Two new sources of reactive gaseous mercury in the free troposphere

Introduction Conclusions References


Introduction
During the last decade the measurements of atmospheric mercury (Hg) have developed substantially.Speciated Hg measurements (including gaseous elemental mercury (GEM), reactive gaseous mercury (RGM), and particle bound mercury (PBM); Introduction

Conclusions References
Tables Figures

Back Close
Full  Landis et al., 2002) are providing important information about mercury sources, transformation and subsequent deposition (e.g.Talbot et al., 2008;Peterson et al., 2009;Gustin and Jaffe, 2010;Sprovieri et al., 2010).Mercury is mainly emitted to the atmosphere in its elemental form, which has a long lifetime allowing global transport (Jaffe et al., 2005;NRC, 2010).Atmospheric oxidation of GEM generates gas-and particle phase oxidized Hg compounds (thought to be primarily inorganic Hg(II) species) that deposit quickly as they are more reactive, more water-soluble, and less volatile (Schroeder and Munthe, 1998;Lin and Pehkonen, 1999;Holmes et al., 2009;NRC, 2010).The spatial distribution of mercury deposition in many instances might thus depend more on atmospheric conditions (e.g.wind direction, oxidant concentrations, temperature) than on proximity to mercury sources (Chand et al., 2008;Sprovieri et al., 2010;Holmes et al., 2010;Rothenberg et al., 2010).
Information on transport and transformation of Hg in the free troposphere (FT) is scarce due to a poor understanding of sources of RGM and lack of long-term measurements at high elevations.The few Hg observations above the boundary layer (BL) demonstrate that the upper troposphere/lower stratosphere (UT/LS) is depleted in GEM and enriched in reactive mercury (RM = RGM + PBM; Murphy et al., 2006;Swartzendruber et al., 2006Swartzendruber et al., , 2008;;Talbot et al., 2007Talbot et al., , 2008;;Lyman and Jaffe, 2011).Oxidation of GEM in the stratosphere (coupled with stratosphere to troposphere transport) is the only known source of RGM above the BL (Murphy et al., 2006;Swartzendruber et al., 2006;Lyman and Jaffe, 2011).However, based on previous studies and calculations this source accounts for only a small fraction of tropospheric RGM (Lyman and Jaffe, 2011).Although halogen species might also play a role in GEM oxidation in the FT, the existing observations and model studies have not clearly proven this.Therefore, the sources and RGM formation mechanisms in the bulk of the global atmosphere remain poorly characterized.Using observations of speciated mercury (GEM, RGM, and PBM), submicron aerosol scattering (σ sp ), trace gases (ozone, O 3 ; carbon monoxide, CO) and meteorology from the Mt.Bachelor Observatory (MBO), we describe in this paper two new sources of RGM in the FT.Introduction

Conclusions References
Tables Figures

Measurements
Speciated Hg (gaseous elemental mercury (GEM), reactive gaseous mercury (RGM), and particle bound mercury (PBM)) measurements have been conducted at MBO since 2005 using a Tekran 2537A Hg vapor analyzer and a Tekran 1130/1135 Hg speciation system.A detailed description of the sampling system can be found in Swartzendruber et al. (2006).Briefly, a Teflon-coated aluminum (Al) cyclone inlet (URG Corp.), coupled to a Teflon-coated Al high-volume inlet (URG), was used for sampling.The flow through the high-volume inlet was maintained between 50 and 100 std l min −1 (standard pressure of 1.01 bar and temperature of 273.15 K).The speciation system sampled 7.5 std l min −1 from the high-volume inlet.A KCl coated annular denuder was used to collect RGM and a quartz fiber filter was used to collect PBM.Downstream of the speciation system the Tekran 2537A sampled an additional 0.65 std l min −1 from the airstream and measured GEM at 5 min intervals.The total flow through the cyclone inlet yielded a particle size cut of 2.5 µm.To improve the RM detection limit, RGM and PBM samples were collected for 3 h prior to analysis.During the RGM and PBM analysis cycle (60 min) GEM was not measured.The Hg instruments were calibrated with the elemental mercury (Hg 0 ) permeation source internal to the Tekran 2537A.The emission rate of the permeation source was verified by injections of Hg 0 from a Tekran Introduction

