Lessons Learned:

1
Lessons Learned One-Atmosphere Photochemical
Modeling in Southeastern U.S.
Presentation from Southern Appalachian Mountains
Initiative to Meeting of Regional Planning
Organizations December 3, 2002
2
SAMI Atmospheric Modeling

• Unique Contributions
• Demonstrated fully-integrated one-atmosphere
  model
• ozone, aerosols, and deposition
• performance comparable to or better than recent
  applications of CMAQ or REMSAD

3
SAMI Atmospheric Model

• In 1997 selected to use
• RAMS-3B meteorological model
• EMS-95 emissions model
• Urban to Regional Multi-scale (URM) air quality
  model
• variable grid (12-km over Southern Appalachian
  Mountains)
• SAPRC chemical mechanism for gases
• ISORROPIA for aerosols
• Reactive Scavenging Module for deposition
• Decoupled Direct Method for sensitivity to
  emissions changes

4
SAMI Atmospheric Modeling Domain
Georgia Institute of Technology
5
SAMI Atmospheric Modeling

• Unique Contributions
• Selected episodes to represent annual and
  seasonal air quality measures
• based on meteorology characterized for 5-year
  period
• 9 episodes in Feb, Mar, Apr, May, Jun, Jul, Aug
• Lesson learned prioritize computational power
  greater spatial resolution or longer time
  periods?

6
SAMI Atmospheric Model Lessons
Learned

• Emissions Inventory uncertainties
• especially NH3, primary OC non-road, area
  sources
• Meteorological Model performance
• clouds and precipitation affect chemistry and
  deposition
• wind speed and direction, mixing heights affect
  transport

7
SAMI Atmospheric Model Lessons
Learned

• Air Quality Observations limited spatially and
  temporally
• PM2.5 data, especially NH4
• wet and dry deposition data
• vertical profiles for initial and boundary
  conditions

8
SAMI Atmospheric Model Lessons Learned

• Photochemical Model Performance
• SO4 and OC best performance (/- 50), largest
  components of PM2.5
• overpredict NO3, soil, and EC small components
• SO4 not fully neutralized by NH4, atmosphere NH4-
  limited
• need better measures NH3, NH4, primary vs
  secondary OC

9
URM Model Performance - Fine Particle Mass
Great Smoky Mtns
Class 5
Class 4
Class 3
Class 2
Class 1
30.0
(mg/m3)
20.0
10.0
Concentration
0
2/09/94 3/24/93 4/26/95 8/04/93 8/07/93
8/11/93 7/12/95 7/31/91 7/15/95
Modeled (left) IMPROVE (right)
SO4
NO3
NH4
ORG
EC
SOIL
2/01/01
10
URM Model Performance Sulfate Fine Particle Mass
50
July 1995
May 1995
May 1993
March 1993
February 1994
July 1991
- 50
June 1992
August 1993
April 1995
(based on data from 3-10 IMPROVE sites in 12, 24,
and 48 km grids)
11
URM Model Performance PM2.5 Mass
50
July 1995
May 1995
May 1993
March 1993
URM Modeled Concentration (mg/m3)
February 1994
July 1991
- 50
June 1992
August 1993
April 1995
IMPROVE Measurements (mg/m3)
(based on data from 3-10 IMPROVE sites in 12, 24,
and 48 km grids)
12
Wet Ammonium Deposition Normalized Percent Bias
(based on data from 9-14 NADP wet deposition
sites in 12, 24, and 48 km grids)
Normalized Percent Bias
13
SAMI Atmospheric Modeling

• Unique Contributions
• To assess effects, used modeled relative change
  in air quality to adjust measured air quality
• visibility
• ozone effects to forests
• acid deposition effects to streams and forests
• Lesson learned bound relative reduction factor
  by model performance

14
SAMI Atmospheric Modeling

• Unique Contributions
• Used direct sensitivity analyses to evaluate
  state contributions to Class I areas
• Decoupled Direct Method (DDM-3D)
• evaluated responses to 10 change in emissions
• Lessons learned
• trust relative contributions rather than absolute
  
• daily source contributions from DDM compare
  favorably to daily back trajectories

