S1 (II) Measurement and understanding of atmospheric mercury processes

A unique set of gaseous mercury measurements in the upper troposphere and lower stratosphere (UT/LS) has been obtained during the monthly CARIBIC (www.caribic-atmospheric.com) flights since 2005. The passenger Airbus 340-600 of Lufthansa covered routes to the Far East, North America, India, and the southern hemisphere. The accompanying measurements of CO, O₃, NOy, H₂O, aerosols, halocarbons, hydrocarbons, greenhouse gases, SF₆ and several other parameters as well as backward trajectories enable a detailed analysis of these measurements. The CARIBIC measurements represent a good approximation of total gaseous mercury (TGM) concentrations as speciation tests have shown. Above the tropopause gaseous mercury concentration always decreases with increasing PV and O₃ which implies its conversion to particle bound mercury. The observation of the lowest TGM concentrations at the highest particle concentrations in the stratosphere provides further evidence for such conversion. We will show how a seasonally dependent conversion rate could be derived using concomitantly measured SF₆ mixing ratios as a timer. Tropospheric mercury data suggest the existence of a decreasing trend on > 90% significance level. The size of this trend is comparable with the trend derived from long-term measurements at Cape Point (South Africa) and Mace Head (Ireland) since 1996.

Exchange of gaseous elemental mercury (GEM) between vegetation and the atmosphere plays a major role in Hg accumulation and cycling within the environment, yet this fundamental component of the terrestrial Hg cycle remains poorly characterized. A widely-accepted conceptual model is that GEM exchange occurs primarily through plant stomata, but recent evidence from whole-plant chamber studies has indicated that non-stomatal uptake may actually be more important. We examine evidence for GEM depositional and emission pathways from previously published studies and present our own data from multi-seasonal measurements made over an uncontaminated high-elevation wetland meadow in Shenandoah National Park, Virginia. In this setting, nighttime deposition (when stomata are closed) and the poor correlation between Hg fluxes and canopy conductance during periods of active vegetation growth suggest that stomatal processes are not the dominant mechanism for ecosystem-level GEM exchange. Strong springtime deposition relative to summer implies that young vegetation is better at scavenging Hg, with the highest deposition occurring at night and possibly via a cuticular pathway. We also present GEM flux data from an ongoing experiment conducted over a forested site in the Piedmont region of Virginia, and interpret these results in the context identifying the dominant vegetation-atmosphere exchange pathways.

The Southeastern Aerosol Research and Characterization (SEARCH) network operates Tekran Hg speciation analyzers at 3 sites in the southeastern U.S.: Yorkville, GA; Birmingham, AL and near Pensacola, FL. Each site routinely measures elemental Hg (Hg⁰), fine particulate Hg (Hg_PM2.5) and reactive gaseous mercury (RGM) every other hour. In addition to the routine analyzer, the FL site has a second Hg analyzer that has been used, over time, to: 1) assess analyzer precision; 2) obtain continuous hourly measurements; 3) estimate coarse particulate Hg concentrations; and 4) evaluate interferences with speciation measurements due to ozone and other trace gases. The experimental design for items 1 and 2 simply involves synchronous or asynchronous operation of collocated analyzers for set periods of time (typically 3-5 weeks per calendar quarter). For item 3, analyzers were first run in synchronous mode with identical PM2.5 particle inlets to fine tune response. The collocated analyzer was then modified to accommodate a PM₁₀ inlet and both analyzers were operated in synchronous mode. Coarse particulate Hg was then inferred based on the difference between analyzers, as follows: Hg_PMcoarse = (RGM_c + Hg_PM10) – (RGM_p +Hg_PM2.5), where subscripts c and p refer to collocated and primary, respectively. Data were collected for 3-4 weeks during each of four successive calendar quarters. For interference testing (item 4), the two Hg analyzers were operated synchronously and the inlet to the collocated analyzer was modified with a side tap to permit addition of a relatively small (50 sccm) drip of potentially interfering gases into the sample stream. The gases used were O₃, SO₂, NO and NO₂, singly and in combination. Concentrations of each gas were such as to raise the total concentration in the sample stream by 25 parts per billion (ppb) to roughly 200 ppb. This method of addition approach allowed us to control interferent concentration without significantly changing the characteristics of the sample air. Ambient concentrations of O₃, SO₂, NO, NO₂ and NO_y were also measured continuously to provide estimates of total sample concentration (i.e., ambient plus drip). Interferent gas was added continuously to the collocated inlet and differences between analyzers (above and beyond analytical uncertainty) were used to infer interference effects. This presentation will present results from experimental configurations 1-4, assess implications of coarse particulate Hg and interferents and provide recommendations for further investigations.

