AIR-SURFACE EXCHANGE OF MERCURY OVER NATURAL AND IMPACTED SURFACES IN ATLANTIC CANADA.

S. Beauchamp, R. Tordon, L. Phinney, A. Pinette (Atmospheric Science Division, Meteorological Service of Canada, Dartmouth, Nova Scotia, Canada, B2Y 2N6; Steve.Beauchamp@ec.gc.ca), A. Rencz (Geological Survey of Canada, Ottawa, Ontario, Canada), J. Dalziel (Dept., Fisheries and Oceans, Dartmouth, Nova Scotia, Canada), H.T.K. Wong (National Water Research Institute, Burlington, Ontario, Canada)

Abstract A potentially important component of the biogeochemical mercury cycle involves the exchange of mercury between the atmosphere and aquatic and terrestrial surfaces. The objective of this study was to quantify air-surface mercury exchange rates, to put boundaries on these rates and provide insight into physical-chemical processes regulating flux. Average daily (24 hr) mercury flux over natural forest soils was low ranging from -0.4 (net deposition) to 2.2 ng/m²/hr (evasion). Low flux rates were partially due to poor light penetration through the forest canopy and lower soil temperatures. Minimum and maximum soil flux rates ranged from -1.3 to 5.7 ng/m²/hr. Soil mercury flux at an open field site was low over undisturbed soil compared to the same soil mixed to a depth of 1-2 meters (0.9 versus 8.0 ng/m²/hr, respectively). Average daily flux over a clear freshwater lake (low dissolved organic carbon) was also relatively low (~0.6 ng/m²/hr) compared to lakes with higher dissolved organic carbon (DOC) content (2.2 to 5.0 ng/m²/hr). Total mercury concentrations in marine waters at a moderately polluted coastal harbor site ([Hg_(t)] = 0.65 to 1.4 ng/l) and mean daily mercury flux rates over salt water were similar to those measured over low DOC lake. The highest daily average mercury flux rates were measured at 2 abandoned gold mine tailings sites (130 and 237 ng/m²/hr, respectively).

Introduction Mercury can be emitted, deposited and re-emitted back into the atmosphere from soil, water and vegetation (Schroeder et al., 1989) thus mercury exchange between the atmosphere and the surface is bi-directional (emission and deposition). Deposition measurements, especially when taken in isolation, do not adequately reflect net mercury loading from the atmosphere. Mercury emissions from diffuse (area) sources are only recently being quantified (Rasmussen et al., 1998). In addition to the need to quantify and set boundaries for air-surface mercury exchange rates in various landscape settings, it is necessary to understand mechanisms controlling exchange processes. Quantification of mercury exchange between the atmosphere and receiving systems are integral to mass balance studies as well as in understanding sources and fate of mercury, understanding biogeochemical processes and subsequently, the role of the atmosphere in relation to mercury in biota. The objectives of this study were to quantify mercury air-surface exchange in representative natural freshwater and terrestrial landscapes as well as at selected sites which have been anthropogenically impacted (disturbed) and to investigate physical and chemical factors regulating air-surface exchange at these sites in support of mercury emission inventory programs, atmospheric modeling, watershed mass balance studies and ecosystem research.

Methods and Materials Air-surface mercury exchange measurements were conducted using a Teflon™ dynamic flux chamber (Carpi and Lindberg, 1998) enclosing a surface area of 0.12 m² (20 cm x 60 cm). Total gaseous mercury (TGM) concentration was measured using a dual channel Tekran™ mercury vapour analyser which sampled for 5 minutes at a flow rate of 1.5 L/min alternating every 10 minutes between ambient and chamber air. Switching between ambient and chamber air was done using a Tekran™ Automated Dual Sampling System (TADS). The flux chamber had a nominal flushing rate of ~ 10 minutes. Mercury flux was calculated as the difference between the mean mercury concentration in ambient air versus air (blank corrected) which had passed through the chamber (Xiao et al., 1991).

Physical and chemical data measurements including meteorology (e.g. wind, temperature, solar insolation), soil/water temperature, soil moisture, soil/water total mercury content were conducted concurrently with air-surface mercury exchange measurements. Meteorological measurements were made using standard meteorological methods and instrumentation. System quality control procedures including analyser calibrations, chamber blanks and temperature, wind and solar radiation sensor calibrations were performed on a regular basis. Blank flux was determined using the complete system of lines, fittings, switches and the chamber. Flux rates presented in this study are blank corrected.

Soil and water samples used for total mercury analysis were collected in Teflon bottles using clean sampling techniques. Total mercury concentration was determined following oxidation by bromine monochloride and UV irradiation. Each sample was pre-reduced with hydroxylamine, transferred to a bubbling apparatus containing stannous chloride and the free mercury was volatilized was then purged through a gold coated sand trap and thermally desorbed into the CVAFS analyser.