Conclusions References
Tables Figures

Back Close
Full 2505 saturated Hg 0 vapor source, using a gas-tight microliter syringe (Hamilton).There are uncertainties about the calibration and interferences in the Tekran RGM measurements (Lyman et al., 2010;Gustin et al., 2012;Ambrose et al., 2012).Nonetheless, until these issues can be definitively resolved, we assume the Tekran RGM measurements represent all oxidized forms of Hg in the atmosphere, similar to most researchers.
Measurements of CO, O 3 and submicron aerosol scattering (σ sp ) are described in Ambrose et al. (2011).Briefly, Ozone and CO concentrations were measured with Dasibi 1008-RS and TECO 48C Trace level Enhanced CO analyzers (Weiss-Penzias et al., 2006;Jaffe et al., 2005;Weiss-Penzias et al., 2006, 2007).Vaisala PTB101B transmitter was used to measure pressure and Campbell Scientific HMP45C probe was used to monitor the ambient temperature (T ) and relative humidity (RH).Introduction

Conclusions References
Tables Figures

Back Close
Full

Back trajectory modeling
We calculated 10-day airmass backward trajectories for each high RGM event to establish the transport history of the associated air masses.Trajectories were calculated using the NOAA Hybrid Single-Particle Lagrangian Integrated Trajectory model (HYS-PLIT, v4.9, Draxler and Rolph, 2012) and global meteorological data from the Global Data Assimilation System (GDAS) archive, having a horizontal resolution of 1 • × 1 • , 3 h time resolution, and a vertical resolution of 23 pressure surfaces between 1000 and 20 hPa.During events trajectories were initialized from the summit of MBO for the hours when the highest RGM/GEM-ratios were observed.

Statistical analysis
Due to error in both x and y variables and the relatively small number of RGM observations during each event, correlations between measured compounds were calculated using the Williamson-York Iterative Bivariate Fit method (Cantrell et al., 2008).

Results
RGM measurements in the FT are scarce and thus the identification of mercury sources and characterization of the global mercury cycle has been a challenge.Transport of RGM from the stratosphere (type 1. UTLS event) is the only known source of RGM above the BL (Murphy et al., 2006;Swartzendruber et al., 2006;Lyman and Jaffe, 2011).During UT/LS type 1 events an increase in RGM and ozone concentrations and a clear correlation between RGM and ozone is typically observed (Swartzendruber et al., 2006).Also, during these events, a clear decrease in GEM, CO and aerosol concentrations is observed, coinciding with the RGM and O 3 enhancements (Swartzendruber et al., 2006).Using mercury (GEM, RGM, PBM), aerosol (scattering and elemental composition) and trace gas (CO, O 3 ) measurements conducted at MBO since 2005 we present here two previously unidentified sources of RGM to the FT.Introduction

Conclusions References
Tables Figures

Back Close
Full Figure 1 shows a conceptual model of the three types of RGM sources to the lower FT (1, UT/LS; 2, Asian long-range transport (ALRT); 3, marine boundary layer (MBL)) observed at MBO. Type 1 (UT/LS) events were identified previously by Swartzendruber et al. (2006) and will not be discussed here.Type 2 ALRT and type 3 MBL events are described in Sects.3.1 and 3.2.

Anthropogenic RGM events (type 2)
First, we present evidence for in-situ oxidation of GEM to RGM (yielding RGM/GEMratios up to 0.18) in anthropogenic pollution plumes that were transported from Asia to the free troposphere over the US Pacific Northwest (source type 2).This is important because oxidation of GEM in long-range transported pollution plumes in the FT has not been observed previously (Jaffe et al., 2005;Swartzendruber et al., 2006;Faïn et al., 2009).Previous studies conducted in urban areas have shown that point source emissions affect the GEM and RGM concentrations in downwind areas (Lynam and Keeler, 2005;Rothenberg et al., 2009).However, these studies were not able to identify in-situ production of RGM.During springtime, meteorological conditions are favorable for transpacific transport and Asian pollution plumes are repeatedly observed at MBO (Jaffe et al., 2005;Weiss-Penzias et al., 2006;Ambrose et al., 2011).Increasing anthropogenic Hg emissions in developing Asian countries (Pacyna et al., 2010;Fu et al., 2012), coupled with formation of RGM during long-range transport will enhance Hg deposition downwind in North America.
We present three clear ALRT events (Fig. 2, Table 1) with total Hg (THg = GEM + RGM+PBM) and RGM concentrations substantially elevated above typical background FT levels.Atmospheric concentrations of Hg species and trace gases, aerosol scattering values, and correlations between main components are shown for each event in Table 1.During ALRT events airmass back trajectories show direct transport from Asia in ≤ 10 days (Fig. 3).In addition, the THg/CO enhancement ratios (on average 0.0069 ng m −3 ppbv −1 ; Fig. 4) were consistent with the ratios measured previously in Asian plumes (Jaffe et al., 2005;Weiss-Penzias et al., 2007).Furthermore, aerosol Introduction