15
DDM Sensitivity Performance
SO2
NOx
NH3
VOCs
Ozone
Good
Good
Gaseous Species
SO2
Good
NH3
Good
SO4
Good
Good
NO3
Good
Poor
Good
Fair
Aerosol Species
NH4
Good
Good
Fair
Good
OC
Good
Good
EC
SOIL
Wet Deposition Species
PM2.5
Good
Good
Poor
Good
SO4
Good
NO3
Poor
Good
Poor
Good
NH4
Poor
Poor
Poor
Poor
16
(No Transcript)
17
(No Transcript)
18
Annual SO4 Fine Particles Response to 10
Reduction in SO2 Emissions from 2010 A2 strategy
-8.0
Non-SAMI states
SAMI states
-6.0
-4.0
SO4 Fine Particle Response ()
-2.0
0.0
Shining Rock, NC
Dolly Sods, WV
Otter Creek, WV
Joyce Kilmer, NC
Linville Gorge, NC
Shenanhoah, VA
Great Smoky Mtn, TN
James River Face, VA
Sipsey, AL
Cohutta, GA
19
Appendix
20
Sulfate Aerosol Normalized Percent Bias
150
100
50
Normalized Percent Bias
0
-50
-100
(based on data from 3-10 IMPROVE sites in 12, 24,
and 48 km grids)
21
Ammonium Aerosol Normalized Percent Bias
120
100
80
60
40
Normalized Mean Bias
20
0
-20
-40
-60
2/9/1994
5/3/1995
8/4/1993
8/7/1993
7/12/1995
7/15/1995
7/19/1999
5/24/1995
5/27/1995
5/12/1993
5/15/1993
3/24/1993
3/27/1993
3/31/1993
2/12/1994
4/26/1995
4/29/1995
6/24/1992
6/27/1992
8/11/1993
7/24/1991
7/27/1991
7/31/1991
(based on data from 3-10 IMPROVE sites in 12, 24,
and 48 km grids)
22
Organic Aerosol Normalized Percent Bias
120
100
80
60
Normalized Percent Bias
40
20
0
-20
-40
-60
2/9/1994
5/3/1995
8/4/1993
8/7/1993
7/12/1995
7/15/1995
7/19/1999
5/24/1995
5/27/1995
5/12/1993
5/15/1993
3/24/1993
3/27/1993
3/31/1993
2/12/1994
4/26/1995
4/29/1995
6/24/1992
6/27/1992
8/11/1993
7/24/1991
7/27/1991
7/31/1991
(based on data from 3-10 IMPROVE sites in 12, 24,
and 48 km grids)
23
Nitrate Aerosol Normalized Percent Bias
250
200
150
Normalized Percent Bias
100
50
0
-50
2/9/1994
5/3/1995
8/4/1993
8/7/1993
7/15/1995
7/19/1999
5/24/1995
5/27/1995
5/12/1993
5/15/1993
3/24/1993
3/27/1993
3/31/1993
2/12/1994
4/26/1995
4/29/1995
6/24/1992
6/27/1992
8/11/1993
7/24/1991
7/27/1991
7/31/1991
7/12/1995
(based on data from 3-10 IMPROVE sites in 12, 24,
and 48 km grids)
24
PM 2.5 Normalized Percent Bias
(based on data from 3-10 IMPROVE aerosol sites in
12, 24, and 48 km grids)
200
150
100
Normalized Bias ()
50
0
-50
-100
2/9/94
8/7/93
4/29/95
7/24/91
7/15/95
8/11/93
8/4/93
3/24/93
3/31/93
5/15/93
5/24/95
6/24/92
7/31/91
4/26/95
5/27/95
6/27/92
7/27/91
7/12/95
7/19/95
2/12/94
3/27/93
5/12/93
5/21/95
Normalized Bias of /- 50 is potential criteria
for aerosol model performance
25
Aerosol Model Performance
URM Model Results vs. Observations for July 15,
1995
40.0
30.0
20.0
PM 2.5 (m g/m3)
10.0
0.0
Modeled on Left, IMPROVE On Right
James River Face, VA
Shenandoah, VA
Great Smoky Mtns.,TN
Shining Rock, NC
Sipsey, AL
Dolly Sods, WV

2/01/01
SO4
NO3
NH4
ORG
EC
SOIL
26
Wet Sulfate Deposition Normalized Percent Bias
(based on data from 9-14 NADP wet deposition
sites in 12, 24, and 48 km grids)
10
5
0
Normalized Percent Bias
-5
-10
-15
-20
-25
-30
1995
1994
Feb 8-15,
May 3,
April 26 -
11, 1993
1995
1995
1993
August 3-
30, 1993
May 23-30,
May 11-18,
July 23-30,
1991
June 24-29,
1992
July 11-18,
March 23-
27
Confidence Levels in SAMI 1990 Base Year
inventory