydrothermal vent. Analysis of the air trap and soil samples is ongoing. This study represents a comprehensive investigation into the emission and fate of volcanic mercury in Central America.

A unique set of mercury measurements in the upper troposphere and lower stratosphere (UT/LS) has been obtained during the monthly CARIBIC (www.caribic-atmospheric.com) intercontinental flights between May 2005 and February 2016. The passenger Airbus 340-600 of Lufthansa covered routes to North America, East and South East Asia, and the southern hemisphere. The accompanying measurements of CO, O3, NOy, H2O, aerosols, halocarbons, hydrocarbons, greenhouse gases, and several other parameters as well as backward trajectories enable a detailed analysis of the measurements.

Post-flight processed data since April 2014 provide new insights into the spatiotemporal distribution and speciation of mercury in the UT/LS. We will compare the data with previous observations, estimate the stratospheric mercury lifetime, and discuss a conceptual model of stratospheric mercury cycle.

Global inventories suggest increasing anthropogenic atmospheric mercury (Hg) emissions in the past two decades, especially in the East and South Asian regions due to growing industrial activities and energy demands. However, observations at Mauna Loa Observatory and sites in North America and Europe showed decreasing trends in atmospheric Hg concentrations from 1990 to present, inconsistent with current global inventories. Such a trend analysis has not been reported from sites in East Asia due to the lack of long-term monitoring data. Here we reported the trend and source-receptor relationship of atmospheric Hg observed at the Lulin Atmospheric Background Station (LABS), a tropical mountain site in central Taiwan (23.47ºN, 120.87ºE, 2862 m a.s.l.), between 2006 and 2016. Concentration-weighted trajectory (CWT) approach was applied to identify source regions of each atmospheric Hg species, including gaseous elemental Hg (GEM), gaseous oxidized Hg (GOM) and particulate Hg (PHg). Southwest and Southeast China and northern Indochina Peninsula were the common major source regions of all species. For GEM, an additional source region was identified extending from northeast China along the coastal region of east China. Moreover, air from South China Sea and the Pacific Ocean could also be enriched with GOM, possibly due to GEM oxidation. Trend analysis of GEM was performed by using the Sen’s slope approach. A significant decreasing trend in GEM concentrations was observed with a rate of -2.1% yr-1 (-0.0336 ng m-3 yr-1) over the 2006-2016 period. This value is similar to those reported from sites in North America and Europe. Further analysis found significant decreasing trend between September and May when LABS is mainly under the influence of air masses from the East Asia continent, but no trend or slightly increasing trend from June to August when marine air masses prevail. Concurrently monitored CO concentrations also showed a significant decreasing trend with a rate of -2.1% yr-1 (-2.904 ppb yr-1) over the same time period. The similar decreasing trends in GEM and CO concentrations suggest that changes in anthropogenic emissions may have played a role. However, other factors, such as changes in re-emission flux, atmospheric chemistry and regional transport pattern, may also contribute. We are still working on detail analysis about the relationships between GEM and other parameters and the trends of GEM concentrations associated with various air mass clusters. Updated results will be reported in the 2017 ICMGP.

In January 2011 NOAAs Air Resources Laboratory assumed oversight of speciated atmospheric mercury measurements at the Mauna Loa Observatory (MLO), and the site joined the Atmospheric Mercury Network (AMNet). MLO is one of six NOAA baseline monitoring stations for the study of the background global atmosphere, and is located at an elevation of 3,397 m on the northern slope of the Mauna Loa volcano on the Big Island of Hawaii. A single Tekran speciation system measures gaseous elemental mercury (GEM), gaseous oxidized mercury (GOM), and particulate-bound mercury (PBM) with nominal 1-hr resolution. Since 2011, measurements of ozone (O3), sulfur dioxide (SO2), and carbon monoxide (CO) were also added.

The site is an ideal high-altitude location from which to monitor the global background atmosphere. Free tropospheric flow at the elevation of MLO is typically from the east and northeast (trade wind circulation), with local upslope/downslope circulation superimposed on the larger-scale flows. Measured mercury species concentrations will be presented with respect to diurnal, seasonal, and annual variations, and trends in mercury concentrations at the site will be discussed. Source regions associated with both high and low mercury concentrations will be investigated through gridded trajectory frequency analysis using NOAAs HYSPLIT model.

Finally, a brief summary of measurement accuracy assessment activities for GOM and PBM will be presented.