Conclusions References
Tables Figures
During ALRT events RGM was correlated with the anthropogenic pollution tracers CO, σ sp and/or O 3 (Table 1), indicating that they have a common origin.Furthermore, RGM and GEM were anti-correlated in the ALRT plumes.The RGM enhancements corresponded with GEM depletions of 0.2-0.4ng m −3 (Fig. 2).This indicates that we are seeing the effects of conversion of GEM to RGM in-situ, rather than primary Asian RGM.Depletion of GEM in anthropogenic pollution plumes has only been seen previously in the BL (Weiss-Penzias et al., 2003;Lynam and Keeler, 2005).Several studies have measured RGM and GEM concentrations in China and in Asian pollution plumes in the Western Pacific BL (Jaffe et al., 2005;Chand et al., 2008;Fu et al., 2012).The RGM/GEM-ratios observed were typically below 0.03.In contrast, for the three ALRT events presented in this paper the peak RGM/GEM-ratio was between 0.15 and 0.18, showing that the measured RGM is a significant fraction of THg (Table 1, Fig. 2).Previous studies have shown that for primary RGM emissions a positive correlation between RGM and GEM is typically observed, whereas for in-situ oxidation an anti-correlation between RGM and GEM is expected (Swartzendruber et al., 2006;Sillman et al., 2007).
The anti-correlation between RGM and GEM (Table 1, Fig. 5) and elevated RGM/GEMratios in ALRT plumes observed at MBO, indicate that the source of RGM during these events was in-situ oxidation during FT transport.Previous studies have shown that the lifetime of RGM is longer in the FT than in the BL due to the absence of removal mechanisms (Munthe et al., 2003;Selin et al., 2007).Therefore, conditions in the FT are expected to be more favorable for accumulation and transport of secondary (e.g.photochemical) RGM.Introduction

Conclusions References
Tables Figures

Back Close
Full

Marine boundary layer events (type 3)
The second new RGM source we have identified (type 3; marine boundary layer) occurs in very clean air masses.The air mass back trajectories show that these air masses had circled above the Pacific Ocean for at least 10 days prior to arriving at MBO (Figs. 1 and 6).Here we present four clear MBL events (Fig. 7, Table 1).During MBL events a simultaneous increase in RGM and decrease in CO, aerosol scattering, and O 3 was observed.The low CO and aerosol scattering values (Fig. 7, Table 1) measured during this time period suggest minimal influence from anthropogenic emissions or from biomass burning (Munthe et al., 2003;Weiss-Penzias et al., 2007).The O 3 and water vapor mixing ratios (∼ 30-45 ppbv and 3-4 g kg −1 , respectively) during the event do not indicate transport from the UT/LS; rather, the composition of these air masses is more consistent with an influence from the clean sub-tropical MBL.The RGM concentrations observed during MBL events were very high (200-700 pg m −3 ) compared to anthropogenic pollution plumes (200-300 pg m −3 ) (Table 1).Such high RGM levels have been measured in the BL near emission sources and locations with high concentrations of oxidants (e.g.those encountered at the Dead Sea; Obrist et al., 2010).However, such high levels have not been previously observed in the FT.Much lower THg concentrations were measured during MBL events (1.2-1.4 ng m −3 ) when compared to anthropogenic pollution plumes (1.7-1.8 ng m −3 ).As for Type 2 events, PBM concentrations were low (< 12 pg m −3 ), amounting to < 15 % of reactive mercury.A clear anti-correlation between RGM and GEM was observed during MBL events (Table 1, Fig. 7) indicating in-situ GEM oxidation.Furthermore, the large RGM/GEM-ratios (0.31-1.05,Table 1) demonstrate an efficient process for GEM oxidation and transport of RGM in the FT.It is notable that the large RGM enhancements in these air masses far exceed those previously observed in UT/LS influenced air masses (Swartzendruber et al., 2006;Ambrose et al., 2011).
Aerosol chemical composition was measured with rotating DRUM impactors during the MBL event observed in May 2011.A small increase in aerosol sea salt (Na + and Introduction