Hg(II) is a semivolatile species that can partition between the gas and particle phases. Because gases and particles are removed from the atmosphere with different efficiencies the phase of mercury has important implications for both the residence time of Hg in the atmosphere and its pattern of deposition. We constrain Hg(II) gas-particle partitioning with long-term observational records of speciated mercury and concurrent data for temperature, relative humidity, and ambient particulate matter (PM). This analysis uses observations from sites in urban, rural, and remote environments across North America. We implement a new mechanistic representation of Hg(II) gas-particle partitioning in a state-of-the-science global 3-D chemical transport model (GEOS-Chem) for multi-year simulations. Results indicate that Hg(II) gas-particle partitioning can be successfully described by a thermochemical parameterization dependent on PM and temperature, but the parameters are significantly different from previously published Hg(II) gas-particle partitioning coefficients derived from observations.

The modeled deposition of mercury is sensitive to the phase of Hg, but also to the parameterizations for wet and dry removal processes. Wet deposition of Hg(II) and HgP in GEOS-Chem is modeled by scavenging from moist convective updrafts as well as rainout and washout by large-scale precipitation [Liu et al., 2001]. We have found that GEOS-Chem has excessive re-evaporation taking place below convective clouds, reflecting lack of information on the vertical distribution of precipitation. To address this problem we have accessed a new NASA global meteorological product, the Modern Era Reanalysis for Research and Applications (MERRA), which improves on the current product notably through better archival of hydrometeorological information. We revised the wet deposition code in GEOS-Chem to exploit the detailed information made available by MERRA. Preliminary results comparing the model to observations from the US Mercury Deposition Network (MDN) indicate an improvement in the simulation of Hg deposition over the Gulf Coast.

Deposition is the most environmentally relevant endpoint for atmospheric mercury because inputs to terrestrial and aquatic ecosystems can become available for methylation and bioaccumulate in organisms. We quantify the effects of gas-particle partitioning of atmospheric mercury on global deposition and assess the impact on oceanic and terrestrial ecosystems.

The development of techniques for measurement of both gaseous elemental (GEM; Hg⁰) and reactive gaseous mercury (RGM; HG²⁺) remains a significant analytical challenge. For RGM, KCl-coated annular denuder sampling provides limited temporal resolution, no information on speciation and may be subject to interferences. Although the CVAFS approach is well established for measurement of Hg⁰, no measurement technique has shown the combination of sensitivity, temporal resolution and precision necessary for direct measurement of Hg⁰ fluxes. In this work we describe experimental approaches using laser induced fluorescence (LIF) that are designed to address some of these issues. We have been investigating the use of denuder sampling coupled with analysis using programmable thermal dissociation (PTD) as route to obtaining information on the chemical speciation of RGM. The technique was tested at a coal fired power plant sampling from stack exhaust gases and from the stack plume, downwind of the stack using an airship. Profiles were consistent with HgCl₂ being the dominant component of the stack gas RGM. Plume sampling suggests that RGM is quantitatively released from the stack and that any reduction of RGM back to Hg⁰ occurs in-plume. Sampling under more controlled conditions at the same plant examined the effects of addition of bromine and ammonia on the PTD profiles of stack RGM. The addition of Br and ammonia appear to result in higher decomposition temperatures in the PTD profiles. We are attempting to reproduce these results under controlled laboratory conditions by doping hydrocarbon flames with mercury and mercury / halogen mixtures and performing in-situ monitoring of Hg⁰ and Hg²⁺ radicals in the flame using LIF. We are also performing denuder sampling from the post flame gases. We are using the thermal dissociation technique to monitor RGM in the unpolluted atmosphere. Both KCl coated annular denuders and uncoated etched tubular denuders were deployed on a Piper Navajo aircraft together with a Tekran 2537B mercury analyzer in flights over the S.E. United States. We will present altitude profiles of Hg⁰, RGM and thermal dissociation profiles of RGM. We see very distinct differences in the thermal dissociation profiles of RGM collected in the marine boundary layer and aloft. We have reported laboratory studies that suggest that sequential two photon LIF has the sensitivity to monitor Hg⁰ at background levels in air. We have now deployed the instrument in a mobile laboratory and we will report the first in-situ measurements of Hg⁰ in ambient air using sequential two photon LIF together with our progress towards high frequency flux measurements using the eddy correlation technique. This work is supported by EPRI, NOAA and NSF.