Most flux measurements were conducted under similar meteorological conditions, usually while under the influence of a high pressure system providing sunny days, moderate temperatures and light wind speeds (Table 1). Ground surface (soil) measurement sites were undisturbed and regionally representative unless otherwise specified. Lake site selection was based on dissolved organic carbon (DOC) content and covered a range from clear water to highly colored water. The CAPMoN site is a forest clearing on the top of a glacial drumlin representative of native glacial till. Mine tailings are remnants of gold mining activities which used the mercury amalgamation process from the mid 1800’s up until the late 1930’s. The Halifax Harbour site represents a marine ecosystem which is moderately impacted by urban sewage and industrial activity.

Results and Discussion Analyser and sensor calibrations were stable and consistently within acceptable operating criterion. Duplicate total gaseous mercury (TGM) measurements from the chamber and ambient air were within 5% and chamber blank fluxes were low (< 0.3 ng/m²/hr) except at the mine tailing sites where blanks reached 7.5 ng/m²/hr due to the presence of fine contaminated dust. Chamber blank fluxes should be small relative to the magnitude of measured fluxes (Wallschlager et al., 1999) but chamber blanks in this study were quantitatively important given low measured flux rates. Blanks at mine tailing sites were high compared to those measured at more pristine sites but were small compared to flux rates measured over the tailings.

Mercury flux was positive (evasive) over forest soils at Puzzle and North Cranberry Lakes throughout the diurnal cycle measurement period (Table 1). Small negative fluxes (deposition) were measured over soils at Big Dam West and at the CAPMoN site. In all cases, the magnitude of air-surface mercury exchange was small (<10 ng/m²/hr). Mercury flux at all sites showed a strong diurnal pattern following the solar cycle. Mercury flux rates conducted over consecutive 24 hr periods were similar at each site indicating that the rates did not vary widely under similar meteorological conditions over short time scales. At forest soil sites where measurements were made in both 1997 and 1999, flux rates were similar between years indicating that flux rates are consistent year to year.

Soil flux rates at Big Dam West were the lowest rates measured over undisturbed terrestrial soils. Flux rates at the other forested sites were consistently positive but also low and did not show a wide range in daily flux rates with maximum diurnal variation of 0.5 to 5.7 ng/m²/hr. Low flux rates and minimal diurnal variability over forested soils is likely due to poor light penetration to the soils under the forest canopy. Average daily radiation at the forest floor remained < 0.02 kW/m² with maxima not exceeding 0.3 kW/m². Solar radiation at the CAPMoN site (forest clearing) was much higher (>0.2 kW/m² and daily maxima ~1.0 kW/m²). Factor analysis of soil flux and meteorological data from the forested sites showed poor relationship between flux and radiation-temperature which may indicate that processes driving mercury flux under a forest canopy may not be primarily meteorological. However, flux at the CAPMoN site with good exposure and comparable soil mercury content was lower than flux measured at Puzzle and North Cranberry forest sites and only slightly greater than that measured at the Big Dam West site. Mean daily mercury flux measured over recently disturbed soil at the CAPMoN site was significantly greater (p<0.01) that flux rates measured over undisturbed soil at the same site (8.0 compared to 0.9 ng/m²/hr; Table 1). Nighttime flux remained positive soil and the diurnal range (5 to 20 ng/m²/hr) was higher at this site than at the forest sites suggesting that there is more potential volatilization of mercury from deeper soils possibly due to weathering losses of surficial soil mercury.

Table 1. Mercury flux and meteorological data at various aquatic and terrestrial sites from Atlantic Canada.

Site (yr) [blank flux ng/m^2/hr]	Surface Type/ Hg Conc.	24 hr Mean Hg Flux ng/m^2/hr. (range)	Air Temp deg C	RH %	Solar Radiation kW/m^2	Soil/ Water Temp deg C	Wind Speed m/s	Wind Dir. deg T
I. TERRESTRIAL
i. Undisturbed	[Hg(t)] mg/g
BDW (97)	forest soil [0.33]	- 0.4 (-1.3 to 1.0)	19.1	90	0.010	15.2	2.6	184
BDW (99) [0.24]	forest soil [0.20]	0.3 (-0.1 to 1.1)	19.6	72	0.010	16.1	0.9	106
PUZ (99) [0.18]	forest soil [0.24]	1.5 (0.5 to 5.7)	15.6	70	0.020	14.0	2.0	236
NCL (97)	forest soil [0.30]	1.9 (1.1 to 3.3)	17.0	88	0.008	15.5	0.2	075
NCL (99) [0.05]	forest soil [0.15]	2.2 (1.8 to 3.0)	18.9	89	0.005	17.5	0.8	227
CAP (97)	Open field [0.10]	0.9 (-0.7 to 3.8)	17.3	85	0.362	18.4	2.5	115
ii. Disturbed
CAP (99) [0.14]	Open field n/a	8.0 (5.1 to 20.2)	20.6	82	0.223	23.5	1.3	169
CARIBOU^a MINES (98)	gold mine tailing	259. (53 to 415.)	20.0	71	0.255	21.8	0.5	050
GOLDENVILLE ^aMINES (98) [15.4]	gold mine tailing [2.13]	237 (116 to 640)	19.6	82	0.181	22.2	0.9	048
II. AQUATIC
i: Freshwater	[Hg(t)] ng/l
BDW (97) n/a	high DOC [6.5]	5.0 (0.5 to 44)	21.8	81	0.250	23.2	2.8	231
NCL (97) n/a	med. DOC [2.8]	2.2 (-0.3 to 13.5)	19.3	96	0.122	20.3	1.1	123
PUZ (99) n/a	low DOC [0.9]	0.6 (-0.1 to 2.5)	21.2	75	0.246	23.4	1.5	201
Ii: Marine	[Hg(t)] ng/l
Halifax Harbor (99)	polluted harbor [0.83]	0.7 (0.1 to 2.3)	14.9	77	0.187	16.7	1.4	166

