reducing uncertainties in modeling atmospheric mercury · • great lakes region with 12 km x 12 km...
TRANSCRIPT
11/23/2009
1
Reducing Uncertainties in Modeling Atmospheric Mercury
Reducing Uncertainties in Modeling Atmospheric Mercury
Tracey HollowayTracey Holloway
Caitlin Littlefield, Jami Morton, Jamie Schauer, Andrew Rutter
University of Wisconsin-Madison
Nelson Institute Center for Sustainability and the Global Environment (SAGE)
R833375
Three different atmospheric mercury species
• Elemental Gaseous Mercury (Hg0) - stays in
Mercury BasicsMercury Basics
• Elemental Gaseous Mercury (Hg ) - stays in atmosphere for 1-1.5 years, global pollutant from global sources
• Reactive Gaseous Mercury (RGHg) - stays in atmosphere for a few days, highly reactive
• Particulate Mercury (PHg) - stays in atmosphere for about a week, removed by precipitation
11/23/2009
2
Mercury BasicsMercury Basics
Mercury BasicsMercury Basics
Hg0 is transported across the globe, while RGHg and PHg are removed
more quickly.
11/23/2009
3
Mercury BasicsMercury Basics
Total wet deposition isTotal wet deposition is known, but species
contribution is unclear.
Mercury BasicsMercury Basics
Total wet deposition isTotal wet deposition is known, but species
contribution is unclear.
11/23/2009
4
Mercury BasicsMercury Basics
Dry deposition is highlyDry deposition is highly uncertain.
Mercury BasicsMercury Basics
Chemical cycling in the atmosphere is uncertain.
11/23/2009
5
Species ContributionSpecies Contribution
•Most mercury emissions as Hg0, followed by RGHg and then PHg
100
75
•Spatial artifacts from data collection methods - by state and country
25
50
Percent Hg0 of Total Annual Emissions
(%)0
Percent RGHg of Total Annual Emissions Percent PHg of Total Annual Emissions
Emissions Meteorology
Modeling SystemsModeling Systems
AssimilatedMeteorology
D t
North AmericanSpeciated Mercuryd C it i E i i
Chemical Transport
Model(CMAQ)
Chemical Transport
Model(CMAQ)
Emissions Processing Model
Regional ClimateModel(WRF)
Dataand Criteria Emissions
ConcentrationsAmbient
ConcentrationsDry
DepositionWet
Deposition
11/23/2009
6
Model Grid & ResolutionModel Grid & Resolution
• Full-year run from January 2003 to January 2004
40
30• U.S. domain with 36 km x 36 km resolution
• Great Lakes region with 12 km x 12 km resolution
• 30 vertical layers in WRF, 15 vertical layers in CMAQ
30
20
10CMAQ• CB05 chemistry and
aerosol module 4, with mercury
Average Temperature for July 2003(°C)
0
Wet DepositionWet Deposition
Mercury Deposition Network (MDN)• Weekly unspeciated total mercury wet deposition• CMAQ generally under-predicts wet deposition
(B ll k d B h 2002 Gb t l 2007 B ll k(Bullock and Brehme, 2002; Gbor et al., 2007; Bullock et al., 2009; Lin et al., 2007)
General agreement across
models, mechanisms andmechanisms and
inputs.
Accuracy or imitation?
11/23/2009
7
Ambient ConcentrationsAmbient Concentrations
• Episodic and/or unspeciated (Sillman et al., 2007; Burke et al., 1995)
Devil’s Lake
, )
• CMAQ captures total mercury concentrations (e.g. Lin and Tao, 2007)
• Comparison with speciated surface concentrations every ..Madison
Milwaukee
surface concentrations every two hours at Devil’s Lake (rural) and Milwaukee (urban) for 2003 and 2004 collected by Jamie Schauer and Andy Rutter
Wet Deposition in the Great LakesWet Deposition in the Great LakesJuly Wet Deposition
y = 0.3859x + 103.18
2500
R2 = 0.2815
500
1000
1500
2000
Mod
eled
(ng/
m2 )
0
500
0 500 1000 1500 2000 2500
MDN (ng/m 2)Observed (ng/m2)
11/23/2009
8
Wet Deposition in the Great LakesWet Deposition in the Great LakesJuly Wet Deposition
y = 0.3859x + 103.18
2500 Reproduces range for total wet deposition as well as previous studies
R2 = 0.2815
500
1000
1500
2000
Mod
eled
(ng/
m2 )
as well as previous studies
Under-predicts Great Lakes weekly wet deposition by a factor of 1.5
Precipitation can explain roughly 50% of wet deposition error
0
500
0 500 1000 1500 2000 2500
MDN (ng/m 2)Observed (ng/m2)
Model reproduces observed values best in summer
Ambient Surface Concentration - RuralAmbient Surface Concentration - RuralHg0
ng/m3
2.5
2
1.5
1
RGHgpg/m3
40
30
20
10
50
Observed Modeled
PHgpg/m340
30
20
10
50
Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May
11/23/2009
9
Ambient Surface Concentration - RuralAmbient Surface Concentration - RuralImported or Emitted?Imported or Emitted?
April RHgpg/m3
July RHgpg/m3pg
Observed Emissions OnlyBase Case Imported Only
Ambient Surface Concentration - UrbanAmbient Surface Concentration - UrbanHg0
ng/m3
5
4
3
2
1
RGHgpg/m3
70
50
30
10
Observed Modeled
PHgpg/m3
20
40
60
80
Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun
11/23/2009
10
Ambient Surface Concentration - UrbanAmbient Surface Concentration - UrbanImported or Emitted?Imported or Emitted?
July RHgpg/m3
Observed Emissions OnlyBase Case Imported Only
Wet DepositionWet DepositionImported or Emitted?Imported or Emitted?
July Wet Deposition
2000
500
1000
1500
Mod
eled
(ng/
m2 )
0
500
0 500 1000 1500 2000
MDN (ng/m 2)
MDN: 2456, 2043Base Case: 807, 973Case IMPT: 766, 916
Case EMIS: 42, 57
Emissions OnlyBase Case Imported Only
Observed (ng/m2)
11/23/2009
11
Dry DepositionDry DepositionImported or Emitted?Imported or Emitted?
100
50
Imported RHg Percent Contribution to Total July Dry Deposition
(%)0
Major Uncertainties…Major Uncertainties…
O di ti t f f ti• Over-prediction at surface for reactive mercury concentrations driven by local chemical production from Hg0
RHg production is too high
• RGHg plumes do not reach rural sitesDry deposition removal likely too fastDry deposition removal likely too fast
• Wet deposition values agree, but are based on erroneous chemistry
“Tuned” wet deposition, not model accuracy
11/23/2009
12
Research Needs
Speciated concentration monitorsSpeciated concentration monitors
Further laboratory study for mercury chemistry and dry deposition
Speciated reporting for wet depositionp p g p
Laboratory and field measurements for dry deposition data
Thank you!Thank you!
Tracey Holloway
Center for Sustainability and the Global Environment (SAGE)
R833375