Regional drivers of Hg loadings informed by spatially and temporally dense observations

Mercury (Hg) is a neurotoxic element that can bioaccumulate in aquatic food webs, posing risks to high trophic level species, including humans. While continental-scale gradients of atmospheric Hg concentration and deposition can be evaluated using national monitoring networks, regional gradients of Hg loadings remain poorly defined. Evaluation of regional Hg gradients has been limited by coarse spatial coverage of observations along with differences in standard operating procedures between measurement sites, hindering quantitative inter-site comparison.

Here, we use Hg observations in the northeastern United States coordinated by the National Atmospheric Deposition Program (NADP) to identify regional variability in Hg loadings and quantify the range of surface, chemical, and physical fluxes that can explain observed seasonal and diurnal trends. Elemental Hg (Hg⁰) concentration measurements between 2014 and 2020 at four key sites are used, two representing the New York City (NYC) metropolitan area (Bronx, NY and New Brunswick, NJ), and two representing rural conditions approximately 350 km north of NYC (Huntington Wildlife Preserve NY and Underhill VT). These measurements are supplemented by wet deposition measurements at several sites throughout the region along with two sites that have directly quantified Hg⁰ surface fluxes over northern forests for greater than one year. We construct a box model representing the NYC metropolitan area that accounts for horizontal advection, boundary layer entrainment, and wet deposition to evaluate the range of surface, chemical, and meteorological fluxes required to reproduce observed diel and seasonal Hg variability. These updated fluxes are then implemented in the GEOS-Chem-Hg chemical transport model, and its impact on regional and global deposition are evaluated.

We find that Hg⁰ concentrations were consistent at rural (1.22 and 1.30 at Huntington and Underhill, respectively) and urban (1.67 and 1.72 at New Brunswick and Bronx, respectively) sites but demonstrated a noteworthy difference across the urban to rural gradient. Interestingly, we find that diel variations differed across region type, with maximum summertime concentrations at urban (rural) sites occurring during the night (day). Furthermore, despite the large average gradient between sites, concentrations at urban and rural sites approached one another during turbulent daylight hours, suggesting a synoptic scale forcing on atmospheric concentrations. Despite obvious spatial and temporal gradients in observations, we find that these patterns are absent from current versions of the GEOS-Chem chemical transport model, underscoring the need to reevaluate surface, chemical, and meteorological fluxes controlling regional gradients of Hg atmospheric loadings.