A State-of-the-Science Hg Redox Mechanism for Atmospheric Models: Constraints from Observations and Global Implications

Hannah M. Horowitz (hmhorow@fas.harvard.edu)1, D. M. Jacob1, H. M. Amos1, D. J. Streets2, Y. Zhang1, T. S. Dibble3, F. Slemr4, E. M. Sunderland1 1. Harvard University; 2. Argonne National Lab; 3. SUNY – ESF; 4. Max Planck Institute for Chemistry

1. Introduction

Selected References: Holmes et al., ACP, 2010 Schmidt et al., JGR, in revision

Streets et al., EST, 2011 Zhang et al., PNAS, in press

4. Results: HgII Production

Mined Hg Fraction accounted for in Streets et al. (2011) atmospheric emissions

Most mercury (Hg) emission is as the elemental gas Hg0, which is transported globally in the atmosphere. It is eventually oxidized to divalent Hg (HgII) salts, which are highly water-soluble, can partition into the aerosol, and are efficiently removed from the atmosphere by wet and dry deposition. HgII can also be reduced back to Hg0. Understanding the redox chemistry of atmospheric Hg is critical to determining the patterns of deposition to surface environments, where Hg can be transformed to toxic and bioaccumulative methylmercury. Here we present a state-of-the-science redox mechanism for use in atmospheric models, and implement it in the GEOS-Chem model to evaluate against atmospheric observations and examine the implications for Hg deposition.

2. Atmospheric Mercury Chemistry Oxidation by Br atoms is known to cause atmospheric mercury depletion events in the Arctic boundary layer in spring. Holmes et al. (2010) suggested that Br atoms could be the main Hg0 oxidant on a global scale, with Br originating mainly from marine biogenic sources. Observations of background tropospheric BrO support this hypothesis (Schmidt et al., 2015). Oxidation of Hg0 to HgII by Br is a two-stage exothermic mechanism where the second stage oxidizing HgI to HgII can be carried out by a number of radical oxidants (Dibble et al., 2012).

Fig. 1: Comparison of regional emissions

between Streets et al. (2011) and improved

inventory (Zhang et al. in press).

Eq. (1-4): Atmospheric Hg oxidation mechanism, with additional radical

oxidants verified theoretically by Dibble et al. (2012) highlighted in red

HgII forms complexes with organic acids that are strong enough for electron transfer during photolysis to reduce HgII. We assume that organic aerosol (OA) is representative of the ensemble of compounds forming photoreducible HgII complexes. We calculate the rate of photoreduction for HgII

(aq) in cloudwater and aqueous aerosols as proportional to the local concentration of OA and to the NO2 photolysis frequency (jNO2). We adjust the reduction rate coefficient (α) in GEOS-Chem to match observed global mean surface Hg0 concentrations and the overall lifetime of Hg0 implied from observations.

X = Br, Cl Y = Br, OH, HO2, NO2, BrO, ClO

1) Hg0 +X+M→HgX+M2) HgX+M→Hg0 + X+M3) HgX+X→Hg0 +X2

4) HgX+Y+M→HgXY+M

5) α jNO2[OA][Hg(aq)

II ] Eq. (5): Representation of aqueous HgII photoreduction rate in clouds and aerosols.

A longstanding conundrum has been the apparent disconnect between increasing global emissions trends and measured declines in atmospheric mercury in North America and Europe. We construct an improved global emission inventory for the period 1990-2010 accounting for emissions from commercial products, emission controls on coal-fired utilities, and trends in ASGM and find a 20% decrease in total Hg emissions and a 30% decrease in anthropogenic Hg0 emissions, with much larger decreases in North America and Europe offsetting the effect of increasing emissions in Asia.

0"

500"

1000"

1500"

1" 1" 1" 1" 1" 1" 2" 2" 2" 2" 2" 2" 3" 3" 3" 3" 3" 3"

Different&trend&in&global&Hg&emissions&than&previously&thought&

1990&

2000&

Hg0&ASGM&Hg0&products&Hg0&other&HgII&Streets&et&al.&(2011)&total&Hg&

Hypothesis testing

•  Byproduct emissions (Streets et al., 2011) •  Commercial emissions (Horowitz et al., 2014) •  Updated Chinese emissions (Zhang et al., 2015) •  Updated Indian emissions (Chakraborty et al., 2013) •  Updated ASGM emissions (Muntean et al., 2014) •  Speciated utilities emissions

Emission Inventories

•  CAMNet atmospheric TGM concentrations •  AMNet atmospheric Hg0 •  CARIBIC free tropospheric Hg0 •  MDN wet deposition flux •  EMEP atmospheric TGM concentrations •  EMEP wet deposition flux

Atmospheric Observations

Y.&Zhang,&Horowitz,&et&al.&in&prep&

2010&

Europe& N.&America& Asia&

1500&

&

&

1000&

&

&

500&

&

&

0&

Hg&emissions,&M

g&a6

1 &

Global&Hg&emissions&now&decrease&from&1990&to&2010&&

We use v9-02 of the global GEOS-Chem Hg model including atmosphere-ocean-land interactions. 3-D atmospheric transport is driven by GEOS-5 assimilated meteorological data from NASA GMAO, degraded to 4˚ x 5˚ horizontal resolution. There are three Hg tracers transported in the atmosphere: Hg0, gas-phase HgII, and aerosol-phase HgII. With the new emissions inventory, declining atmospheric concentrations can be explained by the phase-out of Hg from commercial products and by shifts in the speciation of Hg emissions driven by air pollution control technologies.