Gaseous elemental mercury (GEM, Hg) emissions are transformed to divalent reactive Hg (RM) forms throughout the troposphere and stratosphere. RM is often operationally quantified as the sum of particle bound Hg (PBM) and gaseous oxidized Hg (GOM). The measurement of GOM and PBM is challenging and under mounting criticism. Here we intercompare six months of automated GOM and PBM measurements using a Tekran® KCl-coated denuder and quartz regenerable particulate filter (rpf) method with RM collected on cation exchange membranes (CEMs) at the high altitude Pic du Midi Observatory. We find that denuder/rpf sampled RM is systematically lower by a factor of 1.3 than RM on CEMs. We observe a significant relationship between GOM (but not PBM) and Tekran® flush blanks suggesting significant loss (32%) of labile GOM from the denuder or inlet. Adding the flush blank to denuder/rpf based RM results in good agreement with CEM based RM (slope=1.01, r2=0.90) suggesting we can correct bias in denuder/rpf RM and denuder based GOM. We provide a bias corrected (*) Pic du Midi dataset for 2012-2014 that shows GOM* and RM* levels in dry free tropospheric air of 198 and 229 pg m-3 which agree well with in-flight observed RM and with model based GOM and RM estimates.

Observations from the Mercury Deposition Network (MDN) at Huntington Wildlife Forest (HWF) suggested that a significant decline in Hg concentrations in precipitation was linked to Hg emission decreases in the United States, especially in the Northeast and Midwest, and yet Hg wet deposition has remained fairly constant over the past two decades. The present study was aimed to investigate how climatic, terrestrial, and anthropogenic factors had influenced the Hg wet deposition flux in upstate New York (NY). To achieve this, an improved Community Multiscale Air Quality (CMAQ) model was employed, which included state-of-the-art Hg and halogen chemistry mechanisms. A base simulation and three sensitivity simulations were conducted. The base simulation used 2010 meteorology, U.S. EPA NEI 2011, and GEOS-Chem initial and boundary conditions (IC, BC). The three sensitivity runs each changed one condition at the time as follows: 1) NEI 2005 Hg anthropogenic emission out of NYS instead of NEI 2011, 2) 2005 meteorology instead of 2010, and 3) no in-state Hg anthropogenic emission. The study period of all the simulations was March November 2010, and the domain covered the northeastern United States at 12 km resolution. As a result, compared with rural areas in NYS, Hg wet deposition and ambient Hg concentrations in urban areas were affected more significantly by in-state anthropogenic Hg emission. The in-state anthropogenic Hg emissions contributed ~25% of Hg wet deposition at urban sites and <1% at rural sites during the study period. In contrast, the out-of-state anthropogenic Hg emissions had slightly higher (~3%) influence on NYS rural areas compared with urban regions. Using 2005 anthropogenic Hg emissions, around twice of those in 2010, out-of-NYS emissions increased the total in-state Hg wet deposition by 16%. Hg wet deposition flux was greatly affected by meteorological conditions, causing changes varying from a 91% decrease to a factor of 5 increase in monthly accumulated wet deposition amounts. The possible affecting meteorological factors included, not limited to, solar radiation, cloud height, wind speed and direction, precipitation, and relative humidity, among which precipitation had the largest effects in most areas.

A modelling study of the emission, transport, transformation and deposition of mercury (Hg) over South Africa has been performed. The study employed a state-of-the-art regional chemical transport model, WRF/Chem with Hg, which is based on the Weather Research and Forecasting (WRF) Model. The study used employed two global anthropogenic emissions databases, two atmospheric mercury oxidation chemistry mechanisms, and different combinations of boundary and initial conditions, in order to gain some insight into the possible level of uncertainty associated with the model output. The model output from all the tested configurations suggest that much of the Hg deposition in some areas of South Africa is due to local emissions, particularly in the Highveld. Hg deposition in other regions of South Africa however is almost totally dominated by long-range transport of Hg from remote sources, in particular the Northern, Western and Eastern Cape provinces. The characteristics of the different climate zones in South Africa, in particular rainfall patterns, all influence the Hg deposition patterns, total depend on deposition fluxes both long-range and local sources, although the relative influence of these sources changes significantly across the country. It proved difficult to reproduce some of the higher gas phase Hg observations, and the reason for this is not immediately clear. It may be due to the time averaged emissions in the model which possibly does not accurately reflect Hg release from industrial installations. It could also be the result of particular meteorological conditions favouring release of Hg from previously contaminated sites. Certainly further investigation is warranted, particularly given the at times significantly high observations (similar to heavily populated and industrialised regions of China), and their unfortunate scarcity. The lack of measurement data for South Africa in general means that it is difficult to make hard and fast comments on the models ability to reproduce atmospheric Hg species concentrations and deposition flux magnitudes. However, the simulations do suggest that most of the Hg released by anthropogenic activities in South Africa is deposited within South African territory, with a lesser fraction transported eastward/south eastward towards the Indian Ocean, potentially influencing Hg deposition in Mozambique and Madagascar.