Conclusions References
Tables Figures

Back Close
Full Cl − ) concentration was measured, further indicating that the air mass likely originated from the marine boundary layer.Sea salt aerosol may be an important sink for RGM in the MBL (Malcolm et al., 2009), thus explaining smaller ambient RGM concentrations observed in the studies conducted in marine regions (Laurier et al., 2003;Laurier and Mason, 2007).

Particle bound mercury (PBM)
Reactive mercury (RM) is defined as the sum of reactive gaseous mercury (RGM) and particle bound mercury (PBM).During the measurement period (2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012) the PBM concentrations were usually low (below 30 pg m −3 ) in the FT.Elevated PBM concentrations (30-45 pg m −3 ) were mainly observed during biomass burning episodes (Finley et al., 2009), when the emissions clearly originated from the boundary layer.During RGM events described in this paper PBM concentrations were low (< 15 pg m −3 ), amounting to less than 15 % of RM.The contribution of PBM is clearly lower than expected based on RM partitioning coefficients derived from measurements in the BL at low altitude sites (Rutter and Schauer, 2007b;Amos et al., 2012).However, previous studies have demonstrated that partitioning of RM (and other semivolatile species) between the gas and aerosol phase depends on conditions such as temperature (Rutter and Schauer, 2007b;Amos et al., 2012); aerosol concentration and composition (Rutter and Schauer, 2007a;Amos et al., 2012); and possibly also relative humidity (Pankow et al., 1994;Xiu et al., 2009;Kim et al., 2012).Furthermore, previous studies suggest that heterogeneous reactions at surfaces potentially play a key role in Hg chemistry (Subir et al., 2012).The conditions in the FT (e.g.dry air; low temperature and pressure; high solar radiation intensity; Jaffe et al., 2005;Weiss-Penzias et al., 2006), are substantially different than in the boundary layer; therefore, RM partitioning behavior likely also differs between the FT and BL.If partitioning of RM to PBM is enhanced in the presence of an aerosol aqueous phase, the lower relative humidity in the FT will likely favor partitioning of RM to RGM more so than in the BL.Also, due to below freezing temperatures in the free troposphere, it is likely that aerosol particles will 29212 Introduction

Conclusions References
Tables Figures

Back Close
Full incorporate ice rather than liquid water.The interfacial chemistry of Hg on ice surfaces is poorly known (Subir et al., 2012), although it appears that the uptake of RGM to ice is less efficient than for liquid water (Sigler et al., 2009).We note that previous studies have also measured low PBM concentrations in the free troposphere (Murphy et al., 2006).

Oxidation mechanisms
Gaseous halogen species, in particular bromine atom (Br q ), chlorine atom (Cl q ), and bromine oxide (BrO q ), may be important oxidants for GEM in the atmosphere (Swarzendruber et al., 2006;Holmes et al., 2009;2010;Subir et al., 2011;Stephens et al., 2012;Tas et al., 2012).In addition, O 3 and hydroxyl radical (HO q ) might also contribute to GEM oxidation, although the associated mechanisms are more uncertain than for halogen atoms (Holmes et al., 2010;Rutter et al., 2012;Subir et al., 2012).The three main sources of RGM observed at MBO represent contrasting air mass types: dry upper tropospheric air with high O 3 and RGM concentrations (type 1); aged Asian anthropogenic emissions with elevated Hg concentrations (type 2); and clean air with background Hg and a potential contribution from natural oceanic emissions (type 3).Thus, it is likely that the oxidation mechanisms converting GEM to RGM are different in each case.Production of RGM in type 1 events (Swarzendruber et al., 2006) is likely associated with Br q chemistry in the UT/LS region (Holmes et al., 2010).The RGM concentrations during ALRT events (type 2) were clearly correlated with anthropogenic pollution tracers (CO or σ sp ) possibly indicating that the oxidant has an anthropogenic origin.The RGM associated with type 2 events could also be produced by halogens, but the mechanism is not clear.Previous studies have indicated that Cl q could be generated from the interaction of pollution plumes with chloride (Cl − ) containing aerosols (Lawler et al., 2009;Thornton et al., 2009).Oxidation via heterogeneous chemistry involving aerosol particles is also a possible mechanism (Subir et al., 2012), and would be consistent with the observed correlation between the RGM and aerosol concentrations.

