Quantitative field measurements of atmospheric gaseous elemental mercury using drone-mounted high-flow rate active...

For this study, gaseous elemental mercury (GEM) concentrations were evaluated quantitatively using a drone-mounted high flow-rate active sampling approach under field atmospheric conditions. Relative to earlier low-flow drone-based sampling systems, the higher sampling flow rate (2.50 L/min) facilitated efficient sample collection within limited flight durations, supporting proof-of-concept validation of quantitative GEM measurements. High flow-rate trapping was achieved using a hand-sized pump, with GEM collected by amalgamation on a gold (Au) trap and analyzed using cold vapor atomic absorption spectroscopy (CVAAS). Field deployment at a domestic background monitoring site yielded GEM concentrations comparable to regional background levels (approximately 1.5 ng/m³), with analyte breakthrough to a secondary trap remaining below the method detection limit (MDL: <0.5 ng/m³). Comparative evaluation against an automated continuous monitoring system employing cold vapor atomic fluorescence spectroscopy (CVAFS) revealed good agreement (R² > 0.989), with no significant deviation or evidence of non-equivalence. Absolute errors were small: both RMSE and MAE were less than 0.2 ng/m³, reduced to 0.122-0.133 ng/m³ after outlier removal. Normalized to the MDL, relative errors were 16-22%, which is consistent with reported uncertainty at low concentrations. Under aerial sampling conditions, including rotor-induced turbulence and platform motion, measurements were correlated strongly (R² = 0.993), and were found to have low absolute errors (RMSE = 0.122 ng/m³; MAE = 0.107 ng/m³) and a stable error structure (RMSE/MAE = 1.14). No significant difference in mean concentrations was detected. Statistical equivalence within ±0.1 ng/m³ (p = 0.018 and 0.010) was confirmed, indicating negligible rotor-wash influence under short-term averaged conditions. Dynamic flight effects were evaluated by comparing hovering and moving sampling. Small errors (RMSE = 0.082 ng/m³; MAE = 0.044 ng/m³) and statistical equivalence within ±0.1 ng/m³ (p = 0.012) were observed at the cultivated site. At a forest site, larger but stable errors (RMSE = 0.210 ng/m³; MAE = 0.160 ng/m³) precluded equivalence, likely because of limited statistical power. Extended field observations across coastal, cultivated, forested, and urban environments revealed an altitude-dependent GEM distribution, characterized by decreasing concentrations with increasing sampling height. Observed spatial patterns and seasonal variation suggest that land-use characteristics, particularly the presence of vegetation, influence near-surface GEM distributions. That finding is consistent with earlier reported bidirectional atmosphere-vegetation exchange processes. Overall, these findings demonstrate the feasibility of high-flow-rate airborne active sampling as a proof-of-concept approach for obtaining quantitative GEM measurements comparable to those obtained using conventional methods. This sampling method has the potential to facilitate exploratory atmospheric mercury observations in difficult-to-access locations and vertical regions.