Note: 1997 data from Boudala et al., (1999); BDW = Big Dam West Lake; PUZ = Puzzle Lake; NCL = North Cranberry Lake; CAP = CAPMoN air chemistry site; ^a mine sites used mercury amalgamation process in gold recovery. Mine tailing soil mercury data from Wong et al., 1999.

Mercury flux from the surface of mine tailings showed strong diurnal cycles of mercury evasion correlated with diurnal meteorology. Daytime mean and daily maximum flux rates over the tailings were two orders of magnitude greater than rates measured over native soils in Kejimkujik National Park (Table 1). Although nighttime flux rates over mine tailings were lower, they remained positive with minimums dropping to only 53 and 116 ng/m²/hr. This was an order of magnitude greater than those measured at Kejimkujik. Chamber flux rates have been found to be strongly influenced by chamber air turnover rates (Lindberg et al., 1999).

Total gaseous mercury concentrations inside the flux chamber averaged 182 and 236 ng/m³ thus, it is possible that diffusion gradient at the soil-air boundary was impeded. Flux measured using this method may be underestimated.

Mercury flux measurements conducted over the surface of 3 freshwater lakes was also bi-directional though negative (depositional) fluxes were low and not significantly different from zero assuming a method detection limit of approximately 0.2 ng/m²/hr. Daily average mercury flux over water was somewhat greater at Big Dam West Lake than were fluxes measured over forest soils (Table 1). However, air-water fluxes in lakes with moderate and low organic content were comparable in magnitude to those measured over forest soils. The daily maximum lake mercury flux was greater with increasing organic content reaching 13.5 and 43 ng/m²/hr over water at Big Dam West and North Cranberry Lakes, respectively. Factor analysis of flux and environmental variables in lakes indicated that mercury air-water exchange processes in fresh waters may be more closely tied to ambient sunlight and water temperature.

Conclusions The magnitude of air-surface mercury exchange in undisturbed areas was small (<10 ng/m²/hr) relative to rates measured at sites contaminated or naturally enriched with mercury (Gustin et al., 1999). Other studies have established relationships between mercury flux and selected environmental parameters such as water temperature (Xiao et al., 1991), soil temperature and radiation (Carpi and Lindberg, 1998) and air temperature (Poissant et al., 1999). Other factors such as mercury speciation may also be important (Carpi and Lindberg, 1997). In this study, flux at forested sites was not closely linked to meteorology. The meteorological link was more apparent with flux measured over fresh waters. Within the limited number of sites measured in this study there was no apparent correlation between substrate mercury content and flux although this is likely due to the narrow range in concentrations observed.

References

Boudala, F. S., Folkins, I., Beauchamp, S.T., Tordon, R.T., Neima, J. and Johnson, B. (1999). Water, Air and Soil Pollution (in press).

Carpi, A., and Lindberg, S.E. (1997). Environ. Sci. Technol., 31:2085-2091.

Carpi, A., and Lindberg, S.E. (1998). Atmos. Environ., 32:873-882.

Gustin, M., Rasmussen, P., Edwards, E., Schroeder, W.H., and Kemp, J. (1999b). J. Geophys. Res., 104 D17:21873-21878.

Lindberg, S.E., Zang, H.,Gustin, M., Vette, A.., Marsik, F.,. Owens, J.., Casimir, A., Ebinghaus, R., Edwards, G., Fitzgerald, C., Kemp, J., Kock, H.H., London, J., Majewski, M., Poissant, L., Pilote, M., Rasmussen,.P., Schaedlich, F., Schneeberger, F., Sommar, J., Turner, R., Wallschlager, D., and Xiao, Z. (1999). J. Geophys. Res., 104 D17:21879-21888.

Poissant, L., Pilote, M., and Casimir, A., (1999). J. Geophys. Res., 104 D17:21845-21858.

Rasmussen, P.E., Edwards, G.C., Kemp, R.J., Fitzgerald-Hubble, C.R. and Schroeder, W.H. (1998). An International Symposium, Metall. Soc. Of the Can. Inst. of Min., and Pet., Montreal, Que.,Canada, May 3-7, 1998.

Schroeder, W.H., Munthe, J., Lindqvist. (1989). Water, Air and Soil Pollution. 48:337-347.

Wallschlagar, D., Turner, R.R., London, J., Ebinghaus, R., Kock, H.H. and Xiao, Z. (1999). J. Geophys. Res., 104 D17:21859-21872.

Xiao, Z., Munthe, J., Schroeder, W.H., and Lindqvist, O. (1991).Tellus, Ser. B, 43:267-279.