Les échanges surface-atmosphère du mercure gazeux dans l'écosystème lac Ontario/fleuve Saint-Laurent Mercury surface-atmosphere gas exchange in lake Ontario/St.Lawrence river ecosystem

Résumé

La volatilisation du mercure des surfaces vers l'atmosphère et les dépôts atmosphériques du mercure sont des phénomènes importants dans la dynamique globale du mercure. Les échanges surface-atmosphère du mercure gazeux dans l'écosystème lac Ontario/fleuve Saint-Laurent sont variables dans le temps et dans l'espace. Bien que le modèle de la double couche montre que la grande partie des écosystèmes aquatiques sont en sursaturation par rapport à la constante d'Henry, des observations in situ, à l'aide de techniques de chambre à flux, montrent que des dépôts gazeux sont également possibles. Les dépôts gazeux du mercure dans l'écosystème lac Ontario/fleuve Saint-Laurent oscillent entre 0 et 4,66 ng/m²/h alors que les valeurs de volatilisation varient entre 0 et 9,28 ng/m²/h. Globalement, la volatilisation médiane du mercure est de 0,77 ng/m²/h alors que le dépôt gazeux médian est d'environ un ordre de grandeur inférieur (0,075 ng/m²/h). En été, l'ensemble des dépôts atmosphériques du mercure semble être mis à contribution lors de la volatilisation du mercure au-dessus des sols. Il semble que la majeure partie de cette volatilisation serait en fait de la ré-émission du mercure vers l'atmosphère. En hiver, seule la portion gazeuse des dépôts de mercure semble être ré-émise vers l'atmosphère. Plusieurs observations dans un marécage fluvial du lac Saint-Pierre (baie Saint-François) montrent que les flux de volatilisation du mercure sont supérieurs en période sèche qu'en période inondée (0,83 vs. 0,52 ng/m²/h). Ainsi, en période d'inondation le mercure réactif disponible pour la volatilisation serait en compétition avec les mécanismes responsables pour la méthylation du mercure (bio-disponible) et/ou la formation du sulfure de mercure (inerte sous forme de cinabre). Cet article a pour objectifs de présenter et discuter les échanges surface-atmosphère du mercure gazeux dans l'écosystème lac Ontario/fleuve Saint-Laurent.

Mots-clés:

• Mercure,
• échanges gazeux,
• lac Ontario,
• fleuve Saint-Laurent

Abstract

This paper presents and discusses mercury surface-atmosphere gas exchange in Lake Ontario/St. Lawrence River ecosystem. Atmospheric sources are recognised to be significant in the cycling of global mercury. Being volatile in its elemental and dimethylated forms, mercury is distributed worldwide. The dominant form of mercury in the atmosphere is gaseous elemental mercury (Hg°) (~ 98%). Cycling of atmospheric mercury proceeds by gas exchange, particle settling or by rain scavenging. Wet and particle Hg deposition mainly involves the oxidised form, i.e. the Hg(II) species, which is relatively immobile. Chemical, photolytic or biological reduction to the elemental form can increase the mobility of mercury.

Transport of mercury from bodies of water to the atmosphere (volatilisation) and atmospheric deposition are significant components for mercury budgets in lakes and rivers. The large majority of aquatic ecosystems studied so far have been found to contain dissolved gaseous mercury at concentrations that were supersaturated relative to the equilibrium values predicted by Henry's law. Evasion of elemental mercury was suggested to occur over the ocean and from inland waters but was only measured directly in a few cases. A few instances of net deposition were observed over inland waters. The dynamic aspect of mercury exchange has to be explored for a better understanding of mercury behaviour in the environment. Part of this understanding can be achieved by quantifying the rate of exchange or fluxes between compartments.

Until recently, mercury fluxes were quantified by estimations using mathematical models or by measuring the mercury accumulation in biological, soil and sediment samples. It is only in the last few years that technical improvements have allowed a more or less direct measurement of mercury fluxes between air/soil and air/water compartments. One of these methods is the dynamic flux chamber, which is a simple and relatively reliable technique. Total Gaseous Mercury (TGM) analysis in this study was achieved with an automatic analyser (Tekranâ 2537A). Briefly, the analytical train of this instrument is based on amalgamation of mercury onto a pure gold surface followed by a thermo-desorption step and analysis by Cold Vapour Atomic Fluorescence Spectrophotometry (CVAFS) (l=253.7 nm). Dual cartridge designs allowed alternate sampling and desorption, resulting in continuous measurements of mercury in the air stream. Both modelling and dynamic flux chamber methods are reported in this paper. The former technique was applied along cruises on Lake Ontario and the Upper St. Lawrence River whereas the latter was applied at various stationary locations along the St. Lawrence River (e.g., pasture, water, wetlands and snow surfaces). The dynamic flux chamber used was built in our laboratory. The chamber consists of a hemispheric stainless steel bowl coated with Teflon®. The open area of the chamber is 0.13 m² and its volume is 10 L. The flow rate into the chamber is 0.09 m³/h. The measurement of mercury in the inlet and outlet air sample ports is achieved by the mercury analyser. A peripheral device using a solenoid valve directs the sample to a specific cartridge of the analyser. Hence, the analyser does sequential measurement of the ports. The mercury gas exchange fluxes across the interfaces (surface-atmosphere) (ng/m²/h) are computed using the mass balance of mercury within the flux chamber.

Mercury flux across the water surface was modelled using a two-layer model. The two-layer model is a convenient but not necessarily a mechanistically accurate model, which depends on empirical relationships reported for other chemicals. The model is based on the saturation of mercury within the layers with respect to Henry's law and the overall mass transfer coefficient (air and water). Since Henry's law for mercury is high, most of the resistance to gas exchange lies in the water film (> 99%). There are two main conditions required to apply the model: (1) the chemical does not undergo any reaction within the layers; and (2) the concentrations at the boundaries of the layers are kept constant long enough that the concentration profile reaches a steady state. The model was applied to estimate the mercury flux (ng/m²/h) during cruises on Lake Ontario and the St. Lawrence River. Mercury surface-atmosphere gas exchanges in Lake Ontario/St. Lawrence River ecosystem vary in space and time. Gaseous mercury deposition in Lake Ontario/St. Lawrence River ecosystem varied between 0 and 4.66 ng/m²/h whereas mercury evasion varied between 0 and 9.28 ng/m²/h. Overall, the median mercury evasion value is 0.77 ng/m²/h which is one order magnitude larger than median deposition (0.075 ng/m²/h). In summertime, total deposition flux over soil surface is counterbalanced by mercury re-emission flux. However, during wintertime, only dry deposition over snow surface is counterbalanced by volatilisation. Hence, mercury snow deposition during wintertime might have a huge impact on the ecosystem especially during the springtime through the influence of the melt water on aquatic biota. Many mercury gas exchange observations in fluvial wetlands in the Lake Saint-Pierre (baie Saint-François) suggested larger emissions over dry wetland than flooded wetland (0.83 vs. 0.52 ng/m²/h). Hence, flooded wetlands offered conditions that contributed to competition between mercury volatilisation and methyl mercury formation or immobilisation as mercury sulphide (cinnabar).

Keywords:

• Mercury,
• gas exchange,
• lake Ontario,
• St. Lawrence River