Conclusions References
Tables Figures

Back Close
Full For type 3 (MBL) events a clear anti-correlation between RGM and O 3 was observed, indicating that O 3 is not likely an oxidant.Also, as these events are only seen in clean air masses, it seems unlikely that the oxidant would have an anthropogenic origin.More likely, reactive halogen chemistry plays an important role in these events.For the type 3 events, we believe that Br radicals are the likely oxidants.This is supported by box model calculations (see Sect. 3.4.1)using estimated Br q mixing ratios for the MBL.In the marine boundary layer, RGM enhancements have been reported (Laurier and Mason, 2007;Chand et al., 2008;Holmes et al., 2009), but these never reached the levels we observed in the FT.A possible explanation for the very high RGM concentrations observed during MBL events at MBO is that RGM could accumulate on aqueous supermicron sea-salt aerosol in the MBL (Rutter and Schauer, 2007a;Malcolm et al., 2009), and then subsequently partition back to the gas phase when lofted to the lower FT.For instance, lower relative humidity in the FT, and subsequent evaporation of the aerosol aqueous phase, could also favor partitioning of water-soluble sea salt-bound PBM to RGM.Furthermore, low relative humidity in the FT would favor partitioning of halogens to the gas phase, due to evaporation of the aerosol aqueous phase and associated decrease in pH (Lawler et al., 2009(Lawler et al., , 2011)).

Box model calculations for type 3 (MBL) events
Based on previous studies, gaseous halogen species, in particular Br q , are assumed to be important GEM oxidants in the MBL (Hedgecock and Pirrone, 2004;Holmes et al., 2009Holmes et al., , 2010;;Subir et al., 2012).Therefore, we used box model calculations to estimate the potential contribution of Br q chemistry to RGM production during type 3 (MBL) events.(2) Conditions used in the simulations are summarized in Table 2 and correspond with the median pressure (896 hPa) and temperature (282 K) along 10-day HYSPLIT backward trajectories that were initiated from MBO (1375 m a.g.l. on the model grid) during the event.The trajectories were initiated at the mid-points of the 3-h RM samples corresponding with the highest measured RGM concentrations.(Note that trajectories for this event leave the boundary layer 3.75-5.25days prior to arrival at MBO.)For a 10 day reaction time, with mean [Br q ] and [HO q ] = 0.035 pptv and 0.5 to 1.5 × 10 6 molecules cm −3 , respectively (here brackets denote mixing ratio or concentration), the modeled [RGM] (142-175 pg m −3 ) accounts for 52-64 % of the peak level (275 pg m −3 ) observed at MBO during the 11 May event (Fig. 8, Table 2).Our results indicate that, for reasonable Br q and HO q concentrations, a large fraction of the peak RGM measured can be explained by Br q chemistry alone.We note that the initial value of

Conclusions
Mercury is an important global pollutant, with a complex atmospheric cycle.Current models provide a rough outline of the global budget, but the atmospheric chemistry is poorly understood.Due to observed differences in behavior (e.g.sources, gas-particle partitioning, and concentration levels) of Hg species in the FT in comparison to the BL, studies in the FT are needed to fully understand the chemical cycle of Hg.Using long-term Hg observations at the Mt.Bachelor Observatory we have identified two new sources of RGM in the atmosphere, which suggest stronger coupling between BL emissions and FT oxidation than previously recognized.For both source types, Hg and its oxidants (or oxidant precursors) are likely first lofted from the BL (Asia, or the Pacific MBL) to the FT.Subsequently, GEM is effectively converted to RGM in the FT on relatively short timescales (several days) during long-range transport to downwind regions.The RGM generated will eventually be entrained back to the BL (Lyman and Gustin, 2009;Weiss-Penzias et al., 2009), deposited and incorporated to biota.We believe also that changes in RM gas-particle partitioning during transport could also contribute to the high observed RGM concentrations during MBL events.Results of this study offer new directions for process oriented studies of the atmospheric mercury cycle.Introduction

