1. A new mechanism for atmospheric mercury redox chemistry: Implications for the global mercury budget§
Hannah M. Horowitz Daniel J. Jacob1, Yanxu Zhang1, Theodore S. Dibble2, Franz Slemr3, Helen M. Amos1, Johan A. Schmidt1, Elizabeth S. Corbitt1, Eloïse A. Marais1, and Elsie M. Sunderland1
1. Harvard University; 2. SUNY – ESF; 3. Max Planck Institute for Chemistry; *now at JISAO, Univ. of Washington
§Horowitz et al., ACP, 2017 (accepted); ACPD doi: /acp
Fig. 5. Model vs. observed log([TGM]) vs. [O3] in the lower stratosphere ([O3]/[CO] > 1.25 mol mol-1) from CARIBIC aircraft.
1. Introduction
4.1 Global Hg Budget and Lifetimes
Mercury (Hg) is emitted to the atmosphere mainly as volatile elemental Hg0. Oxidation to water-soluble HgII plays a major role in Hg deposition to ecosystems. Here, we implement a new mechanism for atmospheric Hg0/HgII redox chemistry in the GEOS-Chem global model and examine the implications for the global atmospheric Hg budget and deposition patterns.
Fig. 2. Annual ( ) zonal mean mixing ratios of Hg0 and HgII in GEOS-Chem, and Hg0 oxidation rates in number density and mixing ratio units.
Slope mismatch previously attributed to too low oxidation; appears decline of TGM with depth into stratosphere is driven more by mixing than chemistry.
2. Chemical Mechanism
Vertical structure of [Hg0] and [HgII] is a function of chemistry. Gross ox. is faster in NH because of higher NO2 concentrations. However, [HgII] higher in the SH than in the NH because of faster NH reduction due to higher OA.
Gas-phase oxidation
Fig. 6. Monthly mean and std. dev. of [TGM]. Model coupled to MITgcm and in a sensitivity simulation to the GEOS-Chem 2-D slab ocean.
Mined Hg
NH: model seasonality driven by Hg0 oxidation and ocean evasion. Two-way coupling to the MITgcm improves seasonality.
Fig. 3. Global budget of tropospheric Hg in GEOS-Chem. HgII includes gaseous and particulate forms. bottom panel: major (>100 Mg a-1 globally) reactions cycling Hg0 and HgII.
Y= OH, HO2, BrO, ClO, Cl
what's new in Hg0 oxidation vs. standard model?
Faster rate coefficient for HgBr dissociation + new HgBr reduction pathway via NO2
Additional 2nd-step oxidants (+NO2, HO2, ClO, BrO, Cl)
Improved 2nd-step oxidation rate coefficients, including new three-body Troe’s expressions determined for NO2 and HO2 (Jiao and Dibble, 2017)
Gas-phase Cl-initiated oxidation (same mechanism as above)
Aqueous phase oxidation: O3, OH, HOCl
SH: Model has strong seasonality (seen in std. v9-02, Song et al., 2015), observed does not. Missing long-range transport of atmospheric Hg from Antarctica (Angot et al., 2016) or Southern Ocean sea ice and productivity dynamics?
Chemical lifetime of Hg0 against oxidation is shorter (2.7 months) and must be compensated by faster reduction to lead to 5.2-month lifetime of TGM against deposition. Lifetime of HgII against reduction is 13 days, shorter than lifetime against deposition (26 days). Hence, atmospheric HgII reduction must take place.
Fraction of deposition to oceans increased to 80% (from 71% in Holmes et al., 2010) due to NO2 and HO2 oxidants, low marine OA for reduction, higher Br (Fig. 1).
5. Implications for Global Hg Deposition
New representation of aqueous HgII photoreduction rate in clouds and aerosols:
We adjust the reduction rate coefficient (α) in GEOS-Chem to match observed global mean surface TGM concentrations.
Fig. 7. Modeled (background) and observed (symbols) annual Hg wet deposition fluxes over North America (MDN) and China (Fu et al., 2015, 2016).
Why organic-mediated HgII reduction?
Evidence from other aqueous environments
Observed atmospheric isotopic fractionation
For full mechanism see Horowitz et al. (accepted), ACP
NO2 and HO2 as 2nd-stage HgBr oxidants shifts HgII production to lower latitudes, enabling model to capture
3. GEOS-Chem Model Setup
Gas-phase Br-initiated Hg oxidation dominates in the troposphere and stratosphere. The most important pathways are NO2 and HO2, due to their greater abundance.
the location of the observed US Gulf Coast maximum, but the magnitude is underestimated and sensitive to reduction scaling. In rural China, wet deposition is observed to be low despite very high TGM concentrations. This is reproduced by
We use v9-02 of the global GEOS-Chem Hg model with the following updates:
atmospheric Hg chemical mechanism (above)
3-D oxidant concentrations from Schmidt et al. (2016)
two-way monthly coupling to the MITgcm (using ocean concentrations to calculate air-sea exchange in GEOS-Chem )
After spin-up, we conduct a 3-year simulation for driven by GEOS-5 assimilated meteorological data at 4˚ x 5˚ horizontal resolution.
the model where it is driven in part by fast HgII reduction due to high OA concentrations.
4.2 Spatial Distribution and Seasonality
More Hg is deposited to the tropical oceans (49% of total HgII deposition) compared to previous versions of GEOS-Chem where debromination of sea salt aerosol drove fast Hg0 oxidation and deposition to the Southern Ocean.
Fig. 4: Global surface TGM concentrations, in ng m-3 STP (p = 1 atm, T = 273 K). Model values (background) vs observations (land sites: diamonds; ship cruises: circles).
Schmidt et al. (2016)
Parrella et al. (2012)
Fig. 8. Annual HgII deposition fluxes in GEOS-Chem.
µg m-2 a-1
Fig. 1: Global distribution of modeled BrO.
Top: tropospheric mean. Bottom: zonal mean. Figure courtesy of Johan Schmidt.
6. Conclusions
Obs.: 1.47±0.27 ng m-3 STP
model: 1.44 ±0.25 ng m-3 STP
r = 0.57
Br is dominant oxidant in troposphere and stratosphere. HO2 and NO2 dominate 2nd-step oxidation pathways. Increased oxidation implies reduction takes place.
Low wet dep. observed over China attributed to fast OA-mediated HgII reduction
Additional 2nd-step oxidants, increased tropical Br, and low marine organics for HgII reduction lead to increased HgII deposition to oceans, especially in the Tropics
We find the relative standard deviation in TGM across sites is very sensitive to the TGM lifetime, unlike the interhemispheric gradient which is impacted by ocean evasion. The model’s ability to reproduce the observed standard deviation across land sites supports our simulated TGM atmospheric lifetime of 5.2 months.
Selected References:
Holmes et al., ACP, 2010
Schmidt et al., JGR, 2016
Acknowledgments:
Jeroen Sonke, Ian Hedgecock, 3 anon reviewers, NSF.
Fu et al., ACP, 2015, 2016
Angot et al., ACP, 2016
Jiao and Dibble, PCCP, 2017
Parrella et al., ACP, 2012
Song et al., ACP, 2015