Trends)in)Hg0)captured)with)updated)regional)Hg)emissions)alone)

[Hg0

], ng

m-3

1990 2000 2010 1990 2000 2010 1990 2000 2010 Climate change Local emission

Global emission

Local emission

Background)–)GEOS\Chem)1990)–)2010)trends)Foreground)–)observed)trends))

Y.'Zhang'et'al.,'submiRed'(PNAS)'Figure 2. Regional trends for 1990-2013 in atmospheric Hg0 concentrations - observations (CAMNet, AMNet, ELA) and GEOS-Chem using our revised anthropogenic emissions inventory for 1990 and 2010.

3. Improved Emissions Inventory & Model Setup

Figure 3. Global total Hg oxidation by pathway (Mg a-1)

Sales

BrO Br ClO NO2 HgCl

Gas-phase Br-initiated Hg oxidation dominates in the troposphere and stratosphere. The most important pathways are NO2 and HO2, due their greater abundance relative to other halogen radicals and OH (Fig. 3).

−50 0 50Latitude

5

10

15

20

25

Altit

ude

New Model Version: Zonal Gross Ox

0.00 0.50 1.00 1.50 2.00 pmol mol−1 y−1

−50 0 50

5

10

15

20

25

−50 0 50Latitude

5

10

15

20

25

Altit

ude

Old Model Version: Zonal Gross Ox

0.00 0.50 1.00 1.50 2.00 pmol mol−1 y−1

−50 0 50

5

10

15

20

25

−50 0 50Latitude

5

10

15

20

25

Altit

ude

Absolute Difference

−0.30 −0.15 0.00 0.15 0.30 pmol mol−1 y−1

−50 0 50

5

10

15

20

25

−50 0 50Latitude

5

10

15

20

25

Altit

ude

Percent Difference

−100 −50 0 50 100 %

−50 0 50

5

10

15

20

25

Most Hg oxidation occurs in the upper troposphere (Fig. 4) due to the distribution of oxidants. Lower altitude oxidation is primarily due to the presence of HO2 and NO2. Br – initiated oxidation also dominates in the stratosphere, consistent with aircraft observations in the UT/LS (not shown). Fig. 4: Zonal mean gross Hg

oxidation.

GC51F-1160

5. Implication for Wet Deposition & Surface Ocean New Model Version: Total Wet Dep

5 10 15 20 ug/m2/y

Old Model Version: Total Wet Dep

5 10 15 20 ug/m2/y

Absolute Difference

−5.00 −2.50 0.00 2.50 5.00 ug/m2/y

Percent Difference

−50 −25 0 25 50 %

New Model Version: Total Wet Dep

5 10 15 20 ug/m2/y

Old Model Version: Total Wet Dep

5 10 15 20 ug/m2/y

Absolute Difference

−5.00 −2.50 0.00 2.50 5.00 ug/m2/y

Percent Difference

−100 −50 0 50 100 %

Annual total wet Hg wet deposition. Percent Change

Figure 6. Hg wet deposition. Percent change is after including improved Hg redox mechanism and emission inventory.

Changes in wet deposition are due to a combination of: •  changes in Br distribution (Schmidt et al. 2015) – less deposition over the

Southern Ocean •  additional second step oxidants like HO2 that have greater abundance in the

tropics – increased deposition to the tropics •  changes in Hg reduction –distributions of HgII are related to spatial pattern of OA Now, differences between Br – initiated oxidation and OH/O3 oxidation are slightly muted with respect to HgII wet deposition.

Dibble et al., ACP, 2012

HgII$Reduc*on:$constraints$from$Hg0$life*me$and$poten*al$mechanisms$

!!!!!!!!!!!!!!!!!!!!!!!!!!!!Model:!10!months$Observa4onal!constraints:!5!–!12!months!

Hg0!life4me!against!oxida4on!+!deposi4on:$

$•  Published$rate$coefficients$for$aqueous$HgII$reduc*on$range$from$1!x!10@7!s@1!to!0.013!s@1!•  we$need:$~1.7!x!10@6!to!~6.4!x!10@6!s@1!global!rate!coefficient!for!aqueous!photoreduc4on$

Hg0$$(3.3$Gg)$

Soil$/$Ocean$surface$

14,000$

8,800$5900$1200$ 6200$ 830$

reduc*on$

oxida*on$

emissions$deposi*on$uptake$

emissions$deposi*on$

HgII$$(0.5$Gg)$

Fluxes$in$Mg$aS1$

Figure 5: Global annual Hg budget. Hg0 lifetime against oxidation + deposition: Model: 7 months. Observational constraints: 5 – 12 months Due to additional and more abundant 2nd step oxidants, HgII reduction is needed to comply with constraint on Hg0 lifetime from observations, even within uncertainty in oxidation reaction rates (~+/- 50%)

•  Observed decreasing trends in atmospheric Hg0 can be explained by changes in domestic emissions – based on recently available data

•  Atmospheric redox reactions determine where emitted Hg will be deposited to surface environments

•  Br is dominant oxidant in troposphere and stratosphere. Abundant radicals (HO2, NO2) dominate overall oxidation pathways

•  HgII reduction pathway is needed, even given uncertainty in Hg0 oxidation rates •  Additional 2nd-step oxidants and improved Br simulation lead to increased HgII

deposition to the Tropics and decreased deposition to Southern Ocean

6. 30- Second Summary

Model compares well to limited wet deposition observations, but other locations are needed to verify redox mechanism.

Hg Wet Deposition, GEOS−Chem 2009 MDN 2006, 2007, 2008, 2009, 2010, 2011, 2012

0

5

10

15

20

ug m−2 y−1

Figure 7. Average 2009-2011 model wet deposi t ion, compared to 2008-2014 MDN data (µg m-2y-1).