Assimilation of gaseous atmospheric mercury – Hg(0) – by vegetation activity is considered the most important deposition process in terrestrial environments. Subsequent mobilization of Hg from soils provides a source of Hg to aquatic ecosystems where it enables production of methylmercury which bioaccumulates in the aquatic food chain posing a threat to human and wildlife health. Previous studies have shown strong seasonality of gaseous atmospheric Hg(0) concentrations with summertime minima and wintertime maxima and strong correlations to CO₂ and attributed these patterns to vegetation uptake of Hg(0) along with CO₂ during growing periods. However, diurnal patterns of Hg(0) were inconsistent and flux measurements of Hg(0) to confirm these linkages are largely missing. Here, we characterized the seasonal and diel dynamics of atmospheric Hg(0) concentrations along with direct ecosystem-level Hg(0) flux measurements by means of a micrometeorological flux-gradient approach. In addition, we measured concentrations and fluxes of other trace gases, in particular CO₂, O₃ along with CO and meteorological variables to assess quantitative relationships among these variables. The study was conducted over a temperate deciduous broadleaf forest located at the Harvard Forest research site located in Petersham, MA, USA for two full years (2018-2020).

Atmospheric Hg concentration patterns both over the canopy as well as above the forest floor showed strong minima from July to October and wintertime maxima for both years, with an average seasonal amplitude of 0.12 ng m^-3 (equivalent of 10.5% of mean annual concentrations). Growing season Hg(0) concentration declines were consistent with observed ecosystem-level Hg(0) deposition, providing direct indication that seasonality in concentrations at this remote site was in fact determined by a regional forest Hg sink. In addition, seasonality of Hg(0) concentrations was strongly CO₂ concentrations and fluxes. Notably, however, seasonal Hg(0) concentration minima occurred later (September) than those of CO₂ (August) and the pronounced forest Hg(0) sink lasted through November, while the CO₂ sink largely ceased by late October. Diel patterns of Hg(0) concentrations, in contrast, were inconsistent with diel patterns of Hg(0) fluxes and patterns of CO₂, with Hg(0) concentrations showing lower values during nighttime and highest concentrations during midday. We attribute the contrasting patterns to boundary layer mixing leading to advection of air masses aloft with enriched Hg(0) concentrations during turbulent daytime planetary boundary layers. Finally, we observed Hg(0) sources attributable to transport of polluted air masses, which was pronounced particularly during the spring season. In conclusion, a regional Hg(0) sink led to strong growing-season atmospheric Hg(0) concentration declines, yet diel patterns and correlations to other trace gases also showed that Hg(0) levels were co-determined by emission sources and boundary layer mixing.