GLOBAL MODELING OF MERCURYWITH Br AS ATMOSPHERIC OXIDANT

Chris D. Holmes and Daniel J. Jacob

and funding from EPRI and NSF

Mercury in polar bear fur

US fish consumption advisories (EPA)

Wyoming ice core

EPA, 2007

RISING MERCURY IN THE ENVIRONMENT

Schuster et al., 2002

Dietz et al., 2006

THE MERCURY CYCLE: MAJOR PROCESSES

Hg(0) Hg(II)

particulate

Hg

burial

SEDIMENTS

uplift

volcanoeserosion

oxidation (~1 y)

reduction

volatilization

Hg(0) Hg(II)oxidation

reduction

deposition

biologicaluptake

ANTHROPOGENIC PERTURBATION:fuel combustion

waste incinerationmining

highly water-soluble

ATMOSPHERE

SOIL/OCEAN

ATMOSPHERIC REDOX CHEMISTRY OF MERCURY

Hg(II)Hg(0)OH, O3

HO2 (aq)

XX

X?Calvert and Lindberg, AE 2005Hynes et al., UNEP 2008

• Oxidation of Hg(0) by OH or O3 is endothermic

, ,

Hg Br M HgBr M

HgBr X M HgBrX M X OH Br Cl

• Oxidation by NO3, BrO, O3 (aq) is probably negligible

Br, ClStandard models

• Oxidation by Br and Cl may be important:

• Atmospheric reduction of Hg(II) is hypothetical

MERCURY DEPLETION EVENTS (MDEs) IN ARCTIC SPRINGARCTAS-A aircraft campaign (April 2008) showed ubiquitous MDEs over sea ice

Hg(0) vs. O3 in near-surface data

• MDEs are confined to below 0.5 km altitude, occur concurrently with ODEs and in presence of soluble bromide

• Mercury depletion is consistent with Hg + Br

Mao et al., Kim et al., submitted

DIURNAL CYCLE OF REACTIVE GASEOUS MERCURY (RGM) IN MARINE BOUNDARY LAYER

Early a.m. rise, midday peak suggests Br chemistry, deposition via sea salt uptake

Hg(0) HgBrBr

T

Br, OHHgBrX

sea-salt aerosol

HgCl32-, HgCl4

2-

deposition

MBL budget

Model predicts that ~80% of Hg(II) in MBL should be in sea salt:

Holmes et al. [2009]

Observed [Laurier et al., 2003]Model Hg(0)+BrModel Hg(0)+OH

Subtropical Pacific cruise data

kinetics from Goodsite et al. [2004]

WHAT DO ATMOSPHERIC DATA TELL US ABOUT GLOBAL Hg(0) OXIDATION?

• Atmospheric Hg lifetime against deposition must be ~ 1 year

– Observed variability of Hg(0)

• Oxidant must be photochemical

– Observed late summer minimum at northern mid-latitudes

• Oxidant must be in gas phase and present in stratosphere

– Hg(II) increase with altitude, Hg(0) depletion in stratosphere

…WHAT DO ATMOSPHERIC DATA TELL US ABOUT GLOBAL Hg(II) REDUCTION?

• If it happens at all it’s mostly in lower troposphere (clouds?)

– RGM increase with altitude, Hg(0) depletion in stratosphere

Oxidation by Br atoms can satisfy these constraints [Holmes et al., 2006]

TROPOSPHERIC BROMINE CHEMISTRYsimulated in GEOS-Chem global chemical transport model

CHBr3 hv, OH

14 days

CH2 Br2

OH

91 days

CH3BrOH

1.1 years

Br BrO BrNO3

HOBrHBrBry

deposition

Justin Parrella, in prep.

GEOS-ChemObserved

CHBr3

440 Gg a-1

CH2Br2

62 Gg a-1

Northern mid-latitudes profiles of short-lived bromocarbons

Sea saltdebromination

0.09 0.8 0.2

5.0 1.5

Mean tropospheric concentrations (ppt) In GEOS-Chem

plankton

industry

GEOS-Chem MODEL OF ATMOSPHERIC MERCURY

• Global 3-D atmospheric simulation driven by GEOS meteorological data and coupled to 2-D dynamic surface ocean and land reservoirs

• Hg(0) oxidation by Br [Donohoue et al., 2005; Goodsite et al., 2004; Balabanov et al. [2005]

• Compare to previous model with Hg(0) oxidation by OH and ozone

Holmes et al., in prep.

(2006)

Streets et al. [2009]

SPECIFICATION OF Br CONCENTRATIONS IN GEOS-Chem Hg MODEL

Zonal mean concentrations (ppt) from bromocarbons + hv, OH simulated by TOMCAT (troposphere) and GMI (stratosphere) with standard gas-phase chemistry

Add 1 ppt BrO in MBL

5 ppt in Arctic spring BL

PREFERENTIAL REGIONS FOR Hg(0) OXIDATION

Annual zonal mean oxidation rates

Hg(0) lifetimeagainst

oxidation0.45 years 0.30 years

Add aqueous-phase photoreduction of Hg(II) in cloudtuned to yield Hg lifetime against deposition of 0.9 years

Holmes et al., in prep.

MODEL EVALUATION AGAINST SURFACE TGM DATATotal gaseous mercury (TGM); model is 2006-2008 annual mean

Hg + BrHg + OH/O3

model: • Unbiased at land sites (r2 =0.88 for Hg+Br, r2 = 0.87 forHg+OH/O3)• Underestimate over N Atlantic is corrected in most recent GEOS-Chem version by using observed subsurface ocean concentrations (Soerensen et al., in prep.)• Hg+Br model has steeper latitudinal gradient

Holmes et al., in prep.

Hg+Br simulation

SEASONAL VARIATION OF TGM

15 sites 3 sites

• Both models reproduce late summer minimum at northern mid-latitudes• Summer maximum at Cape Point is due to ocean emission• Only Hg+Br model can simulate polar spring depletion, summer rebound• Only Hg+Br model can simulate high-RGM subsidence events over Antarctica

Holmes et al., in prep.

VERTICAL PROFILES OF TGM

• Uniform in troposphere, dropping in stratosphere• Arctic spring observations show much faster drop in stratosphere than elsewhere – underestimate of halogen oxidants?

Holmes et al., in prep.

WET DEPOSITION FLUX PATTERNS

MDN and EMEP annual means (2006-2008)Observations as symbols, model as background Seasonal variation

• Hg +Br simulation is too low over Gulf of Mexico in summer – missing Br source in subtropics?• Model is too high at northerly sites in winter – insufficient scavenging by snow?

Holmes et al., in prep.

Hg+Br model

MODEL DEPOSITION PATTERNS DEPEND ON OXIDANT

Hg+Br

Hg+OH/O3

Holmes et al., in prep.

Annual total Hg(II) deposition fluxOxidation by Br causes greater deposition to SH oceans

Environmental implications depend on cycling through land and ocean reservoirs;Development of a fully coupled atmosphere-ocean-land model is underway