1. Building a Global Modeling Capability for Mercury with GEOS-CHEM
Constraining the global budget of mercury and atmospheric processes
Providing boundary conditions for CMAQ
Understanding the behavior of mercury in the Arctic
Tracing pathways of intercontinental mercury pollution
Evaluating the impact of climate change on mercury pathways
Noelle Eckley Selin,
Rokjin J. Park, Daniel J. Jacob

2. THE MERCURY CYCLE: CURRENT
ATMOSPHERE
5000
Wet & Dry
Deposition
2600
Anthropogenic
Emissions
2400
Land emissions
1600
Net Oceanic
Evasion
1500
Net
Wet & Dry
Deposition
1900
( )
( )
( )
( )
( )
SURFACE SOILS
1,000,000
OCEAN
289,000
River
200
Extraction from deep reservoirs
2400
( )
Quantities in Mg/year
Uncertainty ranges in parentheses
Adapted from Mason & Sheu, 2002
Net burial
200

3. ? OH Hg0 Gaseous Hg2+ 1.7 ng/m3 Phase O3 Aqueous Hg0 Phase Hg2+ SO3
k=8.7(+/-2.8) x cm3 s-1 (Sommar et al. 2001)
One measured value in literature OH
Hg0
1.7 ng/m3
Gaseous
Phase
Hg2+
pg/m3
Oxidation O3
k=3(+/-2) x cm3 s-1 (Hall 1995)
Reported rate constants up to k=1.7 x cm3 s-1
Henry’s Constant
0.11 M/atm
Henry’s Constant
1.4x106 M/atm
Oxidation
Hg0
Aqueous
Phase ?
Hg2+
HO2
SO3
Reduction
k= x 104 M-1 s-1 (Pehkonen & Lin 1998)
Shouldn’t occur (Gårdfeldt & Jonsson 2003)
k= (+/ ) s-1 (vanLoon et al. 2000)
Occurs only where high sulfur, low chlorine
Particulate Phase
HgP
1-100 pg/m3

4. What does this mean for global modeling?
Use observations from latitudinal gradient, seasonal cycles, and short-term variability to constrain uncertainties
Potential for application of inverse modeling?
GEOS-CHEM: 2 simulations
“Original” simulation: best guess from the published literature
“Improved” simulation: adjust oxidations to latitudinal gradient and observations

5. MERCURY BUDGET IN GEOS-CHEM
ATMOSPHERE: 4621
τ = 0.82 yr
τ = 21 days
τ = 3.4 days
Hg0
4272
Via OH: 2769
Hg(II)
347
Hg(P) 2
k=1.98 x cm3 s-1
Via O3: 2444
k=3 x cm3 s-1
1500
2000
775
1446
204
Land Re-emissions
Ocean Emissions
2227
160
Anthropogenic Emissions
500
Dry Deposition 43
3673
Wet Deposition
Land (Natural) Emissions
Wet Deposition
Inventories in Mg
Rates in Mg/yr
Dry Deposition

6. Measured
Improved GEOS-CHEM
Original GEOS-CHEM

7. Comparing Model with Measurements: Hemispheric Average TGM
Ratio of NH/SH in measurements: / (Temme et al. 2003)
Ratio of NH/SH in optimized GEOS-CHEM simulation: 1.49
Shows that Hg lifetime in GEOS-CHEM is realistic
Lamborg et al. 2002
GEOS-CHEM

8. TGM: Model vs. Measurements
Guiyang, China:
Measured: 9.00
Modeled: 2.98
Underestimate of sources
In Asia? +
Model is high at northern midlatitudes: overestimate of sources?

9. Wet Deposition: Model vs. Measurements
High Hg deposition in
tropical regions;
Gradient with latitude
Overestimate of deposition:
Reduction in sources needed (14%)?

10. Overestimate between 7-25% (depending on season): overestimate of sources?

11. Future plans for GEOS-CHEM Hg simulation
Land and ocean re-emission parameterization: tracing emissions from source to receptor
Source tag maintained through
deposition and reemission process
Emissions
“tagged” by
source and
region
Chemistry and
Deposition
Reemission
Deposition
“tagged” by
source and
region
Source Region
Receptor
Region
Land or Ocean
Surface
Ocean emissions model: collaboration with
Sarah Strode, Lyatt Univ. of Washington

12. Re-emission Modeling in GEOS-CHEM
Wet and Dry Deposition
Emissions
Emissions
Historical Deposition
Lifetime of “new mercury: weeks to months
“Old Mercury” soil concentrations initialized based on historical deposition patterns of natural, anthropogenic sources
New Mercury
Lifetime of “old mercury”: about 1000 yrs
Old Mercury
920,000 preindustrial
80,000 postindustrial
Quantities in Mg