The Western Airborne Contaminants Assessment Project (WACAP) showed that fish in eight National Parks of the western U.S. had mercury(Hg) concentrations that exceeded the threshold for fish eating wildlife (www.nature.nps.gov/air/Studies/air_toxics/wacap.cfm). These observations led to the development of this study focused on investigating air Hg concentrations and potential dry deposition using newly developed passive samplers and surrogate surfaces. The primary question being addressed is whether local, regional or global sources are responsible for the Hg measured in fish in western parks.

To investigate this question passive and surrogate surface samplers that measure reactive gaseous mercury (RGM) concentrations and deposition, respectively, are being deployed along a transect from the coast of California to the eastern edge of Nevada. Passive samplers for ozone are also being deployed simultaneously. In addition, the sampling locations have ancillary meteorological and air quality data available that will be used along with back trajectory analyses to better understand the source of air interacting with each site. Air Hg concentrations will be measured at select locations using a Tekran 2537a/1120/1135 speciation system for 4-6 weeks to validate sampler data. Sampling locations are, from west to east, Point Reyes National Seashore, CA; Elkhorn Slough, CA; Lick Observatory on Mt. Hamilton, CA ; Yosemite National Park, CA; Sequoia & Kings Canyon National Park, CA; and Great Basin National Park, NV. Investigation of elevation gradients in RGM concentration and deposition within select parks during sampling intensives will allow us to better understand the sources of Hg to park ecosystems. Data collection started in August of 2010 and will extend to November of 2011. Thus far, the lowest RGM deposition and relative concentrations have been observed at the low elevation coastal sites, Elkhorn Slough and Point Reyes NP. The highest values have been recorded at Lick Observatory, a high elevation coastal site. Elemental Hg and RGM concentrations collected using a Tekran 2537A/1130 system during August and September 2010 at Great Basin NP had mean concentrations of 1.48 + 0.62 ng/m³ and 68 + 51 pg/m³ respectively.

Dynamic flux chambers (DFCs) have been widely employed for Hg^o flux measurement over soils. However, DFCs of different sizes, shapes and sampling flow rates can yield distinct measured fluxes under the same environmental conditions for a given soil substrate. The main reason is that different DFC designs and sampling flows can create largely varied flow conditions, which in turn changes the mass transfer of Hg^o from soil surface to air. In addition, the flow conditions inside DFCs may be complicated and are currently poorly understood. Such a lack of understanding makes it difficult to apply the measured fluxes in atmospheric models for estimating the magnitude of air-surface exchange because the relationship between measured fluxes and flow conditions has not been established quantitatively.

In this study, we investigated the flow structure inside DFCs using computational fluid dynamics (CFD) techniques. It was found that the flows inside several DFCs in use are not uniform, causing non-uniform mass fluxes over the measured surface. Furthermore, the surface shear stress is not linearly proportional to the sampling flow rate, and the distribution of shear stress can change significantly at different flow rates. Based on the results of CFD calculations, a DFC design that can produce a steady and uniform air flow over the measured surface was proposed. The flow conditions of the proposed DFC design were evaluated mathematically under different sampling flow rates. The relationship between the sampling flow rate and friction velocity, a flow parameter widely employed in atmospheric models, was also developed for various surface roughness. CFD simulations using the proposed DFC design showed that increasing sampling flow rate increases the measured flux until it asymptotically approaches a flux value that represents mass transfer limitation.

A major source of uncertainty in atmospheric modelling of mercury cycling is the lack of knowledge of mercury chemistry at various surfaces. Oxidized mercury compounds exhibit low volatility and water solubility and thus can “stick” to numerous surfaces, e.g. air/water, gas/aerosol, air/snow interfaces, present in the atmosphere and the ecosystem. Moreover, adsorption of mercury compounds on to the glass surface of reaction chambers, commonly used in determining the rate constant of gas-phase mercury reactions, leads to uncertainty in the kinetic parameter. However, the adsorption mechanism of mercury and related compounds on to these surfaces is not well understood and characterized. Herein we present our investigation of interfacial mercury chemistry using a series of state-of-the-art surface facilities. Adsorption properties of gaseous and dissolved oxidized mercury compounds on both natural and laboratory surfaces will be presented. We will further discuss the impact of our results on mercury chemical schemes parameterization in the atmospheric global circulation models. A laser based surface specific spectroscopic technique which holds the potential to investigate interfacial and heterogeneous mercury reactions will be elaborated.