Variability-based constraints on regional gaseous elemental mercury fluxes

Abstract: Global efforts to prevent future human and environmental exposure to mercury (Hg) require a quantitative understanding of its environmental fate, which is mediated by the atmospheric transport of gaseous elemental mercury (Hg⁰). Hg⁰ is exchanged with the atmosphere by spatially diffuse surface fluxes (i.e., emissions and deposition) that remain poorly constrained due to a limited number of measurements. Here, we present a Bayesian inversion approach that uses day-to-day concentration variability at a measurement site as a constraint to estimate the magnitude of regionally dominant fluxes. 

We first use this inversion framework to estimate Hg⁰ deposition (i.e., vegetative assimilation) in a deciduous forest in Massachusetts, USA where ecosystem-scale flux measurements are available, finding that this approach demonstrates that litterfall underestimates regional Hg⁰ uptake. We then use this approach for two additional case studies, first to estimate Hg⁰ biomass burning emissions in Northern Territory, Australia, and then to estimate Peruvian artisanal and small-scale gold mining (ASGM) emissions using observations from a nearby mountaintop site. We estimate Northern Territory biomass burning emissions of 5.3 Mg (5^th-95^th percentiles: 2.5-15.3 Mg) in 2014 and 2.4 Mg (5^th-95^th percentiles: 0.6-11.9 Mg) in 2015. We also estimate total ASGM emissions in southern Peru to be 33.6 Mg (5^th-95^th percentiles: 16.3-72.3 Mg) but find that our estimates are sensitive to meteorological inputs and prior flux estimates at this site. 

Our results highlight the notable impact of modeling choices on posterior fluxes, indicating that these uncertainties should be considered in future atmospheric inversions for Hg⁰.