Conclusions References
Tables Figures

Back Close
Full     −3 and a factor of ∼ 20-70 above typical background FT levels) were observed during MBL events.The GEM concentrations were higher during ALRT events (THg = 1.65-1.80ng m −3 ; 1 nanogram per cubic meter ≈ 0.12 parts per trillion by volume at 1.01 bar and 273 K).In general, the MBL events (type 3) show stronger inverse correlations between RGM and GEM compared to ALRT events (type 2).Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Figure 8 shows simulated Br-initiated oxidation of Hg 0 by the reaction mechanism including reactions Eqs.(1)-(4) (with rate data fromGoodsite et al., 2004;  Donohue et al., 2006;Holmes et al., 2010) for conditions representative of the 11

[
THg] in the simulations (1.19 ng m −3 -the concentration corresponding with the RGM peak) is quite low for the MBL and lower FT in the Northern Hemisphere(Lindberg et al., 2007;Ambrose et al., 2011).Therefore, the low measured value of[THg]  at the center of the event is consistent with loss of a substantial quantity of RM upwind of MBO.An alternative model scenario aimed at reproducing [GEM], [RGM], and [THg] at the center of the event could include higher initial [THg] (with a value closer to the mean background of ∼ 1.5-1.7 ng m −3 ; Lindberg et al., 2007; Ambrose et al., 2011) and a pathway for RGM removal.Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | NRC: Global Sources of Local Pollution: an Assessment of Long-Range Transport of Key Air Pollutants to and from the United States, The National Academies Press, Washington, DC, USA, 2010.Obrist, D., Tas, E., Peleg, M., Matveev, V., Faïn, X., Asaf, D., and Luria, M.: Bromine-induced oxidation of mercury in the mid-latitude atmosphere, Nat.Geosci., 4, 22-26, 2010Discussion Paper | Discussion Paper | Discussion Paper | Weiss-Penzias, P., Jaffe, D., Swartzendruber, P., Hafner, W., Chand, D., and Prestbo, E.: Quantifying Asian and biomass burning sources of mercury using the Hg/CO ratio in pollution plumes observed at the Mount Bachelor observatory, Atmos.Environ., 41, 4366-4379, 2007.Weiss-Penzias, P., Gustin, M. S., and Lyman, S. N.: Observations of speciated atmospheric Discussion Paper | Discussion Paper | Discussion Paper |

Fig. 1 .Fig. 2 .
Fig. 1.Sources of RGM in the free troposphere as measured at MBO: 1, UT/LS (Swartzendruber et al., 2006; Ambrose et al., 2011) ; 2, ALRT (this paper); and 3, clean marine boundary layer (MBL) air masses processed in the FT above the Pacific Ocean (this paper).The HYS-PLIT airmass back trajectories of three events (UT/LS: 22 April 2006; ALRT: 17 April 2008; MBL: 9 May 2007) are used to show a typical pathway of airmasses during each event type.Trajectories are colored by height.The typical chemical composition of airmasses (ozone (O 3 ), CO, and particulate matter (PM) concentrations) are shown for the areas where the trajectories originate.Key features of the Hg observations at MBO are also shown.

Fig. 5 .
Fig. 5.A negative correlation was observed between RGM and GEM during high-RGM events.Elevated total mercury (THg) and RGM concentrations (peak values of 200-700 pg m −3 , elevated by up to ∼ 0.2-0.7 ng m−3 and a factor of ∼ 20-70 above typical background FT levels) were observed during MBL events.The GEM concentrations were higher during ALRT events (THg = 1.65-1.80ng m −3 ; 1 nanogram per cubic meter ≈ 0.12 parts per trillion by volume at 1.01 bar and 273 K).In general, the MBL events (type 3) show stronger inverse correlations between RGM and GEM compared to ALRT events (type 2).

Table 1 .
Chemical characteristics of ALRT (type 2) and MBL (type 3) high-RGM events observed at MBO.Some CO data are missing for this event, which would tend to reduce the statistical significance of correlations a The event time period includes the maximum RGM enhancement and up to 14 h preceding and following a discernible increase in RGM associated with the peak enhancement.Correlation parameters were calculated for this time period (n = 5-15), while the mean and maximum values for each listed chemical parameter correspond with only the times for which RGM was > 10 % of Max RGM (n = 2-6).b Pearson correlation coefficient; * : p < 0.05; * * : p < 0.01; no star: p > 0.05.c