Improved atmospheric mercury simulation using updated gas-particle partition and organic aerosol concentrations.

The gaseous or particulate forms of divalent mercury (Hg) significantly impact the spatial distribution of atmospheric mercury concentration and deposition flux (FLX). In the new nested-grid GEOS-Chem model, we try to modify the Hg gas-particle partitioning relationship with synchronous and hourly observations at four sites in China. Observations of gaseous oxidized Hg (GOM), particulate-bound Hg (PBM), and PM were used to derive an empirical gas-particle partitioning coefficient as a function of temperature (T) and organic aerosol (OA) concentrations under different relative humidity (RH). Results showed that with increasing RH, the dominant process of Hg gas-particle partitioning changed from physical adsorption to chemical desorption. And the dominant factor of Hg gas-particle partitioning changed from T to OA concentrations. We thus improved the simulated OA concentration field by introducing intermediate-volatility and semi-volatile organic compounds (I/SVOCs) emission inventory into the model framework and refining the volatile distributions of I/SVOCs according to new filed tests in the recent literatures. Finally, normalized mean biases (NMBs) of monthly gaseous element mercury (GEM), GOM, PBM, WFLX were reduced from -33%-29%, 95%-300%, 64%-261%, 117%-122% to -13%-0%, -20%-80%, -31%-50%, -17%-23%. The improved model explains 69%-98% of the observed atmospheric Hg decrease during 2013-2020 and can serve as a useful tool to evaluate the effectiveness of the Minamata Convention on Mercury.